summaryrefslogtreecommitdiffstats
path: root/llvm/test/CodeGen/Mips/msa/compare.ll
diff options
context:
space:
mode:
authorDaniel Sanders <daniel.sanders@imgtec.com>2014-03-12 11:54:00 +0000
committerDaniel Sanders <daniel.sanders@imgtec.com>2014-03-12 11:54:00 +0000
commitdf221545793e222ab7870054eb7a220183f6ce7a (patch)
tree5af6bce093b816aaaac2099f6fd17746454c7323 /llvm/test/CodeGen/Mips/msa/compare.ll
parentbd58580cb81ed26eb0ae6ffb3270adfb27b44f6e (diff)
downloadbcm5719-llvm-df221545793e222ab7870054eb7a220183f6ce7a.tar.gz
bcm5719-llvm-df221545793e222ab7870054eb7a220183f6ce7a.zip
[mips] BSEL's and BINS[RL] operands are reversed compared to the vselect node used in the pattern.
Summary: Correct the match patterns and the lowerings that made the CodeGen tests pass despite the mistakes. The original testcase that discovered the problem was SingleSource/UnitTests/SignlessType/factor.c in test-suite. During review, we also found that some of the existing CodeGen tests were incorrect and fixed them: * bitwise.ll: In bsel_v16i8 the IfSet/IfClear were reversed because bsel and bmnz have different operand orders and the test didn't correctly account for this. bmnz goes 'IfClear, IfSet, CondMask', while bsel goes 'CondMask, IfClear, IfSet'. * vec.ll: In the cases where a bsel is emitted as a bmnz (they are the same operation with a different input tied to the result) the operands were in the wrong order. * compare.ll and compare_float.ll: The bsel operand order was correct for a greater-than comparison, but a greater-than comparison instruction doesn't exist. Lowering this operation inverts the condition so the IfSet/IfClear need to be swapped to match. The differences between BSEL, BMNZ, and BMZ and how they map to/from vselect are rather confusing. I've therefore added a note to MSA.txt to explain this in a single place in addition to the comments that explain each case. Reviewers: matheusalmeida, jacksprat Reviewed By: matheusalmeida Differential Revision: http://llvm-reviews.chandlerc.com/D3028 llvm-svn: 203657
Diffstat (limited to 'llvm/test/CodeGen/Mips/msa/compare.ll')
-rw-r--r--llvm/test/CodeGen/Mips/msa/compare.ll34
1 files changed, 20 insertions, 14 deletions
diff --git a/llvm/test/CodeGen/Mips/msa/compare.ll b/llvm/test/CodeGen/Mips/msa/compare.ll
index 6408d7ba09f..87ca1482da8 100644
--- a/llvm/test/CodeGen/Mips/msa/compare.ll
+++ b/llvm/test/CodeGen/Mips/msa/compare.ll
@@ -761,7 +761,8 @@ define void @bsel_s_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
%4 = icmp sgt <8 x i16> %1, %2
; CHECK-DAG: clt_s.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
%5 = select <8 x i1> %4, <8 x i16> %1, <8 x i16> %3
- ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+ ; Note that IfSet and IfClr are swapped since the condition is inverted
+ ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
store <8 x i16> %5, <8 x i16>* %d
; CHECK-DAG: st.h [[R4]], 0($4)
@@ -782,7 +783,8 @@ define void @bsel_s_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
%4 = icmp sgt <4 x i32> %1, %2
; CHECK-DAG: clt_s.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
%5 = select <4 x i1> %4, <4 x i32> %1, <4 x i32> %3
- ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+ ; Note that IfSet and IfClr are swapped since the condition is inverted
+ ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
store <4 x i32> %5, <4 x i32>* %d
; CHECK-DAG: st.w [[R4]], 0($4)
@@ -803,7 +805,8 @@ define void @bsel_s_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
%4 = icmp sgt <2 x i64> %1, %2
; CHECK-DAG: clt_s.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
%5 = select <2 x i1> %4, <2 x i64> %1, <2 x i64> %3
- ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+ ; Note that IfSet and IfClr are swapped since the condition is inverted
+ ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
store <2 x i64> %5, <2 x i64>* %d
; CHECK-DAG: st.d [[R4]], 0($4)
@@ -846,7 +849,8 @@ define void @bsel_u_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
%4 = icmp ugt <8 x i16> %1, %2
; CHECK-DAG: clt_u.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
%5 = select <8 x i1> %4, <8 x i16> %1, <8 x i16> %3
- ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+ ; Note that IfSet and IfClr are swapped since the condition is inverted
+ ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
store <8 x i16> %5, <8 x i16>* %d
; CHECK-DAG: st.h [[R4]], 0($4)
@@ -867,7 +871,8 @@ define void @bsel_u_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
%4 = icmp ugt <4 x i32> %1, %2
; CHECK-DAG: clt_u.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
%5 = select <4 x i1> %4, <4 x i32> %1, <4 x i32> %3
- ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+ ; Note that IfSet and IfClr are swapped since the condition is inverted
+ ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
store <4 x i32> %5, <4 x i32>* %d
; CHECK-DAG: st.w [[R4]], 0($4)
@@ -888,7 +893,8 @@ define void @bsel_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
%4 = icmp ugt <2 x i64> %1, %2
; CHECK-DAG: clt_u.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
%5 = select <2 x i1> %4, <2 x i64> %1, <2 x i64> %3
- ; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
+ ; Note that IfSet and IfClr are swapped since the condition is inverted
+ ; CHECK-DAG: bsel.v [[R4]], [[R3]], [[R1]]
store <2 x i64> %5, <2 x i64>* %d
; CHECK-DAG: st.d [[R4]], 0($4)
@@ -906,7 +912,7 @@ define void @bseli_s_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
%3 = icmp sgt <16 x i8> %1, %2
; CHECK-DAG: clt_s.b [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
+ %4 = select <16 x i1> %3, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, <16 x i8> %1
; CHECK-DAG: bseli.b [[R4]], [[R1]], 1
store <16 x i8> %4, <16 x i8>* %d
; CHECK-DAG: st.b [[R4]], 0($4)
@@ -925,7 +931,7 @@ define void @bseli_s_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
%3 = icmp sgt <8 x i16> %1, %2
; CHECK-DAG: clt_s.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
+ %4 = select <8 x i1> %3, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, <8 x i16> %1
; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 1
; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
store <8 x i16> %4, <8 x i16>* %d
@@ -945,7 +951,7 @@ define void @bseli_s_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
%3 = icmp sgt <4 x i32> %1, %2
; CHECK-DAG: clt_s.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
+ %4 = select <4 x i1> %3, <4 x i32> <i32 1, i32 1, i32 1, i32 1>, <4 x i32> %1
; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 1
; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
store <4 x i32> %4, <4 x i32>* %d
@@ -965,7 +971,7 @@ define void @bseli_s_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
%3 = icmp sgt <2 x i64> %1, %2
; CHECK-DAG: clt_s.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
+ %4 = select <2 x i1> %3, <2 x i64> <i64 1, i64 1>, <2 x i64> %1
; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 1
; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
store <2 x i64> %4, <2 x i64>* %d
@@ -985,7 +991,7 @@ define void @bseli_u_v16i8(<16 x i8>* %d, <16 x i8>* %a, <16 x i8>* %b,
; CHECK-DAG: ld.b [[R2:\$w[0-9]+]], 0($6)
%3 = icmp ugt <16 x i8> %1, %2
; CHECK-DAG: clt_u.b [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <16 x i1> %3, <16 x i8> %1, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>
+ %4 = select <16 x i1> %3, <16 x i8> <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1>, <16 x i8> %1
; CHECK-DAG: bseli.b [[R4]], [[R1]], 1
store <16 x i8> %4, <16 x i8>* %d
; CHECK-DAG: st.b [[R4]], 0($4)
@@ -1004,7 +1010,7 @@ define void @bseli_u_v8i16(<8 x i16>* %d, <8 x i16>* %a, <8 x i16>* %b,
; CHECK-DAG: ld.h [[R2:\$w[0-9]+]], 0($6)
%3 = icmp ugt <8 x i16> %1, %2
; CHECK-DAG: clt_u.h [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <8 x i1> %3, <8 x i16> %1, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>
+ %4 = select <8 x i1> %3, <8 x i16> <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1>, <8 x i16> %1
; CHECK-DAG: ldi.h [[R3:\$w[0-9]+]], 1
; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
store <8 x i16> %4, <8 x i16>* %d
@@ -1024,7 +1030,7 @@ define void @bseli_u_v4i32(<4 x i32>* %d, <4 x i32>* %a, <4 x i32>* %b,
; CHECK-DAG: ld.w [[R2:\$w[0-9]+]], 0($6)
%3 = icmp ugt <4 x i32> %1, %2
; CHECK-DAG: clt_u.w [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <4 x i1> %3, <4 x i32> %1, <4 x i32> <i32 1, i32 1, i32 1, i32 1>
+ %4 = select <4 x i1> %3, <4 x i32> <i32 1, i32 1, i32 1, i32 1>, <4 x i32> %1
; CHECK-DAG: ldi.w [[R3:\$w[0-9]+]], 1
; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
store <4 x i32> %4, <4 x i32>* %d
@@ -1044,7 +1050,7 @@ define void @bseli_u_v2i64(<2 x i64>* %d, <2 x i64>* %a, <2 x i64>* %b,
; CHECK-DAG: ld.d [[R2:\$w[0-9]+]], 0($6)
%3 = icmp ugt <2 x i64> %1, %2
; CHECK-DAG: clt_u.d [[R4:\$w[0-9]+]], [[R2]], [[R1]]
- %4 = select <2 x i1> %3, <2 x i64> %1, <2 x i64> <i64 1, i64 1>
+ %4 = select <2 x i1> %3, <2 x i64> <i64 1, i64 1>, <2 x i64> %1
; CHECK-DAG: ldi.d [[R3:\$w[0-9]+]], 1
; CHECK-DAG: bsel.v [[R4]], [[R1]], [[R3]]
store <2 x i64> %4, <2 x i64>* %d
OpenPOWER on IntegriCloud