[llvm] [RISCV] Be more aggressive about forming floating point constants (PR #68433)
via llvm-commits
llvm-commits at lists.llvm.org
Fri Oct 6 10:47:40 PDT 2023
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-risc-v
<details>
<summary>Changes</summary>
We were being very conservative about forming floating point constants via an integer materialization sequence and a fmv. With the default threshold of 2, we'd only do so if the bit sequence could be produced with a single instruction (LUI, ADDI, or sometimes BSETI).
This change removes the separate threshold entirely, and ties the floating point expansion costing to the integer costing threshold. The effect of this is that the default threshold increases by 2, and that more sequences are materialized via integers - avoiding constant pool loads. This is sufficient to cover all constants of types fp16, bf16, and fp32. Many f64 constants are covered as well, but not all.
One downside of this change is that double constants for ELEN=64 configurations on rv32 can't be formed via the same integer sequences. This causes a forking in the test coverage which is more than a tad ugly. Ideas on how to reduce this, or restructure tests to avoid it are more than welcome.
---
Patch is 1.05 MiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/68433.diff
66 Files Affected:
- (modified) llvm/lib/Target/RISCV/RISCVISelLowering.cpp (+2-8)
- (modified) llvm/test/CodeGen/RISCV/bfloat-convert.ll (+118-77)
- (modified) llvm/test/CodeGen/RISCV/bfloat-imm.ll (+5-4)
- (modified) llvm/test/CodeGen/RISCV/calling-conv-half.ll (+27-18)
- (modified) llvm/test/CodeGen/RISCV/codemodel-lowering.ll (+8-7)
- (modified) llvm/test/CodeGen/RISCV/double-convert.ll (+54-44)
- (modified) llvm/test/CodeGen/RISCV/double-imm.ll (+41-11)
- (modified) llvm/test/CodeGen/RISCV/double-intrinsics.ll (+30-24)
- (modified) llvm/test/CodeGen/RISCV/double-round-conv.ll (+25-20)
- (modified) llvm/test/CodeGen/RISCV/double-zfa.ll (+74-30)
- (modified) llvm/test/CodeGen/RISCV/float-convert.ll (+52-46)
- (modified) llvm/test/CodeGen/RISCV/float-imm.ll (+8-6)
- (modified) llvm/test/CodeGen/RISCV/float-round-conv-sat.ll (+50-40)
- (modified) llvm/test/CodeGen/RISCV/half-convert.ll (+529-447)
- (modified) llvm/test/CodeGen/RISCV/half-imm.ll (+20-23)
- (modified) llvm/test/CodeGen/RISCV/half-intrinsics.ll (+30-24)
- (modified) llvm/test/CodeGen/RISCV/half-round-conv-sat.ll (+200-170)
- (modified) llvm/test/CodeGen/RISCV/half-round-conv.ll (+195-180)
- (modified) llvm/test/CodeGen/RISCV/half-zfa-fli.ll (+36-24)
- (modified) llvm/test/CodeGen/RISCV/repeated-fp-divisors.ll (+3-2)
- (modified) llvm/test/CodeGen/RISCV/rvv/ceil-vp.ll (+578-284)
- (modified) llvm/test/CodeGen/RISCV/rvv/double-round-conv.ll (+96-64)
- (modified) llvm/test/CodeGen/RISCV/rvv/fceil-constrained-sdnode.ll (+156-78)
- (modified) llvm/test/CodeGen/RISCV/rvv/fceil-sdnode.ll (+140-70)
- (modified) llvm/test/CodeGen/RISCV/rvv/ffloor-constrained-sdnode.ll (+156-78)
- (modified) llvm/test/CodeGen/RISCV/rvv/ffloor-sdnode.ll (+140-70)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-ceil-vp.ll (+572-279)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fceil-constrained-sdnode.ll (+156-78)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-ffloor-constrained-sdnode.ll (+156-78)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-floor-vp.ll (+1124-279)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fnearbyint-constrained-sdnode.ll (+15-58)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp-shuffles.ll (+75-37)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp.ll (+630-112)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fp2i-sat.ll (+58-52)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fround-costrained-sdnode.ll (+156-78)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-fround.ll (+266-72)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-froundeven-constrained-sdnode.ll (+156-78)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-froundeven.ll (+266-72)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-ftrunc-constrained-sdnode.ll (+18-68)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-nearbyint-vp.ll (+24-267)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-rint-vp.ll (+512-249)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-round-vp.ll (+1124-279)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundeven-vp.ll (+24-275)
- (modified) llvm/test/CodeGen/RISCV/rvv/fixed-vectors-roundtozero-vp.ll (+24-275)
- (modified) llvm/test/CodeGen/RISCV/rvv/floor-vp.ll (+578-284)
- (modified) llvm/test/CodeGen/RISCV/rvv/fnearbyint-constrained-sdnode.ll (+18-76)
- (modified) llvm/test/CodeGen/RISCV/rvv/fnearbyint-sdnode.ll (+18-68)
- (modified) llvm/test/CodeGen/RISCV/rvv/fptosi-sat.ll (-36)
- (modified) llvm/test/CodeGen/RISCV/rvv/fptoui-sat.ll (+12-8)
- (modified) llvm/test/CodeGen/RISCV/rvv/frint-sdnode.ll (+124-62)
- (modified) llvm/test/CodeGen/RISCV/rvv/fround-costrained-sdnode.ll (+156-78)
- (modified) llvm/test/CodeGen/RISCV/rvv/fround-sdnode.ll (+140-70)
- (modified) llvm/test/CodeGen/RISCV/rvv/froundeven-constrained-sdnode.ll (+18-76)
- (modified) llvm/test/CodeGen/RISCV/rvv/froundeven-sdnode.ll (+18-68)
- (modified) llvm/test/CodeGen/RISCV/rvv/ftrunc-constrained-sdnode.ll (+18-68)
- (modified) llvm/test/CodeGen/RISCV/rvv/ftrunc-sdnode.ll (+18-60)
- (modified) llvm/test/CodeGen/RISCV/rvv/half-round-conv.ll (+77-322)
- (modified) llvm/test/CodeGen/RISCV/rvv/nearbyint-vp.ll (+36-293)
- (modified) llvm/test/CodeGen/RISCV/rvv/rint-vp.ll (+1020-258)
- (modified) llvm/test/CodeGen/RISCV/rvv/round-vp.ll (+36-282)
- (modified) llvm/test/CodeGen/RISCV/rvv/roundeven-vp.ll (+36-282)
- (modified) llvm/test/CodeGen/RISCV/rvv/roundtozero-vp.ll (+36-282)
- (modified) llvm/test/CodeGen/RISCV/rvv/vfma-vp-combine.ll (-25)
- (modified) llvm/test/CodeGen/RISCV/rvv/vreductions-fp-sdnode.ll (+4-4)
- (modified) llvm/test/CodeGen/RISCV/rvv/vsetvli-insert-crossbb.ll (+12-12)
- (modified) llvm/test/CodeGen/RISCV/zfbfmin.ll (+3-2)
``````````diff
diff --git a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
index bd4150c87eabbd0..4d71798a967a144 100644
--- a/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
+++ b/llvm/lib/Target/RISCV/RISCVISelLowering.cpp
@@ -68,12 +68,6 @@ static cl::opt<unsigned> NumRepeatedDivisors(
"transformation to multiplications by the reciprocal"),
cl::init(2));
-static cl::opt<int>
- FPImmCost(DEBUG_TYPE "-fpimm-cost", cl::Hidden,
- cl::desc("Give the maximum number of instructions that we will "
- "use for creating a floating-point immediate value"),
- cl::init(2));
-
RISCVTargetLowering::RISCVTargetLowering(const TargetMachine &TM,
const RISCVSubtarget &STI)
: TargetLowering(TM), Subtarget(STI) {
@@ -2026,10 +2020,10 @@ bool RISCVTargetLowering::isFPImmLegal(const APFloat &Imm, EVT VT,
// Building an integer and then converting requires a fmv at the end of
// the integer sequence.
- const int Cost =
+ const unsigned Cost =
1 + RISCVMatInt::getIntMatCost(Imm.bitcastToAPInt(), Subtarget.getXLen(),
Subtarget.getFeatureBits());
- return Cost <= FPImmCost;
+ return Cost <= Subtarget.getMaxBuildIntsCost();
}
// TODO: This is very conservative.
diff --git a/llvm/test/CodeGen/RISCV/bfloat-convert.ll b/llvm/test/CodeGen/RISCV/bfloat-convert.ll
index 8a0c4240d161bfb..5debb81e6d47447 100644
--- a/llvm/test/CodeGen/RISCV/bfloat-convert.ll
+++ b/llvm/test/CodeGen/RISCV/bfloat-convert.ll
@@ -55,11 +55,12 @@ define i16 @fcvt_si_bf16_sat(bfloat %a) nounwind {
; CHECK32ZFBFMIN-NEXT: fcvt.s.bf16 fa5, fa0
; CHECK32ZFBFMIN-NEXT: feq.s a0, fa5, fa5
; CHECK32ZFBFMIN-NEXT: neg a0, a0
-; CHECK32ZFBFMIN-NEXT: lui a1, %hi(.LCPI1_0)
-; CHECK32ZFBFMIN-NEXT: flw fa4, %lo(.LCPI1_0)(a1)
; CHECK32ZFBFMIN-NEXT: lui a1, 815104
-; CHECK32ZFBFMIN-NEXT: fmv.w.x fa3, a1
-; CHECK32ZFBFMIN-NEXT: fmax.s fa5, fa5, fa3
+; CHECK32ZFBFMIN-NEXT: fmv.w.x fa4, a1
+; CHECK32ZFBFMIN-NEXT: fmax.s fa5, fa5, fa4
+; CHECK32ZFBFMIN-NEXT: lui a1, 290816
+; CHECK32ZFBFMIN-NEXT: addi a1, a1, -512
+; CHECK32ZFBFMIN-NEXT: fmv.w.x fa4, a1
; CHECK32ZFBFMIN-NEXT: fmin.s fa5, fa5, fa4
; CHECK32ZFBFMIN-NEXT: fcvt.w.s a1, fa5, rtz
; CHECK32ZFBFMIN-NEXT: and a0, a0, a1
@@ -71,12 +72,13 @@ define i16 @fcvt_si_bf16_sat(bfloat %a) nounwind {
; RV32ID-NEXT: slli a0, a0, 16
; RV32ID-NEXT: fmv.w.x fa5, a0
; RV32ID-NEXT: feq.s a0, fa5, fa5
-; RV32ID-NEXT: lui a1, %hi(.LCPI1_0)
-; RV32ID-NEXT: flw fa4, %lo(.LCPI1_0)(a1)
-; RV32ID-NEXT: lui a1, 815104
-; RV32ID-NEXT: fmv.w.x fa3, a1
-; RV32ID-NEXT: fmax.s fa5, fa5, fa3
; RV32ID-NEXT: neg a0, a0
+; RV32ID-NEXT: lui a1, 815104
+; RV32ID-NEXT: fmv.w.x fa4, a1
+; RV32ID-NEXT: fmax.s fa5, fa5, fa4
+; RV32ID-NEXT: lui a1, 290816
+; RV32ID-NEXT: addi a1, a1, -512
+; RV32ID-NEXT: fmv.w.x fa4, a1
; RV32ID-NEXT: fmin.s fa5, fa5, fa4
; RV32ID-NEXT: fcvt.w.s a1, fa5, rtz
; RV32ID-NEXT: and a0, a0, a1
@@ -86,12 +88,13 @@ define i16 @fcvt_si_bf16_sat(bfloat %a) nounwind {
; CHECK64ZFBFMIN: # %bb.0: # %start
; CHECK64ZFBFMIN-NEXT: fcvt.s.bf16 fa5, fa0
; CHECK64ZFBFMIN-NEXT: feq.s a0, fa5, fa5
-; CHECK64ZFBFMIN-NEXT: lui a1, %hi(.LCPI1_0)
-; CHECK64ZFBFMIN-NEXT: flw fa4, %lo(.LCPI1_0)(a1)
-; CHECK64ZFBFMIN-NEXT: lui a1, 815104
-; CHECK64ZFBFMIN-NEXT: fmv.w.x fa3, a1
-; CHECK64ZFBFMIN-NEXT: fmax.s fa5, fa5, fa3
; CHECK64ZFBFMIN-NEXT: neg a0, a0
+; CHECK64ZFBFMIN-NEXT: lui a1, 815104
+; CHECK64ZFBFMIN-NEXT: fmv.w.x fa4, a1
+; CHECK64ZFBFMIN-NEXT: fmax.s fa5, fa5, fa4
+; CHECK64ZFBFMIN-NEXT: lui a1, 290816
+; CHECK64ZFBFMIN-NEXT: addi a1, a1, -512
+; CHECK64ZFBFMIN-NEXT: fmv.w.x fa4, a1
; CHECK64ZFBFMIN-NEXT: fmin.s fa5, fa5, fa4
; CHECK64ZFBFMIN-NEXT: fcvt.l.s a1, fa5, rtz
; CHECK64ZFBFMIN-NEXT: and a0, a0, a1
@@ -105,12 +108,13 @@ define i16 @fcvt_si_bf16_sat(bfloat %a) nounwind {
; RV64ID-NEXT: slli a0, a0, 16
; RV64ID-NEXT: fmv.w.x fa5, a0
; RV64ID-NEXT: feq.s a0, fa5, fa5
-; RV64ID-NEXT: lui a1, %hi(.LCPI1_0)
-; RV64ID-NEXT: flw fa4, %lo(.LCPI1_0)(a1)
-; RV64ID-NEXT: lui a1, 815104
-; RV64ID-NEXT: fmv.w.x fa3, a1
-; RV64ID-NEXT: fmax.s fa5, fa5, fa3
; RV64ID-NEXT: neg a0, a0
+; RV64ID-NEXT: lui a1, 815104
+; RV64ID-NEXT: fmv.w.x fa4, a1
+; RV64ID-NEXT: fmax.s fa5, fa5, fa4
+; RV64ID-NEXT: lui a1, 290816
+; RV64ID-NEXT: addi a1, a1, -512
+; RV64ID-NEXT: fmv.w.x fa4, a1
; RV64ID-NEXT: fmin.s fa5, fa5, fa4
; RV64ID-NEXT: fcvt.l.s a1, fa5, rtz
; RV64ID-NEXT: and a0, a0, a1
@@ -158,12 +162,13 @@ define i16 @fcvt_ui_bf16(bfloat %a) nounwind {
define i16 @fcvt_ui_bf16_sat(bfloat %a) nounwind {
; CHECK32ZFBFMIN-LABEL: fcvt_ui_bf16_sat:
; CHECK32ZFBFMIN: # %bb.0: # %start
-; CHECK32ZFBFMIN-NEXT: lui a0, %hi(.LCPI3_0)
-; CHECK32ZFBFMIN-NEXT: flw fa5, %lo(.LCPI3_0)(a0)
-; CHECK32ZFBFMIN-NEXT: fcvt.s.bf16 fa4, fa0
-; CHECK32ZFBFMIN-NEXT: fmv.w.x fa3, zero
-; CHECK32ZFBFMIN-NEXT: fmax.s fa4, fa4, fa3
-; CHECK32ZFBFMIN-NEXT: fmin.s fa5, fa4, fa5
+; CHECK32ZFBFMIN-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK32ZFBFMIN-NEXT: fmv.w.x fa4, zero
+; CHECK32ZFBFMIN-NEXT: fmax.s fa5, fa5, fa4
+; CHECK32ZFBFMIN-NEXT: lui a0, 292864
+; CHECK32ZFBFMIN-NEXT: addi a0, a0, -256
+; CHECK32ZFBFMIN-NEXT: fmv.w.x fa4, a0
+; CHECK32ZFBFMIN-NEXT: fmin.s fa5, fa5, fa4
; CHECK32ZFBFMIN-NEXT: fcvt.wu.s a0, fa5, rtz
; CHECK32ZFBFMIN-NEXT: ret
;
@@ -171,38 +176,41 @@ define i16 @fcvt_ui_bf16_sat(bfloat %a) nounwind {
; RV32ID: # %bb.0: # %start
; RV32ID-NEXT: fmv.x.w a0, fa0
; RV32ID-NEXT: slli a0, a0, 16
-; RV32ID-NEXT: lui a1, %hi(.LCPI3_0)
-; RV32ID-NEXT: flw fa5, %lo(.LCPI3_0)(a1)
+; RV32ID-NEXT: fmv.w.x fa5, a0
+; RV32ID-NEXT: fmv.w.x fa4, zero
+; RV32ID-NEXT: fmax.s fa5, fa5, fa4
+; RV32ID-NEXT: lui a0, 292864
+; RV32ID-NEXT: addi a0, a0, -256
; RV32ID-NEXT: fmv.w.x fa4, a0
-; RV32ID-NEXT: fmv.w.x fa3, zero
-; RV32ID-NEXT: fmax.s fa4, fa4, fa3
-; RV32ID-NEXT: fmin.s fa5, fa4, fa5
+; RV32ID-NEXT: fmin.s fa5, fa5, fa4
; RV32ID-NEXT: fcvt.wu.s a0, fa5, rtz
; RV32ID-NEXT: ret
;
; CHECK64ZFBFMIN-LABEL: fcvt_ui_bf16_sat:
; CHECK64ZFBFMIN: # %bb.0: # %start
-; CHECK64ZFBFMIN-NEXT: lui a0, %hi(.LCPI3_0)
-; CHECK64ZFBFMIN-NEXT: flw fa5, %lo(.LCPI3_0)(a0)
-; CHECK64ZFBFMIN-NEXT: fcvt.s.bf16 fa4, fa0
-; CHECK64ZFBFMIN-NEXT: fmv.w.x fa3, zero
-; CHECK64ZFBFMIN-NEXT: fmax.s fa4, fa4, fa3
-; CHECK64ZFBFMIN-NEXT: fmin.s fa5, fa4, fa5
+; CHECK64ZFBFMIN-NEXT: fcvt.s.bf16 fa5, fa0
+; CHECK64ZFBFMIN-NEXT: fmv.w.x fa4, zero
+; CHECK64ZFBFMIN-NEXT: fmax.s fa5, fa5, fa4
+; CHECK64ZFBFMIN-NEXT: lui a0, 292864
+; CHECK64ZFBFMIN-NEXT: addi a0, a0, -256
+; CHECK64ZFBFMIN-NEXT: fmv.w.x fa4, a0
+; CHECK64ZFBFMIN-NEXT: fmin.s fa5, fa5, fa4
; CHECK64ZFBFMIN-NEXT: fcvt.lu.s a0, fa5, rtz
; CHECK64ZFBFMIN-NEXT: ret
;
; RV64ID-LABEL: fcvt_ui_bf16_sat:
; RV64ID: # %bb.0: # %start
-; RV64ID-NEXT: lui a0, %hi(.LCPI3_0)
-; RV64ID-NEXT: flw fa5, %lo(.LCPI3_0)(a0)
; RV64ID-NEXT: fmv.x.w a0, fa0
; RV64ID-NEXT: slli a0, a0, 48
; RV64ID-NEXT: srli a0, a0, 48
; RV64ID-NEXT: slli a0, a0, 16
+; RV64ID-NEXT: fmv.w.x fa5, a0
+; RV64ID-NEXT: fmv.w.x fa4, zero
+; RV64ID-NEXT: fmax.s fa5, fa5, fa4
+; RV64ID-NEXT: lui a0, 292864
+; RV64ID-NEXT: addi a0, a0, -256
; RV64ID-NEXT: fmv.w.x fa4, a0
-; RV64ID-NEXT: fmv.w.x fa3, zero
-; RV64ID-NEXT: fmax.s fa4, fa4, fa3
-; RV64ID-NEXT: fmin.s fa5, fa4, fa5
+; RV64ID-NEXT: fmin.s fa5, fa5, fa4
; RV64ID-NEXT: fcvt.lu.s a0, fa5, rtz
; RV64ID-NEXT: ret
start:
@@ -492,8 +500,9 @@ define i64 @fcvt_l_bf16_sat(bfloat %a) nounwind {
; RV32IZFBFMIN-NEXT: # %bb.1: # %start
; RV32IZFBFMIN-NEXT: mv a2, a1
; RV32IZFBFMIN-NEXT: .LBB10_2: # %start
-; RV32IZFBFMIN-NEXT: lui a1, %hi(.LCPI10_0)
-; RV32IZFBFMIN-NEXT: flw fa5, %lo(.LCPI10_0)(a1)
+; RV32IZFBFMIN-NEXT: lui a1, 389120
+; RV32IZFBFMIN-NEXT: addi a1, a1, -1
+; RV32IZFBFMIN-NEXT: fmv.w.x fa5, a1
; RV32IZFBFMIN-NEXT: flt.s a3, fa5, fs0
; RV32IZFBFMIN-NEXT: beqz a3, .LBB10_4
; RV32IZFBFMIN-NEXT: # %bb.3:
@@ -502,9 +511,9 @@ define i64 @fcvt_l_bf16_sat(bfloat %a) nounwind {
; RV32IZFBFMIN-NEXT: feq.s a1, fs0, fs0
; RV32IZFBFMIN-NEXT: neg a4, a1
; RV32IZFBFMIN-NEXT: and a1, a4, a2
+; RV32IZFBFMIN-NEXT: neg a2, s0
+; RV32IZFBFMIN-NEXT: and a0, a2, a0
; RV32IZFBFMIN-NEXT: neg a2, a3
-; RV32IZFBFMIN-NEXT: neg a3, s0
-; RV32IZFBFMIN-NEXT: and a0, a3, a0
; RV32IZFBFMIN-NEXT: or a0, a2, a0
; RV32IZFBFMIN-NEXT: and a0, a4, a0
; RV32IZFBFMIN-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
@@ -531,8 +540,9 @@ define i64 @fcvt_l_bf16_sat(bfloat %a) nounwind {
; R32IDZFBFMIN-NEXT: # %bb.1: # %start
; R32IDZFBFMIN-NEXT: mv a2, a1
; R32IDZFBFMIN-NEXT: .LBB10_2: # %start
-; R32IDZFBFMIN-NEXT: lui a1, %hi(.LCPI10_0)
-; R32IDZFBFMIN-NEXT: flw fa5, %lo(.LCPI10_0)(a1)
+; R32IDZFBFMIN-NEXT: lui a1, 389120
+; R32IDZFBFMIN-NEXT: addi a1, a1, -1
+; R32IDZFBFMIN-NEXT: fmv.w.x fa5, a1
; R32IDZFBFMIN-NEXT: flt.s a3, fa5, fs0
; R32IDZFBFMIN-NEXT: beqz a3, .LBB10_4
; R32IDZFBFMIN-NEXT: # %bb.3:
@@ -541,9 +551,9 @@ define i64 @fcvt_l_bf16_sat(bfloat %a) nounwind {
; R32IDZFBFMIN-NEXT: feq.s a1, fs0, fs0
; R32IDZFBFMIN-NEXT: neg a4, a1
; R32IDZFBFMIN-NEXT: and a1, a4, a2
+; R32IDZFBFMIN-NEXT: neg a2, s0
+; R32IDZFBFMIN-NEXT: and a0, a2, a0
; R32IDZFBFMIN-NEXT: neg a2, a3
-; R32IDZFBFMIN-NEXT: neg a3, s0
-; R32IDZFBFMIN-NEXT: and a0, a3, a0
; R32IDZFBFMIN-NEXT: or a0, a2, a0
; R32IDZFBFMIN-NEXT: and a0, a4, a0
; R32IDZFBFMIN-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
@@ -572,8 +582,9 @@ define i64 @fcvt_l_bf16_sat(bfloat %a) nounwind {
; RV32ID-NEXT: # %bb.1: # %start
; RV32ID-NEXT: mv a2, a1
; RV32ID-NEXT: .LBB10_2: # %start
-; RV32ID-NEXT: lui a1, %hi(.LCPI10_0)
-; RV32ID-NEXT: flw fa5, %lo(.LCPI10_0)(a1)
+; RV32ID-NEXT: lui a1, 389120
+; RV32ID-NEXT: addi a1, a1, -1
+; RV32ID-NEXT: fmv.w.x fa5, a1
; RV32ID-NEXT: flt.s a3, fa5, fs0
; RV32ID-NEXT: beqz a3, .LBB10_4
; RV32ID-NEXT: # %bb.3:
@@ -665,30 +676,59 @@ define i64 @fcvt_lu_bf16(bfloat %a) nounwind {
}
define i64 @fcvt_lu_bf16_sat(bfloat %a) nounwind {
-; CHECK32ZFBFMIN-LABEL: fcvt_lu_bf16_sat:
-; CHECK32ZFBFMIN: # %bb.0: # %start
-; CHECK32ZFBFMIN-NEXT: addi sp, sp, -16
-; CHECK32ZFBFMIN-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
-; CHECK32ZFBFMIN-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
-; CHECK32ZFBFMIN-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
-; CHECK32ZFBFMIN-NEXT: lui a0, %hi(.LCPI12_0)
-; CHECK32ZFBFMIN-NEXT: flw fa5, %lo(.LCPI12_0)(a0)
-; CHECK32ZFBFMIN-NEXT: fcvt.s.bf16 fa0, fa0
-; CHECK32ZFBFMIN-NEXT: flt.s a0, fa5, fa0
-; CHECK32ZFBFMIN-NEXT: neg s0, a0
-; CHECK32ZFBFMIN-NEXT: fmv.w.x fa5, zero
-; CHECK32ZFBFMIN-NEXT: fle.s a0, fa5, fa0
-; CHECK32ZFBFMIN-NEXT: neg s1, a0
-; CHECK32ZFBFMIN-NEXT: call __fixunssfdi at plt
-; CHECK32ZFBFMIN-NEXT: and a0, s1, a0
-; CHECK32ZFBFMIN-NEXT: or a0, s0, a0
-; CHECK32ZFBFMIN-NEXT: and a1, s1, a1
-; CHECK32ZFBFMIN-NEXT: or a1, s0, a1
-; CHECK32ZFBFMIN-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
-; CHECK32ZFBFMIN-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
-; CHECK32ZFBFMIN-NEXT: lw s1, 4(sp) # 4-byte Folded Reload
-; CHECK32ZFBFMIN-NEXT: addi sp, sp, 16
-; CHECK32ZFBFMIN-NEXT: ret
+; RV32IZFBFMIN-LABEL: fcvt_lu_bf16_sat:
+; RV32IZFBFMIN: # %bb.0: # %start
+; RV32IZFBFMIN-NEXT: addi sp, sp, -16
+; RV32IZFBFMIN-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: fsw fs0, 4(sp) # 4-byte Folded Spill
+; RV32IZFBFMIN-NEXT: fcvt.s.bf16 fs0, fa0
+; RV32IZFBFMIN-NEXT: fmv.w.x fa5, zero
+; RV32IZFBFMIN-NEXT: fle.s a0, fa5, fs0
+; RV32IZFBFMIN-NEXT: neg s0, a0
+; RV32IZFBFMIN-NEXT: fmv.s fa0, fs0
+; RV32IZFBFMIN-NEXT: call __fixunssfdi at plt
+; RV32IZFBFMIN-NEXT: and a0, s0, a0
+; RV32IZFBFMIN-NEXT: lui a2, 391168
+; RV32IZFBFMIN-NEXT: addi a2, a2, -1
+; RV32IZFBFMIN-NEXT: fmv.w.x fa5, a2
+; RV32IZFBFMIN-NEXT: flt.s a2, fa5, fs0
+; RV32IZFBFMIN-NEXT: neg a2, a2
+; RV32IZFBFMIN-NEXT: or a0, a2, a0
+; RV32IZFBFMIN-NEXT: and a1, s0, a1
+; RV32IZFBFMIN-NEXT: or a1, a2, a1
+; RV32IZFBFMIN-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: flw fs0, 4(sp) # 4-byte Folded Reload
+; RV32IZFBFMIN-NEXT: addi sp, sp, 16
+; RV32IZFBFMIN-NEXT: ret
+;
+; R32IDZFBFMIN-LABEL: fcvt_lu_bf16_sat:
+; R32IDZFBFMIN: # %bb.0: # %start
+; R32IDZFBFMIN-NEXT: addi sp, sp, -16
+; R32IDZFBFMIN-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
+; R32IDZFBFMIN-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
+; R32IDZFBFMIN-NEXT: fsd fs0, 0(sp) # 8-byte Folded Spill
+; R32IDZFBFMIN-NEXT: fcvt.s.bf16 fs0, fa0
+; R32IDZFBFMIN-NEXT: fmv.w.x fa5, zero
+; R32IDZFBFMIN-NEXT: fle.s a0, fa5, fs0
+; R32IDZFBFMIN-NEXT: neg s0, a0
+; R32IDZFBFMIN-NEXT: fmv.s fa0, fs0
+; R32IDZFBFMIN-NEXT: call __fixunssfdi at plt
+; R32IDZFBFMIN-NEXT: and a0, s0, a0
+; R32IDZFBFMIN-NEXT: lui a2, 391168
+; R32IDZFBFMIN-NEXT: addi a2, a2, -1
+; R32IDZFBFMIN-NEXT: fmv.w.x fa5, a2
+; R32IDZFBFMIN-NEXT: flt.s a2, fa5, fs0
+; R32IDZFBFMIN-NEXT: neg a2, a2
+; R32IDZFBFMIN-NEXT: or a0, a2, a0
+; R32IDZFBFMIN-NEXT: and a1, s0, a1
+; R32IDZFBFMIN-NEXT: or a1, a2, a1
+; R32IDZFBFMIN-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
+; R32IDZFBFMIN-NEXT: lw s0, 8(sp) # 4-byte Folded Reload
+; R32IDZFBFMIN-NEXT: fld fs0, 0(sp) # 8-byte Folded Reload
+; R32IDZFBFMIN-NEXT: addi sp, sp, 16
+; R32IDZFBFMIN-NEXT: ret
;
; RV32ID-LABEL: fcvt_lu_bf16_sat:
; RV32ID: # %bb.0: # %start
@@ -696,11 +736,12 @@ define i64 @fcvt_lu_bf16_sat(bfloat %a) nounwind {
; RV32ID-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
; RV32ID-NEXT: sw s0, 8(sp) # 4-byte Folded Spill
; RV32ID-NEXT: sw s1, 4(sp) # 4-byte Folded Spill
-; RV32ID-NEXT: lui a0, %hi(.LCPI12_0)
-; RV32ID-NEXT: flw fa5, %lo(.LCPI12_0)(a0)
; RV32ID-NEXT: fmv.x.w a0, fa0
; RV32ID-NEXT: slli a0, a0, 16
; RV32ID-NEXT: fmv.w.x fa0, a0
+; RV32ID-NEXT: lui a0, 391168
+; RV32ID-NEXT: addi a0, a0, -1
+; RV32ID-NEXT: fmv.w.x fa5, a0
; RV32ID-NEXT: flt.s a0, fa5, fa0
; RV32ID-NEXT: neg s0, a0
; RV32ID-NEXT: fmv.w.x fa5, zero
diff --git a/llvm/test/CodeGen/RISCV/bfloat-imm.ll b/llvm/test/CodeGen/RISCV/bfloat-imm.ll
index cd4e960b5a062a0..ed2c6f59f8fd576 100644
--- a/llvm/test/CodeGen/RISCV/bfloat-imm.ll
+++ b/llvm/test/CodeGen/RISCV/bfloat-imm.ll
@@ -1,14 +1,15 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 2
; RUN: llc -mtriple=riscv32 -mattr=+experimental-zfbfmin -verify-machineinstrs \
-; RUN: -target-abi ilp32f < %s | FileCheck %s
+; RUN: -target-abi ilp32f < %s | FileCheck --check-prefixes=CHECK %s
; RUN: llc -mtriple=riscv64 -mattr=+experimental-zfbfmin -verify-machineinstrs \
-; RUN: -target-abi lp64f < %s | FileCheck %s
+; RUN: -target-abi lp64f < %s | FileCheck --check-prefixes=CHECK %s
define bfloat @bfloat_imm() nounwind {
; CHECK-LABEL: bfloat_imm:
; CHECK: # %bb.0:
-; CHECK-NEXT: lui a0, %hi(.LCPI0_0)
-; CHECK-NEXT: flh fa0, %lo(.LCPI0_0)(a0)
+; CHECK-NEXT: lui a0, 4
+; CHECK-NEXT: addi a0, a0, 64
+; CHECK-NEXT: fmv.h.x fa0, a0
; CHECK-NEXT: ret
ret bfloat 3.0
}
diff --git a/llvm/test/CodeGen/RISCV/calling-conv-half.ll b/llvm/test/CodeGen/RISCV/calling-conv-half.ll
index 6587f0c8c5af7bf..ac0aa7b620c616f 100644
--- a/llvm/test/CodeGen/RISCV/calling-conv-half.ll
+++ b/llvm/test/CodeGen/RISCV/calling-conv-half.ll
@@ -415,8 +415,9 @@ define i32 @caller_half_on_stack() nounwind {
; RV32-ILP32F: # %bb.0:
; RV32-ILP32F-NEXT: addi sp, sp, -16
; RV32-ILP32F-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
-; RV32-ILP32F-NEXT: lui a0, %hi(.LCPI3_0)
-; RV32-ILP32F-NEXT: flw fa0, %lo(.LCPI3_0)(a0)
+; RV32-ILP32F-NEXT: lui a0, 1048565
+; RV32-ILP32F-NEXT: addi a0, a0, -1792
+; RV32-ILP32F-NEXT: fmv.w.x fa0, a0
; RV32-ILP32F-NEXT: li a0, 1
; RV32-ILP32F-NEXT: li a1, 2
; RV32-ILP32F-NEXT: li a2, 3
@@ -434,8 +435,9 @@ define i32 @caller_half_on_stack() nounwind {
; RV64-LP64F: # %bb.0:
; RV64-LP64F-NEXT: addi sp, sp, -16
; RV64-LP64F-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
-; RV64-LP64F-NEXT: lui a0, %hi(.LCPI3_0)
-; RV64-LP64F-NEXT: flw fa0, %lo(.LCPI3_0)(a0)
+; RV64-LP64F-NEXT: lui a0, 1048565
+; RV64-LP64F-NEXT: addi a0, a0, -1792
+; RV64-LP64F-NEXT: fmv.w.x fa0, a0
; RV64-LP64F-NEXT: li a0, 1
; RV64-LP64F-NEXT: li a1, 2
; RV64-LP64F-NEXT: li a2, 3
@@ -453,8 +455,9 @@ define i32 @caller_half_on_stack() nounwind {
; RV32-ILP32ZFHMIN: # %bb.0:
; RV32-ILP32ZFHMIN-NEXT: addi sp, sp, -16
; RV32-ILP32ZFHMIN-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
-; RV32-ILP32ZFHMIN-NEXT: lui a0, %hi(.LCPI3_0)
-; RV32-ILP32ZFHMIN-NEXT: flh fa0, %lo(.LCPI3_0)(a0)
+; RV32-ILP32ZFHMIN-NEXT: lui a0, 5
+; RV32-ILP32ZFHMIN-NEXT: addi a0, a0, -1792
+; RV32-ILP32ZFHMIN-NEXT: fmv.h.x fa0, a0
; RV32-ILP32ZFHMIN-NEXT: li a0, 1
; RV32-ILP32ZFHMIN-NEXT: li a1, 2
; RV32-ILP32ZFHMIN-NEXT: li a2, 3
@@ -472,8 +475,9 @@ define i32 @caller_half_on_stack() nounwind {
; RV64-LP64ZFHMIN: # %bb.0:
; RV64-LP64ZFHMIN-NEXT: addi sp, sp, -16
; RV64-LP64ZFHMIN-NEXT: sd ra, 8(sp) # 8-byte Folded Spill
-; RV64-LP64ZFHMIN-NEXT: lui a0, %hi(.LCPI3_0)
-; RV64-LP64ZFHMIN-NEXT: flh fa0, %lo(.LCPI3_0)(a0)
+; RV64-LP64ZFHMIN-NEXT: lui a0, 5
+; RV64-LP64ZFHMIN-NEXT: addi a0, a0, -1792
+; RV64-LP64ZFHMIN-NEXT: fmv.h.x fa0, a0
; RV64-LP64ZFHMIN-NEXT: li a0, 1
; RV64-LP64ZFHMIN-NEXT: li a1, 2
; RV64-LP64ZFHMIN-NEXT: li a2, 3
@@ -511,33 +515,38 @@ define half @callee_half_ret() nounwind {
;
; RV64IF-LABEL: callee_half_ret:
; RV64IF: # %bb.0:
-; RV64IF-NEXT: lui a0, %hi(.LCPI4_0)
-; RV64IF-NEXT: flw fa5, %lo(.LCPI4_0)(a0)
+; RV64IF-NEXT: lui a0, 1048564
+; RV64IF-NEXT: addi a0, a0, -1024
+; RV64IF-NEXT: fmv.w.x fa5, a0
; RV64IF-NEXT: fmv.x.w a0, fa5
; RV64IF-NEXT: ret
;
; RV32-ILP32F-LABEL: callee_half_ret:
; RV32-ILP32F: # %bb.0:
-; RV32-ILP32F-NEXT: lui a0, %hi(.LCPI4_0)
-; RV32-ILP32F-NEXT: flw fa0, %lo(.LCPI4_0)(a0)
+; RV32-ILP32F-NEXT: lui a0, 1048564
+; RV32-ILP32F-NEXT: addi a0, a0, -1024
+; RV32-ILP32F-NEXT: fmv.w.x fa0, a0
; RV32-ILP32F-NEXT: ret
;
; RV64-LP64F-LABEL: callee_half_ret:
; RV64-LP64F: # %bb.0:
-; RV64-LP64F-NEXT: lui a0, %hi(.LCPI4_0)
-; RV64-LP64F-NEXT: flw fa0, %lo(.LCPI4_0)(a0)
+; RV64-LP64F-NEXT: lui a0, 1048564
+; RV64-LP64F-NEXT: addi a0, a0, -1024
+; RV64-LP64F-NEXT: fmv.w.x fa0, a0
; RV64-LP64F-NEXT: ret
;
; RV32-ILP32ZFHMIN-LABEL: callee_half_ret:
; RV32-ILP32ZFHMIN: # %bb.0:
-; RV32-ILP32ZFHMIN-NEXT: lui a0, %hi(.LCPI4_0)
-; RV32-ILP32ZFHMIN-NEXT: flh fa0, %lo(.LCPI4_0)(a0)
+; RV32-ILP32ZFHMIN-NEXT: li a0, 15
+; RV32-ILP32ZFHMIN-NEXT: slli a0, a0, 10
+; RV32-ILP32ZFHMIN-NEXT: fmv.h.x fa0, a0
; RV32-ILP32ZFHMIN-NEXT: ret
;
; RV64-LP64ZFHMIN-LABEL: callee_half_ret:
; RV64-LP64ZFHMIN: # %bb.0:
-; RV64-LP64ZFHMIN-NEXT: lui a0, %hi(.LCPI4_0)
-; RV64-LP64ZFHMIN-NEXT: flh fa0, %lo(.LCPI4_0)(a0)
+; RV64-LP64ZFHMIN-NEXT: li a0, 15
+; RV64-LP64ZFHMIN-NEXT: slli a0, a0, 10
+; RV64-LP64ZFHMIN-NEXT: fmv.h.x fa0, a0
; RV64-LP64ZFHMIN-NEXT: ret
ret half 1.0
}
diff --git a/llvm/test/CodeGen/RISCV/codemodel-lowering.ll b/llvm/test/CodeGen/RISCV/codemodel-lowering.ll
index 617155b31976187..562f2fd0c270cf7 100644
--- a/llvm/test/CodeGen/RISCV/codemodel-lowering.ll
+++ b/llvm/test/CodeGen/RISCV/codemodel-lowering.ll
@@ -124,16 +124,17 @@ indirectgoto:
define float @lower_constantpool(float %a) nounwind {
; RV32I-SMALL-LABEL: lower_constantpool:
; RV32I-SMALL: # %bb.0:
-; RV32I-SMALL-NEXT: lui a0, %hi(.LCPI3_0)
-; RV32I-SMA...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/68433
More information about the llvm-commits
mailing list