Author: rafael
Date: Mon Aug 27 05:18:20 2007
New Revision: 41479
URL: http://llvm.org/viewvc/llvm-project?rev=41479&view=rev
Log:
call libc memcpy/memset if array size is bigger then threshold.
Coping 100MB array (after a warmup) shows that glibc 2.6.1 implementation on
x86-64 (core 2) is 30% fa
Author: baldrick
Date: Mon Aug 27 07:53:48 2007
New Revision: 41480
URL: http://llvm.org/viewvc/llvm-project?rev=41480&view=rev
Log:
Turn on exception handling code generation.
Modified:
llvm-gcc-4.0/trunk/gcc/llvm-backend.cpp
Modified: llvm-gcc-4.0/trunk/gcc/llvm-backend.cpp
URL:
http://ll
Author: baldrick
Date: Mon Aug 27 07:56:28 2007
New Revision: 41481
URL: http://llvm.org/viewvc/llvm-project?rev=41481&view=rev
Log:
Turn on exception handling codegen.
Modified:
llvm-gcc-4.2/trunk/gcc/llvm-backend.cpp
Modified: llvm-gcc-4.2/trunk/gcc/llvm-backend.cpp
URL:
http://llvm.org/v
On Aug 27, 2007, at 3:18 AM, Rafael Espindola wrote:
> Coping 100MB array (after a warmup) shows that glibc 2.6.1
> implementation on
> x86-64 (core 2) is 30% faster (from 0.270917s to 0.188079s)
Please record this information in a comment? It is likely other
people will want to revise this
Author: djg
Date: Mon Aug 27 09:50:10 2007
New Revision: 41482
URL: http://llvm.org/viewvc/llvm-project?rev=41482&view=rev
Log:
Add explicit keywords and remove spurious trailing semicolons.
Modified:
llvm/trunk/include/llvm/ADT/FoldingSet.h
llvm/trunk/include/llvm/CodeGen/LiveInterval.h
Author: baldrick
Date: Mon Aug 27 10:47:50 2007
New Revision: 41484
URL: http://llvm.org/viewvc/llvm-project?rev=41484&view=rev
Log:
There is an impedance matching problem between LLVM and
gcc exception handling: if an exception unwinds through
an invoke, then execution must branch to the invoke's
Author: baldrick
Date: Mon Aug 27 10:51:13 2007
New Revision: 41485
URL: http://llvm.org/viewvc/llvm-project?rev=41485&view=rev
Log:
Filters are now specified by using the number of type infos
plus one. Always append a catch-all to the selector call,
unless it is pointless.
Modified:
llvm-gc
Author: baldrick
Date: Mon Aug 27 11:08:37 2007
New Revision: 41486
URL: http://llvm.org/viewvc/llvm-project?rev=41486&view=rev
Log:
Fix a mismatch between can_throw_external_1 (CTE) and
foreach_reachable_handler (FRH): CTE should visit the
same handlers as FRE (indeed it could have been implement
Author: baldrick
Date: Mon Aug 27 11:11:09 2007
New Revision: 41487
URL: http://llvm.org/viewvc/llvm-project?rev=41487&view=rev
Log:
Filters are now specified using the number of type infos
plus one. If an eh.selector call isn't guaranteed to
match, append a catch-all. Also, some trivial cleanup
Author: djg
Date: Mon Aug 27 11:11:48 2007
New Revision: 41488
URL: http://llvm.org/viewvc/llvm-project?rev=41488&view=rev
Log:
Change comments to refer to @malloc and @free instead of %malloc and %free.
Modified:
llvm/trunk/lib/Transforms/IPO/RaiseAllocations.cpp
Modified: llvm/trunk/lib/Tr
Author: djg
Date: Mon Aug 27 11:26:13 2007
New Revision: 41489
URL: http://llvm.org/viewvc/llvm-project?rev=41489&view=rev
Log:
If the source and destination pointers in an llvm.memmove are known
to not alias each other, it can be translated as an llvm.memcpy.
Added:
llvm/trunk/test/CodeGen/X
Author: baldrick
Date: Mon Aug 27 11:30:05 2007
New Revision: 41490
URL: http://llvm.org/viewvc/llvm-project?rev=41490&view=rev
Log:
Now that we don't output cleanups by default, the action
offset needs to be adjusted in this test.
Modified:
llvm/trunk/test/CodeGen/Generic/2007-06-06-Critical
Author: djg
Date: Mon Aug 27 11:32:11 2007
New Revision: 41491
URL: http://llvm.org/viewvc/llvm-project?rev=41491&view=rev
Log:
Make DAGCombiner's global alias analysis query more precise in the case
where both pointers have non-zero offsets.
Modified:
llvm/trunk/lib/CodeGen/SelectionDAG/DAGC
On Aug 27, 2007, at 9:26 AM, Dan Gohman wrote:
> Author: djg
> Date: Mon Aug 27 11:26:13 2007
> New Revision: 41489
>
> URL: http://llvm.org/viewvc/llvm-project?rev=41489&view=rev
> Log:
> If the source and destination pointers in an llvm.memmove are known
> to not alias each other, it can be tra
Author: baldrick
Date: Mon Aug 27 11:47:32 2007
New Revision: 41494
URL: http://llvm.org/viewvc/llvm-project?rev=41494&view=rev
Log:
Fix this test: llvm-gcc-4.2 optimizes almost everything away,
resulting in zero matches, while llvm-gcc-4.1 manages to
remove one pointless selector changing the num
> This is nifty, but shouldn't this be done at the LLVM IR level? I
> can't think of cases where lowering would create new memmove calls.
I'm not aware of any cases where memmove vs. memcpy matters before
codegen currently. If it were done at the LLVM IR level, would it be
too trivial to be in
Author: rafael
Date: Mon Aug 27 12:48:26 2007
New Revision: 41502
URL: http://llvm.org/viewvc/llvm-project?rev=41502&view=rev
Log:
Add a comment about using libc memset/memcpy or generating inline code.
Modified:
llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
Modified: llvm/trunk/lib/Target/X
On Aug 17, 2007, at 5:28 PM, Raul Fernandes Herbster wrote:
2007/8/17, Evan Cheng <[EMAIL PROTECTED]>:
Very good progress. Thanks!
Comments inline.
Evan
...
Please use abort() instead so it does what's expected in non-debug
build.
OK
...
i s -> is :-)
OK
Also, why the name "ARMCom
Author: asl
Date: Mon Aug 27 13:44:54 2007
New Revision: 41504
URL: http://llvm.org/viewvc/llvm-project?rev=41504&view=rev
Log:
Teach gfortran about LLVM options
Modified:
llvm-gcc-4.2/trunk/gcc/fortran/lang-specs.h
Modified: llvm-gcc-4.2/trunk/gcc/fortran/lang-specs.h
URL:
http://llvm.org/
Author: greened
Date: Mon Aug 27 14:04:21 2007
New Revision: 41506
URL: http://llvm.org/viewvc/llvm-project?rev=41506&view=rev
Log:
Update InvokeInst to work like CallInst
Modified:
llvm/trunk/include/llvm/Instructions.h
llvm/trunk/include/llvm/Support/LLVMBuilder.h
llvm/trunk/lib/As
Author: greened
Date: Mon Aug 27 14:04:17 2007
New Revision: 41505
URL: http://llvm.org/viewvc/llvm-project?rev=41505&view=rev
Log:
Update for new InvokeInst constructor
Modified:
llvm-gcc-4.0/trunk/gcc/llvm-convert.cpp
Modified: llvm-gcc-4.0/trunk/gcc/llvm-convert.cpp
URL:
http://llvm.org
Author: greened
Date: Mon Aug 27 14:05:22 2007
New Revision: 41507
URL: http://llvm.org/viewvc/llvm-project?rev=41507&view=rev
Log:
Update for new InvokeInst constructor
Modified:
llvm-gcc-4.2/trunk/gcc/llvm-convert.cpp
Modified: llvm-gcc-4.2/trunk/gcc/llvm-convert.cpp
URL:
http://llvm.org
Hi Bruno,
Sparc backend also has a delay slot filler that does exactly the same
thing. Can you change this into a target independent pass instead
(then you just need to add a target hook to emit nop)?
Evan
On Aug 17, 2007, at 6:50 PM, Bruno Cardoso Lopes wrote:
> Author: bruno
> Date: Fri A
Author: baldrick
Date: Mon Aug 27 14:58:09 2007
New Revision: 41509
URL: http://llvm.org/viewvc/llvm-project?rev=41509&view=rev
Log:
After further thought and testing, it seems that
it is foreach_reachable_handler that should be
modified to agree with can_throw_external_1, and
not the other way ro
Dan,
The memmove tests you added are failing on x86 darwin.
Here is the output from one of the tests:
lattner% llvm-as < test/CodeGen/X86/memmove-0.ll | llc -march=x86
.text
.align 4,0x90
.globl _foo
_foo:
subl$12, %esp
movl20(%esp), %eax
Rafael,
test/CodeGen/X86/2004-02-12-Memcpy.llx is failing on x86 darwin:
Here is the output in case it helps:
lattner% llvm-as < test/CodeGen/X86/2004-02-12-Memcpy.llx | llc -
march=x86
.text
.align 4,0x90
.globl _main
_main:
pushl %edi
pushl
Author: djg
Date: Mon Aug 27 15:54:48 2007
New Revision: 41510
URL: http://llvm.org/viewvc/llvm-project?rev=41510&view=rev
Log:
Add explicit triples to avoid default behavior that varies by host.
Modified:
llvm/trunk/test/CodeGen/X86/memmove-0.ll
llvm/trunk/test/CodeGen/X86/memmove-1.ll
Author: dpatel
Date: Mon Aug 27 16:34:31 2007
New Revision: 41516
URL: http://llvm.org/viewvc/llvm-project?rev=41516&view=rev
Log:
Use simpler test to filter loops.
Modified:
llvm/trunk/lib/Transforms/Scalar/LoopIndexSplit.cpp
Modified: llvm/trunk/lib/Transforms/Scalar/LoopIndexSplit.cpp
URL
Author: void
Date: Mon Aug 27 16:53:31 2007
New Revision: 41519
URL: http://llvm.org/viewvc/llvm-project?rev=41519&view=rev
Log:
Creating llvm-1200-10 branch
Added:
llvm/tags/Apple/llvm-1200-10/
- copied from r41518, llvm/trunk/
___
llvm-comm
Author: void
Date: Mon Aug 27 16:53:25 2007
New Revision: 41518
URL: http://llvm.org/viewvc/llvm-project?rev=41518&view=rev
Log:
Creating llvm-gcc-1200-10 branch
Added:
llvm-gcc-4.0/tags/llvm-gcc-1200-10/
- copied from r41517, llvm-gcc-4.0/trunk/
___
Hi Duncan,
Could you back this patch out please? It's breaking builds on Darwin.
-bw
On 8/27/07, Duncan Sands <[EMAIL PROTECTED]> wrote:
> Author: baldrick
> Date: Mon Aug 27 07:56:28 2007
> New Revision: 41481
>
> URL: http://llvm.org/viewvc/llvm-project?rev=41481&view=rev
> Log:
> Turn on exce
On Aug 27, 2007, at 10:46 AM, Dan Gohman wrote:
>> This is nifty, but shouldn't this be done at the LLVM IR level? I
>> can't think of cases where lowering would create new memmove calls.
>
> I'm not aware of any cases where memmove vs. memcpy matters before
> codegen currently. If it were done
Thank you. I'm going to be working on it this week and hope to have
the failure fixed soon.
-bw
On Aug 27, 2007, at 5:53 AM, Duncan Sands wrote:
> Author: baldrick
> Date: Mon Aug 27 07:53:48 2007
> New Revision: 41480
>
> URL: http://llvm.org/viewvc/llvm-project?rev=41480&view=rev
> Log:
> Tu
On Aug 24, 2007, at 6:31 AM, Rafael Espindola wrote:
> According to table 3.1 of http://www.x86-64.org/documentation/abi.pdf,
> the i64 and f64 types should have 8 byte alignment.
>
> The attached patch changes this for targets that are not darwin.
Dan, Evan, is this ok? Evan, doesn't darwin foll
Author: lattner
Date: Mon Aug 27 23:23:55 2007
New Revision: 41522
URL: http://llvm.org/viewvc/llvm-project?rev=41522&view=rev
Log:
Cut off crazy computation. This helps PR1622 slightly.
Modified:
llvm/trunk/lib/Transforms/Scalar/InstructionCombining.cpp
Modified: llvm/trunk/lib/Transforms/
Author: bruno
Date: Tue Aug 28 00:04:41 2007
New Revision: 41525
URL: http://llvm.org/viewvc/llvm-project?rev=41525&view=rev
Log:
Added methods to record SPOffsets from LowerFORMAL_ARGUMENTS
Modified:
llvm/trunk/lib/Target/Mips/MipsMachineFunction.h
Modified: llvm/trunk/lib/Target/Mips/MipsM
Author: bruno
Date: Tue Aug 28 00:08:16 2007
New Revision: 41527
URL: http://llvm.org/viewvc/llvm-project?rev=41527&view=rev
Log:
Changed stack allocation On LowerFORMAL_ARGUMENTS.
Added comments about new stack allocation.
Expand SelectCC for i32 results
Modified:
llvm/trunk/lib/Target/Mips/
Author: bruno
Date: Tue Aug 28 00:13:42 2007
New Revision: 41529
URL: http://llvm.org/viewvc/llvm-project?rev=41529&view=rev
Log:
Added method to get Mips register numbers
Changed the stack frame layout, StackGrowsUp fits better to Mips strange stack.
Stack offset calculation bug fixed!
Modified:
Changes in directory llvm-www/pubs:
index.html updated: 1.57 -> 1.58
---
Log message:
add a paper
---
Diffs of the changes: (+5 -0)
index.html |5 +
1 files changed, 5 insertions(+)
Index: llvm-www/pubs/index.html
diff -u llvm-www/pubs/index.html:1.57 llvm-www/pubs/index.html:1.
Changes in directory llvm-www/pubs:
2007-08-16-TRANSACT-Tanger-Slides.pdf added (r1.1)
2007-08-16-TRANSACT-Tanger.html added (r1.1)
2007-08-16-TRANSACT-Tanger.pdf added (r1.1)
---
Log message:
Add Torvald's STM paper.
---
Diffs of the changes: (+60 -0)
2007-08-16-TRANSACT-Tanger-Slides.pdf
Author: bruno
Date: Tue Aug 28 00:06:17 2007
New Revision: 41526
URL: http://llvm.org/viewvc/llvm-project?rev=41526&view=rev
Log:
Mask directive completed with CalleeSave info
Comments for Mips directives added.
Modified:
llvm/trunk/lib/Target/Mips/MipsAsmPrinter.cpp
llvm/trunk/lib/Target
Author: baldrick
Date: Tue Aug 28 01:36:41 2007
New Revision: 41534
URL: http://llvm.org/viewvc/llvm-project?rev=41534&view=rev
Log:
Turn off eh codegen because it breaks the x86
darwin build.
Modified:
llvm-gcc-4.0/trunk/gcc/llvm-backend.cpp
Modified: llvm-gcc-4.0/trunk/gcc/llvm-backend.cpp
Author: baldrick
Date: Tue Aug 28 01:40:16 2007
New Revision: 41535
URL: http://llvm.org/viewvc/llvm-project?rev=41535&view=rev
Log:
Turn off eh codegen because it breaks the x86
darwin build.
Modified:
llvm-gcc-4.2/trunk/gcc/llvm-backend.cpp
Modified: llvm-gcc-4.2/trunk/gcc/llvm-backend.cpp
43 matches
Mail list logo
