Commit Graph

91467 Commits

Author SHA1 Message Date
Nadav Rotem
3fe91a4453 Fix a grammar mistake, and add a line about the two phases that the BB/SLP vectorizers have (top-down and bottom-up).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179566 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 22:21:25 +00:00
Nadav Rotem
96e0b96a91 Fix the internal link.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179565 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 22:11:07 +00:00
Nadav Rotem
74cd12b73a Update the release notes about the vectorizers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179564 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 22:10:39 +00:00
Nadav Rotem
e9a4411db4 SLPVectorizer: Make it a function pass and add code for hoisting the vector-gather sequence out of loops.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179562 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 22:00:26 +00:00
Rafael Espindola
c9363ef67a Fix silly typo that broke big endian hosts.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179551 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 20:13:59 +00:00
Rafael Espindola
92b0d0ec4f Fix endianness on some MSVC versions.
Looks like it was evaluating undef == undef to true.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179549 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 19:28:45 +00:00
Tom Stellard
9a256300f8 R600/SI: Emit config values in register value pairs.
Instead of emitting config values in a predefined order, the code
emitter will now emit a 32-bit register index followed by the 32-bit
config value.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179546 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 17:51:35 +00:00
Tom Stellard
bf1efe6421 R600/SI: Emit configuration value in the .AMDGPU.config ELF section
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179545 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 17:51:30 +00:00
Tom Stellard
3a63bf27c5 R600: Emit ELF formatted code rather than raw ISA.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179544 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 17:51:21 +00:00
Jim Grosbach
467116a1c8 Fix a typo in comment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179542 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 17:40:48 +00:00
Jim Grosbach
0437ef6510 Simplify the MCInst operator iterator declaration.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179541 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 17:40:45 +00:00
John Criswell
24dcc2042e Grammar and punctuation fixes.
No content changes.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179540 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 17:38:06 +00:00
Rafael Espindola
0851ce0d8e Try to fix the mingw builds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179536 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 16:46:43 +00:00
Arnold Schwaighofer
f971916768 Fix bit size of v64i8 and v32i16 vector types.
Patch by Cameron McInally <cameron.mcinally@nyu.edu>.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179535 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 16:11:25 +00:00
Rafael Espindola
c7f424d338 Remove getters now that we can specialize structs on the host endianness.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179534 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 16:08:02 +00:00
Tim Northover
d354644eb1 Avoid outputting temporary test file into source tree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179532 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 15:49:13 +00:00
Rafael Espindola
0034c1f3df Remove unused function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179530 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 15:13:10 +00:00
Rafael Espindola
21a01d1ea8 Make the host endianness check an integer constant expression.
I will remove the isBigEndianHost function once I update clang.

The ifdef logic is designed to
* not use configure/cmake to avoid breaking -arch i686 -arch ppc.
* default to little endian
* be as small as possible

It looks like sys/endian.h is the preferred header on most modern BSD systems,
but it is better to change this in a followup patch as machine/endian.h is
available on FreeBSD, OpenBSD, NetBSD and OS X.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179527 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 14:44:24 +00:00
Andy Gibbs
604b3573f9 Replace uses of the deprecated std::auto_ptr with OwningPtr.
This is a rework of the broken parts in r179373 which were subsequently reverted in r179374 due to incompatibility with C++98 compilers.  This version should be ok under C++98.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179520 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 12:06:32 +00:00
Tim Northover
eaa752f5e1 Enable all targets by default on Visual Studio.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179518 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 11:53:05 +00:00
Eric Christopher
60d20a81fc Revert "Recommit r179497 after fixing uninitialized variable." until
I can fix the testcases here:

http://lab.llvm.org:8011/builders/clang-native-arm-cortex-a9/builds/6952

This reverts commit r179512 due to testcases specifying triples
that they didn't actually mean and causing failures on other platforms.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179513 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 07:31:37 +00:00
Eric Christopher
fdf9624f3c Recommit r179497 after fixing uninitialized variable.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179512 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 07:07:21 +00:00
Nadav Rotem
2f7ce4522d Document our desire to enable the loop vectorizer on -Os in future releases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179511 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 05:56:55 +00:00
Nadav Rotem
fc175d97ff Docs: merge the description of the BB and SLP vectorizers and document the -fslp-vectorize-aggressive flag.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179510 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 05:53:23 +00:00
Nadav Rotem
1129a832e6 Add an option -vectorize-slp-aggressive for running the BB vectorizer. Make -fslp-vectorize run the slp-vectorizer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179508 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 05:39:58 +00:00
Nadav Rotem
8849838965 Rename the slp-vectorizer clang/llvm flags. No functionality change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179505 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 04:54:42 +00:00
Nadav Rotem
09616565dd SLPVectorizer: Add support for vectorizing trees that start at compare instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179504 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 04:25:27 +00:00
Jia Liu
fd975b0762 fix include path in doc Extending LLVM
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179503 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 03:26:13 +00:00
Hal Finkel
00e86ad167 Mark all PPC comparison instructions as not having side effects
Now that the CR spilling issues have been resolved, we can remove the
unmodeled-side-effect attributes from the comparison instructions (and also
mark them as isCompare). By allowing these, by default, to have unmodeled side
effects, we were hiding problems with CR spilling; but everything seems much
happier now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179502 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 02:37:46 +00:00
Hal Finkel
fb6fe0aea2 Fix PPC64 CR spill location for callee-saved registers
This fixes an ABI bug for non-Darwin PPC64. For the callee-saved condition
registers, the spill location is specified relative to the stack pointer (SP +
8). However, this is not relative to the SP after the new stack frame is
established, but instead relative to the caller's stack pointer (it is stored
into the linkage area of the parent's stack frame).

So, like with the link register, we don't directly spill the CRs with other
callee-saved registers, but just mark them to be spilled during prologue
generation.

In practice, this reverts r179457 for PPC64 (but leaves it in place for PPC32).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179500 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-15 02:07:05 +00:00
Eric Christopher
f1216abf7e Revert "Remove some unused triple and data layout."
This reverts commit r179497 and the accompanying commit as it broke random platforms that aren't osx.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179499 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 23:35:36 +00:00
Eric Christopher
199ff9cdd7 Remove some unused triple and data layout.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179498 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 23:32:44 +00:00
Eric Christopher
10f8d2bb5b If we've specified a triple on the command line then go ahead
and use that as the default triple for the module and target
data layout.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179497 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 23:32:40 +00:00
Nico Rieck
ef1762b6a1 Use object file specific section type for initial text section
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179494 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 21:18:36 +00:00
David Majnemer
024d943bca Reorders two transforms that collide with each other
One performs: (X == 13 | X == 14) -> X-13 <u 2
The other: (A == C1 || A == C2) -> (A & ~(C1 ^ C2)) == C1

The problem is that there are certain values of C1 and C2 that
trigger both transforms but the first one blocks out the second,
this generates suboptimal code.

Reordering the transforms should be better in every case and
allows us to do interesting stuff like turn:
  %shr = lshr i32 %X, 4
  %and = and i32 %shr, 15
  %add = add i32 %and, -14
  %tobool = icmp ne i32 %add, 0

into:
  %and = and i32 %X, 240
  %tobool = icmp ne i32 %and, 224


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179493 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 21:15:43 +00:00
Nadav Rotem
687a9dfcb9 Make the command line triple match the module triple.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179492 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 20:13:05 +00:00
Benjamin Kramer
e197486908 Miscellaneous cleanups for VecUtils.h
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179483 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 09:33:08 +00:00
Nadav Rotem
a15dedba28 Document the SLP infrastructure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179480 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 07:42:25 +00:00
Nadav Rotem
0774629936 SLP: Document the scalarization cost method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179479 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 07:22:22 +00:00
Nadav Rotem
420820056d Document the decision to assume that the cost of floats is twice as much as integers.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179478 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 05:55:18 +00:00
Jakob Stoklund Olesen
d9f88da7b3 Use i32 for all SPARC shift amounts, even in 64-bit mode.
Test case by llvm-stress.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179477 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 05:48:50 +00:00
Nadav Rotem
1f098af364 Remove unused function attributes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179476 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 05:47:04 +00:00
Nadav Rotem
ab105ae95f SLPVectorizer: Add support for trees that don't start at binary operators, and add the cost of extracting values from the roots of the tree.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179475 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 05:15:53 +00:00
Jakob Stoklund Olesen
618eda7a60 Add support for the abs64 SPARC v9 code model.
For when 16 TB just isn't enough.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179474 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 05:10:36 +00:00
Jakob Stoklund Olesen
87ce01739b Add support for the SPARC v9 abs44 code model.
This is the default model for non-PIC 64-bit code. It supports
text+data+bss linked anywhere in the low 16 TB of the address space.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179473 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 04:57:51 +00:00
Jakob Stoklund Olesen
cab0abd03d Use target flags for printing SPARC asm operands.
64-bit code models need multiple relocations that can't be inferred from
the opcode like they can in 32-bit code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179472 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 04:35:19 +00:00
Jakob Stoklund Olesen
2693210656 Also put target flags on SPARC constant pool references.
Constant pool entries are accessed exactly the same way as global
variables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179471 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 04:35:16 +00:00
Nadav Rotem
f7eaf29cf7 SLPVectorizer: add initial support for reduction variable vectorization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179470 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 03:22:20 +00:00
Jakob Stoklund Olesen
ef596e1a80 Fix patterns for 64-bit pointers.
This fixes the pic32 code model for SPARC v9.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179469 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 01:53:23 +00:00
Jakob Stoklund Olesen
0ec587e26c Add target flags to SPARC address operands.
SDNodes and MachineOperands get target flags representing the %hi() and
%lo() assembly annotations that eventually become relocations.

Also define flags to be used by the 64-bit code models.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179468 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-14 01:33:32 +00:00