ildjit-all-ml Mailing List for ILDJIT
Brought to you by:
campanoni
Menu
▾
▴
ildjit-all-ml — This mailing list is attacched to ILDJIT project.
You can subscribe to this list here.
| 2008 |
Jan
|
Feb
|
Mar
(6) |
Apr
(3) |
May
(8) |
Jun
(17) |
Jul
(5) |
Aug
(10) |
Sep
(28) |
Oct
(33) |
Nov
(18) |
Dec
(22) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2009 |
Jan
(11) |
Feb
(9) |
Mar
(8) |
Apr
(3) |
May
(1) |
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2010 |
Jan
|
Feb
|
Mar
|
Apr
(2) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2012 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(2) |
Nov
|
Dec
|
|
From: Simone C. <sim...@gm...> - 2012-10-30 16:02:19
|
Hi, I have just released some codetools with some improvements. The documentation is also improved to describe how to recompile programs starting from a previously-generated code. Simone Campanoni |
|
From: Simone C. <xa...@ee...> - 2012-10-25 22:55:00
|
Hey folks, ILDJIT 1.0.0 is out there! The main changes compared to 0.8 are: - integration with LLVM! (yes, you read it! now ILDJIT uses LLVM as backend!) - several minor bugs have been fixed - internal methods are now part of the plugins that compose the CLR (i.e. clrplugins). - a significant extension to the API available for codetools Thanks Simone Campanoni |
|
From: Simone C. <xa...@ee...> - 2012-04-10 13:30:50
|
Hi all,
I am glad to announce the new release of ILDJIT 0.8.0 called
"Sloth-mania"!
There are several significant improvements in this release including:
- a tremendous extension to the public API that external tools can use
- a significant improvement to the documentation of the project
- a restructuring to the backend API
Thanks,
Simone Campanoni
|
|
From: Simone C. <xa...@ee...> - 2010-04-29 13:48:57
|
Hi all,
the new release "Sloth on trampolines" (0.3.0) is out!
Thanks to everyone for your contribution
Simone Campanoni
|
|
From: xan <cam...@el...> - 2010-04-02 19:22:23
|
Hi all,
we are planning to release a new version of ILDJIT compiler.
The new features are:
- support for CLI generics (Luca Rocchini)
- re-design of the interface of libiljitmm (Luca Rocchini)
- Bugs fixing within the Reflection module (Luca Rocchini)
- Bug fixing of Libjit code generatore (Michele Tartara and me)
- Various bugs fixing (everyone)
- Forwarding of CLI types to every IR variable (me)
- CFG with a single exit node (Michele Tartara)
- Introduction of IRLIBRARYCALL and IRNATIVECALL instead of IRNCALL
- New memory alias analyzer (Luca Rocchini and me)
We would like to have a name for that release, so go ahead, we want to
hear your proposals (please send to this mailing list).
Moreover, we have opened a request to Ubuntu distribution in order to
include ILDJIT within the next release (Lucid); if you want it too, go
to the website https://bugs.launchpad.net/ubuntu/+bug/553445 and click
on "Nominate for release
<https://bugs.launchpad.net/ubuntu/+bug/553445/+nominate>"
I would like to thank everyone has contributed to the ILDJIT project.
Sincerely,
Simone Campanoni
|
|
From: Kirill K. <kir...@gm...> - 2009-06-12 18:46:41
|
Hello everyone I have updated both the web site of libJIT Linear Scan (www.libjit.com) and its documentation. 1. Benchmarks: http://libjit-linear-scan-register-allocator.googlecode.com/files/benchmarks_of_libJIT_Linear_Scan_Register_Allocator_0.1.2.5%2B.pdf 2. Diagrams and figures: http://libjit-linear-scan-register-allocator.googlecode.com/files/schemes_of_libJIT_Linear_Scan_Register_Allocator_0.1.2.5%2B.pdf 3. Source code package. Last useful code, and changes for ARM architecture from the repository have been applied to libJIT Linear Scan: http://libjit-linear-scan-register-allocator.googlecode.com/files/libjit-0.1.2.5%2B.tar.gz Many other papers are at http://www.libjit.com/downloads/list (most of them are in Russian language for this time). Thanks, Kirill |
|
From: xan <cam...@el...> - 2009-05-05 17:21:03
|
Hi all,
the new policy has been introduced for the regression test within
the CVS server: every not running program has the suffix ".exe.TODO" and
every running program has the suffix ".exe".
For this reason, when you check your commit running the regression test,
you should see that every program runs correctly.
Simone
|
|
From: xan <cam...@el...> - 2009-04-27 14:07:08
|
Hi all,
currently the regression test contains 997 programs where 967 works
correctly.
Here is the output you should expect:
Regression test without optimization algorithm
Working progress: 73/997
testClrType.exe
Test ./Running/ILDJIT/InternalMethods/Reflection/testClrType.exe
is failed
Working progress: 110/997
ReflectiontestMethod.exe
Test ./Running/ILDJIT/Reflection/ReflectiontestMethod.exe is
failed
Working progress: 151/997
JGFSparseMatmultModified.exe
Test ./Running/JGrande/section2/3/JGFSparseMatmultModified.exe
is failed
Working progress: 162/997
bitcnts_small.exe
Test ./Running/MiBench/bitcnts_small.exe is
failed
Working progress: 163/997
qsort_small.exe
Test ./Running/MiBench/qsort_small.exe is
failed
Working progress: 200/997
test-126.exe
Test ./Running/Mono/mcs/tests/test-126.exe is
failed
Working progress: 201/997
test-126_ext.exe
Test ./Running/Mono/mcs/tests/test-126_ext.exe is
failed
Working progress: 203/997
test-12.exe
Test ./Running/Mono/mcs/tests/test-12.exe is
failed
Working progress: 220/997
test-148.exe
Test ./Running/Mono/mcs/tests/test-148.exe is
failed
Working progress: 284/997
test-220.exe
Test ./Running/Mono/mcs/tests/test-220.exe is
failed
Working progress: 506/997
test-525.exe
Test ./Running/Mono/mcs/tests/test-525.exe is
failed
Working progress: 800/997 ctor-bench.exe
Test ./Running/Mono/mono/mono/benchmark/ctor-bench.exe is failed
Working progress: 801/997 divun.exe
Test ./Running/Mono/mono/mono/benchmark/divun.exe is failed
Working progress: 814/997 life.exe
Test ./Running/Mono/mono/mono/benchmark/life.exe is failed
Working progress: 822/997 pinvoke.exe
Test ./Running/Mono/mono/mono/benchmark/pinvoke.exe is failed
Working progress: 823/997 readonly-byte-array.exe
Test ./Running/Mono/mono/mono/benchmark/readonly-byte-array.exe
is failed
Working progress: 831/997 sbperf1.exe
Test ./Running/Mono/mono/mono/benchmark/sbperf1.exe is failed
Working progress: 832/997 sbperf2.exe
Test ./Running/Mono/mono/mono/benchmark/sbperf2.exe is failed
Working progress: 837/997 valuetype-hash-equals.exe
Test
./Running/Mono/mono/mono/benchmark/valuetype-hash-equals.exe is failed
Working progress: 846/997 arraylist-clone.exe
Test ./Running/Mono/mono/mono/tests/arraylist-clone.exe is failed
Working progress: 853/997 box.exe
Test ./Running/Mono/mono/mono/tests/box.exe is failed
Working progress: 885/997 intptrcast.exe
Test ./Running/Mono/mono/mono/tests/intptrcast.exe is failed
Working progress: 899/997 mutexes.exe
Test ./Running/Mono/mono/mono/tests/mutexes.exe is failed
Working progress: 925/997 test-12.exe
Test ./Running/Mono/mono/mono/tests/test-12.exe is failed
Working progress: 935/997 thread5.exe
Test ./Running/Mono/mono/mono/tests/thread5.exe is failed
Working progress: 936/997 thread-dump.exe
Test ./Running/Mono/mono/mono/tests/thread-dump.exe is failed
Working progress: 940/997 time.exe
Test ./Running/Mono/mono/mono/tests/time.exe is failed
Working progress: 945/997 vtype.exe
Test ./Running/Mono/mono/mono/tests/vtype.exe is failed
Working progress: 986/997 clock.exe
Test ./Running/TestC/clock.exe is failed
Working progress: 988/997 file.exe
Test ./Running/TestC/file.exe is failed
Working progress: 997/997 Sha1.exe
Tests succesfully run: 967/997
Some test is failed
Simone Campanoni
|
|
From: Massimiliano G. <dar...@gm...> - 2009-04-25 12:52:50
|
hi, modules libiljitu, libiljitmm and iljit has been updated to fix some bug into Reflection. Changes are committed in CVS max |
|
From: Kirill K. <kir...@gm...> - 2009-04-04 13:15:22
|
0.1.2.5 + / 0.1.2 1/2 version release (code name: "libJIT-ON-TESTOSTERONE") * main branch + libJIT-linear-scan-register-allocator * Add optimization levels for IA-32 from 0 to 4 (jit_set_optimization_level) * Add a new specialized jit_abi_internal * Add brand new optimized object code generator (level 1, 2, 3, 4 of optimization) * Various low-level machine dependent optimizations and tricks * Aggressive optimization of division by integer constants as by Torbjorn Granlund and Peter L. Montgomery in "Division By Invariant Integers using Multiplication" * Add primitive code generators for MMX/Streaming SIMD Extensions/SSE/SSE2/SSE3 and others * Use SIMD SSE/SSE3/SSE3 for floating point values and operations (level 1, 2, 3, 4 of optimization) * Data-flow and control-flow based analysis (level 1, 2, 3, 4 of optimization) * Fast liveness analysis (level 2 of optimization) * Dead-code elimination (level 4 of optimization) * Full liveness analysis (level 3, 4 of optimization) * Linear scan register allocator algorithm (level 2) * Bin packing register allocator algorithm (level 3, 4) * Tested on DotGNU Portable.NET Common Language Runtime / a Microsoft Common Intermediate Language Virtual Machine (a special "unofficial" research release version <http://code.google.com/p/libjit-linear-scan-register-allocator> <jit-cil.net>) Download is at: http://libjit-linear-scan-register-allocator.googlecode.com/files/libjit-0.1.2.5%2B.tar.gz Thanks, Kirill |
|
From: Max <man...@gm...> - 2009-03-30 18:18:49
|
Hi, iljit has been updated about vararg support and in order to fix a bug about String. Changes have been commited to CVS Max. |
|
From: Kirill K. <kir...@gm...> - 2009-03-28 08:34:15
|
Hello Everyone I have registered a libjit.com and jit-cil.net domain with pointers to my experimental code there at the moment. Right now there is not much in plan of documentation, but it will get much more information latter with everyones help. Please feel free to give your ideas or suggestions. Thanks, Kirill |
|
From: Massimiliano G. <dar...@gm...> - 2009-03-27 19:31:51
|
hello,
update for the implementation of the method
GetManifestResourceStream. The changes are in the packages: libiljitu,
libiljitmm, iljit
max
|
|
From: xan <cam...@el...> - 2009-03-26 23:09:22
|
Hi,
ILDJIT 0.0.3 has been released.
I want to thank to everybody has made contributions on this project
during the entire last year.
Simone Campanoni
|
|
From: Kirill K. <kir...@gm...> - 2009-03-26 15:08:26
|
Hello I explore various register allocation techniques in libJIT, and will add support to experiment with them from ILDJIT. Let me know your ideas about this. Thanks, Kirill |
|
From: xan <cam...@el...> - 2009-03-13 07:40:26
|
Hi all,
CVS is now working :-)
Simone
|
|
From: xan <cam...@el...> - 2009-03-12 08:37:09
|
Hi all,
cvs is down for every project on sourceforge that has the name
starting with "i" :-(
Basically, they said that the CVS server will be re-start tomorrow.
Simone
|
|
From: Kirill K. <kir...@gm...> - 2009-03-05 13:48:16
|
Hi, I have read on the website, that you have the following targets: Replacing the current code generation and register allocation library, Libjit, with a more flexible and optimizing infrastructure based on llvm or with a brand new one; Extending support to other high level languages (C, Python); Extending the level of optimization available in ILDJIT; You can try my code at http://code.google.com/p/libjit-linear-scan-register-allocator/. It has more optimizations done, for example try set_optimization_level(1) and set_optimization_level(2). The first one uses a linear-scan register allocator, the second one a bin packing register allocator, and dead-code elimination. It should be very easy to extend this code for any other optimization you want. For example, I enabled dead-code elimination in a matter of a week with 40 lines of codes, which has given 70% of improvement in PNetMark benchmark, for Logic. Thanks, Kirill |
|
From: xan <cam...@el...> - 2009-02-28 15:00:32
|
Hi,
a new package has been inserted inside the ILDJIT project:
libcompilermemorymanager.
This package manages the dynamic memory needed by the compiler and not
by the running application.
iljit has been upgraded in order to link with the new package.
Changes are in CVS.
Simone
|
|
From: Stefano A. <ste...@gm...> - 2009-02-27 11:22:06
|
Hi, modules iljit and irprofiler has been updated in order to perform method specialization. Changes have been commited to CVS Stefano |
|
From: xan <cam...@el...> - 2009-02-18 11:26:18
|
Hi,
iljit, libiljitmm have been upgraded in order to fix a bug inside
the CIL->IR translator for the call sites where the return value is a
TypedReference.
Moreover, they have been upgraded in order to fix a bug inside the
Decimal internal methods.
Some new tests are added to the regression test Tester.
Changes are in CVS.
Simone
|
|
From: xan <cam...@el...> - 2009-02-17 15:27:44
|
Ciao a tutti, vi ricordo il seguente seminario che ci sara' domani. ========================================================================================================================================================== Industrial perspectives on multiprocessor systems on chip In this seminar, the research and development activities of the Advanced Computing group at STMicroelectronics will be presented, with a focus on single chip multiprocessors and the required system software to support such platforms. Key STMicroelectronics personnel will present an embedded, massively parallel multiprocessor architecture aimed at improving manufacturability for complex System on Chips at 32nm (and beyond) for the consumer market. The key challenges of the HW/SW architecture, programming tools and platform modeling will be discussed. About the speakers Eric Flamand is a Director of Advanced Computing at STMicroelectronics. Diego Melpignano is Program Manager at STMicroelectronics, leading R&D activities in the embedded computing domain. Eric Flamand and Diego Melpignano STMicroelectronics 18/02/2009 14:30 Politecnico di Milano E.G.6 |
|
From: xan <cam...@el...> - 2009-02-10 09:07:01
|
Hi,
iljit, libiljitu and libiljitir have been upgraded
Changes are in CVS.
Simone
|
|
From: xan <cam...@el...> - 2009-02-08 12:31:50
|
Hi,
iljit has been upgraded in order to fix a bug on the reflection
module as well as improving the quality of the assembly generated.
Changes are in CVS.
SImone
|
|
From: xan <cam...@el...> - 2009-02-07 10:32:00
|
Hi,
new releases of iljit, libiljit* are available on sourceforge.
The new ILDJIT supports more C# and C programs (833 of the regression
test) as well as it is able to run faster than before.
Changes are in CVS.
Simone
|
