Menu
▾
▴
tbb-users — application development using TBB; TBB deployment, etc.
You can subscribe to this list here.
| 2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(4) |
Sep
(3) |
Oct
(15) |
Nov
(8) |
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2008 |
Jan
(4) |
Feb
(5) |
Mar
(9) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(3) |
| 2009 |
Jan
(1) |
Feb
(2) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2011 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(1) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Farzaneh G. <far...@gm...> - 2011-07-13 09:04:48
|
Hi,
I have wrote a program by tbb. but runtime is almost the same for sequential
and parallel version
I want to make sure that my sequential program uses only 1 core (I have a
corei5 cpu)
I'm not sure how to do this and where is the problem (runtime)
my program computes shortest path. My graph is stored as adjacency list. I
have an array as input quires.
in sequential program I used a for loop for answering to queries.
in parallel program I used the same code and replace the simple loop by a
parallel loop
*sequential code:*
for(int i=0;i<q;i++)
X[i]=shortest_path(S[i],T[i],i+q);
*parallel loop definition: *
class parallel_shortest_path{
int * i1, * i2;
int *sp;
public:
parallel_shortest_path(int *I1, int *I2,int *SP)
: i1(I1), i2(I2), sp(SP) { }
void operator() (const blocked_range<size_t> &r) const {
for(size_t SPI=r.begin(); SPI!=r.end();SPI++)
if(SPI<q)
sp[SPI]=shortest_path(i1[SPI],i2[SPI],SPI);
}
};
*using parallel loop:*
parallel_for(blocked_range<size_t>(0,q,1000),parallel_shortest_path(S,T,X)
);
I usesd *task_scheduler_init init(); *too.
but runtime does not change if I pass it no input or with inputs 1, 2 or 4!
I'd really appreciate it if someone could help me.
Where is my problem?
how could I make sure my program is running on how many cores?
and how could I get speed up? (it's just answering a number of queries)
Thanks in Advance,
--
Farzaneh Ghayour
M.Sc. Student , Software Engineering
Department of Computer Engineering
Sharif University of Technology
|
|
From: Reed, R. W <rob...@in...> - 2009-02-12 00:55:26
|
I don't know whether anyone has tried TBB with Watcom. This is the first time I've heard such a request. Insofar as Intel distributes in their commercially aligned and stable releases 32-bit Windows binary libraries, these MAY be compatible with the Watcom link objects. I looked at the available documents for the 1.8 release but could not tell whether the C++ language support is robust enough to swallow the TBB C++ headers. A note I saw from a Watcom release from last spring suggested there were serious limitations in the basic language support (from http://www.azillionmonkeys.com/qed/watfaq.shtml): A06. As of version 11.0x the following list applies: Supported features - bool Keyword - mutable Keyword - explicit Keyword - Namespaces - Run-Time Type Information (RTTI) - New Cast Syntax - Exception Specification Not supported features - typename Keyword - Member Templates - New Template Specialization Syntax - Partial Specialization - Template Function Call Syntax - Method try Blocks TBB uses several of these features, and if 1.8 does not support them, it will be tough to get TBB programs compiled using Watcom. --Robert -----Original Message----- From: Childs, D L [mailto:ii...@um...] Sent: Tuesday, February 10, 2009 10:19 AM To: tbb...@li... Subject: [Tbb-users] Watcom Is there any way to use the TBB library with Watcom 1.8? -- Dave__________________________________________________________________ D L Childs * INTEGRATED INFORMATION SYSTEMS XSP TECHNOLOGY http://xsp.xegesis.org * ii...@um... * (734) 668-8951 ------------------------------------------------------------------------------ Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM) software. With Adobe AIR, Ajax developers can use existing skills and code to build responsive, highly engaging applications that combine the power of local resources and data with the reach of the web. Download the Adobe AIR SDK and Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com _______________________________________________ Tbb-users mailing list Tbb...@li... https://lists.sourceforge.net/lists/listinfo/tbb-users |
|
From: Childs, D L <ii...@um...> - 2009-02-10 18:19:21
|
Is there any way to use the TBB library with Watcom 1.8?
--
Dave__________________________________________________________________
D L Childs * INTEGRATED INFORMATION SYSTEMS
XSP TECHNOLOGY http://xsp.xegesis.org * ii...@um... * (734) 668-8951
|
|
From: Reed, R. W <rob...@in...> - 2009-01-13 17:49:50
|
Adrien, though you weren't ranting about GPL licensing regarding TBB specifically, we have noticed that our FAQ on the subject has gotten a bit decrepit of late. We were delayed by the holidays but have work in progress to rewrite specific sections of the FAQ to address some of the questions about dual licensing and related topics and make it easier to find them. Look for that content to change in the future. Thanks for your continuing interest in TBB. -----Original Message----- From: Adrien Guillon [mailto:aj....@gm...] Sent: Thursday, December 11, 2008 9:38 AM To: how to apply TBB to develop scalable parallel applications Subject: Re: [Tbb-users] Misconceptions about TBB I agree. I've asked the FSF if they are willing to create a FAQ about the GPL with runtime exception. Otherwise, the licensing will remain a mystery (I wouldn't trust anyone claiming to be a lawyer on a mailing list, but I would trust an official FAQ from the FSF). Part of my interest in this license is not about TBB, but for my own open source work. I don't want this to become a licensing thread, just a thread for addressing common misconceptions about TBB to generate a good FAQ. AJ On Thu, Dec 11, 2008 at 12:03 PM, Robison, Arch <arc...@in...> wrote: > Unless there are legal experts on this mailing list, the mailing list can do no better than the forum on the topic of licensing. > > - Arch Robison (Intel) > ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Tbb-users mailing list Tbb...@li... https://lists.sourceforge.net/lists/listinfo/tbb-users |
|
From: Adrien G. <aj....@gm...> - 2008-12-11 17:38:07
|
I agree. I've asked the FSF if they are willing to create a FAQ about the GPL with runtime exception. Otherwise, the licensing will remain a mystery (I wouldn't trust anyone claiming to be a lawyer on a mailing list, but I would trust an official FAQ from the FSF). Part of my interest in this license is not about TBB, but for my own open source work. I don't want this to become a licensing thread, just a thread for addressing common misconceptions about TBB to generate a good FAQ. AJ On Thu, Dec 11, 2008 at 12:03 PM, Robison, Arch <arc...@in...> wrote: > Unless there are legal experts on this mailing list, the mailing list can do no better than the forum on the topic of licensing. > > - Arch Robison (Intel) > |
|
From: Robison, A. <arc...@in...> - 2008-12-11 17:03:49
|
Unless there are legal experts on this mailing list, the mailing list can do no better than the forum on the topic of licensing. - Arch Robison (Intel) -----Original Message----- From: Adrien Guillon [mailto:aj....@gm...] Sent: Wednesday, December 10, 2008 9:54 PM To: how to apply TBB to develop scalable parallel applications Subject: [Tbb-users] Misconceptions about TBB I began a conversation on the TBB forums, which can be found here: http://software.intel.com/en-us/forums/intel-threading-building-blocks/topic/62235/ It has been suggested that this conversation should not be continued on the forums, and to respect this request I have moved the discussion to this mailing list where it is more appropriate. I would like to clarify that my intention is to directly address misconceptions that I have encountered. My intention is to provide individuals with an opportunity to bring out misconceptions about TBB and have them discussed. Here is a verbatim copy of the dialogue on the forum: ======== AJ Posted: -------- I've talked about TBB a lot, and I mean, a lot. With all this TBB-talking, I've also encountered some misconceptions. Tonight, I was pushed over the edge while talking online. I'm tired of addressing the same misconceptions over and over. So, I'm here to put these misconceptions out in the open, partially as therapy for myself so I don't go insane when I hear them again :-) Please feel free to post misconceptions you've heard to this thread, I want to get this all out in the open. And if I'm wrong on one of these, please let me know... it would mean I would be the one with the misconception, and that would just be crazy :-) Misconception #1: TBB is not open source. This is kinda silly, but it comes up for some very strange reason. I don't know why. It says right on the main page for TBB: threadingbuildingblocks.org. TBB is released under GPLv2 with the runtime exception. The runtime exception is not taking away rights, look at libstdc++... it's released under the same license, and you trust that right? So download the source code, you'll find the license is in order. It's open source. Now, I think people get very confused with this dual-licensing thing. I myself am very confused as to why there is a commercial version. I've met a number of people who think they have to buy TBB to use it. I'm not a lawyer, so maybe I've really missed some secret thing in the license that says it's not really open... but not that I've seen. I'm not entirely sure about the usage of TBB in commercial products, and I wish that someone would just come out and say what is intended in this regard. The FAQ on the site doesn't help, it refers to a defunct page for libstdc++. My understanding is that the runtime exception means you can use GPLv2 to link against TBB in your commercial programs... obviously if you change TBB itself you have to contribute your changes back... but that's fair. On a side note, I also used to think it was silly to assign my copyright to Intel for contributions. The FSF does that with it's software too. With TBB it's a bit strange because there is a commercial version, and I wouldn't want my code going into a commercial version either. One of my goals this next week is to setup a TBB svn repository, and people who are wary of putting patches up here can patch the unofficial SVN... and keep their own copyright. Misconception #2: TBB only works on Intel hardware. Not that I've seen. I haven't tried it on AMD hardware personally but I know others who have. It may be that there are details about AMD chips that aren't handled elegantly by TBB, but I'm not sure about these issues. There are only a couple of assembly files in all of TBB, and they can be ported to other systems. So nope, I don't see anything that ties TBB to one platform or another. Take a look at the code yourself, if you see something let me know and I'll post a patch myself, after coming to this forum and asking a thousand questions. Perhaps AMD handles barriers in some strange way, I really don't know... but if you think this is the case feel free to patch the code, and respond to this post. I'd like to know where this Intel-only code is too. There might be some truth to this with PowerPC which I understand does barriers in a strange way, but I saw powerpc is supported in the release notes... so... Misconception #3: C is better than C++, hence TBB is a waste Believe it or not, I've met people online who actually believe TBB would have been better implemented in C. Without starting a language war, for the C people out there... C++ can be just as fast as C if carefully crafted. TBB is carefully crafted. It also uses advanced language features of C++ internally, so that we get this really cool library. A TBB-like library in C would be an absolute nightmare to work with. This doesen't mean you can't use TBB within your C code, but that's a different story. I'm not sure what else to say to this one, but I've encountered it. TBB is in C++, and integrates in a really cool manner. Misconceptions you've heard? Feel free to respond to this thread with other misconceptions. Let's get it all out in the open, and if we need to let's make an FAQ to deal with it, let's do it. If you read this thread hoping for something more interesting, my apologies. I want to get all these misconceptions out, and discussed so that they're over with and I don't have to hear them anymore. I can just refer them here. ======== ======== Raf Schietekat Posted: -------- "I wouldn't want my code going into a commercial version either" Why not? "Misconception #2: TBB only works on Intel hardware" The releases you can download from Intel now support almost a third of the architectures targeted in my "Additions to atomic<T>" patch (in no particular order x86/x64, Itanium, Alpha, ARM, MIPS, POWER/PowerPC, PA-RISC, SPARC, ESA/390-z/Architecture). ======== ======== Anton Pegushin (Intel) Posted: -------- Misconception #4 (if no one beats me): If I use parallel algorithms I don't have to care about thread safety. It's been a number of times that I heard people say that using TBB tasks directly is a lot harder than using parallel algorithms, plus (!) in that case user has to take care of thread safety in his application. This is not all true. Allthough generally using tasks directly is more complex, it's not the only scenario in which user needs to think about thread safety. Implementing a function object in the thread safe manner is still a TBB user responsibility and both race conditions and dead locks are possible while using parallel_for or parallel_reduce. And btw, hope this discussion continues and we hear more of those. ======== ======== Dmitriy Vyukov Posted: -------- Misconception #5: TBB is all-powerful and will help me with any kind of application. Sorry, TBB is only for "computations". TBB will be of little help for "unstructured" parallelism. TBB will be of little help for general-purpose client software (non number crunching). TBB won't help with server/middle-ware/systems software. ======== ======== Dmitriy Vyukov Posted: -------- Misconception #6: TBB is so smart so I just have to get linear speedup the very first time. Sorry, the default is performance degradation. Period. Even with TBB. If you are smart and will try hard then, maybe, you will achieve linear speedup. ======== ======== Dmitriy Vyukov Posted: -------- --- Quoting - Dmitriy Vyukov --- Misconception #5: TBB is all-powerful and will help me with any kind of application. Sorry, TBB is only for "computations". TBB will be of little help for "unstructured" parallelism. TBB will be of little help for general-purpose client software (non number crunching). TBB won't help with server/middle-ware/systems software." --- END QUOTE --- This one must go especially to James Reinders ;) According to my investigations only ~20-40% of developers are dealing with number-crunching and are ready to express the algorithm with tasks. All others are still barehanded in the multicore era... ======== ======== dez123 Posted: -------- --- Quoting - AJ --- Misconception #1: TBB is not open source. I'm not entirely sure about the usage of TBB in commercial products, and I wish that someone would just come out and say what is intended in this regard. --- END QUOTE --- Hi Aj, just to clarify this one once more - i AM confused with this dual-licence :( - Do I need a commercial version of TBB in order to be able to sell my product using it (suppose I am not modifying anything in TBB sources)? Otherwise I would compile the sources and distribute DLLs together with my software. Just need a clear statement. I would expect in "Which Intel TBB license is right for your needs?" chapter on TBB frontpage a simple sentence and not something like "If your legal counsel is comfortable with your use of software under the Intel TBB open source license and you do not require commercial support services, please download the latest version of open source Intel TBB...". I am not a native speaker and not a lawyer, but if I get it right, it is either allowed to use OSS-TBB in my commercial software or not allowed, independent of my legal councel being comfortable with it. And nother thing: what do I actually buy for the money apart from Premium Support? Updates? I can get them anyway from the TBB homepage (which if same as OSS version anyway) ... And if this is the case, wouldn't it be better to sell "premium support" and not "the product"? Don't get me wrong, it is not the matter of not paying $299 for the copy, I simply want to know clear this up for me. Thanks a lot! ======== ======== Dmitriy Vyukov Posted: -------- --- Quoting - dez123 --- Hi Aj, just to clarify this one once more - i AM confused with this dual-licence :( - Do I need a commercial version of TBB in order to be able to sell my product using it (suppose I am not modifying anything in TBB sources)? Otherwise I would compile the sources and distribute DLLs together with my software. Just need a clear statement. I would expect in "Which Intel TBB license is right for your needs?" chapter on TBB frontpage a simple sentence and not something like "If your legal counsel is comfortable with your use of software under the Intel TBB open source license and you do not require commercial support services, please download the latest version of open source Intel TBB...". I am not a native speaker and not a lawyer, but if I get it right, it is either allowed to use OSS-TBB in my commercial software or not allowed, independent of my legal councel being comfortable with it. And nother thing: what do I actually buy for the money apart from Premium Support? Updates? I can get them anyway from the TBB homepage (which if same as OSS version anyway) ... And if this is the case, wouldn't it be better to sell "premium support" and not "the product"? Don't get me wrong, it is not the matter of not paying $299 for the copy, I simply want to know clear this up for me. Thanks a lot! --- END QUOTE --- First of all, you are mixing up two independent things: (1) open-sourceness and (2) licencing. Open-source is when you can see the source. No more and less. For example there are paid libraries with open sources. TBB is definitely open-source, just because you can download the sources. TBB is dual-licenced. And free licence is GPLv2 with the runtime exception. If you comfortable with it, you can use it. ======== ======== Dmitriy Vyukov Posted: -------- --- Quoting - Dmitriy Vyukov --- First of all, you are mixing up two independent things: (1) open-sourceness and (2) licencing. Open-source is when you can see the source. No more and less. For example there are paid libraries with open sources. TBB is definitely open-source, just because you can download the sources. --- END QUOTE --- Sorry for that, it seems that it is me who mixing up something... --- Quoting - Dmitriy Vyukov --- TBB is dual-licenced. And free licence is GPLv2 with the runtime exception. If you comfortable with it, you can use it. --- END QUOTE --- When you buy TBB you also get library distributed under different licence (which has nothing in common with GPL, I believe). ======== ======== AJ Posted: -------- It seems to me that you would want the paid product if you intend to change TBB internally, and not tell anyone? Generally you see dual-licensing with something like the GPL, where you can't really release a closed-source product. Perhaps the best thing here is to ask the nice people at the FSF for their advice :-) ======== ======== Alexey Kukanov (Intel) Posted: -------- --- Quoting - AJ --- It seems to me that you would want the paid product if you intend to change TBB internally, and not tell anyone? --- END QUOTE --- Far from truth. Commercial TBB packages do not include sources (request an evaluation license to download one, and check). If you take sources from the TBB site, you get it under the terms of open-source license, no matter whether you have bought a copy of TBB or not. It's not my business why some customers want to buy commercial TBB licenses, but it would be stupid to lose them, would not it? That's just one reason why we provide TBB as a commercial product as well. Also the TBB team is not in a business of interpreting and explaining licensing terms. That's the business of lawyers, so if someone does not understand something about GPL with runtime exception, or any other license, he/she should ask lawyers. And this forum is not the place to ask them, neither will they provide an answer here. So let's better stick to technical topics at this forum. ======== ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ Tbb-users mailing list Tbb...@li... https://lists.sourceforge.net/lists/listinfo/tbb-users |
|
From: Adrien G. <aj....@gm...> - 2008-12-11 03:53:55
|
I began a conversation on the TBB forums, which can be found here: http://software.intel.com/en-us/forums/intel-threading-building-blocks/topic/62235/ It has been suggested that this conversation should not be continued on the forums, and to respect this request I have moved the discussion to this mailing list where it is more appropriate. I would like to clarify that my intention is to directly address misconceptions that I have encountered. My intention is to provide individuals with an opportunity to bring out misconceptions about TBB and have them discussed. Here is a verbatim copy of the dialogue on the forum: ======== AJ Posted: -------- I've talked about TBB a lot, and I mean, a lot. With all this TBB-talking, I've also encountered some misconceptions. Tonight, I was pushed over the edge while talking online. I'm tired of addressing the same misconceptions over and over. So, I'm here to put these misconceptions out in the open, partially as therapy for myself so I don't go insane when I hear them again :-) Please feel free to post misconceptions you've heard to this thread, I want to get this all out in the open. And if I'm wrong on one of these, please let me know... it would mean I would be the one with the misconception, and that would just be crazy :-) Misconception #1: TBB is not open source. This is kinda silly, but it comes up for some very strange reason. I don't know why. It says right on the main page for TBB: threadingbuildingblocks.org. TBB is released under GPLv2 with the runtime exception. The runtime exception is not taking away rights, look at libstdc++... it's released under the same license, and you trust that right? So download the source code, you'll find the license is in order. It's open source. Now, I think people get very confused with this dual-licensing thing. I myself am very confused as to why there is a commercial version. I've met a number of people who think they have to buy TBB to use it. I'm not a lawyer, so maybe I've really missed some secret thing in the license that says it's not really open... but not that I've seen. I'm not entirely sure about the usage of TBB in commercial products, and I wish that someone would just come out and say what is intended in this regard. The FAQ on the site doesn't help, it refers to a defunct page for libstdc++. My understanding is that the runtime exception means you can use GPLv2 to link against TBB in your commercial programs... obviously if you change TBB itself you have to contribute your changes back... but that's fair. On a side note, I also used to think it was silly to assign my copyright to Intel for contributions. The FSF does that with it's software too. With TBB it's a bit strange because there is a commercial version, and I wouldn't want my code going into a commercial version either. One of my goals this next week is to setup a TBB svn repository, and people who are wary of putting patches up here can patch the unofficial SVN... and keep their own copyright. Misconception #2: TBB only works on Intel hardware. Not that I've seen. I haven't tried it on AMD hardware personally but I know others who have. It may be that there are details about AMD chips that aren't handled elegantly by TBB, but I'm not sure about these issues. There are only a couple of assembly files in all of TBB, and they can be ported to other systems. So nope, I don't see anything that ties TBB to one platform or another. Take a look at the code yourself, if you see something let me know and I'll post a patch myself, after coming to this forum and asking a thousand questions. Perhaps AMD handles barriers in some strange way, I really don't know... but if you think this is the case feel free to patch the code, and respond to this post. I'd like to know where this Intel-only code is too. There might be some truth to this with PowerPC which I understand does barriers in a strange way, but I saw powerpc is supported in the release notes... so... Misconception #3: C is better than C++, hence TBB is a waste Believe it or not, I've met people online who actually believe TBB would have been better implemented in C. Without starting a language war, for the C people out there... C++ can be just as fast as C if carefully crafted. TBB is carefully crafted. It also uses advanced language features of C++ internally, so that we get this really cool library. A TBB-like library in C would be an absolute nightmare to work with. This doesen't mean you can't use TBB within your C code, but that's a different story. I'm not sure what else to say to this one, but I've encountered it. TBB is in C++, and integrates in a really cool manner. Misconceptions you've heard? Feel free to respond to this thread with other misconceptions. Let's get it all out in the open, and if we need to let's make an FAQ to deal with it, let's do it. If you read this thread hoping for something more interesting, my apologies. I want to get all these misconceptions out, and discussed so that they're over with and I don't have to hear them anymore. I can just refer them here. ======== ======== Raf Schietekat Posted: -------- "I wouldn't want my code going into a commercial version either" Why not? "Misconception #2: TBB only works on Intel hardware" The releases you can download from Intel now support almost a third of the architectures targeted in my "Additions to atomic<T>" patch (in no particular order x86/x64, Itanium, Alpha, ARM, MIPS, POWER/PowerPC, PA-RISC, SPARC, ESA/390-z/Architecture). ======== ======== Anton Pegushin (Intel) Posted: -------- Misconception #4 (if no one beats me): If I use parallel algorithms I don't have to care about thread safety. It's been a number of times that I heard people say that using TBB tasks directly is a lot harder than using parallel algorithms, plus (!) in that case user has to take care of thread safety in his application. This is not all true. Allthough generally using tasks directly is more complex, it's not the only scenario in which user needs to think about thread safety. Implementing a function object in the thread safe manner is still a TBB user responsibility and both race conditions and dead locks are possible while using parallel_for or parallel_reduce. And btw, hope this discussion continues and we hear more of those. ======== ======== Dmitriy Vyukov Posted: -------- Misconception #5: TBB is all-powerful and will help me with any kind of application. Sorry, TBB is only for "computations". TBB will be of little help for "unstructured" parallelism. TBB will be of little help for general-purpose client software (non number crunching). TBB won't help with server/middle-ware/systems software. ======== ======== Dmitriy Vyukov Posted: -------- Misconception #6: TBB is so smart so I just have to get linear speedup the very first time. Sorry, the default is performance degradation. Period. Even with TBB. If you are smart and will try hard then, maybe, you will achieve linear speedup. ======== ======== Dmitriy Vyukov Posted: -------- --- Quoting - Dmitriy Vyukov --- Misconception #5: TBB is all-powerful and will help me with any kind of application. Sorry, TBB is only for "computations". TBB will be of little help for "unstructured" parallelism. TBB will be of little help for general-purpose client software (non number crunching). TBB won't help with server/middle-ware/systems software." --- END QUOTE --- This one must go especially to James Reinders ;) According to my investigations only ~20-40% of developers are dealing with number-crunching and are ready to express the algorithm with tasks. All others are still barehanded in the multicore era... ======== ======== dez123 Posted: -------- --- Quoting - AJ --- Misconception #1: TBB is not open source. I'm not entirely sure about the usage of TBB in commercial products, and I wish that someone would just come out and say what is intended in this regard. --- END QUOTE --- Hi Aj, just to clarify this one once more - i AM confused with this dual-licence :( - Do I need a commercial version of TBB in order to be able to sell my product using it (suppose I am not modifying anything in TBB sources)? Otherwise I would compile the sources and distribute DLLs together with my software. Just need a clear statement. I would expect in "Which Intel TBB license is right for your needs?" chapter on TBB frontpage a simple sentence and not something like "If your legal counsel is comfortable with your use of software under the Intel TBB open source license and you do not require commercial support services, please download the latest version of open source Intel TBB...". I am not a native speaker and not a lawyer, but if I get it right, it is either allowed to use OSS-TBB in my commercial software or not allowed, independent of my legal councel being comfortable with it. And nother thing: what do I actually buy for the money apart from Premium Support? Updates? I can get them anyway from the TBB homepage (which if same as OSS version anyway) ... And if this is the case, wouldn't it be better to sell "premium support" and not "the product"? Don't get me wrong, it is not the matter of not paying $299 for the copy, I simply want to know clear this up for me. Thanks a lot! ======== ======== Dmitriy Vyukov Posted: -------- --- Quoting - dez123 --- Hi Aj, just to clarify this one once more - i AM confused with this dual-licence :( - Do I need a commercial version of TBB in order to be able to sell my product using it (suppose I am not modifying anything in TBB sources)? Otherwise I would compile the sources and distribute DLLs together with my software. Just need a clear statement. I would expect in "Which Intel TBB license is right for your needs?" chapter on TBB frontpage a simple sentence and not something like "If your legal counsel is comfortable with your use of software under the Intel TBB open source license and you do not require commercial support services, please download the latest version of open source Intel TBB...". I am not a native speaker and not a lawyer, but if I get it right, it is either allowed to use OSS-TBB in my commercial software or not allowed, independent of my legal councel being comfortable with it. And nother thing: what do I actually buy for the money apart from Premium Support? Updates? I can get them anyway from the TBB homepage (which if same as OSS version anyway) ... And if this is the case, wouldn't it be better to sell "premium support" and not "the product"? Don't get me wrong, it is not the matter of not paying $299 for the copy, I simply want to know clear this up for me. Thanks a lot! --- END QUOTE --- First of all, you are mixing up two independent things: (1) open-sourceness and (2) licencing. Open-source is when you can see the source. No more and less. For example there are paid libraries with open sources. TBB is definitely open-source, just because you can download the sources. TBB is dual-licenced. And free licence is GPLv2 with the runtime exception. If you comfortable with it, you can use it. ======== ======== Dmitriy Vyukov Posted: -------- --- Quoting - Dmitriy Vyukov --- First of all, you are mixing up two independent things: (1) open-sourceness and (2) licencing. Open-source is when you can see the source. No more and less. For example there are paid libraries with open sources. TBB is definitely open-source, just because you can download the sources. --- END QUOTE --- Sorry for that, it seems that it is me who mixing up something... --- Quoting - Dmitriy Vyukov --- TBB is dual-licenced. And free licence is GPLv2 with the runtime exception. If you comfortable with it, you can use it. --- END QUOTE --- When you buy TBB you also get library distributed under different licence (which has nothing in common with GPL, I believe). ======== ======== AJ Posted: -------- It seems to me that you would want the paid product if you intend to change TBB internally, and not tell anyone? Generally you see dual-licensing with something like the GPL, where you can't really release a closed-source product. Perhaps the best thing here is to ask the nice people at the FSF for their advice :-) ======== ======== Alexey Kukanov (Intel) Posted: -------- --- Quoting - AJ --- It seems to me that you would want the paid product if you intend to change TBB internally, and not tell anyone? --- END QUOTE --- Far from truth. Commercial TBB packages do not include sources (request an evaluation license to download one, and check). If you take sources from the TBB site, you get it under the terms of open-source license, no matter whether you have bought a copy of TBB or not. It's not my business why some customers want to buy commercial TBB licenses, but it would be stupid to lose them, would not it? That's just one reason why we provide TBB as a commercial product as well. Also the TBB team is not in a business of interpreting and explaining licensing terms. That's the business of lawyers, so if someone does not understand something about GPL with runtime exception, or any other license, he/she should ask lawyers. And this forum is not the place to ask them, neither will they provide an answer here. So let's better stick to technical topics at this forum. ======== |
|
From: Kevin F. <ke...@or...> - 2008-03-30 04:08:43
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: A.J. Guillon, founder of the YetiSim open source project (http://www.yetisim.org) will be presenting a TBB-related poster session at the Spring Simulation Multiconference 2008 in Ottawa, Canada (April 14-17). A.J.'s session is titled "YetiSim: A C++ Simulation Framework Executing State Diagrams Instead of Coroutines." The related paper will be published in the conference journal. Here are the links: http://www.scs.org/confernc/springsim/springsim08/prelimProgram/Poster/2.html http://www.scs.org/confernc/springsim/springsim08/springsim08.htm Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - Intel has proposed TBB tutorials for this year's OSCON. The OSCON schedule should be public soon. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/03/24/threading-building-blocks-and-linux-distributions-part-1/ - "Threading Building Blocks and Linux Distributions, Part 1" - discusses TBB packages development in Debian, Ubuntu, and Fedora. http://softwareblogs.intel.com/2008/03/26/threading-building-blocks-and-linux-distributions-part-2/ - "Threading Building Blocks and Linux Distributions, Part 2" - discusses TBB packages in other Linux distributions, and in FreeBSD. http://softwareblogs.intel.com/2008/03/27/help-how-do-i-find-customer-support-at-intel/ - "HELP! How do I find customer support at Intel?" - a general post informing people about how to find help from Intel. http://softwareblogs.intel.com/2008/03/28/threading-building-blocks-on-wikipedia/ - "Threading Building Blocks on Wikipedia" - talks about the TBB page on Wikipedia, and invites developers who are using TBB to contribute to the page by talking about their own TBB-related projects. -------------------------- TBB Forum Activity http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. This week's forum activity included the following new topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251487.aspx - "Ideas for using the Intel TBB task scheduler with Intel OpenCv" - (2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251494.aspx - "Task scheduler init Segfaults" - (4 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251492.aspx - "Assert in initMemoryManager under PPC64" - (5 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251591.aspx - "TBB I/O Support?" - (2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251664.aspx - "Intel C++ Profe. Compiler - TBB support" - (2 posts) Discussion continued in these topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251136.aspx - "concurrent_hash_map::erase" - (2 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250439.aspx - "MinGW Support" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250849.aspx - " Memory Management Strategy" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251040.aspx - "TBB in interactive application" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250938.aspx - "volatile and memory model" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250057.aspx - "Problems using parallel_scan" - (2 new posts) --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 Several new articles should be available online soon. --------------------------- Polls: The March TBB poll (submitted by A.J. Guillon of http://www.yetisim.org) is continuing. Here's the poll and options: At what project level are you currently applying TBB? * Maintaining software that uses TBB * Developing new software that applies TBB * Modifying existing software to use TBB * Designing new software that will apply TBB * Working at multiple levels on multiple TBB projects * Just getting started (learning about TBB) This poll will run through March. If you haven't voted yet, please visit the http://www.ThreadingBuildingBlocks.org and do so! ---------------------------- Latest TBB Releases: There were no new TBB releases in the past week. The current lastest releases of TBB are: Commercial Aligned Release tbb20_017oss - http://threadingbuildingblocks.org/ver.php?fid=96 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080319oss - http://threadingbuildingblocks.org/ver.php?fid=101 ------------------------------------ That's this week's TBB Weekly Summary post. This is my last TBB weekly summary post (for now). I'm moving on to another project for O'Reilly. Still, I intend to continue as a member of the TBB community. So, I'll be around if you want to contact me. I wish all of you happy multithreaded programming long into the future! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Adrien G. <aj....@gm...> - 2008-03-24 16:41:34
|
Thanks, I have a lot to read about now and learn from your detailed responses. I'm hoping to receive my multi-core programming book pack soon, and that it will have sufficient detail for me to learn more about these issues. AJ On Mon, Mar 24, 2008 at 12:06 PM, Robison, Arch <arc...@in...> wrote: > To Robert's excellent summary, I'll add pipeline effects. With a Pentium > 4 or Core 2 processor, all instructions before an atomic operation must > complete before any instructions after the atomic operation can start. For > long pipelines, that can be a major hit. And there are pipeline replays > involved, so the penalty is a multiple of the pipeline length. > > > > For classic PRAM<http://pages.cs.wisc.edu/%7Etvrdik/2/html/Section2.html#PRAM>machines, lock-free algorithms are faster. The trouble is, there are no > PRAM machines in real life. > > > > For priority queues, we've tried both skiplist and heap based versions. > Google "priority queue concurrent" and you will find ample literature. Be > warned that some of the literature reports results for grossly > oversubscribed machines, or for older machines where caches were less of an > issue. > > > > - Arch > ------------------------------ > > *From:* tbb...@li... [mailto: > tbb...@li...] *On Behalf Of *Reed, Robert W > *Sent:* Friday, March 21, 2008 4:29 PM > *To:* 'tbb...@li...'; tbb...@li... > > *Subject:* Re: [Tbb-develop] TBB Priority Queue > > > > It all has to do with cache lines. And is probably more appropriate a > topic for tbb-users. I've added it to the list. If we continue this > discussion, we should probably do it over there. > > > > The hit from atomic operations is not as severe as when using a mutex, but > not negligible. A cache line on current Intel(R) Architecture is 64 bytes. If > Processing Element A does an atomic operation on 4 bytes of Cache Line 1, it > doesn't need to write the 4 bytes all the way to memory. By the MESI > protocol for cache coherence, all it has to have is Exclusive access to the > cache line. The "atomic" part is the guarantee that PE-A either already has > Line 1 in Modified state or can convert it from Exclusive or Shared state to > Modified without any other intervening operations. In MESI, the process for > converting from Shared to Modified involves a broadcast called Read For > Ownership, which essentially forces all other PEs, if they have that cache > line in Shared state, to mark it Invalid. All this involves various snoop > traffic as the PEs exchange information about their cache lines. If > Processing Element B wants to twiddle any of the bytes in Cache Line 1, it > must contend with PE-A for ownership of the cache line. In MESI, PE-B's RFO > would be aborted so that PE-A could grab the bus and write Line 1 back to > memory. PE-B would try again, doing another RFO as the first step in its > read-modify-write cycle. And there're other complications like > write-through versus write-back caches. So contention is still an issue. > > _______________________________________________________________________ > > Robert Reed Rob...@in... > > Intel SSG/ Developer Products Division/ Performance, Analysis and > > Threading Lab/ Technical Consulting > > Dept. JF1-15, 2111 NE 25th Ave phone 503-264-9624 > > Hillsboro, OR 97124 mobile 503-830-1530 fax 503-264-9227 > > > > The only thing that saves us from the bureaucracy is inefficiency. An > > efficient bureaucracy is the greatest threat to liberty. > > --Eugene McCarthy > > ________________________________________________________________________ > > > > > ------------------------------ > > *From:* tbb...@li... [mailto: > tbb...@li...] *On Behalf Of *Adrien Guillon > *Sent:* Friday, March 21, 2008 12:42 PM > *To:* for people interested in developing TBB itself > *Subject:* Re: [Tbb-develop] TBB Priority Queue > > > > The part that puzzled me was how multiple atomic operations could cause > performance bottlenecks. > > As for priority queues by their nature having internal dependencies which > prohibit efficient concurrent behavior, it's quite possible. I am currently > devoting some effort to the study of efficient parallel containers. I > haven't searched for texts on the subject yet, rather I am waiting for my > multi-core programming books from Intel Press to understand more about what > the containers should look like and how they should behave for good > performance. > > Once I understand what I'm looking for, I will attempt to design some > better behaved containers, and perhaps understand the TBB containers better. > > AJ > > On Fri, Mar 21, 2008 at 3:22 PM, Reed, Robert W <rob...@in...> > wrote: > > I'm not sure I agree, Adrien. Publishing bad code doesn't necessarily > provide any lessons to the reader. If there was a demonstration of how to > avoid the shoals of bad code provided in such a release, there'd be a > positive lesson there. If the answer is that priority queues by their very > nature have too many internal dependencies to permit concurrent thread > access, then you might have to accept that priority queue accesses are > serial events. You put a global lock around your STL priority queue and try > to minimize the use of it. The only reason I could see for wanting such > failed implementations is the expectation that you could do better. > Personally I don't see the value there. > > _______________________________________________________________________ > > Robert Reed Rob...@in... > > Intel SSG/ Developer Products Division/ Performance, Analysis and > > Threading Lab/ Technical Consulting > > Dept. JF1-15, 2111 NE 25th Ave phone 503-264-9624 > > Hillsboro, OR 97124 mobile 503-830-1530 fax 503-264-9227 > > > > Imagine the Creator as a low comedian, and at once the world becomes > > explicable. > > --H.L. Mencken > > ________________________________________________________________________ > ------------------------------ > > *From:* tbb...@li... [mailto: > tbb...@li...] *On Behalf Of *Adrien Guillon > *Sent:* Friday, March 21, 2008 8:36 AM > *To:* for people interested in developing TBB itself > *Subject:* [Tbb-develop] TBB Priority Queue > > > > It was mentioned on the forums that a TBB priority queue was implemented > before, but that the performance was horrible due to many atomic operations. > > Could we have this code made public so that we can learn from your > mistakes without making them ourselves? > > AJ > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Tbb-develop mailing list > Tbb...@li... > https://lists.sourceforge.net/lists/listinfo/tbb-develop > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Tbb-develop mailing list > Tbb...@li... > https://lists.sourceforge.net/lists/listinfo/tbb-develop > > |
|
From: Robison, A. <arc...@in...> - 2008-03-24 16:14:15
|
To Robert's excellent summary, I'll add pipeline effects. With a Pentium 4 or Core 2 processor, all instructions before an atomic operation must complete before any instructions after the atomic operation can start. For long pipelines, that can be a major hit. And there are pipeline replays involved, so the penalty is a multiple of the pipeline length. For classic PRAM <http://pages.cs.wisc.edu/~tvrdik/2/html/Section2.html#PRAM> machines, lock-free algorithms are faster. The trouble is, there are no PRAM machines in real life. For priority queues, we've tried both skiplist and heap based versions. Google "priority queue concurrent" and you will find ample literature. Be warned that some of the literature reports results for grossly oversubscribed machines, or for older machines where caches were less of an issue. - Arch ________________________________ From: tbb...@li... [mailto:tbb...@li...] On Behalf Of Reed, Robert W Sent: Friday, March 21, 2008 4:29 PM To: 'tbb...@li...'; tbb...@li... Subject: Re: [Tbb-develop] TBB Priority Queue It all has to do with cache lines. And is probably more appropriate a topic for tbb-users. I've added it to the list. If we continue this discussion, we should probably do it over there. The hit from atomic operations is not as severe as when using a mutex, but not negligible. A cache line on current Intel(r) Architecture is 64 bytes. If Processing Element A does an atomic operation on 4 bytes of Cache Line 1, it doesn't need to write the 4 bytes all the way to memory. By the MESI protocol for cache coherence, all it has to have is Exclusive access to the cache line. The "atomic" part is the guarantee that PE-A either already has Line 1 in Modified state or can convert it from Exclusive or Shared state to Modified without any other intervening operations. In MESI, the process for converting from Shared to Modified involves a broadcast called Read For Ownership, which essentially forces all other PEs, if they have that cache line in Shared state, to mark it Invalid. All this involves various snoop traffic as the PEs exchange information about their cache lines. If Processing Element B wants to twiddle any of the bytes in Cache Line 1, it must contend with PE-A for ownership of the cache line. In MESI, PE-B's RFO would be aborted so that PE-A could grab the bus and write Line 1 back to memory. PE-B would try again, doing another RFO as the first step in its read-modify-write cycle. And there're other complications like write-through versus write-back caches. So contention is still an issue. _______________________________________________________________________ Robert Reed Rob...@in... Intel SSG/ Developer Products Division/ Performance, Analysis and Threading Lab/ Technical Consulting Dept. JF1-15, 2111 NE 25th Ave phone 503-264-9624 Hillsboro, OR 97124 mobile 503-830-1530 fax 503-264-9227 The only thing that saves us from the bureaucracy is inefficiency. An efficient bureaucracy is the greatest threat to liberty. --Eugene McCarthy ________________________________________________________________________ ________________________________ From: tbb...@li... [mailto:tbb...@li...] On Behalf Of Adrien Guillon Sent: Friday, March 21, 2008 12:42 PM To: for people interested in developing TBB itself Subject: Re: [Tbb-develop] TBB Priority Queue The part that puzzled me was how multiple atomic operations could cause performance bottlenecks. As for priority queues by their nature having internal dependencies which prohibit efficient concurrent behavior, it's quite possible. I am currently devoting some effort to the study of efficient parallel containers. I haven't searched for texts on the subject yet, rather I am waiting for my multi-core programming books from Intel Press to understand more about what the containers should look like and how they should behave for good performance. Once I understand what I'm looking for, I will attempt to design some better behaved containers, and perhaps understand the TBB containers better. AJ On Fri, Mar 21, 2008 at 3:22 PM, Reed, Robert W <rob...@in...> wrote: I'm not sure I agree, Adrien. Publishing bad code doesn't necessarily provide any lessons to the reader. If there was a demonstration of how to avoid the shoals of bad code provided in such a release, there'd be a positive lesson there. If the answer is that priority queues by their very nature have too many internal dependencies to permit concurrent thread access, then you might have to accept that priority queue accesses are serial events. You put a global lock around your STL priority queue and try to minimize the use of it. The only reason I could see for wanting such failed implementations is the expectation that you could do better. Personally I don't see the value there. _______________________________________________________________________ Robert Reed Rob...@in... Intel SSG/ Developer Products Division/ Performance, Analysis and Threading Lab/ Technical Consulting Dept. JF1-15, 2111 NE 25th Ave phone 503-264-9624 Hillsboro, OR 97124 mobile 503-830-1530 fax 503-264-9227 Imagine the Creator as a low comedian, and at once the world becomes explicable. --H.L. Mencken ________________________________________________________________________ ________________________________ From: tbb...@li... [mailto:tbb...@li...] On Behalf Of Adrien Guillon Sent: Friday, March 21, 2008 8:36 AM To: for people interested in developing TBB itself Subject: [Tbb-develop] TBB Priority Queue It was mentioned on the forums that a TBB priority queue was implemented before, but that the performance was horrible due to many atomic operations. Could we have this code made public so that we can learn from your mistakes without making them ourselves? AJ ------------------------------------------------------------------------ - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Tbb-develop mailing list Tbb...@li... https://lists.sourceforge.net/lists/listinfo/tbb-develop |
|
From: Kevin F. <ke...@or...> - 2008-03-22 05:57:28
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: A.J. Guillon, founder of the YetiSim open source project (http://www.yetisim.org) will be presenting a TBB-related poster session at the Spring Simulation Multiconference 2008 in Ottawa, Canada (April 14-17). A.J.'s session is titled "YetiSim: A C++ Simulation Framework Executing State Diagrams Instead of Coroutines." The related paper will be published in the conference journal. Here are the links: http://www.scs.org/confernc/springsim/springsim08/prelimProgram/Poster/2.html http://www.scs.org/confernc/springsim/springsim08/springsim08.htm Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - Intel has proposed TBB tutorials for this year's OSCON. The OSCON schedule should be public within a few weeks. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/03/17/superlinearity-is-impossible-we-just-dont-always-think-correctly/ - "Superlinearity Is Impossible; We Just Don't Always Think Correctly" - discusses the much talked about issue of "superlinearity" -- is it possible for a multithreaded application to have performance that exceeds single-threaded performance by a factor greater than the number of processors/cores? http://softwareblogs.intel.com/2008/03/18/the-multicore-race-continues-who-how-and-why/ - "The Multicore Race Continues: Who, How, and Why" - talks about the Intel and Microsoft partnership for research on new software engineering techniques for addressing multicore computers. http://softwareblogs.intel.com/2008/03/19/superlinearity-and-algorithmic-complexity-or-my-interesting-conversation-with-herb-sutter/ - "Superlinearity and Algorithmic Complexity; or, My Interesting Conversation with Herb Sutter" - Herb Sutter contacted me to chat about superlinearity (his articles in Dr. Dobb's Journal were the inspiration for my blog posts); this post summarizes our conversation and talks about my subsequent thoughts on superlinearity. As always, comments and suggestions are welcome! In particular, if you're working on a TBB-related project, and you have a Web site related to the project, let me know soon, and I'll write a blog about your project. -------------------------- TBB Forum Activity http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. This week's forum activity included the following new topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250904.aspx - "parallel_reduce - associative problem" - (3 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250938.aspx - "volatile and memory model" - (1 post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250923.aspx - "gdb backtrace - concurrent_vector issue or my code?" - (5 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250967.aspx - "Performance Tuning YetiSim For Parallel Processing" - (2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251040.aspx - "TBB in interactive application" - (3 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251041.aspx - "R6034 runtime error!!!" - (5 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250982.aspx - "query queuing_rw_mutex lock state (test for reader/writer state)" - (13 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250898.aspx - "Perhaps a tbb::concurrent_set?" - (4 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251025.aspx - "Scheduler per process?" - (9 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30251136.aspx - "concurrent_hash_map::erase" - (13 posts) Discussion continued in these topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250057.aspx - "Problems using parallel_scan" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250490.aspx - " Porblems compiling TBB with ICC10 and VS2008" - (2 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250837.aspx - "concurrent_vector pop_back() support" - (3 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30245979.aspx - "Using TDD with Cygwin, G++ and Eclipse CDT on windows" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250748.aspx - "concurrent_vector remove / erase elements" - (8 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249564.aspx - "VC++ 2008 Express" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30245197.aspx - "a question to start using TBB" - (2 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250849.aspx - " Memory Management Strategy" - (6 new posts) --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 Several new articles should be available online soon. --------------------------- Polls: The March TBB poll (submitted by A.J. Guillon of http://www.yetisim.org) is continuing. Here's the poll and options: At what project level are you currently applying TBB? * Maintaining software that uses TBB * Developing new software that applies TBB * Modifying existing software to use TBB * Designing new software that will apply TBB * Working at multiple levels on multiple TBB projects * Just getting started (learning about TBB) This poll will run through March. If you haven't voted yet, please visit the http://www.ThreadingBuildingBlocks.org and do so! ---------------------------- Latest TBB Releases: There were no new TBB releases in the past week. The current lastest releases of TBB are: Commercial Aligned Release tbb20_017oss - http://threadingbuildingblocks.org/ver.php?fid=96 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080311oss - http://threadingbuildingblocks.org/ver.php?fid=100 ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Reed, R. W <rob...@in...> - 2008-03-21 21:35:10
|
It all has to do with cache lines. And is probably more appropriate a topic for tbb-users. I've added it to the list. If we continue this discussion, we should probably do it over there.
The hit from atomic operations is not as severe as when using a mutex, but not negligible. A cache line on current Intel(r) Architecture is 64 bytes. If Processing Element A does an atomic operation on 4 bytes of Cache Line 1, it doesn't need to write the 4 bytes all the way to memory. By the MESI protocol for cache coherence, all it has to have is Exclusive access to the cache line. The "atomic" part is the guarantee that PE-A either already has Line 1 in Modified state or can convert it from Exclusive or Shared state to Modified without any other intervening operations. In MESI, the process for converting from Shared to Modified involves a broadcast called Read For Ownership, which essentially forces all other PEs, if they have that cache line in Shared state, to mark it Invalid. All this involves various snoop traffic as the PEs exchange information about their cache lines. If Processing Element B wants to twiddle any of the bytes in Cache Line 1, it must contend with PE-A for ownership of the cache line. In MESI, PE-B's RFO would be aborted so that PE-A could grab the bus and write Line 1 back to memory. PE-B would try again, doing another RFO as the first step in its read-modify-write cycle. And there're other complications like write-through versus write-back caches. So contention is still an issue.
_______________________________________________________________________
Robert Reed Rob...@in...
Intel SSG/ Developer Products Division/ Performance, Analysis and
Threading Lab/ Technical Consulting
Dept. JF1-15, 2111 NE 25th Ave phone 503-264-9624
Hillsboro, OR 97124 mobile 503-830-1530 fax 503-264-9227
The only thing that saves us from the bureaucracy is inefficiency. An
efficient bureaucracy is the greatest threat to liberty.
--Eugene McCarthy
________________________________________________________________________
________________________________
From: tbb...@li... [mailto:tbb...@li...] On Behalf Of Adrien Guillon
Sent: Friday, March 21, 2008 12:42 PM
To: for people interested in developing TBB itself
Subject: Re: [Tbb-develop] TBB Priority Queue
The part that puzzled me was how multiple atomic operations could cause performance bottlenecks.
As for priority queues by their nature having internal dependencies which prohibit efficient concurrent behavior, it's quite possible. I am currently devoting some effort to the study of efficient parallel containers. I haven't searched for texts on the subject yet, rather I am waiting for my multi-core programming books from Intel Press to understand more about what the containers should look like and how they should behave for good performance.
Once I understand what I'm looking for, I will attempt to design some better behaved containers, and perhaps understand the TBB containers better.
AJ
On Fri, Mar 21, 2008 at 3:22 PM, Reed, Robert W <rob...@in...<mailto:rob...@in...>> wrote:
I'm not sure I agree, Adrien. Publishing bad code doesn't necessarily provide any lessons to the reader. If there was a demonstration of how to avoid the shoals of bad code provided in such a release, there'd be a positive lesson there. If the answer is that priority queues by their very nature have too many internal dependencies to permit concurrent thread access, then you might have to accept that priority queue accesses are serial events. You put a global lock around your STL priority queue and try to minimize the use of it. The only reason I could see for wanting such failed implementations is the expectation that you could do better. Personally I don't see the value there.
_______________________________________________________________________
Robert Reed Rob...@in...<mailto:Rob...@in...>
Intel SSG/ Developer Products Division/ Performance, Analysis and
Threading Lab/ Technical Consulting
Dept. JF1-15, 2111 NE 25th Ave phone 503-264-9624
Hillsboro, OR 97124 mobile 503-830-1530 fax 503-264-9227
Imagine the Creator as a low comedian, and at once the world becomes
explicable.
--H.L. Mencken
________________________________________________________________________
________________________________
From: tbb...@li...<mailto:tbb...@li...> [mailto:tbb...@li...<mailto:tbb...@li...>] On Behalf Of Adrien Guillon
Sent: Friday, March 21, 2008 8:36 AM
To: for people interested in developing TBB itself
Subject: [Tbb-develop] TBB Priority Queue
It was mentioned on the forums that a TBB priority queue was implemented before, but that the performance was horrible due to many atomic operations.
Could we have this code made public so that we can learn from your mistakes without making them ourselves?
AJ
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Tbb-develop mailing list
Tbb...@li...<mailto:Tbb...@li...>
https://lists.sourceforge.net/lists/listinfo/tbb-develop
|
|
From: Kevin F. <ke...@or...> - 2008-03-21 16:14:07
|
Hi, I am planning to write some blogs about TBB and Linux distributions that offer TBB packages. I am currently aware of TBB packages on Debian and Ubuntu, and I found a page that indicates TBB will be available in a future Fedora release. Are there other Linux distributions that have packaged TBB? Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-03-15 20:09:48
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: A.J. Guillon, founder of the YetiSim open source project (http://www.yetisim.org) will be presenting a TBB-related poster session at the Spring Simulation Multiconference 2008 in Ottawa, Canada (April 14-17). A.J.'s session is titled "YetiSim: A C++ Simulation Framework Executing State Diagrams Instead of Coroutines." The related paper will be published in the conference journal. Here are the links: http://www.scs.org/confernc/springsim/springsim08/prelimProgram/Poster/2.html http://www.scs.org/confernc/springsim/springsim08/springsim08.htm Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - Intel has proposed TBB tutorials for this year's OSCON. The OSCON schedule should be public within a few weeks. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/03/10/hacking-threading-building-blocks-into-cygwin-part-3/ - "Hacking Threading Building Blocks into Cygwin, Part 3" - the final post in my series about modifying TBB in order to buld and use it in the Cygwin environment. http://softwareblogs.intel.com/2008/03/13/threading-building-blocks-20080226-development-release/ - "Threading Building Blocks 20080226 Development Release" - talks about the primary changes and features in the tbb20_20080226oss development release. http://softwareblogs.intel.com/2008/03/14/threading-building-blocks-early-march-2008-development-releases/ - "Threading Building Blocks Early March 2008 Development Releases" - talks about the primary changes and features in the tbb20_20080304oss and tbb20_20080311 development releases. As always, comments and suggestions are welcome! In particular, if you're working on a TBB-related project, and you have a Web site related to the project, let me know soon, and I'll write a blog about your project. -------------------------- TBB Forum Activity http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. This week's forum activity included the following new topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250542.aspx - " parallel_reduce : Exception in summing" - (3 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250490.aspx - " Porblems compiling TBB with ICC10 and VS2008" - (2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250537.aspx - "Using Tasks as a child of parallel_for" - (6 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250745.aspx - "concurrent_vector parallel push_back operations" - (4 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250748.aspx - "concurrent_vector remove / erase elements" - (5 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250773.aspx - "concurrent_hash_map parallel operations and complexity" - (2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250837.aspx - "concurrent_vector pop_back() support" - (1 post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250849.aspx - " Memory Management Strategy" - (1 post) Discussion continued in these topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250439.aspx - "MinGW Support" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250244.aspx - "problems using parallel_do" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247611.aspx - "parallel_for alternative?" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249841.aspx - " Google Summer of Code 2008!!" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30245979.aspx - "Using TDD with Cygwin, G++ and Eclipse CDT on windows" - (1 new post) --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 A new article, "Getting Started with TBB on Windows," should be available online soon. --------------------------- Polls: The March TBB poll (submitted by A.J. Guillon of http://www.yetisim.org) is continuing. Here's the poll and options: At what project level are you currently applying TBB? * Maintaining software that uses TBB * Developing new software that applies TBB * Modifying existing software to use TBB * Designing new software that will apply TBB * Working at multiple levels on multiple TBB projects * Just getting started (learning about TBB) This poll will run through March. If you haven't voted yet, please visit the http://www.ThreadingBuildingBlocks.org and do so! ---------------------------- Latest TBB Releases: The past week included a new TBB development release: tbb20_20080311oss. The current lastest releases of TBB are: Commercial Aligned Release tbb20_017oss - http://threadingbuildingblocks.org/ver.php?fid=96 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080311oss - http://threadingbuildingblocks.org/ver.php?fid=100 ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-03-08 18:53:31
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: A.J. Guillon, founder of the YetiSim open source project (http://www.yetisim.org) will be presenting a TBB-related poster session at the Spring Simulation Multiconference 2008 in Ottawa, Canada (April 14-17). A.J.'s session is titled "YetiSim: A C++ Simulation Framework Executing State Diagrams Instead of Coroutines." The related paper will be published in the conference journal. Here are the links: http://www.scs.org/confernc/springsim/springsim08/prelimProgram/Poster/2.html http://www.scs.org/confernc/springsim/springsim08/springsim08.htm Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - Intel has proposed TBB tutorials for this year's OSCON. The OSCON schedule should be public within a few weeks. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/03/01/yetisim-paper-to-be-presented-at-springsim08/ - "YetiSim Paper to Be Presented at SpringSim'08" - talks about the TBB-related paper that will be presented at the SpringSim'08 conference by Adrien Guillon and Deborah Loach. http://softwareblogs.intel.com/2008/03/02/threading-building-blocks-downloadable-extras/ - "Threading Building Blocks Downloadable Extras" - talks about the new "Extras" download area on the TBB Downloads tab. http://softwareblogs.intel.com/2008/03/04/why-a-simple-test-can-get-parallel-slowdown/ - "Why a simple test can get parallel slowdown" - Alexey Kukanov's discussion of how and why multithreading simple problems can lead to a performance reduction instead of a performance increase. http://softwareblogs.intel.com/2008/03/05/threading-building-blocks-and-cygwin-part-1/ - "Threading Building Blocks and Cygwin, Part 1" - talks about a experimental effort to build and use TBB in the Cygwin environment. http://softwareblogs.intel.com/2008/03/05/hacking-threading-building-blocks-into-cygwin-part-2/ - "Hacking Threading Building Blocks into Cygwin, Part 2" - continues the TBB in Cygwin experiment. As always, comments and suggestions are welcome! In particular, if you're working on a TBB-related project, and you have a Web site related to the project, let me know soon, and I'll write a blog about your project. -------------------------- TBB Forum Activity http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. This week's forum activity included the following new topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250242.aspx - "Asserting binary representations for atomic<T>" - (new topic: 3 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250057.aspx - "Problems using parallel_scan" - (new topic: 4 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250244.aspx - "problems using parallel_do" - (new topic: 3 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250439.aspx - "MinGW Support" - (new topic: 1 post) Discussion continued in these topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30241983.aspx - "TBB On Solaris" - (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249884.aspx - " performance of parallel_for" - (4 new posts) --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 A new article, "Getting Started with TBB on Windows," should be available online soon. --------------------------- Polls: The March TBB poll (submitted by A.J. Guillon of http://www.yetisim.org) is continuing. Here's the poll and options: At what project level are you currently applying TBB? * Maintaining software that uses TBB * Developing new software that applies TBB * Modifying existing software to use TBB * Designing new software that will apply TBB * Working at multiple levels on multiple TBB projects * Just getting started (learning about TBB) This poll will run through March. If you haven't voted yet, please visit the http://www.ThreadingBuildingBlocks.org and do so! ---------------------------- Latest TBB Releases: There were no new TBB releases in the past week. The current lastest releases of TBB are: Commercial Aligned Release tbb20_017oss - http://threadingbuildingblocks.org/ver.php?fid=96 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080122oss - http://threadingbuildingblocks.org/ver.php?fid=94 ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-03-01 18:10:11
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: A.J. Guillon, founder of the YetiSim open source project (http://www.yetisim.org) will be presenting a TBB-related poster session at the Spring Simulation Multiconference 2008 in Ottawa, Canada (April 14-17). A.J.'s session is titled "YetiSim: A C++ Simulation Framework Executing State Diagrams Instead of Coroutines." The related paper will be published in the conference journal. Here are the links: http://www.scs.org/confernc/springsim/springsim08/prelimProgram/Poster/2.html http://www.scs.org/confernc/springsim/springsim08/springsim08.htm Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I'm cancelling the 45-minute session I had proposed for OSCON. However, there will still be a TBB presence at the conference: Intel has proposed a tutorial session. The OSCON schedule should be public within a few weeks. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/02/24/under-the-hood-count_strings/ - "Under the hood: count_strings" by Robert Reed - talks about Robert's recent experiments where he applied the Intel Thread Profiler was to TBB code. http://softwareblogs.intel.com/2008/02/25/threading-building-blocks-template-for-visual-studio-now-available/ - "Threading Building Blocks Template for Visual Studio Now Available" - in this post, I point people to the template developed by the "Parallelism Panorama" blog owner (http://llpanorama.wordpress.org) for Visual Studio Express Edition. The template simplifies the process of configuring Visual Studio Express for developing TBB applications. http://softwareblogs.intel.com/2008/02/26/threading-building-blocks-commercial-aligned-version-20_017/ - "Threading Building Blocks Commercial Aligned Version 20_017" - in this post, I analyze the changes between the previous TBB commercial aligned version (20_014) and the new 20_017 commercial aligned version that was released last week. http://softwareblogs.intel.com/2008/02/29/poll-result-tbb-algorithms-are-the-most-widely-used-component/ - "Poll Result: TBB Algorithms Are the Most Widely Used Component" - in this post, I report on the results of the February TBB poll and introduce the March poll. As always, comments and suggestions are welcome! In particular, if you're working on a TBB-related project, and you have a Web site related to the project, let me know soon, and I'll write a blog about your project. -------------------------- TBB Forum Activity http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. This week's forum activity included the following new topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249610.aspx - "Contribution: parallel_for_group" (new topic: 6 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249564.aspx - "VC++ 2008 Express" (new topic: 5 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249688.aspx - "Possible Bugs in TBB Memory Allocators" (new topic: 2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249768.aspx - "How can i get the best grainsize?" (new topic: 2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249841.aspx - "Google Summer of Code 2008!!" (new topic: 1 post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249843.aspx - "Contribution: STL Containers Wrapped for cache_aligned and scalable allocators" (new topic: 1 post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249849.aspx - "Contribution: Owning Container" (new topic: 2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249891.aspx - "Threading Building Blocks 2.0 update 2 available" (new topic: 1 post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249897.aspx - "Development Progress on TBB I/O Support?" (new topic: 2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249865.aspx - "scalable_malloc error" (new topic: 5 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249884.aspx - "performance of parallel_for" (new topic: 5 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30250011.aspx - "Microsoft Visual Studio TBB plug-in available" (new topic: 1 post) Discussion continued in these topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249498.aspx - "TBB Memory Allocator" (2 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249288.aspx - "Abusing parallel_while" (3 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30245197.aspx - "a question to start using TBB" (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30241983.aspx - "TBB On Solaris" (1 new post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30238853.aspx - "TBB's status with Operating System Vendors (OSVs)" - (1 new post) --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 --------------------------- Polls: The March TBB poll (submitted by A.J. Guillon of http://www.yetisim.org) has started! Here's the poll and options: At what project level are you currently applying TBB? * Maintaining software that uses TBB * Developing new software that applies TBB * Modifying existing software to use TBB * Designing new software that will apply TBB * Working at multiple levels on multiple TBB projects * Just getting started (learning about TBB) This poll will run through March. If you haven't voted yet, please visit the http://www.ThreadingBuildingBlocks.org and do so! ---------------------------- Latest TBB Releases: A new TBB commercial aligned release was posted on the TBB download site this week: http://threadingbuildingblocks.org/ver.php?fid=96 You can read about the changes in this release in my blog post "Threading Building Blocks Commercial Aligned Version 20_017" http://softwareblogs.intel.com/2008/02/26/threading-building-blocks-commercial-aligned-version-20_017/ The current lastest releases of TBB are: Commercial Aligned Release tbb20_017oss - http://threadingbuildingblocks.org/ver.php?fid=96 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080122oss - http://threadingbuildingblocks.org/ver.php?fid=94 ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-02-28 02:22:02
|
Hi, In my weekly updates I have a paragraph that asks anyone who might want to write an article about their work with TBB, or a tutorial that covers a TBB component or two in detail, to contact me. I'm writing tonight because I have a new deadline for any articles people might like to submit: I'll need to receive all articles by March 15 or so. Again, the articles should be around 2000 words; the pay is $800 US. If you have been considering writing an article, and you're sure you could complete it by my mid-March deadline, please contact me immediately, so we can discuss your proposed topic and get your work underway. Regards, Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-02-23 03:12:09
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: This week on the #tbb IRC channel (FreeNode.net) we had a visit from a University of Michigan undergraduate Computer Science student who had just finished a mid-term exam on concurrency and TBB. More evidence of the impact TBB is having in colleges and universities. I posted a blog about this titled "Studying for TBB Mid-Terms": http://softwareblogs.intel.com/2008/02/22/studying-for-tbb-mid-terms/ Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I will be attending OSCON in July. Still awaiting word on the proposal I submitted for a 45 minute session that will feature an application that has a Python GUI and a TBB-threaded computation engine. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following post: http://softwareblogs.intel.com/2008/02/22/studying-for-tbb-mid-terms/ - "Studying for TBB Mid-Terms" - a brief post talking about TBB in education, specifically the new course at University of Michigan we heard about when one of the students visited the #tbb IRC channel a few days ago. As always, comments and suggestions are welcome! -------------------------- TBB Forum Activity http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. This week's forum activity included the following new topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249424.aspx "Why const body in parallel_for implementation?" (new topic: 2 posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249469.aspx "An Interesting Find: Qt Concurrent" (new topic: 1 post) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249498.aspx "TBB Memory Allocator" (new topic: 2 posts) Discussion continued in these topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249288.aspx "Abusing parallel_while" (6 new posts) http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30249226.aspx "Contribution process fails to upload file" (4 new posts) --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 I continue to actively seek authors for Threading Building Blocks articles. Articles should be about TBB itself (a study of one or more components) or about an application that uses TBB. The articles should be about 2000 words long. The pay is $800 US. If you're interested in proposing an article, contact me and I'll send you my author's package, as well as answer any questions you may have about writing TBB articles. --------------------------- Polls: The February TBB poll has one more week to go. If you haven't voted, please do so. Here's the poll and options: Which class of TBB component have you applied the most thus far? * Algorithms (parallel_for, parallel_reduce, pipeline, etc.) * Containers (concurrent_vector, concurrent_queue, etc.) * Memory allocators * Mutexes and atomic * Tasker * Most or all of the above * None of the above (I'm just getting started with TBB) This poll will run through February. As always, I'm seeking help with coming up with the next poll. If you've got an idea for a poll related to TBB or scalable technologies, etc., please send it along! ---------------------------- Latest TBB Releases: There were no new TBB releases this week. The current lastest releases of TBB are: Commercial Aligned Release tbb20_014oss - http://threadingbuildingblocks.org/ver.php?fid=75 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080122oss - http://threadingbuildingblocks.org/ver.php?fid=94 See my TBB Release Matrix post for basic information about the differences between the releases (does not yet include the tbb20_20071218oss, tbb20_20080115oss, or tbb20_20080122 Development Releases): http://softwareblogs.intel.com/2007/12/07/threading-building-blocks-open-source-release-versions-matrix/ ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-02-16 17:00:14
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: Since my last blog post, I've been thinking a lot about whether "superlinear speedups" are possible -- i.e., the notion that solving a problem using N processors/cores can result in a greater than N times performance increase -- or whether any apparent superlinear performance increase is really just an illusion. I'd be interested in hearing what others think about this topic. See my blog post "Can Parallelism Achieve Superlinear Performance Gains?" if you're interested in joining the discussion: http://softwareblogs.intel.com/2008/02/15/can-parallelism-achieve-superlinear-performance-gains/ Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I will be attending OSCON in July. I've submitting a proposal for a 45 minute session that will feature an application that has a Python GUI and a TBB-threaded computation engine. Some of my recent blog posts talk about my research in this area. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/02/09/nice-tutorial-getting-started-with-tbb-on-windows/ "Nice Tutorial: Getting Started with TBB on Windows" - this post talks about and points readers to a nice brief tutorial on how to get TBB installed and configured on Windows systems. http://softwareblogs.intel.com/2008/02/13/threading-building-blocks-early-2008-development-releases/ "Threading Building Blocks Early 2008 Development Releases" - in this post, I reviewed the three Development releases that have been posted to the TBB Download thus far in 2008. http://softwareblogs.intel.com/2008/02/15/can-parallelism-achieve-superlinear-performance-gains/ "Can Parallelism Achieve Superlinear Performance Gains?" - this post talks about a recent article in Dr. Dobb's Journal in which Herb Sutter writes about achieving superlinear speedups when certain types of application are run on multicore systems. As always, comments and suggestions are welcome! -------------------------- TBB Forum Activity http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. This week's forum activity included the following topics: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30248972.aspx "TBB performance on Itanium2" - a conversation about TBB performance on 32 and 128 processor Altix Itanium2 systems. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30248401.aspx "How to install TBB library in Ubuntu" - discusses issues encountered when trying to install TBB in Ubuntu on a system that also has the Boost libraries installed. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30248657.aspx "Question about TBB building" - a TBB user had some questions about building TBB, which were answered by a member of the TBB development team. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30248684.aspx "Development Journey Using TBB" - questions from a TBB beginner who would like to parallelize Dijkstra's shortest path sequential algorithm using TBB. --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 I continue to actively seek authors for Threading Building Blocks articles. Articles should be about TBB itself (a study of one or more components) or about an application that uses TBB. The articles should be about 2000 words long. The pay is $800 US. If you're interested in proposing an article, contact me and I'll send you my author's package, as well as answer any questions you may have about writing TBB articles. --------------------------- Polls: The February TBB poll continues, with a decent response thus far. If you haven't voted, please do so. Here's the poll and options: Which class of TBB component have you applied the most thus far? * Algorithms (parallel_for, parallel_reduce, pipeline, etc.) * Containers (concurrent_vector, concurrent_queue, etc.) * Memory allocators * Mutexes and atomic * Tasker * Most or all of the above * None of the above (I'm just getting started with TBB) This poll will run through February. As always, I'm seeking help with coming up with the next poll. If you've got an idea for a poll related to TBB or scalable technologies, etc., please send it along! ---------------------------- Latest TBB Releases: There were no new TBB releases this week. The current lastest releases of TBB are: Commercial Aligned Release tbb20_014oss - http://threadingbuildingblocks.org/ver.php?fid=75 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080122oss - http://threadingbuildingblocks.org/ver.php?fid=94 See my TBB Release Matrix post for basic information about the differences between the releases (does not yet include the tbb20_20071218oss, tbb20_20080115oss, or tbb20_20080122 Development Releases): http://softwareblogs.intel.com/2007/12/07/threading-building-blocks-open-source-release-versions-matrix/ ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-02-10 05:15:13
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: http://conferences.oreilly.com/money/ - I attended the Money:Tech Conference this past Wednesday and Thursday in New York, NY. It was remarkable to see how much of Wall Street trading strategy and trade execution is now automated. I've posted two blogs related to Money:Tech thus far, and will be referencing my experiences from the conference in other posts in the future. Conference news update: http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I will be attending OSCON in July. I've submitting a proposal for a 45 minute session that will feature an application that has a Python GUI and a TBB-threaded computation engine. Some of my recent blog posts talk about my research in this area. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/02/04/swarm-algorithms-better-pictures-and-multicore-computers/ "Swarm Algorithms, Better Pictures, and Multicore Computers" - talks about algorithms I heard about through a Dr. Dobbs Journal newsletter I subscribe to, and why I think this type of algorithm exemplefies the benefits of multithreading for desktop applications. http://softwareblogs.intel.com/2008/02/07/is-data-parallelism-in-our-future/ "Is Data Parallelism in our future?" - Intel's Clay Breshears talks about data parallelism and some of the capabilities TBB offers when you start thinking about future many-core computers. http://softwareblogs.intel.com/2008/02/07/pattern-recognition-money-algorithms-and-web-20/ "Pattern Recognition, Money, Algorithms, and Web 2.0" - my blog post written after attending the first day of the Money:Tech conference. http://softwareblogs.intel.com/2008/02/09/the-problem-with-algorithms-and-money-and-life/ "The Problem with Algorithms and Money (and Life?)" - my blog post after attending the second and final day of the Money:Tech conference. -------------------------- TBB Forum Activity: http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247563.aspx "Locking elements in a matrix" - this ongoing discussion continued with three more posts this week. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30248401.aspx "How to install TBB library in Ubuntu" - discusses issues encountered when trying to install TBB in Ubuntu on a system that also has the Boost libraries installed. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247611.aspx "parallel_for alternative?" - this ongoing discussion continued. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30248604.aspx "Introducing tbbcommunity.org" - this post introduces a new site with the intention of gathering TBB resources and documentation to supplement what's available on threadingbuildingblocks.org. --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 I continue to actively seek authors for Threading Building Blocks articles. Articles should be about TBB itself (a study of one or more components) or about an application that uses TBB. The articles should be about 2000 words long. The pay is $800 US. If you're interested in proposing an article, contact me and I'll send you my author's package, as well as answer any questions you may have about writing TBB articles. --------------------------- Polls: The February TBB poll continues, with a decent response thus far. If you haven't voted, please do so. Here's the poll and options: Which class of TBB component have you applied the most thus far? * Algorithms (parallel_for, parallel_reduce, pipeline, etc.) * Containers (concurrent_vector, concurrent_queue, etc.) * Memory allocators * Mutexes and atomic * Tasker * Most or all of the above * None of the above (I'm just getting started with TBB) This poll will run through February. As always, I'm seeking help with coming up with the next poll. If you've got an idea for a poll related to TBB or scalable technologies, etc., please send it along! ---------------------------- Latest TBB Releases: There were no new TBB releases this week. The current lastest releases of TBB are: Commercial Aligned Release tbb20_014oss - http://threadingbuildingblocks.org/ver.php?fid=75 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080122oss - http://threadingbuildingblocks.org/ver.php?fid=94 See my TBB Release Matrix post for basic information about the differences between the releases (does not yet include the tbb20_20071218oss, tbb20_20080115oss, or tbb20_20080122 Development Releases): http://softwareblogs.intel.com/2007/12/07/threading-building-blocks-open-source-release-versions-matrix/ ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-02-02 18:23:59
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: Conference news update: http://conferences.oreilly.com/money/ - Money:Tech Conference (Feb 6-7, New York, NY) - I will be traveling to New York Tuesday evening to attend the O'Reilly Money:Tech Conference. I'll let you know about the conference in a series of blogs starting on Wednesday. http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I will be attending OSCON in July. I've submitting a proposal for a 45 minute session that will feature an application that has a Python GUI and a TBB-threaded computation engine. Some of my recent blog posts talk about my research in this area. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/01/28/integrating-python-and-threading-building-blocks-part-3/ "Integrating Python and Threading Building Blocks, Part 3" - continues the discussion of my work on integrating a Python front end with C++ using the Boost.Python library. The ultimate objective is to create an application that has a Python GUI that runs a C++ mathematical analysis that has been threaded using TBB. http://softwareblogs.intel.com/2008/01/31/abstracting-thread-local-storage/ "Abstracting Thread Local Storage" (Arch Robison) - In this post, TBB architect Arch Robison talks about some ideas he has for possible future enhancements to TBB involving thread local storage. http://softwareblogs.intel.com/2008/02/02/poll-result-expectations-for-2008s-most-important-advance/ "Poll Result: Expectations for 2008's Most Important Advance" - this post presents the results of January's TBB polls, talks about the ongoing debates about the impact of multicore processors, and introduces the February TBB poll. -------------------------- TBB Forum Activity: http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247002.aspx "How to get thread ID in TBB?" - this discussion continued this week, and was one of the reasons Arch Robison wrote his "Abstracting Thread Local Storage" post (see above). http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247611.aspx "parallel_for alternative?" - this was the most active discussion in the past week. The discussion is about a specific application situation, and includes code snippets. --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 I continue to actively seek authors for Threading Building Blocks articles. Articles should be about TBB itself (a study of one or more components) or about an application that uses TBB. The articles should be about 2000 words long. The pay is $800 US. If you're interested in proposing an article, contact me and I'll send you my author's package, as well as answer any questions you may have about writing TBB articles. --------------------------- Polls: A new TBB poll was just launched: : Which class of TBB component have you applied the most thus far? * Algorithms (parallel_for, parallel_reduce, pipeline, etc.) * Containers (concurrent_vector, concurrent_queue, etc.) * Memory allocators * Mutexes and atomic * Tasker * Most or all of the above * None of the above (I'm just getting started with TBB) This poll will run through February. As always, I'm seeking help with coming up with the next poll. If you've got an idea for a poll related to TBB or scalable technologies, etc., please send it along! ---------------------------- Latest TBB Releases: This week saw a the second TBB release of 2008, the tbb20_20080122oss development release. This release is the result of the prior week's development work on TBB. The current lastest releases of TBB are: Commercial Aligned Release tbb20_014oss - http://threadingbuildingblocks.org/ver.php?fid=75 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080122oss - http://threadingbuildingblocks.org/ver.php?fid=94 See my TBB Release Matrix post for basic information about the differences between the releases (does not yet include the tbb20_20071218oss, tbb20_20080115oss, or tbb20_20080122 Development Releases): http://softwareblogs.intel.com/2007/12/07/threading-building-blocks-open-source-release-versions-matrix/ ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-01-27 02:48:58
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of TBB blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: One of my blogs from last week: http://softwareblogs.intel.com/2008/01/17/interesting-multicore-crisis-graph-and-analysis/ was noticed by Alexander Wolfe, who writes the "Wolfe's Den" blog on http://www.InformationWeek.com. Alexander wrote a post titled "Intel Blog Warns Of Multicore Crisis" in response to my post, and many people posted comments (mostly doubting that a "multicore crisis" will ever happen). Here's Alexander's post: http://www.informationweek.com/blog/main/archives/2008/01/intel_blog_warn.html Conference news update: http://conferences.oreilly.com/money/ - Money:Tech Conference (Feb 6-7, New York, NY) - I will be attending the upcoming O'Reilly Money:Tech Conference. This conference looks very relevant to TBB, now that I've studied the schedule. Much of the discussion and presentation will involve mathematical models used for automated trading or for providing analysis to traders. Clearly the faster these models can be run, the greater the advantage for executing trades before your competitor notices the trend or pricing anomaly. I'll of course tell you all about the conference in a series of blogs. http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I will be attending OSCON in July. I'm submitting a proposal for a 45 minute session that will feature an application that has a Python GUI and a TBB-threaded computation engine. Some of my recent blog posts talk about my research in this area. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). You can also see the most recent TBB blog posts on the http://www.ThreadingBuildingBlocks.org home page. This past week included the following posts: http://softwareblogs.intel.com/2008/01/20/threading-building-blocks-tbb20_20071218oss-development-release/ "Threading Building Blocks tbb20_20071218oss Development Release" - presents the major modifications in the tbb20_20071218oss development release, including the new recursive_mutex component. http://softwareblogs.intel.com/2008/01/21/integrating-python-and-threading-building-blocks-part-1/ "Integrating Python and Threading Building Blocks, Part 1" - outlines the initial steps in my effort to create an application that has a Python GUI and runs a TBB-threaded computation engine, namely the installation of the Boost library. http://softwareblogs.intel.com/2008/01/25/integrating-python-and-threading-building-blocks-part-2/ "Integrating Python and Threading Building Blocks, Part 2" - continues the discussion by describing the compilation and testing of the Boost.Python library. -------------------------- TBB Forum Activity: http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. You can also see the most recent TBB forum posts on the http://www.ThreadingBuildingBlocks.org home page. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247346.aspx "Installing tbb in windows" - this discussion covered many aspects of installing TBB on Windows; it was the most active discussion on the TBB forum this past week. http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247563.aspx "Locking elements in a matrix" - this discussion was about how to implement locking specific elements within a matrix. In addition, there were questions and discussion about a parallel_for alternative and additional discussion about applying TBB in a simple application. --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 I continue to actively seek authors for Threading Building Blocks articles. Articles should be about TBB itself (a study of one or more components) or about an application that uses TBB. The articles should be about 2000 words long. The pay is $800 US. If you're interested in proposing an article, contact me and I'll send you my author's package, as well as answer any questions you may have about writing TBB articles. --------------------------- Polls: The current TBB poll is: The most important software development technology advance in 2008 will involve * Web development technologies/frameworks * Scalable server technologies * Desktop client/application technologies * Other * Don't know This poll will run through January. As always, I'm seeking help with coming up with the next poll. If you've got an idea for a poll related to TBB or scalable technologies, etc., please send it along! ---------------------------- Latest TBB Releases: This week saw a the second TBB release of 2008, the tbb20_20080122oss development release. This release is the result of the prior week's development work on TBB. The current lastest releases of TBB are: Commercial Aligned Release tbb20_014oss - http://threadingbuildingblocks.org/ver.php?fid=75 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080122oss - http://threadingbuildingblocks.org/ver.php?fid=94 See my TBB Release Matrix post for basic information about the differences between the releases (does not yet include the tbb20_20071218oss, tbb20_20080115oss, or tbb20_20080122 Development Releases): http://softwareblogs.intel.com/2007/12/07/threading-building-blocks-open-source-release-versions-matrix/ ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-01-20 05:03:56
|
TBB users: This message summarizes the activity and news related to Threading Building Blocks open source in the past week. It includes a listing of my blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you can click through to it using the links in this email. ------------ News: http://conferences.oreilly.com/money/ - Money:Tech Conference (Feb 6-7, New York, NY) - It's now confirmed that I will be attending the upcoming O'Reilly Money:Tech Conference. This conference looks very relevant to TBB, now that I've studied the schedule. Much of the discussion and presentation will involve mathematical models used for automated trading or for providing analysis to traders. Clearly the faster these models can be run, the greater the advantage for executing trades before your competitor notices the trend or pricing anomaly. I'll of course tell you all about the conference in a series of blogs. http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I will be attending OSCON in July and probably presenting one or two sessions (45 minutes each). One or both of the sessions will be technical. I'm working on my ideas now (the deadline for submitting proposals is Feb 4). One of the experiments I'm working on, as I plan an OSCON session, involves using Python and the Boost Python interface to create a Python/C++ application where a Python user interface controls a C++ computation library that is multithreaded using TBB. ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). This past week included the following posts: http://softwareblogs.intel.com/2008/01/14/upcoming-moneytech-conference-to-highlight-algorithmic-trading/ - "Upcoming Money:Tech Conference To Highlight Algorithmic Trading" (14 Jan 2008) - talked about the upcoming Money:Tech Conference, describing why I think it's relevant for TBB users; lists some of the key sessions that will present algorithms and computational methods for trading and financial market analysis. http://softwareblogs.intel.com/2008/01/16/open-source-threading-building-blocks-resources-update/ - "Open Source Threading Building Blocks Resources Update" - details some of the changes that have been put into place on the ThreadingBuildingBlocks.org site and related sites; discusses the new open source versions of the TBB documentation. http://softwareblogs.intel.com/2008/01/17/interesting-multicore-crisis-graph-and-analysis/ - "Interesting Multicore Crisis Graph and Analysis" - discusses an interesting blog post that includes a graph of historic CPU clock speeds, clearly pointing out the advent of the "multicore crisis". -------------------------- TBB Forum Activity: http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - brings you to the TBB Forums home page, where you will find the most recent activity immediately beneath the "Topics" heading. http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - This week, discussion of AJ Guillon's proposed parallel for_each construct continued In addition, there were questions and discussion about obtaining thread ID, applying TBB in various environments, and the activity of the scheduler in a particular pipeline application: http://softwarecommunity.intel.com/isn/Community/en-US/forums/thread/30247160.aspx --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 I continue to actively seek authors for Threading Building Blocks articles. Articles should be about TBB itself (a study of one or more components) or about an application that uses TBB. The articles should be about 2000 words long. The pay is $800 US. If you're interested in proposing an article, contact me and I'll send you my author's package, as well as answer any questions you may have about writing TBB articles. --------------------------- Polls: The current TBB poll is: The most important software development technology advance in 2008 will involve * Web development technologies/frameworks * Scalable server technologies * Desktop client/application technologies * Other * Don't know This poll will run through January. As always, I'm seeking help with coming up with the next poll. If you've got an idea for a poll related to TBB or scalable technologies, etc., please send it along! ---------------------------- Latest TBB Releases: This week saw a the first TBB release of 2008, the tbb20_20080115oss development release. The current lastest releases of TBB are: Commercial Aligned Release tbb20_014oss - http://threadingbuildingblocks.org/ver.php?fid=75 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20080115oss - http://threadingbuildingblocks.org/ver.php?fid=93 See my TBB Release Matrix post for basic information about the differences between the releases (does not yet include the tbb20_20071218oss or tbb20_20080115oss Development Releases): http://softwareblogs.intel.com/2007/12/07/threading-building-blocks-open-source-release-versions-matrix/ ------------------------------------ That's this week's TBB Weekly Summary post. If you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-01-12 17:57:51
|
TBB users: This is the first of a new mailing I'm planning to write each week, summarizing the activity and news related to Threading Building Blocks open source in the past week. I'll include a listing of my blog posts, articles, a summary of activity on the TBB forum, and anything else that seems noteworthy. This way, if you missed something that's of interest to you, you'll be able to click through to it using the links in this email. ------------ News: http://conferences.oreilly.com/money/ - Money:Tech Conference (Feb 6-7, New York, NY) - I've requested admission to the upcoming O'Reilly Money:Tech Conference. After looking at the schedule, I decided that this conference is very relevant to TBB. Much of the discussion and presentation will involve mathematical models used for automated trading or for providing analysis to traders. Clearly the faster these models can be run, the greater the advantage for executing trades before your competitor notices the trend or pricing anomaly. If I get my ticket to the conference, I'll of course tell all about it in a series of blogs. http://conferences.oreilly.com/oscon - OSCON (July 21-25, Portland, OR) - I will be attending OSCON in July and probably presenting one or two sessions (45 minutes each). One or both of the sessions will be technical. I'm working on my ideas now (the deadline for submitting proposals is Feb 4). ---------------------- Blog Posts: http://softwareblogs.intel.com/category/osstbb/ - This is the link to the TBB blog, with the latest entries at the top of the page (I include this link in case email formatting breaks the links for the individual posts). This past week included the following posts: http://softwareblogs.intel.com/2008/01/07/installing-threading-building-blocks-on-debian-etch/ - "Installing Threading Building Blocks on Debian Etch" (7 Jan 2008) - continues my series on bringing TBB onto my Debian 4.0 (Etch) machine using Roberto Sanchez's Etch TBB packages and Debian apt. http://softwareblogs.intel.com/2008/01/09/applying-threading-building-blocks-on-debian-etch/ - "Applying Threading Building Blocks on Debian Etch (9 Jan 2008) - talks about the steps that were necessary to get TBB operational on my Debian 4.0 (Etch) machine, and presents some results from running the TBB examples. My Debian machine is a single-core machine, but the processor has hyperthreading. TBB improved performance in most cases when the example applications were run with more than 1 thread. http://softwareblogs.intel.com/2008/01/11/threading-building-blocks-packaged-into-ubuntu-hardy-heron/ - "Threading Building Blocks Packaged into Ubuntu Hardy Heron" (11 Jan 2008) - announced the good news that the Debian TBB packages will be available in the Ubuntu Hardy Heron release (due in late April). Thanks go to Sadiq Jaffer for making the request, and obviously also to the Athena Capital Research team for creating the Debian packages. -------------------------- TBB Forum Activity: http://softwarecommunity.intel.com/isn/Community/en-US/forums/2471/ShowForum.aspx - on the TBB forum, there was substantial discussion of AJ Guillon's proposed parallel for_each construct. AJ also posted an entry about this on his YetiSim.org blog: http://blogs.yetisim.org/2008/01/09/parallel-for_each-algorithm-template/ --------------------------- Articles: TBB articles are published on the http://www.devx.com/go-parallel site. The most recent article is "The Challenges of Developing Multithreaded Processing Pipelines" by Ryan Bloom: http://www.devx.com/go-parallel/Article/36229 I continue to actively seek authors for Threading Building Blocks articles. Articles should be about TBB itself (a study of one or more components) or about an application that uses TBB. The articles should be about 2000 words long. The pay is $800 US. If you're interested in proposing an article, contact me and I'll send you my author's package, as well as answer any questions you may have about writing TBB articles. --------------------------- Polls: The current TBB poll is: The most important software development technology advance in 2008 will involve * Web development technologies/frameworks * Scalable server technologies * Desktop client/application technologies * Other * Don't know This poll will run through January. As always, I'm seeking help with coming up with the next poll. If you've got an idea for a poll related to TBB or scalable technologies, etc., please send it along! ---------------------------- Latest TBB Releases: Commercial Aligned Release tbb20_014oss - http://threadingbuildingblocks.org/ver.php?fid=75 Stable Release tbb20_20070927 - http://threadingbuildingblocks.org/ver.php?fid=84 Development Release tbb20_20071218oss - http://threadingbuildingblocks.org/ver.php?fid=86 See my TBB Release Matrix post for basic information about the differences between the releases (does not yet include the tbb20_20071218oss Development Release): http://softwareblogs.intel.com/2007/12/07/threading-building-blocks-open-source-release-versions-matrix/ ------------------------------------ So, that's my first TBB Weekly Summary post. Obviously, I'm hoping this is useful. If receiving this is terribly annoying, or if you have suggestions for other information I should include in the weekly post, let me know. Have a good week! Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |
|
From: Kevin F. <ke...@or...> - 2008-01-05 02:50:42
|
Hi, I've put a new poll onto the Threading Building Blocks front page -- http://www.ThreadingBuildingBlocks.org asking people to predict the area of software development technology that will see the most significant advance in 2008. The specific question/statement is: The most important software development technology advance in 2008 will involve and the choices are: * Web development technologies/frameworks * Scalable server technologies * Desktop client/application technologies * Other * Don't know If this sounds interesting, please post your vote! I'll be tallying and talking about the results around the end of January. Kevin -- Kevin Farnham O'Reilly Media, TBB Open Source Community http://www.ThreadingBuildingBlocks.org Mailing Lists: http://sourceforge.net/mail/?group_id=200923 FreeNode.net IRC Network: #tbb |