[llvm-commits] CVS: llvm/lib/Target/X86/X86InstrSSE.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/lib/Target/X86: X86InstrSSE.td updated: 1.48 -> 1.49 --- Log message: More logical ops patterns --- Diffs of the changes: (+106 -0) X86InstrSSE.td | 106 + 1 files changed, 106 insertions(+) Index: llvm/lib

[llvm-commits] CVS: llvm/include/llvm/IntrinsicsX86.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/include/llvm: IntrinsicsX86.td updated: 1.6 -> 1.7 --- Log message: Add support for _mm_cmp{cc}_ss and _mm_cmp{cc}_ps intrinsics --- Diffs of the changes: (+4 -106) IntrinsicsX86.td | 110 ++- 1 files changed, 4

[llvm-commits] CVS: llvm/lib/Target/X86/X86InstrSSE.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/lib/Target/X86: X86InstrSSE.td updated: 1.47 -> 1.48 --- Log message: Add support for _mm_cmp{cc}_ss and _mm_cmp{cc}_ps intrinsics --- Diffs of the changes: (+35 -9) X86InstrSSE.td | 44 +++- 1 files changed, 35 insertions(

[llvm-commits] CVS: llvm/include/llvm/IntrinsicsX86.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/include/llvm: IntrinsicsX86.td updated: 1.5 -> 1.6 --- Log message: Add 128-bit pmovmskb intrinsic support. --- Diffs of the changes: (+2 -0) IntrinsicsX86.td |2 ++ 1 files changed, 2 insertions(+) Index: llvm/include/llvm/IntrinsicsX86.td diff -u llvm/incl

[llvm-commits] CVS: llvm/lib/Target/X86/X86InstrSSE.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/lib/Target/X86: X86InstrSSE.td updated: 1.46 -> 1.47 --- Log message: Add 128-bit pmovmskb intrinsic support. --- Diffs of the changes: (+12 -7) X86InstrSSE.td | 19 --- 1 files changed, 12 insertions(+), 7 deletions(-) Index: llvm/lib/Target/X

[llvm-commits] CVS: llvm/include/llvm/IntrinsicsX86.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/include/llvm: IntrinsicsX86.td updated: 1.4 -> 1.5 --- Log message: Change SSE pack operation definitions to fit what the intrinsics expected. For example, packsswb actually creates a v16i8 from a pair of v8i16. But since the intrinsic specification forces the output t

[llvm-commits] CVS: llvm/lib/Target/X86/X86InstrSSE.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/lib/Target/X86: X86InstrSSE.td updated: 1.45 -> 1.46 --- Log message: Change SSE pack operation definitions to fit what the intrinsics expected. For example, packsswb actually creates a v16i8 from a pair of v8i16. But since the intrinsic specification forces the output

[llvm-commits] CVS: llvm/include/llvm/IntrinsicsX86.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/include/llvm: IntrinsicsX86.td updated: 1.3 -> 1.4 --- Log message: Add SSE2 integer pack with saturation intrinsics. --- Diffs of the changes: (+10 -0) IntrinsicsX86.td | 10 ++ 1 files changed, 10 insertions(+) Index: llvm/include/llvm/IntrinsicsX86.

[llvm-commits] CVS: llvm/lib/Target/X86/X86ISelLowering.cpp X86ISelLowering.h X86InstrSSE.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/lib/Target/X86: X86ISelLowering.cpp updated: 1.144 -> 1.145 X86ISelLowering.h updated: 1.47 -> 1.48 X86InstrSSE.td updated: 1.44 -> 1.45 --- Log message: - Added some SSE2 128-bit packed integer ops. - Added SSE2 128-bit integer pack with signed saturation ops. - Added

[llvm-commits] CVS: llvm/test/Regression/CodeGen/X86/vec_shuffle.ll

2006-03-29 Thread Evan Cheng
Changes in directory llvm/test/Regression/CodeGen/X86: vec_shuffle.ll updated: 1.3 -> 1.4 --- Log message: Add a pshufhw test case. --- Diffs of the changes: (+25 -0) vec_shuffle.ll | 25 + 1 files changed, 25 insertions(+) Index: llvm/test/Regression/CodeGen/X8

[llvm-commits] CVS: llvm/lib/Target/X86/X86ISelLowering.cpp X86InstrSSE.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/lib/Target/X86: X86ISelLowering.cpp updated: 1.143 -> 1.144 X86InstrSSE.td updated: 1.43 -> 1.44 --- Log message: Need to special case splat after all. Make the second operand of splat vector_shuffle undef. --- Diffs of the changes: (+18 -15) X86ISelLowering.cpp |

[llvm-commits] CVS: llvm/test/Regression/CodeGen/X86/vec_splat.ll

2006-03-29 Thread Evan Cheng
Changes in directory llvm/test/Regression/CodeGen/X86: vec_splat.ll updated: 1.1 -> 1.2 --- Log message: Use unpcklpd for v2f64 splat. --- Diffs of the changes: (+1 -1) vec_splat.ll |2 +- 1 files changed, 1 insertion(+), 1 deletion(-) Index: llvm/test/Regression/CodeGen/X86/vec_spla

[llvm-commits] CVS: llvm/lib/Target/X86/X86InstrSSE.td

2006-03-29 Thread Evan Cheng
Changes in directory llvm/lib/Target/X86: X86InstrSSE.td updated: 1.42 -> 1.43 --- Log message: Floating point logical operation patterns should match bit_convert. Or else integer vector logical operations would match andp{s|d} instead of pand. --- Diffs of the changes: (+53 -29) X86InstrS