AMD family 17h (znver1) scheduler model update.
Summary:
This patch enables the following:
1) Regex based Instruction itineraries for integer instructions.
2) The instructions are grouped as per the nature of the instructions
(move, arithmetic, logic, Misc, Control Transfer).
3) FP instructions and their itineraries are added which includes values
for SSE4A, BMI, BMI2 and SHA instructions.
Patch by Ganesh Gopalasubramanian
Reviewers: RKSimon, craig.topper
Subscribers: vprasad, shivaram, ddibyend, andreadb, javed.absar, llvm-commits
Differential Revision: https://reviews.llvm.org/D36617
llvm-svn: 312237
diff --git a/llvm/test/CodeGen/X86/sse-schedule.ll b/llvm/test/CodeGen/X86/sse-schedule.ll
index e982cc9..03fb192 100644
--- a/llvm/test/CodeGen/X86/sse-schedule.ll
+++ b/llvm/test/CodeGen/X86/sse-schedule.ll
@@ -56,7 +56,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vaddps (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fadd <4 x float> %a0, %a1
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = fadd <4 x float> %1, %2
@@ -110,7 +110,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vaddss %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vaddss (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fadd float %a0, %a1
%2 = load float, float *%a2, align 4
%3 = fadd float %1, %2
@@ -168,7 +168,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vandps %xmm1, %xmm0, %xmm0 # sched: [1:0.25]
; ZNVER1-NEXT: vandps (%rdi), %xmm0, %xmm0 # sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = bitcast <4 x float> %a0 to <4 x i32>
%2 = bitcast <4 x float> %a1 to <4 x i32>
%3 = and <4 x i32> %1, %2
@@ -230,7 +230,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vandnps %xmm1, %xmm0, %xmm0 # sched: [1:0.25]
; ZNVER1-NEXT: vandnps (%rdi), %xmm0, %xmm0 # sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = bitcast <4 x float> %a0 to <4 x i32>
%2 = bitcast <4 x float> %a1 to <4 x i32>
%3 = xor <4 x i32> %1, <i32 -1, i32 -1, i32 -1, i32 -1>
@@ -298,7 +298,7 @@
; ZNVER1-NEXT: vcmpeqps %xmm1, %xmm0, %xmm1 # sched: [3:1.00]
; ZNVER1-NEXT: vcmpeqps (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
; ZNVER1-NEXT: vorps %xmm0, %xmm1, %xmm0 # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fcmp oeq <4 x float> %a0, %a1
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = fcmp oeq <4 x float> %a0, %2
@@ -355,7 +355,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vcmpeqss %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vcmpeqss (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = insertelement <4 x float> undef, float %a0, i32 0
%2 = insertelement <4 x float> undef, float %a1, i32 0
%3 = call <4 x float> @llvm.x86.sse.cmp.ss(<4 x float> %1, <4 x float> %2, i8 0)
@@ -478,7 +478,7 @@
; ZNVER1-NEXT: andb %al, %dl # sched: [1:0.25]
; ZNVER1-NEXT: orb %cl, %dl # sched: [1:0.25]
; ZNVER1-NEXT: movzbl %dl, %eax # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call i32 @llvm.x86.sse.comieq.ss(<4 x float> %a0, <4 x float> %a1)
%2 = load <4 x float>, <4 x float> *%a2, align 4
%3 = call i32 @llvm.x86.sse.comieq.ss(<4 x float> %a0, <4 x float> %2)
@@ -542,7 +542,7 @@
; ZNVER1-NEXT: vcvtsi2ssl %edi, %xmm0, %xmm0 # sched: [5:1.00]
; ZNVER1-NEXT: vcvtsi2ssl (%rsi), %xmm1, %xmm1 # sched: [12:1.00]
; ZNVER1-NEXT: vaddss %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = sitofp i32 %a0 to float
%2 = load i32, i32 *%a1, align 4
%3 = sitofp i32 %2 to float
@@ -605,7 +605,7 @@
; ZNVER1-NEXT: vcvtsi2ssq %rdi, %xmm0, %xmm0 # sched: [5:1.00]
; ZNVER1-NEXT: vcvtsi2ssq (%rsi), %xmm1, %xmm1 # sched: [12:1.00]
; ZNVER1-NEXT: vaddss %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = sitofp i64 %a0 to float
%2 = load i64, i64 *%a1, align 8
%3 = sitofp i64 %2 to float
@@ -668,7 +668,7 @@
; ZNVER1-NEXT: vcvtss2si (%rdi), %eax # sched: [12:1.00]
; ZNVER1-NEXT: vcvtss2si %xmm0, %ecx # sched: [5:1.00]
; ZNVER1-NEXT: addl %ecx, %eax # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = insertelement <4 x float> undef, float %a0, i32 0
%2 = call i32 @llvm.x86.sse.cvtss2si(<4 x float> %1)
%3 = load float, float *%a1, align 4
@@ -734,7 +734,7 @@
; ZNVER1-NEXT: vcvtss2si (%rdi), %rax # sched: [12:1.00]
; ZNVER1-NEXT: vcvtss2si %xmm0, %rcx # sched: [5:1.00]
; ZNVER1-NEXT: addq %rcx, %rax # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = insertelement <4 x float> undef, float %a0, i32 0
%2 = call i64 @llvm.x86.sse.cvtss2si64(<4 x float> %1)
%3 = load float, float *%a1, align 4
@@ -800,7 +800,7 @@
; ZNVER1-NEXT: vcvttss2si (%rdi), %eax # sched: [12:1.00]
; ZNVER1-NEXT: vcvttss2si %xmm0, %ecx # sched: [5:1.00]
; ZNVER1-NEXT: addl %ecx, %eax # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fptosi float %a0 to i32
%2 = load float, float *%a1, align 4
%3 = fptosi float %2 to i32
@@ -863,7 +863,7 @@
; ZNVER1-NEXT: vcvttss2si (%rdi), %rax # sched: [12:1.00]
; ZNVER1-NEXT: vcvttss2si %xmm0, %rcx # sched: [5:1.00]
; ZNVER1-NEXT: addq %rcx, %rax # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fptosi float %a0 to i64
%2 = load float, float *%a1, align 4
%3 = fptosi float %2 to i64
@@ -918,7 +918,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vdivps %xmm1, %xmm0, %xmm0 # sched: [15:1.00]
; ZNVER1-NEXT: vdivps (%rdi), %xmm0, %xmm0 # sched: [22:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fdiv <4 x float> %a0, %a1
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = fdiv <4 x float> %1, %2
@@ -972,7 +972,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vdivss %xmm1, %xmm0, %xmm0 # sched: [15:1.00]
; ZNVER1-NEXT: vdivss (%rdi), %xmm0, %xmm0 # sched: [22:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fdiv float %a0, %a1
%2 = load float, float *%a2, align 4
%3 = fdiv float %1, %2
@@ -1025,8 +1025,8 @@
; ZNVER1-LABEL: test_ldmxcsr:
; ZNVER1: # BB#0:
; ZNVER1-NEXT: movl %edi, -{{[0-9]+}}(%rsp) # sched: [1:0.50]
-; ZNVER1-NEXT: vldmxcsr -{{[0-9]+}}(%rsp) # sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: vldmxcsr -{{[0-9]+}}(%rsp) # sched: [100:?]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = alloca i32, align 4
%2 = bitcast i32* %1 to i8*
store i32 %a0, i32* %1
@@ -1082,7 +1082,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vmaxps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vmaxps (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %a0, <4 x float> %a1)
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = call <4 x float> @llvm.x86.sse.max.ps(<4 x float> %1, <4 x float> %2)
@@ -1137,7 +1137,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vmaxss %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vmaxss (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.max.ss(<4 x float> %a0, <4 x float> %a1)
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = call <4 x float> @llvm.x86.sse.max.ss(<4 x float> %1, <4 x float> %2)
@@ -1192,7 +1192,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vminps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vminps (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %a0, <4 x float> %a1)
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = call <4 x float> @llvm.x86.sse.min.ps(<4 x float> %1, <4 x float> %2)
@@ -1247,7 +1247,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vminss %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vminss (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.min.ss(<4 x float> %a0, <4 x float> %a1)
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = call <4 x float> @llvm.x86.sse.min.ss(<4 x float> %1, <4 x float> %2)
@@ -1310,7 +1310,7 @@
; ZNVER1-NEXT: vmovaps (%rdi), %xmm0 # sched: [8:0.50]
; ZNVER1-NEXT: vaddps %xmm0, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vmovaps %xmm0, (%rsi) # sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = load <4 x float>, <4 x float> *%a0, align 16
%2 = fadd <4 x float> %1, %1
store <4 x float> %2, <4 x float> *%a1, align 16
@@ -1364,7 +1364,7 @@
; ZNVER1-LABEL: test_movhlps:
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vunpckhpd {{.*#+}} xmm0 = xmm1[1],xmm0[1] sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = shufflevector <4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 6, i32 7, i32 2, i32 3>
ret <4 x float> %1
}
@@ -1428,7 +1428,7 @@
; ZNVER1-NEXT: vmovhpd {{.*#+}} xmm1 = xmm1[0],mem[0] sched: [8:0.50]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vpextrq $1, %xmm0, (%rdi) # sched: [8:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = bitcast x86_mmx* %a2 to <2 x float>*
%2 = load <2 x float>, <2 x float> *%1, align 8
%3 = shufflevector <2 x float> %2, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
@@ -1488,7 +1488,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vunpcklpd {{.*#+}} xmm0 = xmm0[0],xmm1[0] sched: [1:0.50]
; ZNVER1-NEXT: vaddps %xmm0, %xmm1, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = shufflevector <4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 0, i32 1, i32 4, i32 5>
%2 = fadd <4 x float> %a1, %1
ret <4 x float> %2
@@ -1549,7 +1549,7 @@
; ZNVER1-NEXT: vmovlpd {{.*#+}} xmm1 = mem[0],xmm1[1] sched: [8:0.50]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vmovlps %xmm0, (%rdi) # sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = bitcast x86_mmx* %a2 to <2 x float>*
%2 = load <2 x float>, <2 x float> *%1, align 8
%3 = shufflevector <2 x float> %2, <2 x float> undef, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
@@ -1600,8 +1600,8 @@
;
; ZNVER1-LABEL: test_movmskps:
; ZNVER1: # BB#0:
-; ZNVER1-NEXT: vmovmskps %xmm0, %eax # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: vmovmskps %xmm0, %eax # sched: [1:1.00]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call i32 @llvm.x86.sse.movmsk.ps(<4 x float> %a0)
ret i32 %1
}
@@ -1652,7 +1652,7 @@
; ZNVER1-LABEL: test_movntps:
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vmovntps %xmm0, (%rdi) # sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
store <4 x float> %a0, <4 x float> *%a1, align 16, !nontemporal !0
ret void
}
@@ -1712,7 +1712,7 @@
; ZNVER1-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero sched: [8:0.50]
; ZNVER1-NEXT: vaddss %xmm0, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vmovss %xmm0, (%rsi) # sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = load float, float* %a0, align 1
%2 = fadd float %1, %1
store float %2, float *%a1, align 1
@@ -1764,7 +1764,7 @@
; ZNVER1-LABEL: test_movss_reg:
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3] sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = shufflevector <4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 4, i32 1, i32 2, i32 3>
ret <4 x float> %1
}
@@ -1824,7 +1824,7 @@
; ZNVER1-NEXT: vmovups (%rdi), %xmm0 # sched: [8:0.50]
; ZNVER1-NEXT: vaddps %xmm0, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vmovups %xmm0, (%rsi) # sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = load <4 x float>, <4 x float> *%a0, align 1
%2 = fadd <4 x float> %1, %1
store <4 x float> %2, <4 x float> *%a1, align 1
@@ -1876,9 +1876,9 @@
;
; ZNVER1-LABEL: test_mulps:
; ZNVER1: # BB#0:
-; ZNVER1-NEXT: vmulps %xmm1, %xmm0, %xmm0 # sched: [5:1.00]
-; ZNVER1-NEXT: vmulps (%rdi), %xmm0, %xmm0 # sched: [12:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: vmulps %xmm1, %xmm0, %xmm0 # sched: [3:0.50]
+; ZNVER1-NEXT: vmulps (%rdi), %xmm0, %xmm0 # sched: [10:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fmul <4 x float> %a0, %a1
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = fmul <4 x float> %1, %2
@@ -1930,9 +1930,9 @@
;
; ZNVER1-LABEL: test_mulss:
; ZNVER1: # BB#0:
-; ZNVER1-NEXT: vmulss %xmm1, %xmm0, %xmm0 # sched: [5:1.00]
-; ZNVER1-NEXT: vmulss (%rdi), %xmm0, %xmm0 # sched: [12:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: vmulss %xmm1, %xmm0, %xmm0 # sched: [3:0.50]
+; ZNVER1-NEXT: vmulss (%rdi), %xmm0, %xmm0 # sched: [10:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fmul float %a0, %a1
%2 = load float, float *%a2, align 4
%3 = fmul float %1, %2
@@ -1990,7 +1990,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vorps %xmm1, %xmm0, %xmm0 # sched: [1:0.25]
; ZNVER1-NEXT: vorps (%rdi), %xmm0, %xmm0 # sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = bitcast <4 x float> %a0 to <4 x i32>
%2 = bitcast <4 x float> %a1 to <4 x i32>
%3 = or <4 x i32> %1, %2
@@ -2046,7 +2046,7 @@
; ZNVER1-LABEL: test_prefetchnta:
; ZNVER1: # BB#0:
; ZNVER1-NEXT: prefetchnta (%rdi) # sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
call void @llvm.prefetch(i8* %a0, i32 0, i32 0, i32 1)
ret void
}
@@ -2109,7 +2109,7 @@
; ZNVER1-NEXT: vrcpps (%rdi), %xmm1 # sched: [12:0.50]
; ZNVER1-NEXT: vrcpps %xmm0, %xmm0 # sched: [5:0.50]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.rcp.ps(<4 x float> %a0)
%2 = load <4 x float>, <4 x float> *%a1, align 16
%3 = call <4 x float> @llvm.x86.sse.rcp.ps(<4 x float> %2)
@@ -2183,7 +2183,7 @@
; ZNVER1-NEXT: vrcpss %xmm0, %xmm0, %xmm0 # sched: [12:0.50]
; ZNVER1-NEXT: vrcpss %xmm1, %xmm1, %xmm1 # sched: [12:0.50]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = insertelement <4 x float> undef, float %a0, i32 0
%2 = call <4 x float> @llvm.x86.sse.rcp.ss(<4 x float> %1)
%3 = load float, float *%a1, align 4
@@ -2251,7 +2251,7 @@
; ZNVER1-NEXT: vrsqrtps (%rdi), %xmm1 # sched: [12:0.50]
; ZNVER1-NEXT: vrsqrtps %xmm0, %xmm0 # sched: [5:0.50]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.rsqrt.ps(<4 x float> %a0)
%2 = load <4 x float>, <4 x float> *%a1, align 16
%3 = call <4 x float> @llvm.x86.sse.rsqrt.ps(<4 x float> %2)
@@ -2322,10 +2322,10 @@
; ZNVER1-LABEL: test_rsqrtss:
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero sched: [8:0.50]
-; ZNVER1-NEXT: vrsqrtss %xmm0, %xmm0, %xmm0 # sched: [12:0.50]
-; ZNVER1-NEXT: vrsqrtss %xmm1, %xmm1, %xmm1 # sched: [12:0.50]
+; ZNVER1-NEXT: vrsqrtss %xmm0, %xmm0, %xmm0 # sched: [5:0.50]
+; ZNVER1-NEXT: vrsqrtss %xmm1, %xmm1, %xmm1 # sched: [5:0.50]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = insertelement <4 x float> undef, float %a0, i32 0
%2 = call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %1)
%3 = load float, float *%a1, align 4
@@ -2381,7 +2381,7 @@
; ZNVER1-LABEL: test_sfence:
; ZNVER1: # BB#0:
; ZNVER1-NEXT: sfence # sched: [1:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
call void @llvm.x86.sse.sfence()
ret void
}
@@ -2438,7 +2438,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,0],xmm1[0,0] sched: [1:0.50]
; ZNVER1-NEXT: vshufps {{.*#+}} xmm0 = xmm0[0,3],mem[0,0] sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = shufflevector <4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 0, i32 0, i32 4, i32 4>
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = shufflevector <4 x float> %1, <4 x float> %2, <4 x i32> <i32 0, i32 3, i32 4, i32 4>
@@ -2501,7 +2501,7 @@
; ZNVER1-NEXT: vsqrtps (%rdi), %xmm1 # sched: [27:1.00]
; ZNVER1-NEXT: vsqrtps %xmm0, %xmm0 # sched: [20:1.00]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.sqrt.ps(<4 x float> %a0)
%2 = load <4 x float>, <4 x float> *%a1, align 16
%3 = call <4 x float> @llvm.x86.sse.sqrt.ps(<4 x float> %2)
@@ -2575,7 +2575,7 @@
; ZNVER1-NEXT: vsqrtss %xmm0, %xmm0, %xmm0 # sched: [27:1.00]
; ZNVER1-NEXT: vsqrtss %xmm1, %xmm1, %xmm1 # sched: [27:1.00]
; ZNVER1-NEXT: vaddps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call <4 x float> @llvm.x86.sse.sqrt.ss(<4 x float> %a0)
%2 = load <4 x float>, <4 x float> *%a1, align 16
%3 = call <4 x float> @llvm.x86.sse.sqrt.ss(<4 x float> %2)
@@ -2629,9 +2629,9 @@
;
; ZNVER1-LABEL: test_stmxcsr:
; ZNVER1: # BB#0:
-; ZNVER1-NEXT: vstmxcsr -{{[0-9]+}}(%rsp) # sched: [1:0.50]
+; ZNVER1-NEXT: vstmxcsr -{{[0-9]+}}(%rsp) # sched: [100:?]
; ZNVER1-NEXT: movl -{{[0-9]+}}(%rsp), %eax # sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = alloca i32, align 4
%2 = bitcast i32* %1 to i8*
call void @llvm.x86.sse.stmxcsr(i8* %2)
@@ -2687,7 +2687,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vsubps %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vsubps (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fsub <4 x float> %a0, %a1
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = fsub <4 x float> %1, %2
@@ -2741,7 +2741,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vsubss %xmm1, %xmm0, %xmm0 # sched: [3:1.00]
; ZNVER1-NEXT: vsubss (%rdi), %xmm0, %xmm0 # sched: [10:1.00]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = fsub float %a0, %a1
%2 = load float, float *%a2, align 4
%3 = fsub float %1, %2
@@ -2859,7 +2859,7 @@
; ZNVER1-NEXT: andb %al, %dl # sched: [1:0.25]
; ZNVER1-NEXT: orb %cl, %dl # sched: [1:0.25]
; ZNVER1-NEXT: movzbl %dl, %eax # sched: [1:0.25]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = call i32 @llvm.x86.sse.ucomieq.ss(<4 x float> %a0, <4 x float> %a1)
%2 = load <4 x float>, <4 x float> *%a2, align 4
%3 = call i32 @llvm.x86.sse.ucomieq.ss(<4 x float> %a0, <4 x float> %2)
@@ -2919,7 +2919,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vunpckhps {{.*#+}} xmm0 = xmm0[2],xmm1[2],xmm0[3],xmm1[3] sched: [1:0.50]
; ZNVER1-NEXT: vunpckhps {{.*#+}} xmm0 = xmm0[2],mem[2],xmm0[3],mem[3] sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = shufflevector <4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = shufflevector <4 x float> %1, <4 x float> %2, <4 x i32> <i32 2, i32 6, i32 3, i32 7>
@@ -2977,7 +2977,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vunpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1] sched: [1:0.50]
; ZNVER1-NEXT: vunpcklps {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1] sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = shufflevector <4 x float> %a0, <4 x float> %a1, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
%2 = load <4 x float>, <4 x float> *%a2, align 16
%3 = shufflevector <4 x float> %1, <4 x float> %2, <4 x i32> <i32 0, i32 4, i32 1, i32 5>
@@ -3035,7 +3035,7 @@
; ZNVER1: # BB#0:
; ZNVER1-NEXT: vxorps %xmm1, %xmm0, %xmm0 # sched: [1:0.25]
; ZNVER1-NEXT: vxorps (%rdi), %xmm0, %xmm0 # sched: [8:0.50]
-; ZNVER1-NEXT: retq # sched: [5:0.50]
+; ZNVER1-NEXT: retq # sched: [1:0.50]
%1 = bitcast <4 x float> %a0 to <4 x i32>
%2 = bitcast <4 x float> %a1 to <4 x i32>
%3 = xor <4 x i32> %1, %2