@@ -465,14 +465,14 @@ vpxor (%rax), %ymm1, %ymm2
465
465
# CHECK-NEXT: 1 3 1.00 vbroadcastss %xmm0, %ymm0
466
466
# CHECK-NEXT: 1 3 1.00 vextracti128 $1, %ymm0, %xmm2
467
467
# CHECK-NEXT: 2 1 1.00 * vextracti128 $1, %ymm0, (%rax)
468
- # CHECK-NEXT: 7 25 3.00 * vgatherdpd %xmm0, (%rax,%xmm1,2), %xmm2
469
- # CHECK-NEXT: 9 26 5 .00 * vgatherdpd %ymm0, (%rax,%xmm1,2), %ymm2
470
- # CHECK-NEXT: 7 25 3 .00 * vgatherdps %xmm0, (%rax,%xmm1,2), %xmm2
471
- # CHECK-NEXT: 14 26 4.00 * vgatherdps %ymm0, (%rax,%ymm1,2), %ymm2
472
- # CHECK-NEXT: 7 22 3.00 * vgatherqpd %xmm0, (%rax,%xmm1,2), %xmm2
473
- # CHECK-NEXT: 9 23 3.00 * vgatherqpd %ymm0, (%rax,%ymm1,2), %ymm2
474
- # CHECK-NEXT: 9 27 5.00 * vgatherqps %xmm0, (%rax,%xmm1,2), %xmm2
475
- # CHECK-NEXT: 9 24 5 .00 * vgatherqps %xmm0, (%rax,%ymm1,2), %xmm2
468
+ # CHECK-NEXT: 7 17 3.00 * vgatherdpd %xmm0, (%rax,%xmm1,2), %xmm2
469
+ # CHECK-NEXT: 9 18 3 .00 * vgatherdpd %ymm0, (%rax,%xmm1,2), %ymm2
470
+ # CHECK-NEXT: 10 19 4 .00 * vgatherdps %xmm0, (%rax,%xmm1,2), %xmm2
471
+ # CHECK-NEXT: 14 21 4.00 * vgatherdps %ymm0, (%rax,%ymm1,2), %ymm2
472
+ # CHECK-NEXT: 7 17 3.00 * vgatherqpd %xmm0, (%rax,%xmm1,2), %xmm2
473
+ # CHECK-NEXT: 9 18 3.00 * vgatherqpd %ymm0, (%rax,%ymm1,2), %ymm2
474
+ # CHECK-NEXT: 9 19 5.00 * vgatherqps %xmm0, (%rax,%xmm1,2), %xmm2
475
+ # CHECK-NEXT: 10 19 4 .00 * vgatherqps %xmm0, (%rax,%ymm1,2), %xmm2
476
476
# CHECK-NEXT: 1 3 1.00 vinserti128 $1, %xmm0, %ymm1, %ymm2
477
477
# CHECK-NEXT: 2 6 0.50 * vinserti128 $1, (%rax), %ymm1, %ymm2
478
478
# CHECK-NEXT: 1 6 0.50 * vmovntdqa (%rax), %ymm0
@@ -568,14 +568,14 @@ vpxor (%rax), %ymm1, %ymm2
568
568
# CHECK-NEXT: 2 9 1.00 * vpermps (%rax), %ymm1, %ymm2
569
569
# CHECK-NEXT: 1 3 1.00 vpermq $1, %ymm0, %ymm2
570
570
# CHECK-NEXT: 2 9 1.00 * vpermq $1, (%rax), %ymm2
571
- # CHECK-NEXT: 1 5 0.50 * vpgatherdd %xmm0, (%rax,%xmm1,2), %xmm2
572
- # CHECK-NEXT: 1 5 0.50 * vpgatherdd %ymm0, (%rax,%ymm1,2), %ymm2
573
- # CHECK-NEXT: 1 5 0.50 * vpgatherdq %xmm0, (%rax,%xmm1,2), %xmm2
574
- # CHECK-NEXT: 1 5 0.50 * vpgatherdq %ymm0, (%rax,%xmm1,2), %ymm2
575
- # CHECK-NEXT: 1 5 0.50 * vpgatherqd %xmm0, (%rax,%xmm1,2), %xmm2
576
- # CHECK-NEXT: 1 5 0.50 * vpgatherqd %xmm0, (%rax,%ymm1,2), %xmm2
577
- # CHECK-NEXT: 1 5 0.50 * vpgatherqq %xmm0, (%rax,%xmm1,2), %xmm2
578
- # CHECK-NEXT: 1 5 0.50 * vpgatherqq %ymm0, (%rax,%ymm1,2), %ymm2
571
+ # CHECK-NEXT: 10 19 4.00 * vpgatherdd %xmm0, (%rax,%xmm1,2), %xmm2
572
+ # CHECK-NEXT: 14 21 4.00 * vpgatherdd %ymm0, (%rax,%ymm1,2), %ymm2
573
+ # CHECK-NEXT: 7 17 3.00 * vpgatherdq %xmm0, (%rax,%xmm1,2), %xmm2
574
+ # CHECK-NEXT: 9 18 3.00 * vpgatherdq %ymm0, (%rax,%xmm1,2), %ymm2
575
+ # CHECK-NEXT: 9 19 5.00 * vpgatherqd %xmm0, (%rax,%xmm1,2), %xmm2
576
+ # CHECK-NEXT: 10 19 4.00 * vpgatherqd %xmm0, (%rax,%ymm1,2), %xmm2
577
+ # CHECK-NEXT: 7 17 3.00 * vpgatherqq %xmm0, (%rax,%xmm1,2), %xmm2
578
+ # CHECK-NEXT: 9 18 3.00 * vpgatherqq %ymm0, (%rax,%ymm1,2), %ymm2
579
579
# CHECK-NEXT: 3 3 2.00 vphaddd %ymm0, %ymm1, %ymm2
580
580
# CHECK-NEXT: 4 9 2.00 * vphaddd (%rax), %ymm1, %ymm2
581
581
# CHECK-NEXT: 3 3 2.00 vphaddsw %ymm0, %ymm1, %ymm2
@@ -776,7 +776,7 @@ vpxor (%rax), %ymm1, %ymm2
776
776
777
777
# CHECK: Resource pressure per iteration:
778
778
# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9]
779
- # CHECK-NEXT: - - 94 .67 58 .67 85 .67 85 .67 13 .00 237 .67 2 .00 1.67
779
+ # CHECK-NEXT: - - 96 .67 60 .67 99 .67 99 .67 21 .00 266 .67 4 .00 1.67
780
780
781
781
# CHECK: Resource pressure by instruction:
782
782
# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8] [9] Instructions:
@@ -786,13 +786,13 @@ vpxor (%rax), %ymm1, %ymm2
786
786
# CHECK-NEXT: - - - - - - - 1.00 - - vextracti128 $1, %ymm0, %xmm2
787
787
# CHECK-NEXT: - - - - 0.33 0.33 1.00 - - 0.33 vextracti128 $1, %ymm0, (%rax)
788
788
# CHECK-NEXT: - - 0.25 0.25 1.00 1.00 1.00 3.25 0.25 - vgatherdpd %xmm0, (%rax,%xmm1,2), %xmm2
789
- # CHECK-NEXT: - - 0.25 0.25 1 .00 1 .00 1.00 5 .25 0.25 - vgatherdpd %ymm0, (%rax,%xmm1,2), %ymm2
790
- # CHECK-NEXT: - - 0.25 0.25 1 .00 1 .00 1.00 3 .25 0.25 - vgatherdps %xmm0, (%rax,%xmm1,2), %xmm2
789
+ # CHECK-NEXT: - - 0.25 0.25 2 .00 2 .00 1.00 3 .25 0.25 - vgatherdpd %ymm0, (%rax,%xmm1,2), %ymm2
790
+ # CHECK-NEXT: - - 0.25 0.25 2 .00 2 .00 1.00 4 .25 0.25 - vgatherdps %xmm0, (%rax,%xmm1,2), %xmm2
791
791
# CHECK-NEXT: - - 0.25 0.25 4.00 4.00 1.00 4.25 0.25 - vgatherdps %ymm0, (%rax,%ymm1,2), %ymm2
792
792
# CHECK-NEXT: - - 0.25 0.25 1.00 1.00 1.00 3.25 0.25 - vgatherqpd %xmm0, (%rax,%xmm1,2), %xmm2
793
793
# CHECK-NEXT: - - 0.25 0.25 2.00 2.00 1.00 3.25 0.25 - vgatherqpd %ymm0, (%rax,%ymm1,2), %ymm2
794
794
# CHECK-NEXT: - - 0.25 0.25 1.00 1.00 1.00 5.25 0.25 - vgatherqps %xmm0, (%rax,%xmm1,2), %xmm2
795
- # CHECK-NEXT: - - 0.25 0.25 1 .00 1 .00 1.00 5 .25 0.25 - vgatherqps %xmm0, (%rax,%ymm1,2), %xmm2
795
+ # CHECK-NEXT: - - 0.25 0.25 2 .00 2 .00 1.00 4 .25 0.25 - vgatherqps %xmm0, (%rax,%ymm1,2), %xmm2
796
796
# CHECK-NEXT: - - - - - - - 1.00 - - vinserti128 $1, %xmm0, %ymm1, %ymm2
797
797
# CHECK-NEXT: - - 0.33 0.33 0.50 0.50 - 0.33 - - vinserti128 $1, (%rax), %ymm1, %ymm2
798
798
# CHECK-NEXT: - - - - 0.50 0.50 - - - - vmovntdqa (%rax), %ymm0
@@ -888,14 +888,14 @@ vpxor (%rax), %ymm1, %ymm2
888
888
# CHECK-NEXT: - - - - 0.50 0.50 - 1.00 - - vpermps (%rax), %ymm1, %ymm2
889
889
# CHECK-NEXT: - - - - - - - 1.00 - - vpermq $1, %ymm0, %ymm2
890
890
# CHECK-NEXT: - - - - 0.50 0.50 - 1.00 - - vpermq $1, (%rax), %ymm2
891
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherdd %xmm0, (%rax,%xmm1,2), %xmm2
892
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherdd %ymm0, (%rax,%ymm1,2), %ymm2
893
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherdq %xmm0, (%rax,%xmm1,2), %xmm2
894
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherdq %ymm0, (%rax,%xmm1,2), %ymm2
895
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherqd %xmm0, (%rax,%xmm1,2), %xmm2
896
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherqd %xmm0, (%rax,%ymm1,2), %xmm2
897
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherqq %xmm0, (%rax,%xmm1,2), %xmm2
898
- # CHECK-NEXT: - - - - 0.50 0.50 - - - - vpgatherqq %ymm0, (%rax,%ymm1,2), %ymm2
891
+ # CHECK-NEXT: - - 0.25 0.25 2.00 2.00 1.00 4.25 0.25 - vpgatherdd %xmm0, (%rax,%xmm1,2), %xmm2
892
+ # CHECK-NEXT: - - 0.25 0.25 4.00 4.00 1.00 4.25 0.25 - vpgatherdd %ymm0, (%rax,%ymm1,2), %ymm2
893
+ # CHECK-NEXT: - - 0.25 0.25 1.00 1.00 1.00 3.25 0.25 - vpgatherdq %xmm0, (%rax,%xmm1,2), %xmm2
894
+ # CHECK-NEXT: - - 0.25 0.25 2.00 2.00 1.00 3.25 0.25 - vpgatherdq %ymm0, (%rax,%xmm1,2), %ymm2
895
+ # CHECK-NEXT: - - 0.25 0.25 1.00 1.00 1.00 5.25 0.25 - vpgatherqd %xmm0, (%rax,%xmm1,2), %xmm2
896
+ # CHECK-NEXT: - - 0.25 0.25 2.00 2.00 1.00 4.25 0.25 - vpgatherqd %xmm0, (%rax,%ymm1,2), %xmm2
897
+ # CHECK-NEXT: - - 0.25 0.25 1.00 1.00 1.00 3.25 0.25 - vpgatherqq %xmm0, (%rax,%xmm1,2), %xmm2
898
+ # CHECK-NEXT: - - 0.25 0.25 2.00 2.00 1.00 3.25 0.25 - vpgatherqq %ymm0, (%rax,%ymm1,2), %ymm2
899
899
# CHECK-NEXT: - - - 0.50 - - - 2.50 - - vphaddd %ymm0, %ymm1, %ymm2
900
900
# CHECK-NEXT: - - - 0.50 0.50 0.50 - 2.50 - - vphaddd (%rax), %ymm1, %ymm2
901
901
# CHECK-NEXT: - - - 0.50 - - - 2.50 - - vphaddsw %ymm0, %ymm1, %ymm2
0 commit comments