[llvm] [RISCV][llvm-mca] Use Sched*MC for Zvk MC instructions and add Zvk tests for P600 (PR #89256)
via llvm-commits
llvm-commits at lists.llvm.org
Thu Apr 18 09:03:46 PDT 2024
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-tools-llvm-mca
Author: Michael Maitland (michaelmaitland)
<details>
<summary>Changes</summary>
This patch adds a commit to use Sched*MC for Zvk MC instructions. Then, there are a series of commits that add llvm-mca tests for P600 for each Zvk extension. Without the initial commit to use. Sched*MC, many of these extensions did not work with llvm-mca because they contained instructions were missing scheduling information entirely. This patch demonstrates that llvm-mca works for Zvk instructions.
This commit will serve as a pre-commit to patches that make sure that the P600 Zvk instructions have the correct behavior. For now, the llvm-mca tests just show the current behavior, even if that isn't the intended behavior longterm. By adding tests first, it will highlight just the behavior diff.
---
Patch is 114.42 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/89256.diff
8 Files Affected:
- (modified) llvm/lib/Target/RISCV/RISCVInstrInfoZvk.td (+13-11)
- (added) llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvbb.s (+472)
- (added) llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvbc.s (+176)
- (added) llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvkg.s (+128)
- (added) llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvkned.s (+296)
- (added) llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvknhb.s (+229)
- (added) llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvksed.s (+152)
- (added) llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvksh.s (+128)
``````````diff
diff --git a/llvm/lib/Target/RISCV/RISCVInstrInfoZvk.td b/llvm/lib/Target/RISCV/RISCVInstrInfoZvk.td
index aac7dc444a2de3..84c4ae859ef3cb 100644
--- a/llvm/lib/Target/RISCV/RISCVInstrInfoZvk.td
+++ b/llvm/lib/Target/RISCV/RISCVInstrInfoZvk.td
@@ -24,11 +24,9 @@ def tuimm5 : RISCVOp, TImmLeaf<XLenVT, [{return isUInt<5>(Imm);}]>;
let hasSideEffects = 0, mayLoad = 0, mayStore = 0 in {
multiclass VCLMUL_MV_V_X<string opcodestr, bits<6> funct6> {
def V : VALUVV<funct6, OPMVV, opcodestr # "." # "vv">,
- Sched<[WriteVIALUV_WorstCase, ReadVIALUV_WorstCase,
- ReadVIALUV_WorstCase, ReadVMask]>;
+ SchedBinaryMC<"WriteVIALUV", "ReadVIALUV", "ReadVIALUV">;
def X : VALUVX<funct6, OPMVX, opcodestr # "." # "vx">,
- Sched<[WriteVIALUX_WorstCase, ReadVIALUV_WorstCase,
- ReadVIALUX_WorstCase, ReadVMask]>;
+ SchedBinaryMC<"WriteVIALUV", "ReadVIALUV", "ReadVIALUV">;
}
class RVInstIVI_VROR<bits<6> funct6, dag outs, dag ins, string opcodestr,
@@ -57,13 +55,13 @@ multiclass VROR_IV_V_X_I<string opcodestr, bits<6> funct6>
def I : RVInstIVI_VROR<funct6, (outs VR:$vd),
(ins VR:$vs2, uimm6:$imm, VMaskOp:$vm),
opcodestr # ".vi", "$vd, $vs2, $imm$vm">,
- Sched<[WriteVIALUI_WorstCase, ReadVIALUV_WorstCase,
- ReadVMask]>;
+ SchedUnaryMC<"WriteVIALUI", "ReadVIALUV">;
}
// op vd, vs2, vs1
class PALUVVNoVm<bits<6> funct6, RISCVVFormat opv, string opcodestr>
- : VALUVVNoVm<funct6, opv, opcodestr> {
+ : VALUVVNoVm<funct6, opv, opcodestr>,
+ SchedUnaryMC<"WriteVIALUI", "ReadVIALUV"> {
let Inst{6-0} = OPC_OP_VE.Value;
}
@@ -71,7 +69,8 @@ class PALUVVNoVm<bits<6> funct6, RISCVVFormat opv, string opcodestr>
class PALUVVNoVmTernary<bits<6> funct6, RISCVVFormat opv, string opcodestr>
: RVInstVV<funct6, opv, (outs VR:$vd_wb),
(ins VR:$vd, VR:$vs2, VR:$vs1),
- opcodestr, "$vd, $vs2, $vs1"> {
+ opcodestr, "$vd, $vs2, $vs1">,
+ SchedBinaryMC<"WriteVIALUV", "ReadVIALUV", "ReadVIALUV"> {
let Constraints = "$vd = $vd_wb";
let vm = 1;
let Inst{6-0} = OPC_OP_VE.Value;
@@ -79,7 +78,8 @@ class PALUVVNoVmTernary<bits<6> funct6, RISCVVFormat opv, string opcodestr>
// op vd, vs2, imm
class PALUVINoVm<bits<6> funct6, string opcodestr, Operand optype>
- : VALUVINoVm<funct6, opcodestr, optype> {
+ : VALUVINoVm<funct6, opcodestr, optype>,
+ SchedUnaryMC<"WriteVIALUV", "ReadVIALUV"> {
let Inst{6-0} = OPC_OP_VE.Value;
let Inst{14-12} = OPMVV.Value;
}
@@ -88,7 +88,8 @@ class PALUVINoVm<bits<6> funct6, string opcodestr, Operand optype>
class PALUVINoVmBinary<bits<6> funct6, string opcodestr, Operand optype>
: RVInstIVI<funct6, (outs VR:$vd_wb),
(ins VR:$vd, VR:$vs2, optype:$imm),
- opcodestr, "$vd, $vs2, $imm"> {
+ opcodestr, "$vd, $vs2, $imm">,
+ SchedBinaryMC<"WriteVIALUV", "ReadVIALUV", "ReadVIALUV"> {
let Constraints = "$vd = $vd_wb";
let vm = 1;
let Inst{6-0} = OPC_OP_VE.Value;
@@ -100,7 +101,8 @@ class PALUVINoVmBinary<bits<6> funct6, string opcodestr, Operand optype>
class PALUVs2NoVmBinary<bits<6> funct6, bits<5> vs1, RISCVVFormat opv,
string opcodestr>
: RVInstV<funct6, vs1, opv, (outs VR:$vd_wb), (ins VR:$vd, VR:$vs2),
- opcodestr, "$vd, $vs2"> {
+ opcodestr, "$vd, $vs2">,
+ SchedBinaryMC<"WriteVIALUV", "ReadVIALUV", "ReadVIALUV"> {
let Constraints = "$vd = $vd_wb";
let vm = 1;
let Inst{6-0} = OPC_OP_VE.Value;
diff --git a/llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvbb.s b/llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvbb.s
new file mode 100644
index 00000000000000..cee5c390d7c54b
--- /dev/null
+++ b/llvm/test/tools/llvm-mca/RISCV/SiFiveP600/zvbb.s
@@ -0,0 +1,472 @@
+# NOTE: Assertions have been autogenerated by utils/update_mca_test_checks.py
+# RUN: llvm-mca -mtriple=riscv64 -mcpu=sifive-p670 -iterations=1 < %s | FileCheck %s
+
+vsetvli zero, zero, e32, mf8, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+vsetvli zero, zero, e32, mf4, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+vsetvli zero, zero, e32, mf2, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+vsetvli zero, zero, e32, m1, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+vsetvli zero, zero, e32, m2, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+vsetvli zero, zero, e32, m4, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+vsetvli zero, zero, e32, m8, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+# Show SEW does not matter
+vsetvli zero, zero, e64, m4, tu, mu
+vandn.vv v4, v8, v12
+vandn.vx v4, v8, a0
+vbrev.v v4, v8
+vbrev8.v v4, v8
+vrev8.v v4, v8
+vclz.v v4, v8
+vctz.v v4, v8
+vcpop.v v4, v8
+vrol.vv v4, v8, v12
+vrol.vx v4, v8, a0
+vror.vv v4, v8, v12
+vror.vx v4, v8, a0
+vror.vi v4, v8, 8
+
+vwsll.vv v4, v8, v12
+vwsll.vx v4, v8, a0
+vwsll.vi v4, v8, 8
+
+# CHECK: Iterations: 1
+# CHECK-NEXT: Instructions: 136
+# CHECK-NEXT: Total Cycles: 182
+# CHECK-NEXT: Total uOps: 136
+
+# CHECK: Dispatch Width: 4
+# CHECK-NEXT: uOps Per Cycle: 0.75
+# CHECK-NEXT: IPC: 0.75
+# CHECK-NEXT: Block RThroughput: 176.0
+
+# CHECK: Instruction Info:
+# CHECK-NEXT: [1]: #uOps
+# CHECK-NEXT: [2]: Latency
+# CHECK-NEXT: [3]: RThroughput
+# CHECK-NEXT: [4]: MayLoad
+# CHECK-NEXT: [5]: MayStore
+# CHECK-NEXT: [6]: HasSideEffects (U)
+
+# CHECK: [1] [2] [3] [4] [5] [6] Instructions:
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e32, mf8, tu, mu
+# CHECK-NEXT: 1 1 0.50 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vclz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vctz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 0.50 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 0.50 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 0.50 vwsll.vi v4, v8, 8
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e32, mf4, tu, mu
+# CHECK-NEXT: 1 1 0.50 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vclz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vctz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 0.50 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 0.50 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 0.50 vwsll.vi v4, v8, 8
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e32, mf2, tu, mu
+# CHECK-NEXT: 1 1 0.50 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vclz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vctz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 0.50 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 0.50 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 0.50 vwsll.vi v4, v8, 8
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e32, m1, tu, mu
+# CHECK-NEXT: 1 1 0.50 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vclz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vctz.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 0.50 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 0.50 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 0.50 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 0.50 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 0.50 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 0.50 vwsll.vi v4, v8, 8
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e32, m2, tu, mu
+# CHECK-NEXT: 1 1 1.00 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 1.00 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 1.00 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 1.00 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 1.00 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 1.00 vclz.v v4, v8
+# CHECK-NEXT: 1 1 1.00 vctz.v v4, v8
+# CHECK-NEXT: 1 1 1.00 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 1.00 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 1.00 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 1.00 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 1.00 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 1.00 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 1.00 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 1.00 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 1.00 vwsll.vi v4, v8, 8
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e32, m4, tu, mu
+# CHECK-NEXT: 1 1 2.00 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 2.00 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 2.00 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vclz.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vctz.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 2.00 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 2.00 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 2.00 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 2.00 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 2.00 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 2.00 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 2.00 vwsll.vi v4, v8, 8
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e32, m8, tu, mu
+# CHECK-NEXT: 1 1 4.00 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 4.00 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 4.00 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 4.00 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 4.00 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 4.00 vclz.v v4, v8
+# CHECK-NEXT: 1 1 4.00 vctz.v v4, v8
+# CHECK-NEXT: 1 1 4.00 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 4.00 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 4.00 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 4.00 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 4.00 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 4.00 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 4.00 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 4.00 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 4.00 vwsll.vi v4, v8, 8
+# CHECK-NEXT: 1 1 1.00 U vsetvli zero, zero, e64, m4, tu, mu
+# CHECK-NEXT: 1 1 2.00 vandn.vv v4, v8, v12
+# CHECK-NEXT: 1 1 2.00 vandn.vx v4, v8, a0
+# CHECK-NEXT: 1 1 2.00 vbrev.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vbrev8.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vrev8.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vclz.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vctz.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vcpop.v v4, v8
+# CHECK-NEXT: 1 1 2.00 vrol.vv v4, v8, v12
+# CHECK-NEXT: 1 1 2.00 vrol.vx v4, v8, a0
+# CHECK-NEXT: 1 1 2.00 vror.vv v4, v8, v12
+# CHECK-NEXT: 1 1 2.00 vror.vx v4, v8, a0
+# CHECK-NEXT: 1 1 2.00 vror.vi v4, v8, 8
+# CHECK-NEXT: 1 6 2.00 vwsll.vv v4, v8, v12
+# CHECK-NEXT: 1 6 2.00 vwsll.vx v4, v8, a0
+# CHECK-NEXT: 1 6 2.00 vwsll.vi v4, v8, 8
+
+# CHECK: Resources:
+# CHECK-NEXT: [0] - SiFiveP600Div
+# CHECK-NEXT: [1] - SiFiveP600FEXQ0
+# CHECK-NEXT: [2] - SiFiveP600FEXQ1
+# CHECK-NEXT: [3] - SiFiveP600FloatDiv
+# CHECK-NEXT: [4] - SiFiveP600IEXQ0
+# CHECK-NEXT: [5] - SiFiveP600IEXQ1
+# CHECK-NEXT: [6] - SiFiveP600IEXQ2
+# CHECK-NEXT: [7] - SiFiveP600IEXQ3
+# CHECK-NEXT: [8.0] - SiFiveP600LDST
+# CHECK-NEXT: [8.1] - SiFiveP600LDST
+# CHECK-NEXT: [9] - SiFiveP600VDiv
+# CHECK-NEXT: [10] - SiFiveP600VEXQ0
+# CHECK-NEXT: [11] - SiFiveP600VEXQ1
+# CHECK-NEXT: [12] - SiFiveP600VFloatDiv
+# CHECK-NEXT: [13] - SiFiveP600VLD
+# CHECK-NEXT: [14] - SiFiveP600VST
+
+# CHECK: Resource pressure per iteration:
+# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8.0] [8.1] [9] [10] [11] [12] [13] [14]
+# CHECK-NEXT: - - - - 8.00 - - - - - - 176.00 176.00 - - -
+
+# CHECK: Resource pressure by instruction:
+# CHECK-NEXT: [0] [1] [2] [3] [4] [5] [6] [7] [8.0] [8.1] [9] [10] [11] [12] [13] [14] Instructions:
+# CHECK-NEXT: - - - - 1.00 - - - - - - - - - - - vsetvli zero, zero, e32, mf8, tu, mu
+# CHECK-NEXT: - - - - - - - - - - - - 1.00 - - - vandn.vv v4, v8, v12
+# CHECK-NEXT: - - - - - - - - - - - 1.00 - - - - vandn.vx v4, v8, a0
+# CHECK-NEXT: - - - - - - - - - - - - 1.00 - - - vbrev.v v4, v8
+# CHECK-NEXT: - - - - - - - - - - - 1.00 - - - - vbrev8.v v4, v8
+# CHECK-NEXT: - - - - - - - - - - - - 1.00 - - - vrev8.v v4, v8
+# CHECK-NEXT: - - - - - - - - - - - 1.00 - - - - vclz.v v4, v8
+# CHECK-NEXT: - - - - - - - - - - - - 1.00 - - - vctz.v v4, v8
+# CHECK-NEXT: - - - - - - - - - - - 1.00 - - - - vcpop.v v4, v8
+# CHECK-NEXT: - - - - - - - - - - - - 1.00 - - - vrol.vv v4, v8, v12
+# CHECK-NEXT: - - - - - - - - - - - 1.00 - - - - ...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/89256
More information about the llvm-commits
mailing list