RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2025-02-19 03:38:26 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
87,805 Commits 6 Branches 6 Tags
Commit Graph

58258 Commits

Author SHA1 Message Date
Nadav Rotem
309d628c4f LoopVectorizer: Fix an endless loop in the code that looks for reductions. 
The bug was in the code that detects PHIs in if-then-else block sequence.

PR14701.

llvm-svn: 171008
 2012-12-24 01:22:06 +00:00
Nadav Rotem
fb56b5fe2e CostModel: Change the default target-independent implementation for finding 
the cost of arithmetic functions. We now assume that the cost of arithmetic
operations that are marked as Legal or Promote is low, but ops that are
marked as custom are higher.

llvm-svn: 171002
 2012-12-23 17:31:23 +00:00
Benjamin Kramer
1add5cc571 LoopVectorize: Fix accidentaly inverted condition. 
llvm-svn: 171001
 2012-12-23 13:21:41 +00:00
Benjamin Kramer
d0a5bcaf53 LoopVectorize: For scalars and void types there is no need to compute vector insert/extract costs. 
Fixes an assert during the build of oggenc in the test suite.

llvm-svn: 171000
 2012-12-23 13:19:18 +00:00
Nadav Rotem
db133d1c52 whitespace 
llvm-svn: 170997
 2012-12-23 07:33:44 +00:00
Nadav Rotem
9942930d57 Rename a function. 
llvm-svn: 170996
 2012-12-23 07:30:09 +00:00
Nadav Rotem
e237376e62 Loop Vectorizer: Update the cost model of scatter/gather operations and make 
them more expensive.

llvm-svn: 170995
 2012-12-23 07:23:55 +00:00
Craig Topper
05145cd465 Remove trailing whitespace. 
llvm-svn: 170991
 2012-12-22 19:15:35 +00:00
Craig Topper
0855d67e26 Remove trailing whitespace 
llvm-svn: 170990
 2012-12-22 18:09:02 +00:00
Jakob Stoklund Olesen
a610e92cf4 Remove a special case that doesn't seem necessary any longer. 
Back when this exception was added, it was skipping a lot more code, but
now it just looks like a premature optimization.

llvm-svn: 170989
 2012-12-22 17:33:22 +00:00
Jakob Stoklund Olesen
00d07e608c Use getNumOperands() instead of Operands.size(). 
The representation of the Operands array is going to change soon so it
can be allocated from a BumpPtrAllocator.

llvm-svn: 170988
 2012-12-22 17:13:06 +00:00
Benjamin Kramer
52d36d0ccd X86: Turn mul of <4 x i32> into pmuludq when no SSE4.1 is available. 
pmuludq is slow, but it turns out that all the unpacking and packing of the
scalarized mul is even slower. 10% speedup on loop-vectorized paq8p.

llvm-svn: 170985
 2012-12-22 16:07:56 +00:00
Benjamin Kramer
a9cc52a6b3 X86: Emit vector sext as shuffle + sra if vpmovsx is not available. 
Also loosen the SSSE3 dependency a bit, expanded pshufb + psra is still better
than scalarized loads. Fixes PR14590.

llvm-svn: 170984
 2012-12-22 11:34:28 +00:00
Bill Wendling
a16e5519db Change 'AttrVal' to 'AttrKind' to better reflect that it's a kind of attribute instead of the value of the attribute. 
llvm-svn: 170972
 2012-12-22 00:37:52 +00:00
Richard Smith
80529530ef Don't call back() on an empty SmallVector. Found by -fsanitize=enum! 
llvm-svn: 170968
 2012-12-22 00:15:13 +00:00
Nadav Rotem
df77ac1ba8 In some cases, due to scheduling constraints we copy the EFLAGS. 
The only way to read the eflags is using push and pop. If we don't
adjust the stack then we run over the first frame index. This is
not something that we want to do, so we have to make sure that
our machine function does not copy the flags. If it does then
we have to emit the prolog that adjusts the stack.

rdar://12896831

llvm-svn: 170961
 2012-12-21 23:48:49 +00:00
Akira Hatanaka
6d30f89eeb [mips] Refactor subword-swap, EXT/INS, load-effective-address and read-hardware 
instructions.

llvm-svn: 170956
 2012-12-21 23:21:32 +00:00
Akira Hatanaka
b41ed048b6 [mips] Refactor SYNC and multiply/divide instructions. 
llvm-svn: 170955
 2012-12-21 23:17:36 +00:00
Akira Hatanaka
6b9b8a5489 [mips] Refactor BAL instructions. 
llvm-svn: 170954
 2012-12-21 23:15:59 +00:00
Akira Hatanaka
4b2123dd9c [mips] Fix encoding of BAL instruction. Also, fix assembler test case which 
was not catching the error.

llvm-svn: 170953
 2012-12-21 23:13:59 +00:00
Akira Hatanaka
abfa7cf8bf [mips] Refactor jump, jump register, jump-and-link and nop instructions. 
llvm-svn: 170952
 2012-12-21 23:03:50 +00:00
Akira Hatanaka
5fb1ba2c90 [mips] Refactor load/store left/right and load-link and store-conditional 
instructions.

llvm-svn: 170950
 2012-12-21 23:01:24 +00:00
Akira Hatanaka
7e5047d48a [mips] Refactor load/store instructions. 
llvm-svn: 170948
 2012-12-21 22:58:55 +00:00
Akira Hatanaka
282d5fff48 [mips] Remove unnecessary isPseudo parameter. 
llvm-svn: 170947
 2012-12-21 22:57:26 +00:00
Akira Hatanaka
607c79d3a2 [mips] Refactor LUI instruction. 
llvm-svn: 170944
 2012-12-21 22:46:07 +00:00
Akira Hatanaka
5953571624 [mips] Refactor count leading zero or one instructions. 
llvm-svn: 170942
 2012-12-21 22:43:58 +00:00
Akira Hatanaka
aae0730c16 [mips] Refactor sign-extension-in-register instructions. 
llvm-svn: 170940
 2012-12-21 22:41:52 +00:00
Akira Hatanaka
28f3fe9725 [mips] Refactor instructions which copy from and to HI/LO registers. 
llvm-svn: 170939
 2012-12-21 22:39:17 +00:00
Akira Hatanaka
22e49f11a7 [mips] Refactor logical NOR instructions. 
llvm-svn: 170937
 2012-12-21 22:35:47 +00:00
Akira Hatanaka
adbd4c7cba [mips] Move instruction definitions in MipsInstrInfo.td. 
llvm-svn: 170936
 2012-12-21 22:33:43 +00:00
Tom Stellard
20b14a00e1 R600: Coding style - remove empty spaces from the beginning of functions 
No functionality change.

llvm-svn: 170923
 2012-12-21 20:12:02 +00:00
Tom Stellard
7257d89cf4 R600: Fix MAX_UINT definition 
Patch by: Vadim Girlin

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 170922
 2012-12-21 20:12:01 +00:00
Tom Stellard
0a1defe552 R600: Add SHADOWCUBE to TEX_SHADOW pattern 
Patch by: Vadim Girlin

Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 170921
 2012-12-21 20:11:59 +00:00
Benjamin Kramer
4a7c7f63ce Cleanup compiler warnings on discarding type qualifiers in casts. Switch to C++ style casts. 
Patch by Saleem Abdulrasool!

Differential Revision: http://llvm-reviews.chandlerc.com/D204

llvm-svn: 170917
 2012-12-21 19:09:53 +00:00
Benjamin Kramer
9297ff5d80 X86: Match pmin/pmax as a target specific dag combine. This occurs during vectorization. 
Part of PR14667.

llvm-svn: 170908
 2012-12-21 17:46:58 +00:00
Roman Divacky
a299a2efa5 Remove duplicate includes. 
llvm-svn: 170902
 2012-12-21 17:06:44 +00:00
Tom Stellard
19e8c4abd6 R600: Expand vec4 INT <-> FP conversions 
llvm-svn: 170901
 2012-12-21 16:33:24 +00:00
Benjamin Kramer
06ec6482f0 X86: Match the SSE/AVX min/max vector ops using a custom node instead of intrinsics 
This is very mechanical, no functionality change. Preparation for PR14667.

llvm-svn: 170898
 2012-12-21 14:04:55 +00:00
Evgeniy Stepanov
ea2d72253e [msan] Remove unreachable blocks before instrumenting a function. 
llvm-svn: 170883
 2012-12-21 11:18:49 +00:00
Nadav Rotem
82f3bb6d1f Add a missing "virtual" keyword. 
llvm-svn: 170842
 2012-12-21 05:02:12 +00:00
Nadav Rotem
ad3bbdd5b0 Enable if-conversion. 
llvm-svn: 170841
 2012-12-21 04:47:54 +00:00
Quentin Colombet
56d5ee357a Add ARM cortex-r5 subtarget. 
llvm-svn: 170840
 2012-12-21 04:35:05 +00:00
Rafael Espindola
cb6155ac8d Don't skip __DWARF, 
Now that we don't merge section and segment names, we don't need to skip the
segment name to get to the section name.

llvm-svn: 170839
 2012-12-21 04:08:03 +00:00
Rafael Espindola
cca2985848 Add a function to get the segment name of a section. 
On MachO, sections also have segment names. When a tool looking at a .o file
prints a segment name, this is what they mean. In reality, a .o has only one
anonymous, segment.

This patch adds a MachO only function to fetch that segment name. I named it
getSectionFinalSegmentName since the main use for the name seems to be inform
the linker with segment this section should go to.

The patch also changes MachOObjectFile::getSectionName to return just the
section name instead of computing SegmentName,SectionName.

The main difference from the previous patch is that it doesn't use
InMemoryStruct. It is extremely dangerous: if the endians match it returns
a pointer to the file buffer, if not, it returns a pointer to an internal buffer
that is overwritten in the next API call.

We should change all of this code to use
support::detail::packed_endian_specific_integral like ELF, but since these
functions only handle strings, they work with big and little endian machines
as is.

I have tested this by installing ubuntu 12.10 ppc on qemu, that is why it took
so long :-)

llvm-svn: 170838
 2012-12-21 03:47:03 +00:00
Evan Cheng
6e9736cf5e Add targets to skip running the GC passes. 
llvm-svn: 170836
 2012-12-21 02:57:04 +00:00
Evan Cheng
69cb91fa21 Every pass deserves a name, even codegenprep. 
llvm-svn: 170831
 2012-12-21 01:48:14 +00:00
Nadav Rotem
3d4c6351cf Improve the X86 cost model for loads and stores. 
llvm-svn: 170830
 2012-12-21 01:33:59 +00:00
Nadav Rotem
ec327f0de1 BB-Vectorizer: Check the cost of the store pointer type 
and not the return type, which is void. A number of test
cases fail after adding the assertion in TTImpl.

llvm-svn: 170828
 2012-12-21 01:24:36 +00:00
Reed Kotler
d8213f114a Call llvm_unreachable instead of assert. 
llvm-svn: 170822
 2012-12-21 00:44:59 +00:00
Nadav Rotem
80fefbe978 Fix a bug in the code that checks if we can vectorize loops while using dynamic 
memory bound checks.  Before the fix we were able to vectorize this loop from
the Livermore Loops benchmark:

for ( k=1 ; k<n ; k++ )
  x[k] = x[k-1] + y[k];

llvm-svn: 170811
 2012-12-21 00:07:35 +00:00