diff options
| -rw-r--r-- | llvm/test/CodeGen/X86/avx512-cvt.ll | 12 | ||||
| -rwxr-xr-x | llvm/test/CodeGen/X86/avx512-schedule.ll | 12 | ||||
| -rw-r--r-- | llvm/test/CodeGen/X86/avx512vl-vec-masked-cmp.ll | 8 |
3 files changed, 15 insertions, 17 deletions
diff --git a/llvm/test/CodeGen/X86/avx512-cvt.ll b/llvm/test/CodeGen/X86/avx512-cvt.ll index 29abb3a8e35..573270296ed 100644 --- a/llvm/test/CodeGen/X86/avx512-cvt.ll +++ b/llvm/test/CodeGen/X86/avx512-cvt.ll @@ -2113,7 +2113,7 @@ define <2 x float> @ubto2f32(<2 x i32> %a) { ; NOVL: # %bb.0: ; NOVL-NEXT: vpxor %xmm1, %xmm1, %xmm1 ; NOVL-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; NOVL-NEXT: vpcmpltuq %zmm1, %zmm0, %k1 +; NOVL-NEXT: vptestmq %zmm0, %zmm0, %k1 ; NOVL-NEXT: vpbroadcastd {{.*}}(%rip), %zmm0 {%k1} {z} ; NOVL-NEXT: vcvtdq2ps %xmm0, %xmm0 ; NOVL-NEXT: vzeroupper @@ -2123,11 +2123,11 @@ define <2 x float> @ubto2f32(<2 x i32> %a) { ; VL: # %bb.0: ; VL-NEXT: vpxor %xmm1, %xmm1, %xmm1 ; VL-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; VL-NEXT: vpcmpltuq %xmm1, %xmm0, %k1 +; VL-NEXT: vptestmq %xmm0, %xmm0, %k1 ; VL-NEXT: vpbroadcastd {{.*}}(%rip), %xmm0 {%k1} {z} ; VL-NEXT: vcvtdq2ps %xmm0, %xmm0 ; VL-NEXT: retq - %mask = icmp ult <2 x i32> %a, zeroinitializer + %mask = icmp ne <2 x i32> %a, zeroinitializer %1 = uitofp <2 x i1> %mask to <2 x float> ret <2 x float> %1 } @@ -2137,7 +2137,7 @@ define <2 x double> @ubto2f64(<2 x i32> %a) { ; NOVL: # %bb.0: ; NOVL-NEXT: vpxor %xmm1, %xmm1, %xmm1 ; NOVL-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; NOVL-NEXT: vpcmpltuq %zmm1, %zmm0, %k1 +; NOVL-NEXT: vptestmq %zmm0, %zmm0, %k1 ; NOVL-NEXT: vpbroadcastd {{.*}}(%rip), %zmm0 {%k1} {z} ; NOVL-NEXT: vcvtudq2pd %ymm0, %zmm0 ; NOVL-NEXT: # kill: def $xmm0 killed $xmm0 killed $zmm0 @@ -2148,11 +2148,11 @@ define <2 x double> @ubto2f64(<2 x i32> %a) { ; VL: # %bb.0: ; VL-NEXT: vpxor %xmm1, %xmm1, %xmm1 ; VL-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] -; VL-NEXT: vpcmpltuq %xmm1, %xmm0, %k1 +; VL-NEXT: vptestmq %xmm0, %xmm0, %k1 ; VL-NEXT: vpbroadcastd {{.*}}(%rip), %xmm0 {%k1} {z} ; VL-NEXT: vcvtudq2pd %xmm0, %xmm0 ; VL-NEXT: retq - %mask = icmp ult <2 x i32> %a, zeroinitializer + %mask = icmp ne <2 x i32> %a, zeroinitializer %1 = uitofp <2 x i1> %mask to <2 x double> ret <2 x double> %1 } diff --git a/llvm/test/CodeGen/X86/avx512-schedule.ll b/llvm/test/CodeGen/X86/avx512-schedule.ll index d83c7cd96c3..1f4e4c9062d 100755 --- a/llvm/test/CodeGen/X86/avx512-schedule.ll +++ b/llvm/test/CodeGen/X86/avx512-schedule.ll @@ -2911,7 +2911,7 @@ define <2 x float> @ubto2f32(<2 x i32> %a) { ; GENERIC: # %bb.0: ; GENERIC-NEXT: vpxor %xmm1, %xmm1, %xmm1 # sched: [1:0.33] ; GENERIC-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] sched: [1:0.50] -; GENERIC-NEXT: vpcmpltuq %xmm1, %xmm0, %k1 # sched: [3:1.00] +; GENERIC-NEXT: vptestmq %xmm0, %xmm0, %k1 # sched: [1:1.00] ; GENERIC-NEXT: vpbroadcastd {{.*}}(%rip), %xmm0 {%k1} {z} # sched: [5:1.00] ; GENERIC-NEXT: vcvtdq2ps %xmm0, %xmm0 # sched: [3:1.00] ; GENERIC-NEXT: retq # sched: [1:1.00] @@ -2920,11 +2920,11 @@ define <2 x float> @ubto2f32(<2 x i32> %a) { ; SKX: # %bb.0: ; SKX-NEXT: vpxor %xmm1, %xmm1, %xmm1 # sched: [1:0.33] ; SKX-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] sched: [1:0.33] -; SKX-NEXT: vpcmpltuq %xmm1, %xmm0, %k1 # sched: [3:1.00] +; SKX-NEXT: vptestmq %xmm0, %xmm0, %k1 # sched: [3:1.00] ; SKX-NEXT: vpbroadcastd {{.*}}(%rip), %xmm0 {%k1} {z} # sched: [7:0.50] ; SKX-NEXT: vcvtdq2ps %xmm0, %xmm0 # sched: [4:0.33] ; SKX-NEXT: retq # sched: [7:1.00] - %mask = icmp ult <2 x i32> %a, zeroinitializer + %mask = icmp ne <2 x i32> %a, zeroinitializer %1 = uitofp <2 x i1> %mask to <2 x float> ret <2 x float> %1 } @@ -2934,7 +2934,7 @@ define <2 x double> @ubto2f64(<2 x i32> %a) { ; GENERIC: # %bb.0: ; GENERIC-NEXT: vpxor %xmm1, %xmm1, %xmm1 # sched: [1:0.33] ; GENERIC-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] sched: [1:0.50] -; GENERIC-NEXT: vpcmpltuq %xmm1, %xmm0, %k1 # sched: [3:1.00] +; GENERIC-NEXT: vptestmq %xmm0, %xmm0, %k1 # sched: [1:1.00] ; GENERIC-NEXT: vpbroadcastd {{.*}}(%rip), %xmm0 {%k1} {z} # sched: [5:1.00] ; GENERIC-NEXT: vcvtudq2pd %xmm0, %xmm0 # sched: [4:1.00] ; GENERIC-NEXT: retq # sched: [1:1.00] @@ -2943,11 +2943,11 @@ define <2 x double> @ubto2f64(<2 x i32> %a) { ; SKX: # %bb.0: ; SKX-NEXT: vpxor %xmm1, %xmm1, %xmm1 # sched: [1:0.33] ; SKX-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1],xmm0[2],xmm1[3] sched: [1:0.33] -; SKX-NEXT: vpcmpltuq %xmm1, %xmm0, %k1 # sched: [3:1.00] +; SKX-NEXT: vptestmq %xmm0, %xmm0, %k1 # sched: [3:1.00] ; SKX-NEXT: vpbroadcastd {{.*}}(%rip), %xmm0 {%k1} {z} # sched: [7:0.50] ; SKX-NEXT: vcvtudq2pd %xmm0, %xmm0 # sched: [5:1.00] ; SKX-NEXT: retq # sched: [7:1.00] - %mask = icmp ult <2 x i32> %a, zeroinitializer + %mask = icmp ne <2 x i32> %a, zeroinitializer %1 = uitofp <2 x i1> %mask to <2 x double> ret <2 x double> %1 } diff --git a/llvm/test/CodeGen/X86/avx512vl-vec-masked-cmp.ll b/llvm/test/CodeGen/X86/avx512vl-vec-masked-cmp.ll index 2a04a215c41..f01f347999e 100644 --- a/llvm/test/CodeGen/X86/avx512vl-vec-masked-cmp.ll +++ b/llvm/test/CodeGen/X86/avx512vl-vec-masked-cmp.ll @@ -25859,8 +25859,7 @@ define i32 @test_cmpm_rnd_zero(<16 x float> %a, <16 x float> %b) { define i8 @mask_zero_lower(<4 x i32> %a) { ; VLX-LABEL: mask_zero_lower: ; VLX: # %bb.0: -; VLX-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; VLX-NEXT: vpcmpltud %xmm1, %xmm0, %k0 +; VLX-NEXT: vptestmd %xmm0, %xmm0, %k0 ; VLX-NEXT: kshiftlb $4, %k0, %k0 ; VLX-NEXT: kmovd %k0, %eax ; VLX-NEXT: # kill: def $al killed $al killed $eax @@ -25869,15 +25868,14 @@ define i8 @mask_zero_lower(<4 x i32> %a) { ; NoVLX-LABEL: mask_zero_lower: ; NoVLX: # %bb.0: ; NoVLX-NEXT: # kill: def $xmm0 killed $xmm0 def $zmm0 -; NoVLX-NEXT: vpxor %xmm1, %xmm1, %xmm1 -; NoVLX-NEXT: vpcmpltud %zmm1, %zmm0, %k0 +; NoVLX-NEXT: vptestmd %zmm0, %zmm0, %k0 ; NoVLX-NEXT: kshiftlw $12, %k0, %k0 ; NoVLX-NEXT: kshiftrw $8, %k0, %k0 ; NoVLX-NEXT: kmovw %k0, %eax ; NoVLX-NEXT: # kill: def $al killed $al killed $eax ; NoVLX-NEXT: vzeroupper ; NoVLX-NEXT: retq - %cmp = icmp ult <4 x i32> %a, zeroinitializer + %cmp = icmp ne <4 x i32> %a, zeroinitializer %concat = shufflevector <4 x i1> %cmp, <4 x i1> zeroinitializer, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3> %cast = bitcast <8 x i1> %concat to i8 ret i8 %cast |
