[X86] Correct the VROUND regular expressions in Znver1 scheduler model to account...

llvm-svn: 328260

diff --git a/llvm/lib/Target/X86/X86ScheduleZnver1.td b/llvm/lib/Target/X86/X86ScheduleZnver1.td
index d2f3dacf38c25788fec7f33b9e2ef8c725ad5624..2ed0008757bac5fb7b2a1ad4a3113f1aa72daa98 100644 (file)
--- a/llvm/lib/Target/X86/X86ScheduleZnver1.td
+++ b/llvm/lib/Target/X86/X86ScheduleZnver1.td
@@ -1610,14 +1610,14 @@ def : InstRW<[ZnWriteVRCPPSLd], (instregex "VRCPPSYm")>;
 def ZnWriteROUNDr : SchedWriteRes<[ZnFPU3]> {
   let Latency = 4;
 }
-def : InstRW<[ZnWriteROUNDr], (instregex "(V?)ROUND(Y?)(S|P)(S|D)r")>;
+def : InstRW<[ZnWriteROUNDr], (instregex "(V?)ROUND(S|P)(S|D)(Y?)r")>;
 
 // v,m,i.
 def ZnWriteROUNDm : SchedWriteRes<[ZnAGU, ZnFPU3]> {
   let Latency = 11;
   let NumMicroOps = 2;
 }
-def : InstRW<[ZnWriteROUNDm], (instregex "(V?)ROUND(Y?)(S|P)(S|D)m")>;
+def : InstRW<[ZnWriteROUNDm], (instregex "(V?)ROUND(S|P)(S|D)(Y?)m")>;
 
 // DPPS.
 // x,x,i / v,v,v,i.

diff --git a/llvm/test/CodeGen/X86/avx-schedule.ll b/llvm/test/CodeGen/X86/avx-schedule.ll
index 7df419d51858464985d52a56e3a8f70eb75a8a72..6a4ab314f06768a0c4a46a99dc5bae1aa31a02d7 100644 (file)
--- a/llvm/test/CodeGen/X86/avx-schedule.ll
+++ b/llvm/test/CodeGen/X86/avx-schedule.ll
@@ -4132,8 +4132,8 @@ define <4 x double> @test_roundpd(<4 x double> %a0, <4 x double> *%a1) {
 ;
 ; ZNVER1-LABEL: test_roundpd:
 ; ZNVER1:       # %bb.0:
-; ZNVER1-NEXT:    vroundpd $7, (%rdi), %ymm1 # sched: [10:1.00]
-; ZNVER1-NEXT:    vroundpd $7, %ymm0, %ymm0 # sched: [3:1.00]
+; ZNVER1-NEXT:    vroundpd $7, (%rdi), %ymm1 # sched: [11:1.00]
+; ZNVER1-NEXT:    vroundpd $7, %ymm0, %ymm0 # sched: [4:1.00]
 ; ZNVER1-NEXT:    vaddpd %ymm1, %ymm0, %ymm0 # sched: [3:1.00]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
   %1 = call <4 x double> @llvm.x86.avx.round.pd.256(<4 x double> %a0, i32 7)
@@ -4196,8 +4196,8 @@ define <8 x float> @test_roundps(<8 x float> %a0, <8 x float> *%a1) {
 ;
 ; ZNVER1-LABEL: test_roundps:
 ; ZNVER1:       # %bb.0:
-; ZNVER1-NEXT:    vroundps $7, (%rdi), %ymm1 # sched: [10:1.00]
-; ZNVER1-NEXT:    vroundps $7, %ymm0, %ymm0 # sched: [3:1.00]
+; ZNVER1-NEXT:    vroundps $7, (%rdi), %ymm1 # sched: [11:1.00]
+; ZNVER1-NEXT:    vroundps $7, %ymm0, %ymm0 # sched: [4:1.00]
 ; ZNVER1-NEXT:    vaddps %ymm1, %ymm0, %ymm0 # sched: [3:1.00]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
   %1 = call <8 x float> @llvm.x86.avx.round.ps.256(<8 x float> %a0, i32 7)