[RISCV] Update SpacemiT-X60 vector permutation instructions latencies #152738
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Mikhail R. Gadelha <[email protected]>
Signed-off-by: Mikhail R. Gadelha <[email protected]>
Signed-off-by: Mikhail R. Gadelha <[email protected]>
mikhailramalho
requested review from
lukel97,
mshockwave,
preames,
wangpc-pp and
zqb-all
Member
|
@llvm/pr-subscribers-backend-risc-v Author: Mikhail R. Gadelha (mikhailramalho) ChangesThis PR adds hardware-measured latencies for all instructions defined in Section 16 of the RVV specification: "Vector Permutation Instructions" to the SpacemiT-X60 scheduling model. Patch is 265.12 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/152738.diff 3 Files Affected:
diff --git a/llvm/lib/Target/RISCV/RISCVSchedSpacemitX60.td b/llvm/lib/Target/RISCV/RISCVSchedSpacemitX60.td
index 5541506ceb1e3..7064459c22ab4 100644
--- a/llvm/lib/Target/RISCV/RISCVSchedSpacemitX60.td
+++ b/llvm/lib/Target/RISCV/RISCVSchedSpacemitX60.td
@@ -683,39 +683,82 @@ foreach mx = SchedMxList in {
}
// 16. Vector Permutation Instructions
+// Slide
foreach mx = SchedMxList in {
defvar IsWorstCase = SMX60IsWorstCaseMX<mx, SchedMxList>.c;
- defm "" : LMULWriteResMX<"WriteVSlideI", [SMX60_VIEU], mx, IsWorstCase>;
+ // Latency for slide up: 4/4/8/16, ReleaseAtCycles is 2/4/8/16
+ defvar VSlideUpLat = ConstValueUntilLMULThenDouble<"M2", 4, mx>.c;
+ defvar VSlideUpOcc = ConstOneUntilMF2ThenDouble<mx>.c;
+ let Latency = VSlideUpLat, ReleaseAtCycles =[VSlideUpOcc] in {
+ defm "" : LMULWriteResMX<"WriteVSlideUpX", [SMX60_VIEU], mx, IsWorstCase>;
+ }
- defm "" : LMULWriteResMX<"WriteVISlide1X", [SMX60_VIEU], mx, IsWorstCase>;
- defm "" : LMULWriteResMX<"WriteVFSlide1F", [SMX60_VFP], mx, IsWorstCase>;
+ // Latency for slide down: 4/5/9/17, ReleaseAtCycles is 3/5/9/17
+ defvar VSlideDownLat = GetLMULValue<[4, 4, 4, 4, 5, 9, 17], mx>.c;
+ defvar VSlideDownOcc = GetLMULValue<[1, 1, 1, 3, 5, 9, 17], mx>.c;
+ let Latency = VSlideDownLat, ReleaseAtCycles =[VSlideDownOcc] in {
+ defm "" : LMULWriteResMX<"WriteVSlideDownX", [SMX60_VIEU], mx, IsWorstCase>;
+ }
+ // The following group slide up and down together, so we use the worst-case
+ // (slide down) for all.
+ let Latency = VSlideDownLat, ReleaseAtCycles =[VSlideDownOcc] in {
+ defm "" : LMULWriteResMX<"WriteVSlideI", [SMX60_VIEU], mx, IsWorstCase>;
+ defm "" : LMULWriteResMX<"WriteVISlide1X", [SMX60_VIEU], mx, IsWorstCase>;
- defm "" : LMULWriteResMX<"WriteVSlideUpX", [SMX60_VIEU], mx, IsWorstCase>;
- defm "" : LMULWriteResMX<"WriteVSlideDownX", [SMX60_VIEU], mx, IsWorstCase>;
+ defm "" : LMULWriteResMX<"WriteVFSlide1F", [SMX60_VFP], mx, IsWorstCase>;
+ }
}
-def : WriteRes<WriteVMovXS, [SMX60_VIEU]>;
-def : WriteRes<WriteVMovSX, [SMX60_VIEU]>;
-
-def : WriteRes<WriteVMovFS, [SMX60_VIEU]>;
-def : WriteRes<WriteVMovSF, [SMX60_VIEU]>;
+// ReleaseAtCycles is 2/2/2/2/2/3/6, but we can't set based on MX for now
+// TODO: Split this into separate WriteRes for each MX
+let Latency = 6, ReleaseAtCycles = [6] in {
+ def : WriteRes<WriteVMovXS, [SMX60_VIEU]>;
+}
-// Gather and Compress
-foreach mx = SchedMxList in {
- foreach sew = SchedSEWSet<mx>.val in {
- defvar IsWorstCase = SMX60IsWorstCaseMXSEW<mx, sew, SchedMxList>.c;
- defm "" : LMULSEWWriteResMXSEW<"WriteVRGatherVV", [SMX60_VIEU], mx, sew, IsWorstCase>;
- defm "" : LMULSEWWriteResMXSEW<"WriteVRGatherEI16VV", [SMX60_VIEU], mx, sew, IsWorstCase>;
- defm "" : LMULSEWWriteResMXSEW<"WriteVCompressV", [SMX60_VIEU], mx, sew, IsWorstCase>;
- }
+// ReleaseAtCycles is 1/1/1/1/1/2/4, but we can't set based on MX for now
+// TODO: Split this into separate WriteRes for each MX
+let Latency = 4, ReleaseAtCycles = [4] in {
+ def : WriteRes<WriteVMovSX, [SMX60_VIEU]>;
+ def : WriteRes<WriteVMovFS, [SMX60_VIEU]>;
+ def : WriteRes<WriteVMovSF, [SMX60_VIEU]>;
}
+// Integer LMUL Gather and Compress
foreach mx = SchedMxList in {
defvar IsWorstCase = SMX60IsWorstCaseMX<mx, SchedMxList>.c;
- defm "" : LMULWriteResMX<"WriteVRGatherVX", [SMX60_VIEU], mx, IsWorstCase>;
- defm "" : LMULWriteResMX<"WriteVRGatherVI", [SMX60_VIEU], mx, IsWorstCase>;
+ defvar VRGatherLat = ConstValueUntilLMULThenDouble<"M2", 4, mx>.c;
+ let Latency = VRGatherLat, ReleaseAtCycles = [ConstOneUntilMF2ThenDouble<mx>.c] in {
+ defm "" : LMULWriteResMX<"WriteVRGatherVX", [SMX60_VIEU], mx, IsWorstCase>;
+ defm "" : LMULWriteResMX<"WriteVRGatherVI", [SMX60_VIEU], mx, IsWorstCase>;
+ }
+
+ foreach sew = SchedSEWSet<mx>.val in {
+ defvar IsWorstCaseSEW = SMX60IsWorstCaseMXSEW<mx, sew, SchedMxList>.c;
+
+ defvar VRGatherVVLat = GetLMULValue<[4, 4, 4, 4, 16, 64, 256], mx>.c;
+ defvar VRGatherVVOcc = GetLMULValue<[1, 1, 1, 4, 16, 64, 256], mx>.c;
+ let Latency = VRGatherVVLat, ReleaseAtCycles = [VRGatherVVOcc] in {
+ defm "" : LMULSEWWriteResMXSEW<"WriteVRGatherVV", [SMX60_VIEU], mx, sew, IsWorstCaseSEW>;
+ }
+ // For sew == 8, latency is half of the other cases, except for the fractional LMULs (const 4 cycles)
+ defvar VRGatherEI16Lat = !if(!eq(sew, 8),
+ GetLMULValue<[4, 4, 4, 8, 32, 128, 256], mx>.c,
+ GetLMULValue<[4, 4, 4, 4, 16, 64, 256], mx>.c);
+ defvar VRGatherEI16Occ = !if(!eq(sew, 8),
+ GetLMULValue<[1, 1, 2, 8, 32, 128, 256], mx>.c,
+ GetLMULValue<[1, 1, 1, 4, 16, 64, 256], mx>.c);
+ let Latency = VRGatherEI16Lat, ReleaseAtCycles = [VRGatherEI16Occ] in {
+ defm "" : LMULSEWWriteResMXSEW<"WriteVRGatherEI16VV", [SMX60_VIEU], mx, sew, IsWorstCaseSEW>;
+ }
+
+ defvar VCompressVLat = GetLMULValue<[4, 4, 4, 4, 10, 36, 136], mx>.c;
+ defvar VCompressVOcc = GetLMULValue<[1, 1, 1, 3, 10, 36, 136], mx>.c;
+ let Latency = VCompressVLat, ReleaseAtCycles = [VCompressVOcc] in {
+ defm "" : LMULSEWWriteResMXSEW<"WriteVCompressV", [SMX60_VIEU], mx, sew, IsWorstCaseSEW>;
+ }
+ }
}
// Others
diff --git a/llvm/test/tools/llvm-mca/RISCV/SpacemitX60/rvv-fp.s b/llvm/test/tools/llvm-mca/RISCV/SpacemitX60/rvv-fp.s
index f59c7987b615b..311a13c9427b1 100644
--- a/llvm/test/tools/llvm-mca/RISCV/SpacemitX60/rvv-fp.s
+++ b/llvm/test/tools/llvm-mca/RISCV/SpacemitX60/rvv-fp.s
@@ -2911,65 +2911,65 @@ vfwsub.wv v8, v16, v24
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m8, tu, mu
# CHECK-NEXT: 1 1 1.00 1 SMX60_VFP VFMADD_VV vfmadd.vv v8, v8, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, mf2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, mf4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m1, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m8, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, mf2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m1, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m8, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m1, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m8, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_F_S vfmv.f.s fs0, v8
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_F_S vfmv.f.s fs0, v8
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, mf2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, mf4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m1, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, m8, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, mf2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m1, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e32, m8, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m1, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m2, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m4, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e64, m8, tu, mu
-# CHECK-NEXT: 1 1 1.00 1 SMX60_VIEU VFMV_S_F vfmv.s.f v8, fs0
+# CHECK-NEXT: 1 4 4.00 4 SMX60_VIEU[4] VFMV_S_F vfmv.s.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, mf2, tu, mu
# CHECK-NEXT: 1 1 1.00 1 SMX60_VFP VFMV_V_F vfmv.v.f v8, fs0
# CHECK-NEXT: 1 1 1.00 U 1 SMX60_IEU,SMX60_IEUA VSETVLI vsetvli t3, zero, e16, mf4, tu, mu
@@ -3763,7 +3763,7 @@ vfwsub.wv v8, v16, v24
#...
[truncated]
|
mshockwave
approved these changes
Oct 30, 2025
| def : WriteRes<WriteVMovFS, [SMX60_VIEU]>; | ||
| def : WriteRes<WriteVMovSF, [SMX60_VIEU]>; | ||
| // ReleaseAtCycles is 2/2/2/2/2/3/6, but we can't set based on MX for now | ||
| // TODO: Split this into separate WriteRes for each MX |
Member
There was a problem hiding this comment.
it wasn't split by MX because vmv.x.s and friends should be independent of LMUL. I think Specmit X60 might be an outlier here and future hardware are probably more likely to have LMUL-independent latency & occupancy.
git-crd
pushed a commit
to git-crd/crd-llvm-project
that referenced
this pull request
Nov 13, 2025
…llvm#152738) This PR adds hardware-measured latencies for all instructions defined in Section 16 of the RVV specification: "Vector Permutation Instructions" to the SpacemiT-X60 scheduling model. --------- Signed-off-by: Mikhail R. Gadelha <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds hardware-measured latencies for all instructions defined in Section 16 of the RVV specification: "Vector Permutation Instructions" to the SpacemiT-X60 scheduling model.