Chris Lattner [Wed, 5 Apr 2006 17:38:50 +0000 (17:38 +0000)]
new testcase that was miscompiled to vspltisb
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27438
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 5 Apr 2006 07:20:06 +0000 (07:20 +0000)]
Handle canonical form of e.g.
vector_shuffle v1, v1, <0, 4, 1, 5, 2, 6, 3, 7>
This is turned into
vector_shuffle v1, <undef>, <0, 0, 1, 1, 2, 2, 3, 3>
by dag combiner.
It would match a {p}unpckl on x86.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27437
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 06:54:42 +0000 (06:54 +0000)]
Make a vector live across blocks have the correct Vec type. This fixes
CodeGen/X86/2006-04-04-CrossBlockCrash.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27436
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 06:54:14 +0000 (06:54 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27435
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 5 Apr 2006 06:11:20 +0000 (06:11 +0000)]
Bogus assert
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27434
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 5 Apr 2006 06:09:26 +0000 (06:09 +0000)]
Fallthrough to expand if a VECTOR_SHUFFLE cannot be custom lowered.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27433
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 5 Apr 2006 06:07:11 +0000 (06:07 +0000)]
Exapnd a VECTOR_SHUFFLE to a BUILD_VECTOR if target asks for it to be expanded
or custom lowering fails.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27432
91177308-0d34-0410-b5e6-
96231b3b80d8
Jeff Cohen [Wed, 5 Apr 2006 05:19:18 +0000 (05:19 +0000)]
Fix more tablegen depedency issues in Visual Studio.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27431
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 5 Apr 2006 04:40:55 +0000 (04:40 +0000)]
Separate out to 2 test cases
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27430
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Wed, 5 Apr 2006 03:31:45 +0000 (03:31 +0000)]
make this test less exacting
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27429
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Wed, 5 Apr 2006 02:42:36 +0000 (02:42 +0000)]
revert this, this is safe, if conservative. leave a note to that effect
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27428
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 5 Apr 2006 01:47:37 +0000 (01:47 +0000)]
Handle v8i16 shuffle that must be broken into a pair of pshufhw / pshuflw.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27427
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Wed, 5 Apr 2006 01:44:57 +0000 (01:44 +0000)]
Add a new shuffle test case that requires pshuflw / pshufhw pair.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27426
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 01:16:22 +0000 (01:16 +0000)]
add vsl
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27425
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 01:15:54 +0000 (01:15 +0000)]
Get the types right, third time is the charm. Add vsl.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27424
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 00:49:48 +0000 (00:49 +0000)]
add vmladduhm
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27423
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 00:49:14 +0000 (00:49 +0000)]
correct the type of two intrinsics, add int_ppc_altivec_vmladduhm
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27422
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 00:03:57 +0000 (00:03 +0000)]
Add m[tf]vscr instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27421
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Wed, 5 Apr 2006 00:03:03 +0000 (00:03 +0000)]
Add m[tf]vscr intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27420
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 23:45:11 +0000 (23:45 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27419
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 23:43:56 +0000 (23:43 +0000)]
Add missing byte merges.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27418
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 23:25:02 +0000 (23:25 +0000)]
Add FP -> Int Conversions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27417
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 23:14:00 +0000 (23:14 +0000)]
add average intrinsics
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27416
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 23:13:21 +0000 (23:13 +0000)]
add average intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27415
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 22:43:55 +0000 (22:43 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27414
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 22:28:35 +0000 (22:28 +0000)]
Fix some broken logic that would cause us to codegen {
2147483647,
2147483647,
2147483647,
2147483647} as 'vspltisb v0, -1'.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27413
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 4 Apr 2006 21:49:39 +0000 (21:49 +0000)]
Added pslldq and psrldq.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27412
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 4 Apr 2006 21:48:31 +0000 (21:48 +0000)]
Added intrinsics to match __builtin_ia32_pslldqi128 and
__builtin_ia32_psrldqi128.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27411
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 4 Apr 2006 19:12:30 +0000 (19:12 +0000)]
Minor fixes + naming changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27410
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 19:05:42 +0000 (19:05 +0000)]
How could this ever have worked?
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27409
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 4 Apr 2006 18:40:36 +0000 (18:40 +0000)]
PSHUF* encoding bugs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27405
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 17:39:56 +0000 (17:39 +0000)]
Make sure to consider alignment of variable sized objects.
This, along with the previous dag combiner fix, fixes
CodeGen/Alpha/2006-04-04-zextload.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27403
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 17:39:18 +0000 (17:39 +0000)]
Do not create ZEXTLOAD's unless we are before legalize or the operation is
legal.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27402
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 17:38:31 +0000 (17:38 +0000)]
New testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27401
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 17:25:31 +0000 (17:25 +0000)]
Ask legalize to promote all vector shuffles to be v16i8 instead of having to
handle all 4 PPC vector types. This simplifies the matching code and allows
us to eliminate a bunch of patterns. This also adds cases we were missing,
such as CodeGen/PowerPC/vec_splat.ll:splat_h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27400
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 17:23:26 +0000 (17:23 +0000)]
* Add supprot for SCALAR_TO_VECTOR operations where the input needs to be
promoted/expanded (e.g. SCALAR_TO_VECTOR from i8/i16 on PPC).
* Add support for targets to request that VECTOR_SHUFFLE nodes be promoted
to a canonical type, for example, we only want v16i8 shuffles on PPC.
* Move isShuffleLegal out of TLI into Legalize.
* Teach isShuffleLegal to allow shuffles that need to be promoted.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27399
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 17:21:22 +0000 (17:21 +0000)]
Move isShuffleLegal from TLI to Legalize.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27398
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 17:20:45 +0000 (17:20 +0000)]
new testcase
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27397
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 06:11:42 +0000 (06:11 +0000)]
Signed shr by a constant is not the same as sdiv by 2^k
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27395
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 4 Apr 2006 03:04:07 +0000 (03:04 +0000)]
cmpps / cmppd encoding bug
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27393
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 01:40:06 +0000 (01:40 +0000)]
Fix the types for these intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27392
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 01:02:22 +0000 (01:02 +0000)]
Constant fold bitconvert(undef)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27391
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 00:25:10 +0000 (00:25 +0000)]
Allow targets to have fine grained control over which types various ops get
promoted to, if they desire.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27389
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Tue, 4 Apr 2006 00:10:53 +0000 (00:10 +0000)]
Compact some intrinsic definitions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27388
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Tue, 4 Apr 2006 00:05:13 +0000 (00:05 +0000)]
Plug in the byte and short splats
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27387
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 23:58:04 +0000 (23:58 +0000)]
Revert accidentally committed hunks.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27386
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 23:55:43 +0000 (23:55 +0000)]
Make sure to mark unsupported SCALAR_TO_VECTOR operations as expand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27385
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 3 Apr 2006 23:49:17 +0000 (23:49 +0000)]
Some SSE1 intrinsics: min, max, sqrt, etc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27384
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 23:14:49 +0000 (23:14 +0000)]
revert previous patch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27383
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 3 Apr 2006 22:30:54 +0000 (22:30 +0000)]
Use movlpd to: store lower f64 extracted from v2f64.
Use movhpd to: store upper f64 extracted from v2f64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27382
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 22:03:29 +0000 (22:03 +0000)]
Force use of a frame-pointer if there is anything on the stack that is aligned
more than the OS keeps the stack aligned.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27381
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 21:39:57 +0000 (21:39 +0000)]
The stack alignment is now computed dynamically, just verify it is correct.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27380
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 21:39:03 +0000 (21:39 +0000)]
Remove unused method
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27379
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 21:38:39 +0000 (21:38 +0000)]
Keep track of max stack alignment as objects are added. Remove an obsolete method.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27378
91177308-0d34-0410-b5e6-
96231b3b80d8
Evan Cheng [Mon, 3 Apr 2006 20:53:28 +0000 (20:53 +0000)]
- More efficient extract_vector_elt with shuffle and movss, movsd, movd, etc.
- Some bug fixes and naming inconsistency fixes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27377
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 19:28:50 +0000 (19:28 +0000)]
Align vectors to the size in bytes, not bits.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27376
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 17:29:28 +0000 (17:29 +0000)]
Add a missing check, this fixes UnitTests/Vector/sumarray.c
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27375
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 17:21:50 +0000 (17:21 +0000)]
Add a missing check, which broke a bunch of vector tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27374
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 17:20:06 +0000 (17:20 +0000)]
shrinkify intrinsics more by using some local classes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27373
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 15:58:28 +0000 (15:58 +0000)]
Add the full set of min/max instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27372
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Mon, 3 Apr 2006 15:43:07 +0000 (15:43 +0000)]
Add some classes to make it easier to define intrinsics. Add min/max intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27371
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Mon, 3 Apr 2006 04:19:17 +0000 (04:19 +0000)]
support x * (c1 + c2) where c1 and c2 are pow2s. special case for c2 == 4
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27370
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Mon, 3 Apr 2006 04:14:39 +0000 (04:14 +0000)]
test powers of 2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27369
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Mon, 3 Apr 2006 03:18:59 +0000 (03:18 +0000)]
mul by const conversion sequences. more coming soon
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27368
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Mon, 3 Apr 2006 03:16:50 +0000 (03:16 +0000)]
back this out
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27367
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Mon, 3 Apr 2006 03:16:09 +0000 (03:16 +0000)]
test some more mul by constant removal
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27366
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Sun, 2 Apr 2006 21:47:07 +0000 (21:47 +0000)]
Make sure mul by constant 5 is turned into a s4addq
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27365
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Sun, 2 Apr 2006 21:42:45 +0000 (21:42 +0000)]
This should be a win of every arch
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27364
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Sun, 2 Apr 2006 21:08:39 +0000 (21:08 +0000)]
This makes McCat/12-IOtest go 8x faster or so
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27363
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Lenharth [Sun, 2 Apr 2006 20:13:57 +0000 (20:13 +0000)]
This will be needed soon
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27362
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Spencer [Sun, 2 Apr 2006 14:34:26 +0000 (14:34 +0000)]
For PR722:
Change the check for llvm-gcc from using LLVMGCCDIR to LLVMGCC. This checks
for the actual tool rather than the directory in which the tool resides. In
the case of this bug, it is possible that the directory exists but that the
tools in that directory do not. This fix should avoid the makefile from
erroneously proceeding without the actual tools being available.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27361
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 07:20:00 +0000 (07:20 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27360
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 06:26:07 +0000 (06:26 +0000)]
Inform the dag combiner that the predicate compares only return a low bit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27359
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 06:19:46 +0000 (06:19 +0000)]
relax assertion
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27358
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 06:15:09 +0000 (06:15 +0000)]
Allow targets to compute masked bits for intrinsics.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27357
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 06:11:11 +0000 (06:11 +0000)]
Add a little dag combine to compile this:
int %AreSecondAndThirdElementsBothNegative(<4 x float>* %in) {
entry:
%tmp1 = load <4 x float>* %in ; <<4 x float>> [#uses=1]
%tmp = tail call int %llvm.ppc.altivec.vcmpgefp.p( int 1, <4 x float> < float 0x7FF8000000000000, float 0.
000000e+00, float 0.
000000e+00, float 0x7FF8000000000000 >, <4 x float> %tmp1 ) ; <int> [#uses=1]
%tmp = seteq int %tmp, 0 ; <bool> [#uses=1]
%tmp3 = cast bool %tmp to int ; <int> [#uses=1]
ret int %tmp3
}
into this:
_AreSecondAndThirdElementsBothNegative:
mfspr r2, 256
oris r4, r2, 49152
mtspr 256, r4
li r4, lo16(LCPI1_0)
lis r5, ha16(LCPI1_0)
lvx v0, 0, r3
lvx v1, r5, r4
vcmpgefp. v0, v1, v0
mfcr r3, 2
rlwinm r3, r3, 27, 31, 31
mtspr 256, r2
blr
instead of this:
_AreSecondAndThirdElementsBothNegative:
mfspr r2, 256
oris r4, r2, 49152
mtspr 256, r4
li r4, lo16(LCPI1_0)
lis r5, ha16(LCPI1_0)
lvx v0, 0, r3
lvx v1, r5, r4
vcmpgefp. v0, v1, v0
mfcr r3, 2
rlwinm r3, r3, 27, 31, 31
xori r3, r3, 1
cntlzw r3, r3
srwi r3, r3, 5
mtspr 256, r2
blr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27356
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 05:43:13 +0000 (05:43 +0000)]
vector casts of casts are eliminable. Transform this:
%tmp = cast <4 x uint> %tmp to <4 x int> ; <<4 x int>> [#uses=1]
%tmp = cast <4 x int> %tmp to <4 x float> ; <<4 x float>> [#uses=1]
into:
%tmp = cast <4 x uint> %tmp to <4 x float> ; <<4 x float>> [#uses=1]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27355
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 05:40:28 +0000 (05:40 +0000)]
vector casts never reinterpret bits
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27354
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 05:37:12 +0000 (05:37 +0000)]
Allow transforming this:
%tmp = cast <4 x uint>* %testData to <4 x int>* ; <<4 x int>*> [#uses=1]
%tmp = load <4 x int>* %tmp ; <<4 x int>> [#uses=1]
to this:
%tmp = load <4 x uint>* %testData ; <<4 x uint>> [#uses=1]
%tmp = cast <4 x uint> %tmp to <4 x int> ; <<4 x int>> [#uses=1]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27353
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 05:30:25 +0000 (05:30 +0000)]
Turn altivec lvx/stvx intrinsics into loads and stores. This allows the
elimination of one load from this:
int AreSecondAndThirdElementsBothNegative( vector float *in ) {
#define QNaN 0x7FC00000
const vector unsigned int testData = (vector unsigned int)( QNaN, 0, 0, QNaN );
vector float test = vec_ld( 0, (float*) &testData );
return ! vec_any_ge( test, *in );
}
Now generating:
_AreSecondAndThirdElementsBothNegative:
mfspr r2, 256
oris r4, r2, 49152
mtspr 256, r4
li r4, lo16(LCPI1_0)
lis r5, ha16(LCPI1_0)
addi r6, r1, -16
lvx v0, r5, r4
stvx v0, 0, r6
lvx v1, 0, r3
vcmpgefp. v0, v0, v1
mfcr r3, 2
rlwinm r3, r3, 27, 31, 31
xori r3, r3, 1
cntlzw r3, r3
srwi r3, r3, 5
mtspr 256, r2
blr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27352
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 05:28:54 +0000 (05:28 +0000)]
Remove done item
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27351
91177308-0d34-0410-b5e6-
96231b3b80d8
Jeff Cohen [Sun, 2 Apr 2006 05:20:53 +0000 (05:20 +0000)]
Fix tablegen related dependencies in Visual Studio.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27350
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 05:06:04 +0000 (05:06 +0000)]
Implement promotion for EXTRACT_VECTOR_ELT, allowing v16i8 multiplies to work with PowerPC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27349
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 03:59:11 +0000 (03:59 +0000)]
add a note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27348
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 03:57:31 +0000 (03:57 +0000)]
Implement the Expand action for binary vector operations to break the binop
into elements and operate on each piece. This allows generic vector integer
multiplies to work on PPC, though the generated code is horrible.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27347
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 03:41:14 +0000 (03:41 +0000)]
Intrinsics that just load from memory can be treated like loads: they don't
have to serialize against each other. This allows us to schedule lvx's
across each other, for example.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27346
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 03:35:30 +0000 (03:35 +0000)]
Adjust the Intrinsics.gen interface a little bit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27345
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 03:35:01 +0000 (03:35 +0000)]
Adjust to change in Intrinsics.gen interface.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27344
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 03:25:57 +0000 (03:25 +0000)]
Constant fold all of the vector binops. This allows us to compile this:
"vector unsigned char mergeLowHigh = (vector unsigned char)
( 8, 9, 10, 11, 16, 17, 18, 19, 12, 13, 14, 15, 20, 21, 22, 23 );
vector unsigned char mergeHighLow = vec_xor( mergeLowHigh, vec_splat_u8(8));"
aka:
void %test2(<16 x sbyte>* %P) {
store <16 x sbyte> cast (<4 x int> xor (<4 x int> cast (<16 x ubyte> < ubyte 8, ubyte 9, ubyte 10, ubyte 11, ubyte 16, ubyte 17, ubyte 18, ubyte 19, ubyte 12, ubyte 13, ubyte 14, ubyte 15, ubyte 20, ubyte 21, ubyte 22, ubyte 23 > to <4 x int>), <4 x int> cast (<16 x sbyte> < sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8, sbyte 8 > to <4 x int>)) to <16 x sbyte>), <16 x sbyte> * %P
ret void
}
into this:
_test2:
mfspr r2, 256
oris r4, r2, 32768
mtspr 256, r4
li r4, lo16(LCPI2_0)
lis r5, ha16(LCPI2_0)
lvx v0, r5, r4
stvx v0, 0, r3
mtspr 256, r2
blr
instead of this:
_test2:
mfspr r2, 256
oris r4, r2, 49152
mtspr 256, r4
li r4, lo16(LCPI2_0)
lis r5, ha16(LCPI2_0)
vspltisb v0, 8
lvx v1, r5, r4
vxor v0, v1, v0
stvx v0, 0, r3
mtspr 256, r2
blr
... which occurs here:
http://developer.apple.com/hardware/ve/calcspeed.html
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27343
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 03:07:27 +0000 (03:07 +0000)]
Add a new -view-legalize-dags command line option
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27342
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 02:53:43 +0000 (02:53 +0000)]
Implement constant folding of bit_convert of arbitrary constant vbuild_vector nodes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27341
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 02:51:27 +0000 (02:51 +0000)]
These entries already exist
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27340
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 02:41:18 +0000 (02:41 +0000)]
Add some missing node names
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27339
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 02:28:52 +0000 (02:28 +0000)]
simplify this method
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27338
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 01:47:20 +0000 (01:47 +0000)]
New note
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27337
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 01:38:28 +0000 (01:38 +0000)]
Constant fold casts from things like <4 x int> -> <4 x uint>, likewise int<->fp.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27336
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 00:43:36 +0000 (00:43 +0000)]
Custom lower all BUILD_VECTOR's so that we can compile vec_splat_u8(8) into
"vspltisb v0, 8" instead of a constant pool load.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27335
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 00:24:45 +0000 (00:24 +0000)]
Prefer larger register classes over smaller ones when a register occurs in
multiple register classes. This fixes PowerPC/2006-04-01-FloatDoubleExtend.ll
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27334
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sun, 2 Apr 2006 00:23:59 +0000 (00:23 +0000)]
New testcase that crashes the compiler.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27333
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Lattner [Sat, 1 Apr 2006 23:17:11 +0000 (23:17 +0000)]
add valuemapper support for inline asm
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@27332
91177308-0d34-0410-b5e6-
96231b3b80d8