RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-22 04:22:57 +02:00
Code Issues Projects Releases Wiki Activity
79,191 Commits 4 Branches 5 Tags 704 MiB
ee2dabebe3
Commit Graph

806 Commits

Author SHA1 Message Date
Bruno Cardoso Lopes
17ae896095 Move code around and add comments 
llvm-svn: 137518
 2011-08-12 21:48:22 +00:00
Bruno Cardoso Lopes
4106caa9af Cleanup: Remove Int_ CVTSS2SI* forms 
llvm-svn: 137297
 2011-08-11 02:52:36 +00:00
Bruno Cardoso Lopes
565ab1542a The following X86 pattern is incorrect: 
def : Pat<(X86Movss VR128:$src1,
                   (bc_v4i32 (v2i64 (load addr:$src2)))),
          (MOVLPSrm VR128:$src1, addr:$src2)>;
This matches a MOVSS dag with a MOVLPS instruction. However, MOVSS will replace only the low 32 bits of the register, while the MOVLPS instruction will replace the low 64 bits. A testcase is added and illustrates the bug and also modified the one that was already present. Patch by Tanya Lattner.

llvm-svn: 137227
 2011-08-10 17:45:17 +00:00
Bruno Cardoso Lopes
7461b930f3 Add v16i16 and v32i8 store patterns 
llvm-svn: 137166
 2011-08-09 22:39:53 +00:00
Bruno Cardoso Lopes
028c6aa951 Use fp unpack instructions to unpack int types. Until we have AVX2, this 
is the best we can do for these patterns. This fix PR10554.

llvm-svn: 137161
 2011-08-09 22:18:37 +00:00
Bruno Cardoso Lopes
633400ee00 Reapply a more appropriate solution than in r137114. AVX supports 
v4f64 = sitofp v4i32. This fix PR10559.
Also add support for v4i32 = fptosi v4f64.

llvm-svn: 137128
 2011-08-09 17:39:13 +00:00
Bruno Cardoso Lopes
d521431558 Add support for avx vector fextend 
llvm-svn: 137105
 2011-08-09 03:04:29 +00:00
Bruno Cardoso Lopes
09a727298f Add AVX versions of 128-bit sitofp and fptosi 
llvm-svn: 137104
 2011-08-09 03:04:25 +00:00
Bruno Cardoso Lopes
1025d1eb3b Add two patterns to match special vmovss and vmovsd cases. Also fix 
the patterns already there to be more strict regarding the predicate.
This fixes PR10558

llvm-svn: 137100
 2011-08-09 01:43:09 +00:00
Bruno Cardoso Lopes
d7eac41193 Make LowerVSETCC aware of AVX types and add patterns to match them. 
llvm-svn: 137090
 2011-08-09 00:46:57 +00:00
Bruno Cardoso Lopes
771876cade Add v4f64 -> v2f32 fp_round support. Also add a testcase to exercise 
the legalizer. This commit together with the two previous ones fixes
PR10495.

llvm-svn: 136654
 2011-08-01 21:54:09 +00:00
Bruno Cardoso Lopes
473d982caf Add v8i32 and v4i64 vpermil patterns 
llvm-svn: 136451
 2011-07-29 01:31:07 +00:00
Bruno Cardoso Lopes
02bbf20b02 Cleanup PALIGNR handling and remove the old palign pattern fragment. 
Also make PALIGNR masks to don't match 256-bits, which isn't supported
It's also a step to solve PR10489

llvm-svn: 136448
 2011-07-29 01:30:59 +00:00
Bruno Cardoso Lopes
e24a043703 Add patterns to generate copies for extract_subvector instead of 
using vextractf128. This will reduce the number of issued instruction
for several avx codes.

llvm-svn: 136323
 2011-07-28 01:26:50 +00:00
Bruno Cardoso Lopes
73945bf79a movd/movq write zeros in the high 128-bit part of the vector. Use 
them to match 256-bit scalar_to_vector+zext.

llvm-svn: 136322
 2011-07-28 01:26:46 +00:00
Bruno Cardoso Lopes
1f63a37172 Add a few patterns to match allzeros without having to use the fp unit. 
Take advantage that the 128-bit vpxor zeros the higher part and use it.
This also fixes PR10491

llvm-svn: 136321
 2011-07-28 01:26:43 +00:00
Bruno Cardoso Lopes
06d8be564f Add SINT_TO_FP and FP_TO_SINT support for v8i32 types. Also move 
a convert pattern close to the instruction definition.

llvm-svn: 136320
 2011-07-28 01:26:39 +00:00
Kevin Enderby
9adbbfffd0 Fix llvm-mc handing of x86 instructions that take 8-bit unsigned immediates. 
llvm-mc gives an "invalid operand" error for instructions that take an unsigned
immediate which have the high bit set such as:
    pblendw $0xc5, %xmm2, %xmm1
llvm-mc treats all x86 immediates as signed values and range checks them.
A small number of x86 instructions use the imm8 field as a set of bits.
This change only changes those instructions and where the high bit is not
ignored.  The others remain unchanged.

llvm-svn: 136287
 2011-07-27 23:01:50 +00:00
Bruno Cardoso Lopes
8830fde434 The vpermilps and vpermilpd have different behaviour regarding the 
usage of the shuffle bitmask. Both work in 128-bit lanes without
crossing, but in the former the mask of the high part is the same
used by the low part while in the later both lanes have independent
masks. Handle this properly and and add support for vpermilpd.

llvm-svn: 136200
 2011-07-27 00:56:34 +00:00
Bruno Cardoso Lopes
e53bb853ea Recognize unpckh* masks and match 256-bit versions. The new versions are 
different from the previous 128-bit because they work in lanes.
Update a few comments and add testcases

llvm-svn: 136157
 2011-07-26 22:03:40 +00:00
Bruno Cardoso Lopes
a493ad3938 Remove now unused patterns. 0 insertions(+), 98 deletions(-) 
llvm-svn: 136109
 2011-07-26 18:22:39 +00:00
Bruno Cardoso Lopes
b24e958ffb Cleanup old matching for PUNPCK* variants 
llvm-svn: 136108
 2011-07-26 18:22:27 +00:00
Bruno Cardoso Lopes
ab40a57cce Add 256-bit isel for movsldup/movshdup 
llvm-svn: 136051
 2011-07-26 02:39:32 +00:00
Bruno Cardoso Lopes
cde45ac9ca Add 128-bit AVX versions of movshdup/mosldup 
llvm-svn: 136048
 2011-07-26 02:39:23 +00:00
Bruno Cardoso Lopes
25698b90e9 Cleanup movsldup/movshdup matching. 
27 insertions(+), 62 deletions(-)

llvm-svn: 136047
 2011-07-26 02:39:13 +00:00
Bruno Cardoso Lopes
c94d6a2d2c Codegen allonesvector better while using AVX: vpcmpeqd + vinsertf128 
This also fixes PR10452

llvm-svn: 136004
 2011-07-25 23:05:32 +00:00
Bruno Cardoso Lopes
f457bc8120 Add remaining 256-bit vector bitcasts. This also fixes PR10451 
llvm-svn: 136003
 2011-07-25 23:05:28 +00:00
Bruno Cardoso Lopes
9380919dc5 - Handle special scalar_to_vector case: splats. Using a native 128-bit 
shuffle before inserting on a 256-bit vector.
- Add AVX versions of movd/movq instructions
- Introduce a few COPY patterns to match insert_subvector instructions.
This turns a trivial insert_subvector instruction into a register copy,
coalescing the xmm into a ymm and avoid emiting on more instruction.

llvm-svn: 136002
 2011-07-25 23:05:25 +00:00
Bruno Cardoso Lopes
5382a1e65e Add v8f32->v8i32 bitcast. Fixes PR10440 
llvm-svn: 135794
 2011-07-22 19:51:02 +00:00
Bruno Cardoso Lopes
3691063149 - Register v16i16 as valid VR256 register class 
- Add more bitcasts for v16i16
- Since 135661 and 135662 already added the splat logic,
just add one more splat test for v16i16

llvm-svn: 135663
 2011-07-21 02:24:08 +00:00
Bruno Cardoso Lopes
ba1a2a9135 Add support for 256-bit versions of VPERMIL instruction. This is a new 
instruction introduced in AVX, which can operate on 128 and 256-bit vectors.
It considers a 256-bit vector as two independent 128-bit lanes. It can permute
any 32 or 64 elements inside a lane, and restricts the second lane to
have the same permutation of the first one. With the improved splat support
introduced early today, adding codegen for this instruction enable more
efficient 256-bit code:

Instead of:
  vextractf128  $0, %ymm0, %xmm0
  punpcklbw %xmm0, %xmm0
  punpckhbw %xmm0, %xmm0
  vinsertf128 $0, %xmm0, %ymm0, %ymm1
  vinsertf128 $1, %xmm0, %ymm1, %ymm0
  vextractf128  $1, %ymm0, %xmm1
  shufps  $1, %xmm1, %xmm1
  movss %xmm1, 28(%rsp)
  movss %xmm1, 24(%rsp)
  movss %xmm1, 20(%rsp)
  movss %xmm1, 16(%rsp)
  vextractf128  $0, %ymm0, %xmm0
  shufps  $1, %xmm0, %xmm0
  movss %xmm0, 12(%rsp)
  movss %xmm0, 8(%rsp)
  movss %xmm0, 4(%rsp)
  movss %xmm0, (%rsp)
  vmovaps (%rsp), %ymm0
We get:
  vextractf128  $0, %ymm0, %xmm0
  punpcklbw %xmm0, %xmm0
  punpckhbw %xmm0, %xmm0
  vinsertf128 $0, %xmm0, %ymm0, %ymm1
  vinsertf128 $1, %xmm0, %ymm1, %ymm0
  vpermilps $85, %ymm0, %ymm0

llvm-svn: 135662
 2011-07-21 01:55:47 +00:00
Bruno Cardoso Lopes
194507cc77 Add aditional patterns for vextractf128 instruction 
llvm-svn: 135660
 2011-07-21 01:55:39 +00:00
Bruno Cardoso Lopes
14c800c1e3 Add aditional patterns for vinsertf128 instruction 
llvm-svn: 135659
 2011-07-21 01:55:36 +00:00
Bruno Cardoso Lopes
e0d5bd467f Move code around. No functionality changes 
llvm-svn: 135657
 2011-07-21 01:55:30 +00:00
Bruno Cardoso Lopes
bdf75dfa28 Be more smart with VCVTSS2SD. Also place the patterns close to the 
definitions.

llvm-svn: 135407
 2011-07-18 18:11:25 +00:00
Bruno Cardoso Lopes
da90f383ab Add AVX 128-bit sqrt versions 
llvm-svn: 135404
 2011-07-18 17:51:40 +00:00
Bruno Cardoso Lopes
d258749f73 Add AVX 128-bit patterns for sint_to_fp 
llvm-svn: 135332
 2011-07-16 00:50:20 +00:00
Bruno Cardoso Lopes
2a23e486ad Add a few patterns for 256-bit bitcasts. No testcases now, they are 
comming together with other tests.

llvm-svn: 135312
 2011-07-15 22:24:17 +00:00
Bruno Cardoso Lopes
d24f039847 Add 256-bit load/store recognition and matching in several places. 
llvm-svn: 135171
 2011-07-14 18:50:58 +00:00
Bruno Cardoso Lopes
c0401dddf7 Make X86ISD::ANDNP more general and Codegen 256-bit VANDNP. A more 
general version of X86ISD::ANDNP also opened the room for a little bit
of refactoring.

llvm-svn: 135088
 2011-07-13 21:36:51 +00:00
Bruno Cardoso Lopes
b98f50da03 The target specific node PANDN name is misleading. That happens because 
it's later selected to a ANDNPD/ANDNPS instruction instead of the PANDN
instruction. Rename it.

llvm-svn: 135087
 2011-07-13 21:36:47 +00:00
Bruno Cardoso Lopes
cb49278ad6 AVX Codegen support for 256-bit versions of vandps, vandpd, vorps, vorpd, vxorps, vxorpd 
llvm-svn: 135023
 2011-07-13 01:15:33 +00:00
Eli Friedman
9765ae0015 Add assembler/disassembler support for non-AVX pclmulqdq. While I'm here, use proper aliases for the pclmullqlqdq and friends. PR10269. 
llvm-svn: 134424
 2011-07-05 18:21:20 +00:00
Eli Friedman
802029c494 Add support for movntil/movntiq mnemonics. Reported on llvmdev. 
llvm-svn: 133759
 2011-06-23 21:07:47 +00:00
Nick Lewycky
8e5c09b7dc Add support for assembling "movq" when it's correct to do so, while continuing 
to emit "movd" across the board to continue supporting a Darwin assembler bug.
This is the reincarnation of r133452.

llvm-svn: 133565
 2011-06-21 22:45:41 +00:00
Bob Wilson
5b04895bb8 Revert r133452: "Emit movq for 64-bit register to XMM register moves..." 
This is breaking compiler-rt and llvm-gcc builds on MacOSX when not using
the integrated assembler.

llvm-svn: 133524
 2011-06-21 17:35:13 +00:00
Nick Lewycky
831fb8200d Emit movq for 64-bit register to XMM register moves, but continue to accept 
movd when assembling.

llvm-svn: 133452
 2011-06-20 18:33:26 +00:00
Bruno Cardoso Lopes
f52f4dd0b8 Add AVX suport for fpextend. 
Original patch by Syoyo Fujita with more comments by me.

llvm-svn: 133153
 2011-06-16 07:03:21 +00:00
Bruno Cardoso Lopes
b6afc5168f Add one more argument to the prefetch intrinsic to indicate whether it's a data 
or instruction cache access. Update the targets to match it and also teach
autoupgrade.

llvm-svn: 132976
 2011-06-14 04:58:37 +00:00
Stuart Hastings
ea8b49dff3 Reapply 132424 with fixes. This fixes PR10068. 
rdar://problem/5993888

llvm-svn: 132606
 2011-06-03 23:53:54 +00:00
Rafael Espindola
1299f014d4 Revert 132424 to fix PR10068. 
llvm-svn: 132479
 2011-06-02 19:57:47 +00:00
Stuart Hastings
9a085fb9d8 Recommit 132404 with fixes. rdar://problem/5993888 
llvm-svn: 132424
 2011-06-01 21:33:14 +00:00
Stuart Hastings
4b33767382 Revert 132404 to appease a buildbot. rdar://problem/5993888 
llvm-svn: 132419
 2011-06-01 19:52:20 +00:00
Stuart Hastings
23f5ceda96 Add support for x86 CMPEQSS and friends. These instructions do a 
floating-point comparison, generate a mask of 0s or 1s, and generally
DTRT with NaNs.  Only profitable when the user wants a materialized 0
or 1 at runtime.  rdar://problem/5993888

llvm-svn: 132404
 2011-06-01 17:17:45 +00:00
Stuart Hastings
fdc9e4af68 FGETSIGN support for x86, using movmskps/pd. Will be enabled with a 
patch to TargetLowering.cpp.  rdar://problem/5660695

llvm-svn: 132388
 2011-06-01 04:39:42 +00:00
Chad Rosier
b87c4a6945 Renamed llvm.x86.sse42.crc32 intrinsics; crc64 doesn't exist. 
crc32.[8|16|32] have been renamed to .crc32.32.[8|16|32] and
crc64.[8|16|32] have been renamed to .crc32.64.[8|64].

llvm-svn: 132163
 2011-05-26 23:13:19 +00:00
Rafael Espindola
98372d430c Don't produce a vmovntdq if we don't have AVX support. 
llvm-svn: 131330
 2011-05-14 00:30:01 +00:00
Bill Wendling
67f5e8f0a7 Replace the "movnt" intrinsics with a native store + nontemporal metadata bit. 
<rdar://problem/8460511>

llvm-svn: 130791
 2011-05-03 21:11:17 +00:00
Eric Christopher
1de0dfaab0 xmm0 is an implicit parameter in this and so shouldn't be in the 
string template.

Fixes rdar://8493866

llvm-svn: 130747
 2011-05-03 01:28:32 +00:00
Chris Lattner
52e40d9b4c clean up after Sean's r127646 patch. 
llvm-svn: 130475
 2011-04-29 05:40:18 +00:00
Bill Wendling
0984f4927e Reapply r129401 with patch for clang. 
llvm-svn: 129419
 2011-04-13 00:36:11 +00:00
Bill Wendling
f6446a0961 Revert r129401 for now. Clang is using the old way of doing things. 
llvm-svn: 129403
 2011-04-12 22:59:27 +00:00
Bill Wendling
f9c9d3e05b Remove the unaligned load intrinsics in favor of using native unaligned loads. 
Now that we have a first-class way to represent unaligned loads, the unaligned
load intrinsics are superfluous.

First part of <rdar://problem/8460511>.

llvm-svn: 129401
 2011-04-12 22:46:31 +00:00
Sean Callanan
a38db2eeda Enabled disassembler support for AVX instructions 
in the instruction tables and fixed a few bugs that
were causing decode conflicts.  Rudimentary tests
are coming up in the next patch.

llvm-svn: 127646
 2011-03-15 01:28:15 +00:00
David Greene
2fd6d03bc9 [AVX] Fix mask predicates for 256-bit UNPCKLPS/D and implement 
missing patterns for them.

      Add a SIMD test subdirectory to hold tests for SIMD instruction
      selection correctness and quality.
'

llvm-svn: 126845
 2011-03-02 17:23:43 +00:00
Joerg Sonnenberger
efa8090e2a Recognize monitor/mwait with explicit register arguments 
llvm-svn: 125805
 2011-02-18 00:48:11 +00:00
David Greene
7de7347ee8 [AVX] Support VSINSERTF128 with more patterns and appropriate 
infrastructure.  This makes lowering 256-bit vectors to 128-bit
vectors simple when 256-bit vector support is not available.

llvm-svn: 124868
 2011-02-04 16:08:29 +00:00
David Greene
2753be260c [AVX] VEXTRACTF128 support. This commit includes patterns for 
matching EXTRACT_SUBVECTOR to VEXTRACTF128 along with support routines
to examine and translate index values.  VINSERTF128 comes next.  With
these two in place we can begin supporting more AVX operations as
INSERT/EXTRACT can be used as a fallback when 256-bit support is not
available.

llvm-svn: 124797
 2011-02-03 15:50:00 +00:00
Chris Lattner
9ba0a83f2b fix a missing shuffle pattern, PR9009. Patch by Artiom Myaskouvskey! 
llvm-svn: 124102
 2011-01-24 03:42:46 +00:00
Chris Lattner
586e7af07d Fix PR8946, a missing reg/reg form of movdqu. 
llvm-svn: 123242
 2011-01-11 17:04:55 +00:00
Chris Lattner
3ef9db5cd4 fix PR8900, a shuffle miscompilation. Patch by Nadav Rotem! 
llvm-svn: 122921
 2011-01-05 22:28:46 +00:00
Nate Begeman
c7dfecb10e Implement feedback from Bruno on making pblendvb an x86-specific ISD node in addition to being an intrinsic, and convert 
lowering to use it.  Hopefully the pattern fragment is doing the right thing with XMM0, looks correct in testing.

llvm-svn: 122277
 2010-12-20 22:04:24 +00:00
Nate Begeman
ef5f3c0fa7 Add support for matching psign & plendvb to the x86 target 
Remove unnecessary pandn patterns, 'vnot' patfrag looks through bitcasts

llvm-svn: 122098
 2010-12-17 22:55:37 +00:00
Nate Begeman
8c00ecd290 Add some missing predicates. 
llvm-svn: 121445
 2010-12-10 00:54:26 +00:00
Nate Begeman
cb6d1c8193 Formalize the notion that AVX and SSE are non-overlapping extensions from the compiler's point of view. Per email discussion, we either want to always use VEX-prefixed instructions or never use them, and are taking "HasAVX" to mean "Always use VEX". Passing -mattr=-avx,+sse42 should serve to restore legacy SSE support when desirable. 
llvm-svn: 121439
 2010-12-10 00:26:57 +00:00
Nate Begeman
4a62a3e229 Add support for AVX to materialize +0.0 when doing scalar FP. 
llvm-svn: 121415
 2010-12-09 21:43:51 +00:00
Benjamin Kramer
851691ddb2 Add patterns for the x86 popcnt instruction. 
- Also adds a new POPCNT subtarget feature that is currently enabled if the target
  supports SSE4.2 (nehalem) or SSE4A (barcelona).

llvm-svn: 120917
 2010-12-04 20:32:23 +00:00
Nate Begeman
deb26223bd Scalar f32/f64 are also subregs of ymm regs 
llvm-svn: 120844
 2010-12-03 21:54:39 +00:00
Eric Christopher
6a21ceab5c Implement a PseudoI class and transfer the sse instructions over to use 
it.

llvm-svn: 120412
 2010-11-30 08:57:23 +00:00
Eric Christopher
f27f0b5234 Rewrite mwait and monitor support and custom lower arguments. 
Fixes PR8573.

llvm-svn: 120404
 2010-11-30 07:20:12 +00:00
Bruno Cardoso Lopes
9f9f796756 Fix PR8211 
llvm-svn: 118445
 2010-11-08 21:24:59 +00:00
Dale Johannesen
b78530f9b0 Fix pastos in handling of AVX cvttsd2si, PR8491. 
Bruno, please review, but I'm pretty sure this is right.
Patch by Alex Mac!

llvm-svn: 117514
 2010-10-28 00:35:54 +00:00
Chris Lattner
72e7e84c3f simplify some map operations. 
llvm-svn: 116014
 2010-10-07 23:57:02 +00:00
Evan Cheng
7c89d70f27 Canonicalize X86ISD::MOVDDUP nodes to v2f64 to make sure all cases match. Also eliminate unneeded isel patterns. rdar://8520311 
llvm-svn: 115977
 2010-10-07 20:50:20 +00:00
Chris Lattner
84846b71af remove the !nameconcat tblgen feature. It "shorthand" and only used in 4 places 
where !cast is just as short.

llvm-svn: 115722
 2010-10-06 00:19:21 +00:00
Chris Lattner
12274b9845 allow !strconcat to take more than two operands to eliminate 
!strconcat(!strconcat(!strconcat(!strconcat

Simplify some x86 td files to use it.

llvm-svn: 115719
 2010-10-05 23:58:18 +00:00
Chris Lattner
5d7d5a81eb distribute the rest of the contents of X86Instr64bit.td out to 
the right places.  X86Instr64bit.td now dies, long live x86-64!

llvm-svn: 115669
 2010-10-05 20:49:15 +00:00
Chris Lattner
9317bf2ed5 move CMOV_FR32 and friends to InstrCompiler, since they are 
pseudo instructions.

Move POPCNT to InstrSSE since they are SSE4 instructions.

llvm-svn: 115603
 2010-10-05 06:41:40 +00:00
Chris Lattner
9c58de2dc4 fix rdar://8490728 - llvm-mc rejects gpr64 form of 'movmskpd' 
llvm-svn: 115029
 2010-09-29 05:05:03 +00:00
Chris Lattner
890c21a20a add assembler support for the cvtsd2sil/cvtsd2siq mnemonics, rdar://8456382 
llvm-svn: 115027
 2010-09-29 04:55:40 +00:00
Chris Lattner
c14d59589c add basic avx support to the disassembler, also teach it about ssmem/sdmem 
operands.

With this done, we can remove the _Int suffixes from the round instructions
without the disassembler blowing up.  This allows the assembler to support
them, implementing rdar://8456376 - llvm-mc rejects 'roundss'

llvm-svn: 115019
 2010-09-29 02:57:56 +00:00
Chris Lattner
f90296b045 add asmparser support for cvttpd2dq by removing some Int_ prefixes. 
Clean up cvttps2dq by removing some redundant implementations of the
same instruction.  rdar://8456382

llvm-svn: 115018
 2010-09-29 02:36:32 +00:00
Chris Lattner
e5c5c8dc1f implement rdar://8456382 - cvtsd2si support, by removing some Int_ prefixes. 
llvm-svn: 115017
 2010-09-29 02:24:57 +00:00
Dale Johannesen
eb807a15a3 Fix typos. 128-bit PSHUFB takes 128-bit memory op. 
v8i16 is not an MMX type; put it where it belongs.

llvm-svn: 113785
 2010-09-13 21:15:43 +00:00
Bruno Cardoso Lopes
49efee5c95 Add one more pattern to fallback movddup 
llvm-svn: 113522
 2010-09-09 18:48:34 +00:00
Dale Johannesen
de53df20d6 Move remaining MMX instructions from SSE to MMX. 
llvm-svn: 113501
 2010-09-09 17:13:07 +00:00
Dale Johannesen
7469923117 Move most MMX instructions (defined as anything that 
uses MMX, even if it also uses other things) from InstrSSE
into InstrMMX.  No (intended) functional change.

llvm-svn: 113462
 2010-09-09 01:02:39 +00:00
Bruno Cardoso Lopes
892c337123 x86 vector shuffle lowering now relies only on target specific 
nodes to emit shuffles and don't do isel mask matching anymore.
- Add the selection of the remaining shuffle opcode (movddup)
- Introduce two new functions to "recognize" where we may get
potential folds and add several comments to them explaining why
they are not yet in the desidered shape.
- Add more patterns to fallback the case where we select
a specific shuffle opcode as if it could fold a load, but it
can't, so remap to a valid instruction.
- Add a couple of FIXMEs to address in the following days once
there's a good solution to the current folding problem.

llvm-svn: 113369
 2010-09-08 17:43:25 +00:00
Dale Johannesen
8354cab2de Add patterns for MMX that use the new intrinsics. 
Enable palignr intrinsic.
These may need adjustment for a new VT in due course.

llvm-svn: 113233
 2010-09-07 18:10:56 +00:00
Bruno Cardoso Lopes
92bb02f722 Remove unused target specific node 
llvm-svn: 113224
 2010-09-07 17:38:55 +00:00
Dale Johannesen
2f4f8f5705 Remove the rest of the nonexistent 64-bit AVX instructions. 
Bruno, please review.

llvm-svn: 113014
 2010-09-03 21:23:00 +00:00
Bruno Cardoso Lopes
b8ce8b7e9f Reapply last harmless part of r112934, the pattern fragment to match X86Unpcklpd 
llvm-svn: 113009
 2010-09-03 20:44:26 +00:00
Daniel Dunbar
26e0e964ab Revert r112934, "- Use specific nodes to match unpckl masks.", which introduced 
some infinite loop and select failures.
 - Apologies for eager reverting, but its branch day.

llvm-svn: 113000
 2010-09-03 19:38:11 +00:00
Bruno Cardoso Lopes
f91bd70e9a AVX doesn't support mm operations neither its instrinsics. 
The AVX versions of PALIGN and PABS* should only exist for
128-bit. Remove the unnecessary stuff.

llvm-svn: 112944
 2010-09-03 02:08:45 +00:00
Bruno Cardoso Lopes
e1ad6555a8 - Use specific nodes to match unpckl masks. 
- Teach getShuffleScalarElt how to handle more target
specific nodes, so the DAGCombine can make use of it.
- Add another hack to avoid the node update problem
during legalization. More description on the comments

llvm-svn: 112934
 2010-09-03 01:24:00 +00:00
Bruno Cardoso Lopes
dcdab94661 become more strict about when it's safe to use X86ISD::MOVLPS 
llvm-svn: 112799
 2010-09-02 02:35:51 +00:00
Bruno Cardoso Lopes
601bf4c6d3 Using target specific nodes for shuffle nodes makes the mask 
check more strict, breaking some cases not checked in the
testsuite, but also exposes some foldings not done before,
as this example:

  movaps  (%rdi), %xmm0
  movaps  (%rax), %xmm1
  movaps  %xmm0, %xmm2
  movss %xmm1, %xmm2
  shufps  $36, %xmm2, %xmm0

now is generated as:

  movaps  (%rdi), %xmm0
  movaps  %xmm0, %xmm1
  movlps  (%rax), %xmm1
  shufps  $36, %xmm1, %xmm0

llvm-svn: 112753
 2010-09-01 22:33:20 +00:00
Bruno Cardoso Lopes
9375b2f67d Use movlps, movlpd, movss and movsd specific nodes instead of pattern matching with movlp pattern fragment 
llvm-svn: 112694
 2010-09-01 05:08:25 +00:00
Bruno Cardoso Lopes
80613a070e Use x86 specific MOVSLDUP node, add more patterns to match it and remove useless load nodes 
llvm-svn: 112661
 2010-08-31 22:35:05 +00:00
Bruno Cardoso Lopes
8fc83b1960 Use x86 specific MOVSHDUP node and add more patterns to match it 
llvm-svn: 112657
 2010-08-31 22:22:11 +00:00
Bruno Cardoso Lopes
6fbe7b9ddd Use MOVLHPS and MOVHLPS x86 nodes whenever possible. Also remove some useless nodes 
llvm-svn: 112642
 2010-08-31 21:15:21 +00:00
Bruno Cardoso Lopes
7939025262 Use pshufhw and pshuflw in more cases and fix getTargetShuffleNode number of arguments 
llvm-svn: 111890
 2010-08-24 01:16:15 +00:00
Bruno Cardoso Lopes
28d9071635 This is the first step towards refactoring the x86 vector shuffle code. The 
general idea here is to have a group of x86 target specific nodes which are
going to be selected during lowering and then directly matched in isel.

The commit includes the addition of those specific nodes and a *bunch* of
patterns, and incrementally we're going to switch between them and what we
have right now. Both the patterns and target specific nodes can change as
we move forward with this work.

llvm-svn: 111691
 2010-08-20 22:55:05 +00:00
Dale Johannesen
3f9c148d0e Revert 110491. While not wrong, it was based on a 
misanalysis and is undesirable.

llvm-svn: 111028
 2010-08-13 18:43:45 +00:00
Bruno Cardoso Lopes
de5f3f5cb6 Improve comment to make explicit why not to touch this could before JIT goes MC 
llvm-svn: 111021
 2010-08-13 17:44:10 +00:00
Eric Christopher
63c83f19a0 Revert last patch and r110954 as I meant to. 
llvm-svn: 111001
 2010-08-13 02:37:50 +00:00
Bruno Cardoso Lopes
350d186d69 Some small clean-up: use of pseudo instructions 
llvm-svn: 110954
 2010-08-12 20:55:18 +00:00
Bruno Cardoso Lopes
7cb26cb8be - Teach SSEDomainFix to switch between different levels of AVX instructions. Here we guess that AVX will have domain issues, so just implement them for consistency and in the future we remove if it's unnecessary. 
- Make foldMemoryOperandImpl aware of 256-bit zero vectors folding and support the 128-bit counterparts of AVX too.
- Make sure MOV[AU]PS instructions are only selected when SSE1 is enabled, and duplicate the patterns to match AVX.
- Add a testcase for a simple 128-bit zero vector creation.

llvm-svn: 110946
 2010-08-12 20:20:53 +00:00
Bruno Cardoso Lopes
99b5298854 Define AVX 128-bit pattern versions of SET0PS/PD. 
llvm-svn: 110937
 2010-08-12 18:20:59 +00:00
Bruno Cardoso Lopes
bb491bd56c Begin to support some vector operations for AVX 256-bit intructions. The long 
term goal here is to be able to match enough of vector_shuffle and build_vector
so all avx intrinsics which aren't mapped to their own built-ins but to
shufflevector calls can be codegen'd. This is the first (baby) step, support
building zeroed vectors.

llvm-svn: 110897
 2010-08-12 02:06:36 +00:00
Bruno Cardoso Lopes
6eb24fd744 Add AVX matching patterns to Packed Bit Test intrinsics. 
Apply the same approach of SSE4.1 ptest intrinsics but
create a new x86 node "testp" since AVX introduces
vtest{ps}{pd} instructions which set ZF and CF depending
on sign bit AND and ANDN of packed floating-point sources.

This is slightly different from what the "ptest" does.
Tests comming with the other 256 intrinsics tests.

llvm-svn: 110744
 2010-08-10 23:25:42 +00:00
Bruno Cardoso Lopes
f1928b60c0 Add AVX movnt{pd,ps,dq} 256-bit intrinsics 
llvm-svn: 110650
 2010-08-10 02:49:24 +00:00
Bruno Cardoso Lopes
f5884c6791 Add AVX movmsk 256-bit intrinsics 
llvm-svn: 110648
 2010-08-10 02:34:56 +00:00
Bruno Cardoso Lopes
2a7ed4b5c9 Support AVX 256-bit load and store intrinsics 
llvm-svn: 110645
 2010-08-10 01:43:16 +00:00
Bruno Cardoso Lopes
1ea37cfa7b Patterns to match AVX cmp instructions 
llvm-svn: 110633
 2010-08-10 00:13:20 +00:00
Bruno Cardoso Lopes
4e8d77892c Add matching patterns for vblend AVX intrinsics 
llvm-svn: 110630
 2010-08-10 00:02:05 +00:00
Bruno Cardoso Lopes
e58d077846 Add VCVTPD2PS, VCVTPS2DQ, VCVTPS2PDY, VCVTTPD2DQY, VCVTTPS2DQ and VCVTPD2DQ 256-bit conversion intrinsics 
llvm-svn: 110608
 2010-08-09 21:51:56 +00:00
Bruno Cardoso Lopes
e7ceec4edf Add patterns to AVX conversions instructions. Do that instead of declaring more intructions whenever is possible, more coming 
llvm-svn: 110605
 2010-08-09 21:24:59 +00:00
Bruno Cardoso Lopes
6a92e01d05 Memory version of vcvtdq2pd intrinsic 
llvm-svn: 110582
 2010-08-09 18:20:14 +00:00
Bruno Cardoso Lopes
0794b8ab3f Patterns to match vinsert, vbroadcast, vmovmask and vcvtdq2pd AVX intrinsics 
llvm-svn: 110580
 2010-08-09 18:03:43 +00:00
Dale Johannesen
23f9086dd3 Use sdmem and sse_load_f64 (etc.) for the vector 
form of CMPSD (etc.)  Matching a 128-bit memory
operand is wrong, the instruction uses only 64 bits
(same as ADDSD etc.)  8193553.

llvm-svn: 110491
 2010-08-07 00:33:42 +00:00
Bruno Cardoso Lopes
5b602f8822 Patterns to match AVX 256-bit vzero intrinsics 
llvm-svn: 110480
 2010-08-06 22:10:01 +00:00
Bruno Cardoso Lopes
821eebf946 Patterns to match AVX 256-bit permutation intrinsics 
llvm-svn: 110468
 2010-08-06 20:03:27 +00:00
Bruno Cardoso Lopes
d186fba555 Patterns to match AVX 256-bit horizontal arithmetic intrinsics 
llvm-svn: 110427
 2010-08-06 02:10:30 +00:00
Bruno Cardoso Lopes
5e9f9c921e Patterns to match AVX 256-bit arithmetic intrinsics 
llvm-svn: 110425
 2010-08-06 01:52:29 +00:00
Bruno Cardoso Lopes
0c0dd2173c Support all 128-bit AVX vector intrinsics. Most part of them I already 
declared during the addition of the assembler support, the additional
changes are:
- Add missing intrinsics
- Move all SSE conversion instructions in X86InstInfo64.td to the SSE.td file.
- Duplicate some patterns to AVX mode.
- Step into PCMPEST/PCMPIST custom inserter and add AVX versions.

llvm-svn: 109878
 2010-07-30 19:54:33 +00:00
Bruno Cardoso Lopes
a80b57e5fc Add AVX version of CLMUL instructions 
llvm-svn: 109248
 2010-07-23 18:41:12 +00:00
Bruno Cardoso Lopes
93fd8bdf6a Fix some AVX instructions which didnt had HasAVX prefix. And also a problem with PINSRW, which was totally wrong because of a typo I introduced previously 
llvm-svn: 109198
 2010-07-23 00:14:54 +00:00
Bruno Cardoso Lopes
7722724eee Add remaining AVX instructions (most of them dealing with GR64 destinations. This complete the assembler support for the general AVX ISA. But we still miss instructions from FMA3 and CLMUL specific feature flags, which are now the next step 
llvm-svn: 109168
 2010-07-22 21:18:49 +00:00
Eric Christopher
4924d5fb93 Custom lower the memory barrier instructions and add support 
for lowering without sse2.  Add a couple of new testcases.

Fixes a few libgomp tests and latent bugs.  Remove a few todos.

llvm-svn: 109078
 2010-07-22 02:48:34 +00:00
Bruno Cardoso Lopes
5920e38cd2 Add more 256-bit forms for a bunch of regular AVX instructions 
Add 64-bit (GR64) versions of some instructions (which are not
described in their SSE forms, but are described in AVX)

llvm-svn: 109063
 2010-07-21 23:53:50 +00:00
Bruno Cardoso Lopes
eea3b7ed83 Add missing AVX convert instructions. Those instructions are not described in their SSE forms (although they exist), but add the AVX forms anyway, so the assembler can benefit from it 
llvm-svn: 109039
 2010-07-21 21:37:59 +00:00
Bruno Cardoso Lopes
1284fdc932 Avoid AVX instructions to be selected instead of its SSE form 
llvm-svn: 109032
 2010-07-21 20:38:42 +00:00
Bruno Cardoso Lopes
d13d8c2562 Add AVX only vzeroall and vzeroupper instructions 
llvm-svn: 109002
 2010-07-21 08:56:24 +00:00
Bruno Cardoso Lopes
c4d93a5a34 Add new AVX vpermilps, vpermilpd and vperm2f128 instructions 
llvm-svn: 108984
 2010-07-21 03:07:42 +00:00
Bruno Cardoso Lopes
a7efb29695 Add new AVX vmaskmov instructions, and also fix the VEX encoding bits to support it 
llvm-svn: 108983
 2010-07-21 02:46:58 +00:00
Bruno Cardoso Lopes
e0dce1c741 Add new AVX vextractf128 instructions 
llvm-svn: 108964
 2010-07-20 23:19:02 +00:00
Bruno Cardoso Lopes
b677cbc9b2 Add new AVX instruction vinsertf128 
llvm-svn: 108892
 2010-07-20 19:44:51 +00:00
Bruno Cardoso Lopes
88869cb4db Add AVX vbroadcast new instruction 
llvm-svn: 108788
 2010-07-20 00:11:13 +00:00
Bruno Cardoso Lopes
4ca44dda21 Add 256-bit vaddsub, vhadd, vhsub, vblend and vdpp instructions! 
llvm-svn: 108769
 2010-07-19 23:32:44 +00:00
Bruno Cardoso Lopes
0616a418b6 Add AVX 256-bit compare instructions and a bunch of testcases 
llvm-svn: 108286
 2010-07-13 22:06:38 +00:00
Bruno Cardoso Lopes
7bc71d2d0a AVX 256-bit conversion instructions 
Add the x86 VEX_L form to handle special cases where VEX_L must be set.

llvm-svn: 108274
 2010-07-13 21:07:28 +00:00
Bruno Cardoso Lopes
ae37153b05 Add AVX 256-bit packed logical forms 
llvm-svn: 108224
 2010-07-13 02:38:35 +00:00
Bruno Cardoso Lopes
495ae629bb Add AVX 256-bit unop arithmetic instructions 
llvm-svn: 108223
 2010-07-13 01:53:31 +00:00
Bruno Cardoso Lopes
185483638b Since AVX is a superset of all SSE versions, only use HasAVX for AVX instructions 
llvm-svn: 108222
 2010-07-13 00:38:47 +00:00
David Greene
d81591ee09 Move some SIMD fragment code into X86InstrFragmentsSIMD so that the 
utility classes can be used from multiple files.  This will aid
transitioning to a new refactored x86 SIMD specification.

llvm-svn: 108213
 2010-07-12 23:41:28 +00:00
Bruno Cardoso Lopes
852e3bf472 Add AVX 256 binary arithmetic instructions 
llvm-svn: 108207
 2010-07-12 23:04:15 +00:00
Bruno Cardoso Lopes
b021506033 More refactoring of basic SSE arith instructions. Open room for 256-bit instructions 
llvm-svn: 108204
 2010-07-12 22:41:32 +00:00
Dan Gohman
e9c4426bb0 Apply the SSE dependence idiom for SSE unary operations to 
SD instructions too, in addition to SS instructions. And
add a comment about it.

llvm-svn: 108191
 2010-07-12 20:46:04 +00:00
Bruno Cardoso Lopes
a4889e6f93 Add AVX 256-bit MOVMSK forms 
llvm-svn: 108184
 2010-07-12 20:06:32 +00:00
Bruno Cardoso Lopes
f4180a9a7b Add AVX 256-bit packed MOVNT variants 
llvm-svn: 108021
 2010-07-09 21:42:42 +00:00
Bruno Cardoso Lopes
6ca8dc935c Add AVX 256-bit unpack and interleave 
llvm-svn: 108017
 2010-07-09 21:20:35 +00:00
Bruno Cardoso Lopes
3676e24b67 Start the support for AVX instructions with 256-bit %ymm registers. A couple of 
notes:
- The instructions are being added with dummy placeholder patterns using some 256
  specifiers, this is not meant to work now, but since there are some multiclasses
  generic enough to accept them,  when we go for codegen, the stuff will be already
  there.
- Add VEX encoding bits to support YMM
- Add MOVUPS and MOVAPS in the first round
- Use "Y" as suffix for those Instructions: MOVUPSYrr, ...
- All AVX instructions in X86InstrSSE.td will move soon to a new X86InstrAVX
  file.

llvm-svn: 107996
 2010-07-09 18:27:43 +00:00
Bruno Cardoso Lopes
8d350872d4 Add AVX AES instructions 
llvm-svn: 107798
 2010-07-07 18:24:20 +00:00
Bruno Cardoso Lopes
6222076cd1 Add AVX SSE4.2 instructions 
llvm-svn: 107752
 2010-07-07 03:39:29 +00:00
Bruno Cardoso Lopes
931471d7e8 Use only one multiclass to pinsrq instructions 
llvm-svn: 107750
 2010-07-07 01:43:01 +00:00
Bruno Cardoso Lopes
65fbd0530f Now that almost all SSE4.1 AVX instructions are added, move code around to more appropriate sections. No functionality changes 
llvm-svn: 107749
 2010-07-07 01:33:38 +00:00
Bruno Cardoso Lopes
675ebe2dc0 Add AVX SSE4.1 insertps, ptest and movntdqa instructions 
llvm-svn: 107747
 2010-07-07 01:14:56 +00:00
Bruno Cardoso Lopes
fa10461265 Add AVX SSE4.1 extractps and pinsr instructions 
llvm-svn: 107746
 2010-07-07 01:01:13 +00:00
Bruno Cardoso Lopes
54c2f858b3 Add AVX SSE4.1 Extract Integer instructions 
llvm-svn: 107740
 2010-07-07 00:07:24 +00:00
Bruno Cardoso Lopes
b9e1c33054 Add the rest of AVX SSE4.1 packed move with sign/zero extend instructions 
llvm-svn: 107723
 2010-07-06 23:15:17 +00:00
Bruno Cardoso Lopes
0c6ec0b068 Add part of AVX SSE4.1 packed move with sign/zero extend instructions 
llvm-svn: 107720
 2010-07-06 23:01:41 +00:00
Bruno Cardoso Lopes
a0b37e839c Add AVX vblendvpd, vblendvps and vpblendvb instructions 
Update VEX encoding to support those new instructions

llvm-svn: 107715
 2010-07-06 22:36:24 +00:00
Chris Lattner
e7c95bcd9e rip out even more sporadic v2f32 support. 
llvm-svn: 107610
 2010-07-05 04:38:33 +00:00
Bill Wendling
689155c673 Revert r107583. I no longer think that this is the way to solve the problem. 
llvm-svn: 107585
 2010-07-04 09:16:57 +00:00
Bill Wendling
8a3ecba7a4 Mark sse_load_f32 and sse_load_f64 as having memory operands 
(SDNPMemOperand). This way when they're morphed the memory operands will be
copied as well.

llvm-svn: 107583
 2010-07-04 08:59:55 +00:00
Bruno Cardoso Lopes
dc16024895 Add AVX SSE4.1 blend, mpsadbw and vdp 
llvm-svn: 107560
 2010-07-03 01:37:03 +00:00
Bruno Cardoso Lopes
9cbb625579 Add AVX SSE4.1 binop (some forms of packed max,min,mul,pack,cmp) instructions 
llvm-svn: 107558
 2010-07-03 01:15:47 +00:00
Bruno Cardoso Lopes
df02d037e4 Add AVX SSE4.1 Horizontal Minimum and Position instruction 
llvm-svn: 107552
 2010-07-03 00:49:21 +00:00
Bruno Cardoso Lopes
e6b70efcb0 Add AVX SSE4.1 round instructions 
llvm-svn: 107549
 2010-07-03 00:37:44 +00:00
Bruno Cardoso Lopes
473863e456 Simple refactoring of SSE4.1 instructions, making room for the AVX forms 
llvm-svn: 107540
 2010-07-02 23:27:59 +00:00
Bruno Cardoso Lopes
4931e183b5 - Add support for the rest of AVX SSE3 instructions 
- Fix VEX prefix to be emitted with 3 bytes whenever VEX_5M
represents a REX equivalent two byte leading opcode

llvm-svn: 107523
 2010-07-02 22:06:54 +00:00
Bruno Cardoso Lopes
c5670fcb23 Shrink down SSE3 code by more multiclass refactoring 
llvm-svn: 107448
 2010-07-01 23:10:49 +00:00
Bruno Cardoso Lopes
c215186088 Shrink down SSE3 code by some multiclass refactoring - 1st part 
llvm-svn: 107438
 2010-07-01 22:33:18 +00:00
Bruno Cardoso Lopes
511e5f47de Move SSE3 Move patterns to a more appropriate section 
Add AVX SSE3 packed horizontal and & sub instructions

llvm-svn: 107405
 2010-07-01 17:35:02 +00:00
Bruno Cardoso Lopes
0a3048e8b9 Add AVX SSE3 packed addsub instructions 
llvm-svn: 107404
 2010-07-01 17:08:18 +00:00
Bruno Cardoso Lopes
c1abe91367 Add AVX SSE3 replicate and convert instructions 
llvm-svn: 107375
 2010-07-01 02:33:39 +00:00
Bruno Cardoso Lopes
956316a3d7 - Add AVX SSE2 Move doubleword and quadword instructions. 
- Add encode bits for VEX_W
- All 128-bit SSE 1 & SSE2 instructions that are described
  in the .td file now have a AVX encoded form already working.

llvm-svn: 107365
 2010-07-01 01:20:06 +00:00
Bruno Cardoso Lopes
7ae1ebd3b4 Move MOVD/MODQ code around, creating sections for each of them 
llvm-svn: 107308
 2010-06-30 18:49:10 +00:00
Bruno Cardoso Lopes
f8855c22be Add AVX SSE2 mask creation and conditional store instructions 
llvm-svn: 107306
 2010-06-30 18:38:10 +00:00
Bruno Cardoso Lopes
6c468039a2 Fix a bug introduced in r107211 where instructions with memory operands are declared as commutable 
llvm-svn: 107300
 2010-06-30 18:06:01 +00:00
Bruno Cardoso Lopes
3c02702830 Add AVX SSE2 packed integer extract/insert instructions 
llvm-svn: 107293
 2010-06-30 17:03:03 +00:00
Bruno Cardoso Lopes
39594cc5d0 Add AVX SSE2 integer unpack instructions 
llvm-svn: 107246
 2010-06-30 04:06:39 +00:00
Bruno Cardoso Lopes
419f8f29c3 Add AVX SSE2 packed integer shuffle instructions 
llvm-svn: 107245
 2010-06-30 03:47:56 +00:00
Bruno Cardoso Lopes
c2f5cd2389 Small refactoring of SSE2 packed integer shuffle instructions 
llvm-svn: 107243
 2010-06-30 03:29:36 +00:00
Bruno Cardoso Lopes
d9acb34aa2 Add AVX SSE2 pack with saturation integer instructions 
llvm-svn: 107241
 2010-06-30 02:30:25 +00:00
Bruno Cardoso Lopes
c470ba9937 Add AVX SSE2 integer packed compare instructions 
llvm-svn: 107240
 2010-06-30 02:21:09 +00:00
Bruno Cardoso Lopes
cfbebb3921 - Add AVX form of all SSE2 logical instructions 
- Add VEX encoding bits to x86 MRM0r-MRM7r

llvm-svn: 107238
 2010-06-30 01:58:37 +00:00
Bruno Cardoso Lopes
2439877e05 Add *several* AVX integer packed binop instructions 
llvm-svn: 107225
 2010-06-29 23:47:49 +00:00
Bruno Cardoso Lopes
b80121d316 Move SSE2 Packed Integer instructions around, and create specific sections for each of them 
llvm-svn: 107211
 2010-06-29 22:12:16 +00:00