AMDGPU/GlobalISel: Select mul24 intrinsics

3 files changed, 77 insertions, 4 deletions

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td b/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td
index 13afcb4cdb9..50c451be4b8 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td
+++ b/llvm/lib/Target/AMDGPU/AMDGPUInstrInfo.td
@@ -290,10 +290,10 @@ def AMDGPUffbl_b32 : SDNode<"AMDGPUISD::FFBL_B32", SDTIntUnaryOp>;
 
 // Signed and unsigned 24-bit multiply. The highest 8-bits are ignore
 // when performing the mulitply. The result is a 32-bit value.
-def AMDGPUmul_u24 : SDNode<"AMDGPUISD::MUL_U24", SDTIntBinOp,
+def AMDGPUmul_u24_impl : SDNode<"AMDGPUISD::MUL_U24", SDTIntBinOp,
   [SDNPCommutative, SDNPAssociative]
 >;
-def AMDGPUmul_i24 : SDNode<"AMDGPUISD::MUL_I24", SDTIntBinOp,
+def AMDGPUmul_i24_impl : SDNode<"AMDGPUISD::MUL_I24", SDTIntBinOp,
   [SDNPCommutative, SDNPAssociative]
 >;
 
@@ -465,3 +465,11 @@ def AMDGPUpk_u16_u32 : PatFrags<(ops node:$src0, node:$src1),
 def AMDGPUfmad_ftz : PatFrags<(ops node:$src0, node:$src1, node:$src2),
   [(int_amdgcn_fmad_ftz node:$src0, node:$src1, node:$src2),
    (AMDGPUfmad_ftz_impl node:$src0, node:$src1, node:$src2)]>;
+
+def AMDGPUmul_u24 : PatFrags<(ops node:$src0, node:$src1),
+  [(int_amdgcn_mul_u24 node:$src0, node:$src1),
+   (AMDGPUmul_u24_impl node:$src0, node:$src1)]>;
+
+def AMDGPUmul_i24 : PatFrags<(ops node:$src0, node:$src1),
+  [(int_amdgcn_mul_i24 node:$src0, node:$src1),
+   (AMDGPUmul_i24_impl node:$src0, node:$src1)]>;
diff --git a/llvm/lib/Target/AMDGPU/VOP2Instructions.td b/llvm/lib/Target/AMDGPU/VOP2Instructions.td
index 45928a54fd9..94e737d7641 100644
--- a/llvm/lib/Target/AMDGPU/VOP2Instructions.td
+++ b/llvm/lib/Target/AMDGPU/VOP2Instructions.td
@@ -467,9 +467,9 @@ defm V_SUB_F32 : VOP2Inst <"v_sub_f32", VOP_F32_F32_F32, fsub>;
 defm V_SUBREV_F32 : VOP2Inst <"v_subrev_f32", VOP_F32_F32_F32, null_frag, "v_sub_f32">;
 defm V_MUL_LEGACY_F32 : VOP2Inst <"v_mul_legacy_f32", VOP_F32_F32_F32, AMDGPUfmul_legacy>;
 defm V_MUL_F32 : VOP2Inst <"v_mul_f32", VOP_F32_F32_F32, fmul>;
-defm V_MUL_I32_I24 : VOP2Inst <"v_mul_i32_i24", VOP_PAT_GEN<VOP_I32_I32_I32, 2>, AMDGPUmul_i24>;
+defm V_MUL_I32_I24 : VOP2Inst <"v_mul_i32_i24", VOP_I32_I32_I32, AMDGPUmul_i24>;
 defm V_MUL_HI_I32_I24 : VOP2Inst <"v_mul_hi_i32_i24", VOP_PAT_GEN<VOP_I32_I32_I32, 2>, AMDGPUmulhi_i24>;
-defm V_MUL_U32_U24 : VOP2Inst <"v_mul_u32_u24", VOP_PAT_GEN<VOP_I32_I32_I32, 2>, AMDGPUmul_u24>;
+defm V_MUL_U32_U24 : VOP2Inst <"v_mul_u32_u24", VOP_I32_I32_I32, AMDGPUmul_u24>;
 defm V_MUL_HI_U32_U24 : VOP2Inst <"v_mul_hi_u32_u24", VOP_PAT_GEN<VOP_I32_I32_I32, 2>, AMDGPUmulhi_u24>;
 defm V_MIN_F32 : VOP2Inst <"v_min_f32", VOP_F32_F32_F32, fminnum_like>;
 defm V_MAX_F32 : VOP2Inst <"v_max_f32", VOP_F32_F32_F32, fmaxnum_like>;
diff --git a/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-amdgcn.mul.u24.mir b/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-amdgcn.mul.u24.mir
new file mode 100644
index 00000000000..b5b9368ed4e
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/GlobalISel/inst-select-amdgcn.mul.u24.mir
@@ -0,0 +1,65 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -march=amdgcn -mcpu=tahiti -run-pass=instruction-select -verify-machineinstrs %s -o -  | FileCheck -check-prefix=GCN %s
+
+---
+name: mul_u24_vsv
+legalized: true
+regBankSelected: true
+tracksRegLiveness: true
+
+body: |
+  bb.0:
+    liveins: $sgpr0, $vgpr0
+    ; GCN-LABEL: name: mul_u24_vsv
+    ; GCN: liveins: $sgpr0, $vgpr0
+    ; GCN: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
+    ; GCN: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; GCN: [[V_MUL_U32_U24_e64_:%[0-9]+]]:vgpr_32 = V_MUL_U32_U24_e64 [[COPY]], [[COPY1]], implicit $exec
+    ; GCN: S_ENDPGM 0, implicit [[V_MUL_U32_U24_e64_]]
+    %0:sgpr(s32) = COPY $sgpr0
+    %1:vgpr(s32) = COPY $vgpr0
+    %2:vgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.mul.u24), %0, %1
+    S_ENDPGM 0, implicit %2
+...
+
+---
+name: mul_u24_vvs
+legalized: true
+regBankSelected: true
+tracksRegLiveness: true
+
+body: |
+  bb.0:
+    liveins: $sgpr0, $vgpr0
+    ; GCN-LABEL: name: mul_u24_vvs
+    ; GCN: liveins: $sgpr0, $vgpr0
+    ; GCN: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; GCN: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr0
+    ; GCN: [[V_MUL_U32_U24_e64_:%[0-9]+]]:vgpr_32 = V_MUL_U32_U24_e64 [[COPY]], [[COPY1]], implicit $exec
+    ; GCN: S_ENDPGM 0, implicit [[V_MUL_U32_U24_e64_]]
+    %0:vgpr(s32) = COPY $vgpr0
+    %1:sgpr(s32) = COPY $sgpr0
+    %2:vgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.mul.u24), %0, %1
+    S_ENDPGM 0, implicit %2
+...
+
+---
+name: mul_u24_vvv
+legalized: true
+regBankSelected: true
+tracksRegLiveness: true
+
+body: |
+  bb.0:
+    liveins: $vgpr0, $vgpr1
+    ; GCN-LABEL: name: mul_u24_vvv
+    ; GCN: liveins: $vgpr0, $vgpr1
+    ; GCN: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
+    ; GCN: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
+    ; GCN: [[V_MUL_U32_U24_e64_:%[0-9]+]]:vgpr_32 = V_MUL_U32_U24_e64 [[COPY]], [[COPY1]], implicit $exec
+    ; GCN: S_ENDPGM 0, implicit [[V_MUL_U32_U24_e64_]]
+    %0:vgpr(s32) = COPY $vgpr0
+    %1:vgpr(s32) = COPY $vgpr1
+    %2:vgpr(s32) = G_INTRINSIC intrinsic(@llvm.amdgcn.mul.u24), %0, %1
+    S_ENDPGM 0, implicit %2
+...