Nadav Rotem [Mon, 24 Dec 2012 10:04:03 +0000 (10:04 +0000)]
CostModel: We have API for checking the costs of known shuffles. This patch adds
support for the insert-subvector and extract-subvector kinds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171027
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Mon, 24 Dec 2012 10:03:57 +0000 (10:03 +0000)]
Added 6 more value types: v32i1, v64i1, v32i16, v32i8, v64i8, v8f64
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171026
91177308-0d34-0410-b5e6-
96231b3b80d8
Elena Demikhovsky [Mon, 24 Dec 2012 09:42:27 +0000 (09:42 +0000)]
Removed "static" from "__jit_debug_descriptor" because "static" adds C++ mangling prefix to this symbol.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171025
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 24 Dec 2012 09:40:33 +0000 (09:40 +0000)]
Some x86 instructions can load/store one of the operands to memory. On SSE, this memory needs to be aligned.
When these instructions are encoded in VEX (on AVX) there is no such requirement. This changes the folding
tables and removes the alignment restrictions from VEX-encoded instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171024
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 24 Dec 2012 09:14:18 +0000 (09:14 +0000)]
LoopVectorizer: When checking for vectorizable types, also check
the StoreInst operands.
PR14705.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171023
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 24 Dec 2012 08:57:47 +0000 (08:57 +0000)]
Change the codegen Cost Model API for shuffeles. This patch removes the API for broadcast and adds a more general API that accepts an enum of known shuffles.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171022
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Mon, 24 Dec 2012 08:52:53 +0000 (08:52 +0000)]
Fix typo in comments
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171021
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 24 Dec 2012 05:51:12 +0000 (05:51 +0000)]
Update the docs of the cost model.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171016
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Mon, 24 Dec 2012 01:24:04 +0000 (01:24 +0000)]
llvm/MC/MCMachObjectWriter.h: ComputeSymbolTable(): Prune one description in the comment. [-Wdocumentation]
/// \param StringIndexMap [out] - Map from symbol names to offsets in the string table.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171010
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Mon, 24 Dec 2012 01:22:06 +0000 (01:22 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171008
91177308-0d34-0410-b5e6-
96231b3b80d8
Dmitri Gribenko [Sun, 23 Dec 2012 18:46:11 +0000 (18:46 +0000)]
Documentation: fix typos reported in PR13866
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171006
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 23 Dec 2012 17:31:23 +0000 (17:31 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171002
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 23 Dec 2012 13:21:41 +0000 (13:21 +0000)]
LoopVectorize: Fix accidentaly inverted condition.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171001
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sun, 23 Dec 2012 13:19:18 +0000 (13:19 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171000
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 23 Dec 2012 09:11:07 +0000 (09:11 +0000)]
We are not ready to estimate the cost of integer expansions based on the number of parts. This test is too noisy.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170999
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Sun, 23 Dec 2012 07:34:51 +0000 (07:34 +0000)]
docs: Add link to external LLVM backend tutorial.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170998
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 23 Dec 2012 07:33:44 +0000 (07:33 +0000)]
whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170997
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 23 Dec 2012 07:30:09 +0000 (07:30 +0000)]
Rename a function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170996
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Sun, 23 Dec 2012 07:23:55 +0000 (07:23 +0000)]
Loop Vectorizer: Update the cost model of scatter/gather operations and make
them more expensive.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170995
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 22 Dec 2012 19:15:35 +0000 (19:15 +0000)]
Remove trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170991
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 22 Dec 2012 18:09:02 +0000 (18:09 +0000)]
Remove trailing whitespace
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170990
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 22 Dec 2012 17:33:22 +0000 (17:33 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170989
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Sat, 22 Dec 2012 17:13:06 +0000 (17:13 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170988
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 22 Dec 2012 16:07:56 +0000 (16:07 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170985
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Sat, 22 Dec 2012 11:34:28 +0000 (11:34 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170984
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 22 Dec 2012 08:22:01 +0000 (08:22 +0000)]
Add a comma to fix the build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170982
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 22 Dec 2012 08:16:17 +0000 (08:16 +0000)]
Use a negative value to represent INVALID_SIMPLE_VALUE_TYPE instead of 256. Its much cheaper for the isSimple() checks to look for values less than 0 rather than a value greater than 255. This shaves ~8k off the size of the llc binary on x86-64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170981
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Sat, 22 Dec 2012 03:08:37 +0000 (03:08 +0000)]
Add vAny and Metadata to the switch in getSizeInBits for consistency since every other enum was listed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170977
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Dunbar [Sat, 22 Dec 2012 00:47:06 +0000 (00:47 +0000)]
[utils] Tweak utils/clang-parse-diagnostics-file to ignore autoconf diagnostics.
- Also, don't print headers if we aren't going to print any diagnostics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170973
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Sat, 22 Dec 2012 00:37:52 +0000 (00:37 +0000)]
Change 'AttrVal' to 'AttrKind' to better reflect that it's a kind of attribute instead of the value of the attribute.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170972
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Sat, 22 Dec 2012 00:31:54 +0000 (00:31 +0000)]
Fix some undefined behavior when parsing YAML input: don't try to compare an
uninitialized value against a default value. Found by -fsanitize=enum.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170970
91177308-0d34-0410-b5e6-
96231b3b80d8
Richard Smith [Sat, 22 Dec 2012 00:15:13 +0000 (00:15 +0000)]
Don't call back() on an empty SmallVector. Found by -fsanitize=enum!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170968
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Fri, 21 Dec 2012 23:48:49 +0000 (23:48 +0000)]
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
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170961
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 23:21:32 +0000 (23:21 +0000)]
[mips] Refactor subword-swap, EXT/INS, load-effective-address and read-hardware
instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170956
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 23:17:36 +0000 (23:17 +0000)]
[mips] Refactor SYNC and multiply/divide instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170955
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 23:15:59 +0000 (23:15 +0000)]
[mips] Refactor BAL instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170954
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 23:13:59 +0000 (23:13 +0000)]
[mips] Fix encoding of BAL instruction. Also, fix assembler test case which
was not catching the error.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170953
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 23:03:50 +0000 (23:03 +0000)]
[mips] Refactor jump, jump register, jump-and-link and nop instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170952
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 23:01:24 +0000 (23:01 +0000)]
[mips] Refactor load/store left/right and load-link and store-conditional
instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170950
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:58:55 +0000 (22:58 +0000)]
[mips] Refactor load/store instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170948
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:57:26 +0000 (22:57 +0000)]
[mips] Remove unnecessary isPseudo parameter.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170947
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:46:07 +0000 (22:46 +0000)]
[mips] Refactor LUI instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170944
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:43:58 +0000 (22:43 +0000)]
[mips] Refactor count leading zero or one instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170942
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:41:52 +0000 (22:41 +0000)]
[mips] Refactor sign-extension-in-register instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170940
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:39:17 +0000 (22:39 +0000)]
[mips] Refactor instructions which copy from and to HI/LO registers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170939
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:35:47 +0000 (22:35 +0000)]
[mips] Refactor logical NOR instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170937
91177308-0d34-0410-b5e6-
96231b3b80d8
Akira Hatanaka [Fri, 21 Dec 2012 22:33:43 +0000 (22:33 +0000)]
[mips] Move instruction definitions in MipsInstrInfo.td.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170936
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 21 Dec 2012 20:12:02 +0000 (20:12 +0000)]
R600: Coding style - remove empty spaces from the beginning of functions
No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170923
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 21 Dec 2012 20:12:01 +0000 (20:12 +0000)]
R600: Fix MAX_UINT definition
Patch by: Vadim Girlin
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170922
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 21 Dec 2012 20:11:59 +0000 (20:11 +0000)]
R600: Add SHADOWCUBE to TEX_SHADOW pattern
Patch by: Vadim Girlin
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170921
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 21 Dec 2012 19:09:53 +0000 (19:09 +0000)]
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
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170917
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Fri, 21 Dec 2012 18:38:09 +0000 (18:38 +0000)]
Add a missing assertion, the null register has no register units.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170916
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 21 Dec 2012 18:11:45 +0000 (18:11 +0000)]
try to unbreak ppc buildbots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170913
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 21 Dec 2012 18:00:08 +0000 (18:00 +0000)]
Teach sort_includes.py to drop duplicated includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170911
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 21 Dec 2012 17:46:58 +0000 (17:46 +0000)]
X86: Match pmin/pmax as a target specific dag combine. This occurs during vectorization.
Part of PR14667.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170908
91177308-0d34-0410-b5e6-
96231b3b80d8
Roman Divacky [Fri, 21 Dec 2012 17:06:44 +0000 (17:06 +0000)]
Remove duplicate includes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170902
91177308-0d34-0410-b5e6-
96231b3b80d8
Tom Stellard [Fri, 21 Dec 2012 16:33:24 +0000 (16:33 +0000)]
R600: Expand vec4 INT <-> FP conversions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170901
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Fri, 21 Dec 2012 14:04:55 +0000 (14:04 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170898
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan Sands [Fri, 21 Dec 2012 12:03:03 +0000 (12:03 +0000)]
Test that a landingpad gets the name provided when it was created (see commit
170318).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170886
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Fri, 21 Dec 2012 11:18:49 +0000 (11:18 +0000)]
[msan] Remove unreachable blocks before instrumenting a function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170883
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Fri, 21 Dec 2012 05:02:12 +0000 (05:02 +0000)]
Add a missing "virtual" keyword.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170842
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Fri, 21 Dec 2012 04:47:54 +0000 (04:47 +0000)]
Enable if-conversion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170841
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Fri, 21 Dec 2012 04:35:05 +0000 (04:35 +0000)]
Add ARM cortex-r5 subtarget.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170840
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 21 Dec 2012 04:08:03 +0000 (04:08 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170839
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Fri, 21 Dec 2012 03:47:03 +0000 (03:47 +0000)]
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 :-)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170838
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 21 Dec 2012 02:57:04 +0000 (02:57 +0000)]
Add targets to skip running the GC passes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170836
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Fri, 21 Dec 2012 01:48:14 +0000 (01:48 +0000)]
Every pass deserves a name, even codegenprep.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170831
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Fri, 21 Dec 2012 01:33:59 +0000 (01:33 +0000)]
Improve the X86 cost model for loads and stores.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170830
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Fri, 21 Dec 2012 01:24:36 +0000 (01:24 +0000)]
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.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170828
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Fri, 21 Dec 2012 00:55:10 +0000 (00:55 +0000)]
Add test case for r170674
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170823
91177308-0d34-0410-b5e6-
96231b3b80d8
Reed Kotler [Fri, 21 Dec 2012 00:44:59 +0000 (00:44 +0000)]
Call llvm_unreachable instead of assert.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170822
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Fri, 21 Dec 2012 00:28:42 +0000 (00:28 +0000)]
docs: More robust image scaling fix.
Hopefully these benchmarks will be updated in the future, so avoid
hardcoding image dimensions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170819
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Fri, 21 Dec 2012 00:20:25 +0000 (00:20 +0000)]
docs: Prevent image scaling.
Tell the image to be its natural size.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170816
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Fri, 21 Dec 2012 00:07:35 +0000 (00:07 +0000)]
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];
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170811
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Fri, 21 Dec 2012 00:03:42 +0000 (00:03 +0000)]
Move these files over to the debug info directory.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170810
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Thu, 20 Dec 2012 23:35:22 +0000 (23:35 +0000)]
docs: Try out nosidebar.
Please squawk if you find this appalling or otherwise don't like it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170803
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Thu, 20 Dec 2012 22:59:36 +0000 (22:59 +0000)]
docs: Cleanup trailing whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170799
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Dec 2012 22:54:05 +0000 (22:54 +0000)]
Require the two-argument MI::addOperand(MF, MO) for dangling instructions.
Instructions that are inserted in a basic block can still be decorated
with addOperand(MO).
Make the two-argument addOperand() function contain the actual
implementation. This function will now always have a valid MF reference
that it can use for memory allocation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170798
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Dec 2012 22:54:02 +0000 (22:54 +0000)]
Add an MF argument to MI::copyImplicitOps().
This function is often used to decorate dangling instructions, so a
context reference is required to allocate memory for the operands.
Also add a corresponding MachineInstrBuilder method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170797
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Dec 2012 22:53:58 +0000 (22:53 +0000)]
Use two-arg addOperand(MF, MO) internally in MachineInstr when possible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170796
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Dec 2012 22:53:55 +0000 (22:53 +0000)]
MachineInstrBuilderize ARM.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170795
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Dec 2012 22:53:53 +0000 (22:53 +0000)]
MachineInstrBuilderize NVPTX.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170794
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Bendersky [Thu, 20 Dec 2012 22:51:52 +0000 (22:51 +0000)]
Fix an unitialized member variable that may have caused sporadic failures
for code that wasn't even in bundling mode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170793
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Thu, 20 Dec 2012 22:49:13 +0000 (22:49 +0000)]
docs: actually indent these consistently
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170792
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Thu, 20 Dec 2012 22:47:41 +0000 (22:47 +0000)]
docs: Indent consistently in code examples.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170791
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Thu, 20 Dec 2012 22:42:20 +0000 (22:42 +0000)]
docs: Improve navigation for Vectorizers.rst
Add links in the intro paragraph.
Add table of contents.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170790
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Thu, 20 Dec 2012 22:24:37 +0000 (22:24 +0000)]
docs: bring back link for reddit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170776
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 20 Dec 2012 21:58:40 +0000 (21:58 +0000)]
Whitespace and 80-column cleanup.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170771
91177308-0d34-0410-b5e6-
96231b3b80d8
Eric Christopher [Thu, 20 Dec 2012 21:58:36 +0000 (21:58 +0000)]
Start splitting out the debug string section handling by moving it
into the DwarfUnits class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170770
91177308-0d34-0410-b5e6-
96231b3b80d8
Sean Silva [Thu, 20 Dec 2012 21:50:41 +0000 (21:50 +0000)]
docs: Make document name congruent with title.
Hopefully nobody has linked to it yet...
OK'd by Nadav.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170768
91177308-0d34-0410-b5e6-
96231b3b80d8
Bill Wendling [Thu, 20 Dec 2012 21:28:43 +0000 (21:28 +0000)]
Some random comment, naming, and format changes.
Rename the AttributeImpl* from Attrs to pImpl to be consistent with other code.
Add comments where none were before. Or doxygen-ify other comments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170767
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Dec 2012 21:12:42 +0000 (21:12 +0000)]
Remove two dead functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170766
91177308-0d34-0410-b5e6-
96231b3b80d8
Bob Wilson [Thu, 20 Dec 2012 21:09:38 +0000 (21:09 +0000)]
Revert "Adding support for llvm.arm.neon.vaddl[su].* and"
This reverts r170694. The operations can be represented in IR without
adding any new intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170765
91177308-0d34-0410-b5e6-
96231b3b80d8
Nadav Rotem [Thu, 20 Dec 2012 20:24:40 +0000 (20:24 +0000)]
LoopVectorize: Fix a bug in the scalarization of instructions.
Before if-conversion we could check if a value is loop invariant
if it was declared inside the basic block. Now that loops have
multiple blocks this check is incorrect.
This fixes External/SPEC/CINT95/099_go/099_go
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170756
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Thu, 20 Dec 2012 19:59:30 +0000 (19:59 +0000)]
On some ARM cpus, flags setting movs with shifter operand, i.e. lsl, lsr, asr,
are more expensive than the non-flag setting variant. Teach thumb2 size
reduction pass to avoid generating them unless we are optimizing for size.
rdar://
12892707
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170728
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Bendersky [Thu, 20 Dec 2012 19:54:02 +0000 (19:54 +0000)]
Change Lit error redirection to FileCheck to a more common syntax since it
can potentially cause some bots to fail.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170726
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Bendersky [Thu, 20 Dec 2012 19:16:57 +0000 (19:16 +0000)]
Add a largish auto-generated test for the aligned bundling feature, along with
the script generating it. The test should never be modified manually. If anyone
needs to change it, please change the script and re-run it.
The script is placed into utils/testgen - I couldn't think of a better place,
and after some discussion on IRC this looked like a logical location.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170720
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Bendersky [Thu, 20 Dec 2012 19:07:30 +0000 (19:07 +0000)]
Tests for the aligned bundling support added in r170718
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170719
91177308-0d34-0410-b5e6-
96231b3b80d8
Eli Bendersky [Thu, 20 Dec 2012 19:05:53 +0000 (19:05 +0000)]
Aligned bundling support. Following the discussion here:
http://lists.cs.uiuc.edu/pipermail/llvmdev/2012-December/056754.html
The proposal and implementation are fully documented here:
https://sites.google.com/a/chromium.org/dev/nativeclient/pnacl/aligned-bundling-support-in-llvm
Tests will follow shortly.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170718
91177308-0d34-0410-b5e6-
96231b3b80d8
Jakob Stoklund Olesen [Thu, 20 Dec 2012 18:46:29 +0000 (18:46 +0000)]
Use MachineInstrBuilder for PHI nodes in SelectionDAGISel.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@170716
91177308-0d34-0410-b5e6-
96231b3b80d8