dipc-users Mailing List for Distributed Inter-Process Comm
Brought to you by:
nikatjef
You can subscribe to this list here.
2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(3) |
Nov
|
Dec
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
2006 |
Jan
|
Feb
(1) |
Mar
(2) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2008 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(1) |
Dec
|
From: Kamran K. <kam...@ho...> - 2006-03-13 21:35:11
|
Hello all, Distributed semaphores, messages, and shared memory segments now work on a network. You can download the DIPC 2.1 Beta 1 package from http://cs.uwindsor.ca/~kamran/downloads.html and run the dipc.install script. Please let me know of any problems. Note: Only the i386 code is in beta state. SPARC, Alpha, m68k, and ARM modifications should be considered alpha code. I have partially modified MIPS, PPC, and X86_64, and I expect them to need more work. Please let me know if you tried DIPC on any of these architectures. Thanks. -Kamran |
From: Kamran K. <kam...@ho...> - 2006-03-12 05:35:11
|
Hello all, System V messages and semaphores in DIPC (Distributed IPC) Alpha 4 now work over a network. The shared memory seems to need more work. You can download the latest version from http://cs.uwindsor.ca/~kamran/downloads.html Thanks. -Kamran |
From: Kamran K. <kam...@ho...> - 2006-02-18 22:18:08
|
Hello everybody, DIPC (Distributed IPC) is a system software that provides distributed access to messages, semaphores, and shared memory segments. It can be used to transparently exchange data between distributed applications. As another step in porting DIPC to Linux 2.6.x, the example programmes in 2.1-alpha3 version of DIPC now work on a single machine (they haven't been tried over a network). Please let me know of your experience if you installed and tested DIPC, especially in a distributed environment. You can download this version of DIPC from: http://cs.uwindsor.ca/~kamran/downloads.html -Kamran |
From: James E. W. <nik...@us...> - 2000-10-04 13:33:29
|
Whooops, I forgot to add the link for the second document http://sourceforge.net/docman/display_doc.php?docid=1019&group_id=10736 Thanx, James |
From: James E. W. <nik...@us...> - 2000-10-04 09:22:16
|
Greetings, I have posted the final information for the first part of the benchmark results on SourceForge. To view the information, please go to the following URL; http://sourceforge.net/docman/display_doc.php?docid=1013&group_id=10736 I will also be sending this HTML page to Tim to add to the DIPC page at http://wallybox.cei.net There is some very very interesting information in the numbers I found. :) Thanx, James |
From: James E. W. <nik...@us...> - 2000-10-03 10:50:07
|
Greetings, Over the last two days, I have been running a series of benchmark tests against my primary workstation. I am running the sysbench and shmbench benchmarks. Before I explain what I am seeing, I should list the hardware I am using. TEST MACHINE IBM Thinkpad 600E PII 633Mhz 128M Ram Linux 2.2.17 Intel EtherExpress/Modem 10/100 card DIPC Referee Machine eTower 266 Cyrix MII 266Mhz 128M Ram Linux 2.2.17 SMC EtherElite 16 The tests consist of 10 runs of each of the following modes; Referee is Test Machine sysbench & syscall compiled with support for DIPC sysbench & syscall compiled without support for DIPC Referee is Remote sysbench & syscall compiled with support for DIPC sysbench & syscall compiled without support for DIPC I am taking the average in each category for all 10 runs in each mode and comparing them. So far, from the data that I have cooked, I am finding (Overall) that an application compiled with DIPC support actually runs faster than the same application with IPC, but not DIPC support. The break down is as follows; Local machine as Referee, App is DIPC is the best Remote machine as Referee, App is DIPC is second best Remote machine as Referee, App is !DIPC is third Local machine as Referee, App is !DIPC is last Next step, after I finish cooking the current data, will be to rebuild my kernel without DIPC support and running a set of baseline tests and merging that data in as well. Once I have finished cooking the data, I will release *ALL* of my results and the raw data that I use. Right off the bat, this looks like some very very exciting information, if my assessment is correct, I think I have found a way to convince Linus Torvalds to incorporate DIPC as an integral part of the Linux kernel, further, I will have found the numbers that will be needed to get some time at OSDLab for some very high power testing. Thanx, James |