| ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py |
| ; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=tahiti < %s | FileCheck -check-prefixes=GCN,GFX6 %s |
| ; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=fiji < %s | FileCheck -check-prefixes=GCN,GFX8 %s |
| ; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=gfx900 < %s | FileCheck -check-prefixes=GCN,GFX9 %s |
| ; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=gfx1010 < %s | FileCheck -check-prefixes=GFX10PLUS,GFX10 %s |
| ; RUN: llc -global-isel -new-reg-bank-select -mtriple=amdgcn-amd-amdpal -mcpu=gfx1100 -mattr=-real-true16 -amdgpu-enable-delay-alu=0 < %s | FileCheck -check-prefixes=GFX10PLUS,GFX11 %s |
| |
| define i8 @v_lshr_i8(i8 %value, i8 %amount) { |
| ; GFX6-LABEL: v_lshr_i8: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_and_b32_e32 v1, 0xff, v1 |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xff, v0 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v0, v1, v0 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i8: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v0, v1, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:BYTE_0 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i8: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_lshrrev_b16_sdwa v0, v1, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:BYTE_0 src1_sel:BYTE_0 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i8: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_and_b32_e32 v1, 0xff, v1 |
| ; GFX10PLUS-NEXT: v_and_b32_e32 v0, 0xff, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b16 v0, v1, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i8 %value, %amount |
| ret i8 %result |
| } |
| |
| define i8 @v_lshr_i8_7(i8 %value) { |
| ; GFX6-LABEL: v_lshr_i8_7: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_bfe_u32 v0, v0, 7, 1 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i8_7: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_mov_b32_e32 v1, 7 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v0, v1, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_0 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i8_7: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_mov_b32_e32 v1, 7 |
| ; GFX9-NEXT: v_lshrrev_b16_sdwa v0, v1, v0 dst_sel:DWORD dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:BYTE_0 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i8_7: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_and_b32_e32 v0, 0xff, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b16 v0, 7, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i8 %value, 7 |
| ret i8 %result |
| } |
| |
| define amdgpu_ps i8 @s_lshr_i8(i8 inreg %value, i8 inreg %amount) { |
| ; GFX6-LABEL: s_lshr_i8: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xff |
| ; GFX6-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: s_lshr_i8: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_and_b32 s0, s0, 0xff |
| ; GFX8-NEXT: s_and_b32 s1, s1, 0xff |
| ; GFX8-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX8-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: s_lshr_i8: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_and_b32 s0, s0, 0xff |
| ; GFX9-NEXT: s_and_b32 s1, s1, 0xff |
| ; GFX9-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i8: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s0, s0, 0xff |
| ; GFX10PLUS-NEXT: s_and_b32 s1, s1, 0xff |
| ; GFX10PLUS-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i8 %value, %amount |
| ret i8 %result |
| } |
| |
| define amdgpu_ps i8 @s_lshr_i8_7(i8 inreg %value) { |
| ; GFX6-LABEL: s_lshr_i8_7: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_bfe_u32 s0, s0, 0x10007 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: s_lshr_i8_7: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_and_b32 s0, s0, 0xff |
| ; GFX8-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX8-NEXT: s_lshr_b32 s0, s0, 7 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: s_lshr_i8_7: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_and_b32 s0, s0, 0xff |
| ; GFX9-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, 7 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i8_7: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s0, s0, 0xff |
| ; GFX10PLUS-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, 7 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i8 %value, 7 |
| ret i8 %result |
| } |
| |
| |
| define i24 @v_lshr_i24(i24 %value, i24 %amount) { |
| ; GCN-LABEL: v_lshr_i24: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_and_b32_e32 v1, 0xffffff, v1 |
| ; GCN-NEXT: v_and_b32_e32 v0, 0xffffff, v0 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, v1, v0 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i24: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_and_b32_e32 v1, 0xffffff, v1 |
| ; GFX10PLUS-NEXT: v_and_b32_e32 v0, 0xffffff, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, v1, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i24 %value, %amount |
| ret i24 %result |
| } |
| |
| define i24 @v_lshr_i24_7(i24 %value) { |
| ; GCN-LABEL: v_lshr_i24_7: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_bfe_u32 v0, v0, 7, 17 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i24_7: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_bfe_u32 v0, v0, 7, 17 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i24 %value, 7 |
| ret i24 %result |
| } |
| |
| define amdgpu_ps i24 @s_lshr_i24(i24 inreg %value, i24 inreg %amount) { |
| ; GCN-LABEL: s_lshr_i24: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_and_b32 s0, s0, 0xffffff |
| ; GCN-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i24: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s0, s0, 0xffffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i24 %value, %amount |
| ret i24 %result |
| } |
| |
| define amdgpu_ps i24 @s_lshr_i24_7(i24 inreg %value) { |
| ; GCN-LABEL: s_lshr_i24_7: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_bfe_u32 s0, s0, 0x110007 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i24_7: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_bfe_u32 s0, s0, 0x110007 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i24 %value, 7 |
| ret i24 %result |
| } |
| |
| define i32 @v_lshr_i32(i32 %value, i32 %amount) { |
| ; GCN-LABEL: v_lshr_i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, v1, v0 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, v1, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i32 %value, %amount |
| ret i32 %result |
| } |
| |
| define i32 @v_lshr_i32_31(i32 %value) { |
| ; GCN-LABEL: v_lshr_i32_31: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, 31, v0 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i32_31: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, 31, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i32 %value, 31 |
| ret i32 %result |
| } |
| |
| define amdgpu_ps i32 @s_lshr_i32(i32 inreg %value, i32 inreg %amount) { |
| ; GCN-LABEL: s_lshr_i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i32 %value, %amount |
| ret i32 %result |
| } |
| |
| define amdgpu_ps i32 @s_lshr_i32_31(i32 inreg %value) { |
| ; GCN-LABEL: s_lshr_i32_31: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s0, 31 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i32_31: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, 31 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i32 %value, 31 |
| ret i32 %result |
| } |
| |
| define amdgpu_ps float @lshr_i32_sv(i32 inreg %value, i32 %amount) { |
| ; GFX6-LABEL: lshr_i32_sv: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: v_lshr_b32_e32 v0, s0, v0 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: lshr_i32_sv: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: v_lshrrev_b32_e64 v0, v0, s0 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: lshr_i32_sv: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: v_lshrrev_b32_e64 v0, v0, s0 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_i32_sv: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e64 v0, v0, s0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i32 %value, %amount |
| %cast = bitcast i32 %result to float |
| ret float %cast |
| } |
| |
| define amdgpu_ps float @lshr_i32_vs(i32 %value, i32 inreg %amount) { |
| ; GCN-LABEL: lshr_i32_vs: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, s0, v0 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_i32_vs: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, s0, v0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i32 %value, %amount |
| %cast = bitcast i32 %result to float |
| ret float %cast |
| } |
| |
| define <2 x i32> @v_lshr_v2i32(<2 x i32> %value, <2 x i32> %amount) { |
| ; GCN-LABEL: v_lshr_v2i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, v2, v0 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v1, v3, v1 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v2i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, v2, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v1, v3, v1 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <2 x i32> %value, %amount |
| ret <2 x i32> %result |
| } |
| |
| define <2 x i32> @v_lshr_v2i32_31(<2 x i32> %value) { |
| ; GCN-LABEL: v_lshr_v2i32_31: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, 31, v0 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v1, 31, v1 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v2i32_31: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, 31, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v1, 31, v1 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <2 x i32> %value, <i32 31, i32 31> |
| ret <2 x i32> %result |
| } |
| |
| define amdgpu_ps <2 x i32> @s_lshr_v2i32(<2 x i32> inreg %value, <2 x i32> inreg %amount) { |
| ; GCN-LABEL: s_lshr_v2i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s0, s2 |
| ; GCN-NEXT: s_lshr_b32 s1, s1, s3 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v2i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s2 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, s3 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <2 x i32> %value, %amount |
| ret <2 x i32> %result |
| } |
| |
| define <3 x i32> @v_lshr_v3i32(<3 x i32> %value, <3 x i32> %amount) { |
| ; GCN-LABEL: v_lshr_v3i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, v3, v0 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v1, v4, v1 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v2, v5, v2 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v3i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, v3, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v1, v4, v1 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v2, v5, v2 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <3 x i32> %value, %amount |
| ret <3 x i32> %result |
| } |
| |
| define amdgpu_ps <3 x i32> @s_lshr_v3i32(<3 x i32> inreg %value, <3 x i32> inreg %amount) { |
| ; GCN-LABEL: s_lshr_v3i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s0, s3 |
| ; GCN-NEXT: s_lshr_b32 s1, s1, s4 |
| ; GCN-NEXT: s_lshr_b32 s2, s2, s5 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v3i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s3 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, s4 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s2, s2, s5 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <3 x i32> %value, %amount |
| ret <3 x i32> %result |
| } |
| |
| define <4 x i32> @v_lshr_v4i32(<4 x i32> %value, <4 x i32> %amount) { |
| ; GCN-LABEL: v_lshr_v4i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, v4, v0 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v1, v5, v1 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v2, v6, v2 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v3, v7, v3 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v4i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, v4, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v1, v5, v1 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v2, v6, v2 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v3, v7, v3 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <4 x i32> %value, %amount |
| ret <4 x i32> %result |
| } |
| |
| define amdgpu_ps <4 x i32> @s_lshr_v4i32(<4 x i32> inreg %value, <4 x i32> inreg %amount) { |
| ; GCN-LABEL: s_lshr_v4i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s0, s4 |
| ; GCN-NEXT: s_lshr_b32 s1, s1, s5 |
| ; GCN-NEXT: s_lshr_b32 s2, s2, s6 |
| ; GCN-NEXT: s_lshr_b32 s3, s3, s7 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v4i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s4 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, s5 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s2, s2, s6 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s3, s3, s7 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <4 x i32> %value, %amount |
| ret <4 x i32> %result |
| } |
| |
| define <5 x i32> @v_lshr_v5i32(<5 x i32> %value, <5 x i32> %amount) { |
| ; GCN-LABEL: v_lshr_v5i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, v5, v0 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v1, v6, v1 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v2, v7, v2 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v3, v8, v3 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v4, v9, v4 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v5i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, v5, v0 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v1, v6, v1 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v2, v7, v2 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v3, v8, v3 |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v4, v9, v4 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <5 x i32> %value, %amount |
| ret <5 x i32> %result |
| } |
| |
| define amdgpu_ps <5 x i32> @s_lshr_v5i32(<5 x i32> inreg %value, <5 x i32> inreg %amount) { |
| ; GCN-LABEL: s_lshr_v5i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s0, s5 |
| ; GCN-NEXT: s_lshr_b32 s1, s1, s6 |
| ; GCN-NEXT: s_lshr_b32 s2, s2, s7 |
| ; GCN-NEXT: s_lshr_b32 s3, s3, s8 |
| ; GCN-NEXT: s_lshr_b32 s4, s4, s9 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v5i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s5 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, s6 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s2, s2, s7 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s3, s3, s8 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s4, s4, s9 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <5 x i32> %value, %amount |
| ret <5 x i32> %result |
| } |
| |
| define <16 x i32> @v_lshr_v16i32(<16 x i32> %value, <16 x i32> %amount) { |
| ; GCN-LABEL: v_lshr_v16i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, v16, v0 |
| ; GCN-NEXT: buffer_load_dword v16, off, s[0:3], s32 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v1, v17, v1 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v2, v18, v2 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v3, v19, v3 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v4, v20, v4 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v5, v21, v5 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v6, v22, v6 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v7, v23, v7 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v8, v24, v8 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v9, v25, v9 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v10, v26, v10 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v11, v27, v11 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v12, v28, v12 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v13, v29, v13 |
| ; GCN-NEXT: v_lshrrev_b32_e32 v14, v30, v14 |
| ; GCN-NEXT: s_waitcnt vmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v15, v16, v15 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10-LABEL: v_lshr_v16i32: |
| ; GFX10: ; %bb.0: |
| ; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10-NEXT: buffer_load_dword v31, off, s[0:3], s32 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v0, v16, v0 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v1, v17, v1 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v2, v18, v2 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v3, v19, v3 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v4, v20, v4 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v5, v21, v5 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v6, v22, v6 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v7, v23, v7 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v8, v24, v8 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v9, v25, v9 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v10, v26, v10 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v11, v27, v11 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v12, v28, v12 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v13, v29, v13 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v14, v30, v14 |
| ; GFX10-NEXT: s_waitcnt vmcnt(0) |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v15, v31, v15 |
| ; GFX10-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX11-LABEL: v_lshr_v16i32: |
| ; GFX11: ; %bb.0: |
| ; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX11-NEXT: scratch_load_b32 v31, off, s32 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v0, v16, v0 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v1, v17, v1 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v2, v18, v2 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v3, v19, v3 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v4, v20, v4 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v5, v21, v5 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v6, v22, v6 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v7, v23, v7 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v8, v24, v8 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v9, v25, v9 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v10, v26, v10 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v11, v27, v11 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v12, v28, v12 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v13, v29, v13 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v14, v30, v14 |
| ; GFX11-NEXT: s_waitcnt vmcnt(0) |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v15, v31, v15 |
| ; GFX11-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <16 x i32> %value, %amount |
| ret <16 x i32> %result |
| } |
| |
| define amdgpu_ps <16 x i32> @s_lshr_v16i32(<16 x i32> inreg %value, <16 x i32> inreg %amount) { |
| ; GCN-LABEL: s_lshr_v16i32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s0, s16 |
| ; GCN-NEXT: s_lshr_b32 s1, s1, s17 |
| ; GCN-NEXT: s_lshr_b32 s2, s2, s18 |
| ; GCN-NEXT: s_lshr_b32 s3, s3, s19 |
| ; GCN-NEXT: s_lshr_b32 s4, s4, s20 |
| ; GCN-NEXT: s_lshr_b32 s5, s5, s21 |
| ; GCN-NEXT: s_lshr_b32 s6, s6, s22 |
| ; GCN-NEXT: s_lshr_b32 s7, s7, s23 |
| ; GCN-NEXT: s_lshr_b32 s8, s8, s24 |
| ; GCN-NEXT: s_lshr_b32 s9, s9, s25 |
| ; GCN-NEXT: s_lshr_b32 s10, s10, s26 |
| ; GCN-NEXT: s_lshr_b32 s11, s11, s27 |
| ; GCN-NEXT: s_lshr_b32 s12, s12, s28 |
| ; GCN-NEXT: s_lshr_b32 s13, s13, s29 |
| ; GCN-NEXT: s_lshr_b32 s14, s14, s30 |
| ; GCN-NEXT: s_lshr_b32 s15, s15, s31 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v16i32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, s17 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s2, s2, s18 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s3, s3, s19 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s4, s4, s20 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s5, s5, s21 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s6, s6, s22 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s7, s7, s23 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s8, s8, s24 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s9, s9, s25 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s10, s10, s26 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s11, s11, s27 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s12, s12, s28 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s13, s13, s29 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s14, s14, s30 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s15, s15, s31 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <16 x i32> %value, %amount |
| ret <16 x i32> %result |
| } |
| |
| define i16 @v_lshr_i16(i16 %value, i16 %amount) { |
| ; GFX6-LABEL: v_lshr_i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_and_b32_e32 v1, 0xffff, v1 |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v0, v1, v0 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v0, v1, v0 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_lshrrev_b16_e32 v0, v1, v0 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b16 v0, v1, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i16 %value, %amount |
| ret i16 %result |
| } |
| |
| define i16 @v_lshr_i16_15(i16 %value) { |
| ; GFX6-LABEL: v_lshr_i16_15: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_bfe_u32 v0, v0, 15, 1 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i16_15: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v0, 15, v0 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i16_15: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_lshrrev_b16_e32 v0, 15, v0 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i16_15: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b16 v0, 15, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i16 %value, 15 |
| ret i16 %result |
| } |
| |
| define amdgpu_ps i16 @s_lshr_i16(i16 inreg %value, i16 inreg %amount) { |
| ; GFX6-LABEL: s_lshr_i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: s_lshr_i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX8-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: s_lshr_i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i16 %value, %amount |
| ret i16 %result |
| } |
| |
| define amdgpu_ps i16 @s_lshr_i16_15(i16 inreg %value) { |
| ; GFX6-LABEL: s_lshr_i16_15: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_bfe_u32 s0, s0, 0x1000f |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: s_lshr_i16_15: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX8-NEXT: s_lshr_b32 s0, s0, 15 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: s_lshr_i16_15: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, 15 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i16_15: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, 15 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i16 %value, 15 |
| ret i16 %result |
| } |
| |
| define amdgpu_ps half @lshr_i16_sv(i16 inreg %value, i16 %amount) { |
| ; GFX6-LABEL: lshr_i16_sv: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: v_lshr_b32_e32 v0, s0, v0 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: lshr_i16_sv: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: v_lshrrev_b16_e64 v0, v0, s0 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: lshr_i16_sv: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: v_lshrrev_b16_e64 v0, v0, s0 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_i16_sv: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_lshrrev_b16 v0, v0, s0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i16 %value, %amount |
| %cast = bitcast i16 %result to half |
| ret half %cast |
| } |
| |
| define amdgpu_ps half @lshr_i16_vs(i16 %value, i16 inreg %amount) { |
| ; GFX6-LABEL: lshr_i16_vs: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v0, s0, v0 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: lshr_i16_vs: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v0, s0, v0 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: lshr_i16_vs: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: v_lshrrev_b16_e32 v0, s0, v0 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_i16_vs: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_lshrrev_b16 v0, s0, v0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i16 %value, %amount |
| %cast = bitcast i16 %result to half |
| ret half %cast |
| } |
| |
| define <2 x i16> @v_lshr_v2i16(<2 x i16> %value, <2 x i16> %amount) { |
| ; GFX6-LABEL: v_lshr_v2i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_and_b32_e32 v2, 0xffff, v2 |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v0, v2, v0 |
| ; GFX6-NEXT: v_and_b32_e32 v2, 0xffff, v3 |
| ; GFX6-NEXT: v_and_b32_e32 v1, 0xffff, v1 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v1, v2, v1 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_v2i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v2, v1, v0 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v0, v1, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1 |
| ; GFX8-NEXT: v_or_b32_e32 v0, v2, v0 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_v2i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v0, v1, v0 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v2i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v0, v1, v0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <2 x i16> %value, %amount |
| ret <2 x i16> %result |
| } |
| |
| define <2 x i16> @v_lshr_v2i16_15(<2 x i16> %value) { |
| ; GFX6-LABEL: v_lshr_v2i16_15: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_bfe_u32 v0, v0, 15, 1 |
| ; GFX6-NEXT: v_bfe_u32 v1, v1, 15, 1 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_v2i16_15: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_mov_b32_e32 v1, 31 |
| ; GFX8-NEXT: v_lshrrev_b32_sdwa v1, v1, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:DWORD |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v0, 15, v0 |
| ; GFX8-NEXT: v_or_b32_e32 v0, v0, v1 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_v2i16_15: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v0, 15, v0 op_sel_hi:[0,1] |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v2i16_15: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v0, 15, v0 op_sel_hi:[0,1] |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <2 x i16> %value, <i16 15, i16 15> |
| ret <2 x i16> %result |
| } |
| |
| define amdgpu_ps i32 @s_lshr_v2i16(<2 x i16> inreg %value, <2 x i16> inreg %amount) { |
| ; GFX6-LABEL: s_lshr_v2i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_and_b32 s1, s1, 0xffff |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s1, s1, s3 |
| ; GFX6-NEXT: s_lshr_b32 s0, s0, s2 |
| ; GFX6-NEXT: s_lshl_b32 s1, s1, 16 |
| ; GFX6-NEXT: s_or_b32 s0, s0, s1 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: s_lshr_v2i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_lshr_b32 s2, s0, 16 |
| ; GFX8-NEXT: s_lshr_b32 s3, s1, 16 |
| ; GFX8-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX8-NEXT: s_lshr_b32 s0, s0, s1 |
| ; GFX8-NEXT: s_lshr_b32 s1, s2, s3 |
| ; GFX8-NEXT: s_lshl_b32 s1, s1, 16 |
| ; GFX8-NEXT: s_or_b32 s0, s0, s1 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: s_lshr_v2i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_and_b32 s2, s0, 0xffff |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, 16 |
| ; GFX9-NEXT: s_lshr_b32 s3, s1, 16 |
| ; GFX9-NEXT: s_lshr_b32 s1, s2, s1 |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, s3 |
| ; GFX9-NEXT: s_pack_ll_b32_b16 s0, s1, s0 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v2i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s2, s0, 0xffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s3, s1, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s2, s1 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s3 |
| ; GFX10PLUS-NEXT: s_pack_ll_b32_b16 s0, s1, s0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <2 x i16> %value, %amount |
| %cast = bitcast <2 x i16> %result to i32 |
| ret i32 %cast |
| } |
| |
| define amdgpu_ps float @lshr_v2i16_sv(<2 x i16> inreg %value, <2 x i16> %amount) { |
| ; GFX6-LABEL: lshr_v2i16_sv: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: v_lshr_b32_e32 v0, s0, v0 |
| ; GFX6-NEXT: v_and_b32_e32 v1, 0xffff, v1 |
| ; GFX6-NEXT: s_and_b32 s0, s1, 0xffff |
| ; GFX6-NEXT: v_lshr_b32_e32 v1, s0, v1 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v1, 16, v1 |
| ; GFX6-NEXT: v_or_b32_e32 v0, v0, v1 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: lshr_v2i16_sv: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_lshr_b32 s1, s0, 16 |
| ; GFX8-NEXT: v_mov_b32_e32 v2, s1 |
| ; GFX8-NEXT: v_lshrrev_b16_e64 v1, v0, s0 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v0, v0, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:DWORD |
| ; GFX8-NEXT: v_or_b32_e32 v0, v1, v0 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: lshr_v2i16_sv: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v0, v0, s0 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_v2i16_sv: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v0, v0, s0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <2 x i16> %value, %amount |
| %cast = bitcast <2 x i16> %result to float |
| ret float %cast |
| } |
| |
| define amdgpu_ps float @lshr_v2i16_vs(<2 x i16> %value, <2 x i16> inreg %amount) { |
| ; GFX6-LABEL: lshr_v2i16_vs: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v0, s0, v0 |
| ; GFX6-NEXT: s_and_b32 s0, s1, 0xffff |
| ; GFX6-NEXT: v_and_b32_e32 v1, 0xffff, v1 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v1, s0, v1 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v1, 16, v1 |
| ; GFX6-NEXT: v_or_b32_e32 v0, v0, v1 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: lshr_v2i16_vs: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_lshr_b32 s1, s0, 16 |
| ; GFX8-NEXT: v_mov_b32_e32 v2, s1 |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v1, s0, v0 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v0, v2, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:DWORD src1_sel:WORD_1 |
| ; GFX8-NEXT: v_or_b32_e32 v0, v1, v0 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: lshr_v2i16_vs: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v0, s0, v0 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_v2i16_vs: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v0, s0, v0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <2 x i16> %value, %amount |
| %cast = bitcast <2 x i16> %result to float |
| ret float %cast |
| } |
| |
| ; FIXME |
| ; define <3 x i16> @v_lshr_v3i16(<3 x i16> %value, <3 x i16> %amount) { |
| ; %result = lshr <3 x i16> %value, %amount |
| ; ret <3 x i16> %result |
| ; } |
| |
| ; define amdgpu_ps <3 x i16> @s_lshr_v3i16(<3 x i16> inreg %value, <3 x i16> inreg %amount) { |
| ; %result = lshr <3 x i16> %value, %amount |
| ; ret <3 x i16> %result |
| ; } |
| |
| define <2 x float> @v_lshr_v4i16(<4 x i16> %value, <4 x i16> %amount) { |
| ; GFX6-LABEL: v_lshr_v4i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_and_b32_e32 v4, 0xffff, v4 |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v0, v4, v0 |
| ; GFX6-NEXT: v_and_b32_e32 v4, 0xffff, v5 |
| ; GFX6-NEXT: v_and_b32_e32 v1, 0xffff, v1 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v1, v4, v1 |
| ; GFX6-NEXT: v_and_b32_e32 v4, 0xffff, v6 |
| ; GFX6-NEXT: v_and_b32_e32 v2, 0xffff, v2 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v2, v4, v2 |
| ; GFX6-NEXT: v_and_b32_e32 v4, 0xffff, v7 |
| ; GFX6-NEXT: v_and_b32_e32 v3, 0xffff, v3 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v3, v4, v3 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v1, 16, v1 |
| ; GFX6-NEXT: v_or_b32_e32 v0, v0, v1 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v1, 16, v3 |
| ; GFX6-NEXT: v_or_b32_e32 v1, v2, v1 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_v4i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v4, v2, v0 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v0, v2, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1 |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v2, v3, v1 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v1, v3, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1 |
| ; GFX8-NEXT: v_or_b32_e32 v0, v4, v0 |
| ; GFX8-NEXT: v_or_b32_e32 v1, v2, v1 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_v4i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v0, v2, v0 |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v1, v3, v1 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v4i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v0, v2, v0 |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v1, v3, v1 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <4 x i16> %value, %amount |
| %cast = bitcast <4 x i16> %result to <2 x float> |
| ret <2 x float> %cast |
| } |
| |
| define amdgpu_ps <2 x i32> @s_lshr_v4i16(<4 x i16> inreg %value, <4 x i16> inreg %amount) { |
| ; GFX6-LABEL: s_lshr_v4i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_and_b32 s1, s1, 0xffff |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s1, s1, s5 |
| ; GFX6-NEXT: s_and_b32 s3, s3, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s0, s0, s4 |
| ; GFX6-NEXT: s_and_b32 s2, s2, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s3, s3, s7 |
| ; GFX6-NEXT: s_lshl_b32 s1, s1, 16 |
| ; GFX6-NEXT: s_lshr_b32 s2, s2, s6 |
| ; GFX6-NEXT: s_or_b32 s0, s0, s1 |
| ; GFX6-NEXT: s_lshl_b32 s1, s3, 16 |
| ; GFX6-NEXT: s_or_b32 s1, s2, s1 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: s_lshr_v4i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_lshr_b32 s4, s0, 16 |
| ; GFX8-NEXT: s_lshr_b32 s6, s2, 16 |
| ; GFX8-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX8-NEXT: s_lshr_b32 s5, s1, 16 |
| ; GFX8-NEXT: s_lshr_b32 s7, s3, 16 |
| ; GFX8-NEXT: s_lshr_b32 s0, s0, s2 |
| ; GFX8-NEXT: s_lshr_b32 s2, s4, s6 |
| ; GFX8-NEXT: s_and_b32 s1, 0xffff, s1 |
| ; GFX8-NEXT: s_lshr_b32 s1, s1, s3 |
| ; GFX8-NEXT: s_lshr_b32 s3, s5, s7 |
| ; GFX8-NEXT: s_lshl_b32 s2, s2, 16 |
| ; GFX8-NEXT: s_or_b32 s0, s0, s2 |
| ; GFX8-NEXT: s_lshl_b32 s2, s3, 16 |
| ; GFX8-NEXT: s_or_b32 s1, s1, s2 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: s_lshr_v4i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_and_b32 s4, s0, 0xffff |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, 16 |
| ; GFX9-NEXT: s_lshr_b32 s5, s2, 16 |
| ; GFX9-NEXT: s_lshr_b32 s2, s4, s2 |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, s5 |
| ; GFX9-NEXT: s_pack_ll_b32_b16 s0, s2, s0 |
| ; GFX9-NEXT: s_and_b32 s2, s1, 0xffff |
| ; GFX9-NEXT: s_lshr_b32 s1, s1, 16 |
| ; GFX9-NEXT: s_lshr_b32 s4, s3, 16 |
| ; GFX9-NEXT: s_lshr_b32 s2, s2, s3 |
| ; GFX9-NEXT: s_lshr_b32 s1, s1, s4 |
| ; GFX9-NEXT: s_pack_ll_b32_b16 s1, s2, s1 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v4i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s4, s0, 0xffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s5, s2, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s2, s4, s2 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s5 |
| ; GFX10PLUS-NEXT: s_and_b32 s4, s1, 0xffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s5, s3, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s3, s4, s3 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, s5 |
| ; GFX10PLUS-NEXT: s_pack_ll_b32_b16 s0, s2, s0 |
| ; GFX10PLUS-NEXT: s_pack_ll_b32_b16 s1, s3, s1 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <4 x i16> %value, %amount |
| %cast = bitcast <4 x i16> %result to <2 x i32> |
| ret <2 x i32> %cast |
| } |
| |
| ; FIXME |
| ; define <5 x i16> @v_lshr_v5i16(<5 x i16> %value, <5 x i16> %amount) { |
| ; %result = lshr <5 x i16> %value, %amount |
| ; ret <5 x i16> %result |
| ; } |
| |
| ; define amdgpu_ps <5 x i16> @s_lshr_v5i16(<5 x i16> inreg %value, <5 x i16> inreg %amount) { |
| ; %result = lshr <5 x i16> %value, %amount |
| ; ret <5 x i16> %result |
| ; } |
| |
| ; define <3 x float> @v_lshr_v6i16(<6 x i16> %value, <6 x i16> %amount) { |
| ; %result = lshr <6 x i16> %value, %amount |
| ; %cast = bitcast <6 x i16> %result to <3 x float> |
| ; ret <3 x float> %cast |
| ; } |
| |
| ; define amdgpu_ps <3 x i32> @s_lshr_v6i16(<6 x i16> inreg %value, <6 x i16> inreg %amount) { |
| ; %result = lshr <6 x i16> %value, %amount |
| ; %cast = bitcast <6 x i16> %result to <3 x i32> |
| ; ret <3 x i32> %cast |
| ; } |
| |
| define <4 x float> @v_lshr_v8i16(<8 x i16> %value, <8 x i16> %amount) { |
| ; GFX6-LABEL: v_lshr_v8i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v8 |
| ; GFX6-NEXT: v_and_b32_e32 v0, 0xffff, v0 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v0, v8, v0 |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v9 |
| ; GFX6-NEXT: v_and_b32_e32 v1, 0xffff, v1 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v1, v8, v1 |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v10 |
| ; GFX6-NEXT: v_and_b32_e32 v2, 0xffff, v2 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v2, v8, v2 |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v11 |
| ; GFX6-NEXT: v_and_b32_e32 v3, 0xffff, v3 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v3, v8, v3 |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v12 |
| ; GFX6-NEXT: v_and_b32_e32 v4, 0xffff, v4 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v4, v8, v4 |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v13 |
| ; GFX6-NEXT: v_and_b32_e32 v5, 0xffff, v5 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v5, v8, v5 |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v14 |
| ; GFX6-NEXT: v_and_b32_e32 v6, 0xffff, v6 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v6, v8, v6 |
| ; GFX6-NEXT: v_and_b32_e32 v8, 0xffff, v15 |
| ; GFX6-NEXT: v_and_b32_e32 v7, 0xffff, v7 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v1, 16, v1 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v7, v8, v7 |
| ; GFX6-NEXT: v_or_b32_e32 v0, v0, v1 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v1, 16, v3 |
| ; GFX6-NEXT: v_or_b32_e32 v1, v2, v1 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v2, 16, v5 |
| ; GFX6-NEXT: v_lshlrev_b32_e32 v3, 16, v7 |
| ; GFX6-NEXT: v_or_b32_e32 v2, v4, v2 |
| ; GFX6-NEXT: v_or_b32_e32 v3, v6, v3 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_v8i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v8, v4, v0 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v0, v4, v0 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1 |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v4, v5, v1 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v1, v5, v1 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1 |
| ; GFX8-NEXT: v_or_b32_e32 v1, v4, v1 |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v4, v6, v2 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v2, v6, v2 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1 |
| ; GFX8-NEXT: v_or_b32_e32 v2, v4, v2 |
| ; GFX8-NEXT: v_lshrrev_b16_e32 v4, v7, v3 |
| ; GFX8-NEXT: v_lshrrev_b16_sdwa v3, v7, v3 dst_sel:WORD_1 dst_unused:UNUSED_PAD src0_sel:WORD_1 src1_sel:WORD_1 |
| ; GFX8-NEXT: v_or_b32_e32 v0, v8, v0 |
| ; GFX8-NEXT: v_or_b32_e32 v3, v4, v3 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_v8i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v0, v4, v0 |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v1, v5, v1 |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v2, v6, v2 |
| ; GFX9-NEXT: v_pk_lshrrev_b16 v3, v7, v3 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v8i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v0, v4, v0 |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v1, v5, v1 |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v2, v6, v2 |
| ; GFX10PLUS-NEXT: v_pk_lshrrev_b16 v3, v7, v3 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <8 x i16> %value, %amount |
| %cast = bitcast <8 x i16> %result to <4 x float> |
| ret <4 x float> %cast |
| } |
| |
| define amdgpu_ps <4 x i32> @s_lshr_v8i16(<8 x i16> inreg %value, <8 x i16> inreg %amount) { |
| ; GFX6-LABEL: s_lshr_v8i16: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_and_b32 s1, s1, 0xffff |
| ; GFX6-NEXT: s_and_b32 s0, s0, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s1, s1, s9 |
| ; GFX6-NEXT: s_and_b32 s3, s3, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s0, s0, s8 |
| ; GFX6-NEXT: s_and_b32 s2, s2, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s3, s3, s11 |
| ; GFX6-NEXT: s_and_b32 s5, s5, 0xffff |
| ; GFX6-NEXT: s_and_b32 s7, s7, 0xffff |
| ; GFX6-NEXT: s_lshl_b32 s1, s1, 16 |
| ; GFX6-NEXT: s_lshr_b32 s2, s2, s10 |
| ; GFX6-NEXT: s_and_b32 s4, s4, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s5, s5, s13 |
| ; GFX6-NEXT: s_and_b32 s6, s6, 0xffff |
| ; GFX6-NEXT: s_lshr_b32 s7, s7, s15 |
| ; GFX6-NEXT: s_or_b32 s0, s0, s1 |
| ; GFX6-NEXT: s_lshl_b32 s1, s3, 16 |
| ; GFX6-NEXT: s_lshr_b32 s4, s4, s12 |
| ; GFX6-NEXT: s_lshr_b32 s6, s6, s14 |
| ; GFX6-NEXT: s_or_b32 s1, s2, s1 |
| ; GFX6-NEXT: s_lshl_b32 s2, s5, 16 |
| ; GFX6-NEXT: s_lshl_b32 s3, s7, 16 |
| ; GFX6-NEXT: s_or_b32 s2, s4, s2 |
| ; GFX6-NEXT: s_or_b32 s3, s6, s3 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: s_lshr_v8i16: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_lshr_b32 s8, s0, 16 |
| ; GFX8-NEXT: s_lshr_b32 s12, s4, 16 |
| ; GFX8-NEXT: s_and_b32 s0, 0xffff, s0 |
| ; GFX8-NEXT: s_lshr_b32 s9, s1, 16 |
| ; GFX8-NEXT: s_lshr_b32 s13, s5, 16 |
| ; GFX8-NEXT: s_lshr_b32 s0, s0, s4 |
| ; GFX8-NEXT: s_lshr_b32 s4, s8, s12 |
| ; GFX8-NEXT: s_and_b32 s1, 0xffff, s1 |
| ; GFX8-NEXT: s_lshr_b32 s10, s2, 16 |
| ; GFX8-NEXT: s_lshr_b32 s14, s6, 16 |
| ; GFX8-NEXT: s_lshr_b32 s1, s1, s5 |
| ; GFX8-NEXT: s_lshr_b32 s5, s9, s13 |
| ; GFX8-NEXT: s_and_b32 s2, 0xffff, s2 |
| ; GFX8-NEXT: s_lshl_b32 s4, s4, 16 |
| ; GFX8-NEXT: s_lshr_b32 s11, s3, 16 |
| ; GFX8-NEXT: s_lshr_b32 s15, s7, 16 |
| ; GFX8-NEXT: s_lshr_b32 s2, s2, s6 |
| ; GFX8-NEXT: s_lshr_b32 s6, s10, s14 |
| ; GFX8-NEXT: s_and_b32 s3, 0xffff, s3 |
| ; GFX8-NEXT: s_or_b32 s0, s0, s4 |
| ; GFX8-NEXT: s_lshl_b32 s4, s5, 16 |
| ; GFX8-NEXT: s_lshr_b32 s3, s3, s7 |
| ; GFX8-NEXT: s_lshr_b32 s7, s11, s15 |
| ; GFX8-NEXT: s_or_b32 s1, s1, s4 |
| ; GFX8-NEXT: s_lshl_b32 s4, s6, 16 |
| ; GFX8-NEXT: s_or_b32 s2, s2, s4 |
| ; GFX8-NEXT: s_lshl_b32 s4, s7, 16 |
| ; GFX8-NEXT: s_or_b32 s3, s3, s4 |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: s_lshr_v8i16: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_and_b32 s8, s0, 0xffff |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, 16 |
| ; GFX9-NEXT: s_lshr_b32 s9, s4, 16 |
| ; GFX9-NEXT: s_lshr_b32 s4, s8, s4 |
| ; GFX9-NEXT: s_lshr_b32 s0, s0, s9 |
| ; GFX9-NEXT: s_pack_ll_b32_b16 s0, s4, s0 |
| ; GFX9-NEXT: s_and_b32 s4, s1, 0xffff |
| ; GFX9-NEXT: s_lshr_b32 s1, s1, 16 |
| ; GFX9-NEXT: s_lshr_b32 s8, s5, 16 |
| ; GFX9-NEXT: s_lshr_b32 s4, s4, s5 |
| ; GFX9-NEXT: s_lshr_b32 s1, s1, s8 |
| ; GFX9-NEXT: s_pack_ll_b32_b16 s1, s4, s1 |
| ; GFX9-NEXT: s_and_b32 s4, s2, 0xffff |
| ; GFX9-NEXT: s_lshr_b32 s2, s2, 16 |
| ; GFX9-NEXT: s_lshr_b32 s5, s6, 16 |
| ; GFX9-NEXT: s_lshr_b32 s4, s4, s6 |
| ; GFX9-NEXT: s_lshr_b32 s2, s2, s5 |
| ; GFX9-NEXT: s_pack_ll_b32_b16 s2, s4, s2 |
| ; GFX9-NEXT: s_and_b32 s4, s3, 0xffff |
| ; GFX9-NEXT: s_lshr_b32 s3, s3, 16 |
| ; GFX9-NEXT: s_lshr_b32 s5, s7, 16 |
| ; GFX9-NEXT: s_lshr_b32 s4, s4, s7 |
| ; GFX9-NEXT: s_lshr_b32 s3, s3, s5 |
| ; GFX9-NEXT: s_pack_ll_b32_b16 s3, s4, s3 |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v8i16: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b32 s8, s0, 0xffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s9, s4, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s4, s8, s4 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s0, s9 |
| ; GFX10PLUS-NEXT: s_and_b32 s8, s1, 0xffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s9, s5, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s5, s8, s5 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s1, s1, s9 |
| ; GFX10PLUS-NEXT: s_pack_ll_b32_b16 s0, s4, s0 |
| ; GFX10PLUS-NEXT: s_pack_ll_b32_b16 s1, s5, s1 |
| ; GFX10PLUS-NEXT: s_and_b32 s4, s2, 0xffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s2, s2, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s5, s6, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s4, s4, s6 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s2, s2, s5 |
| ; GFX10PLUS-NEXT: s_and_b32 s5, s3, 0xffff |
| ; GFX10PLUS-NEXT: s_lshr_b32 s3, s3, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s6, s7, 16 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s5, s5, s7 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s3, s3, s6 |
| ; GFX10PLUS-NEXT: s_pack_ll_b32_b16 s2, s4, s2 |
| ; GFX10PLUS-NEXT: s_pack_ll_b32_b16 s3, s5, s3 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <8 x i16> %value, %amount |
| %cast = bitcast <8 x i16> %result to <4 x i32> |
| ret <4 x i32> %cast |
| } |
| |
| define i64 @v_lshr_i64(i64 %value, i64 %amount) { |
| ; GFX6-LABEL: v_lshr_i64: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_lshr_b64 v[0:1], v[0:1], v2 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i64: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b64 v[0:1], v2, v[0:1] |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i64: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_lshrrev_b64 v[0:1], v2, v[0:1] |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i64: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[0:1], v2, v[0:1] |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i64 %value, %amount |
| ret i64 %result |
| } |
| |
| define i64 @v_lshr_i64_63(i64 %value) { |
| ; GCN-LABEL: v_lshr_i64_63: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, 31, v1 |
| ; GCN-NEXT: v_mov_b32_e32 v1, 0 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i64_63: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, 31, v1 |
| ; GFX10PLUS-NEXT: v_mov_b32_e32 v1, 0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i64 %value, 63 |
| ret i64 %result |
| } |
| |
| define i64 @v_lshr_i64_33(i64 %value) { |
| ; GCN-LABEL: v_lshr_i64_33: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_lshrrev_b32_e32 v0, 1, v1 |
| ; GCN-NEXT: v_mov_b32_e32 v1, 0 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i64_33: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b32_e32 v0, 1, v1 |
| ; GFX10PLUS-NEXT: v_mov_b32_e32 v1, 0 |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i64 %value, 33 |
| ret i64 %result |
| } |
| |
| define i64 @v_lshr_i64_32(i64 %value) { |
| ; GCN-LABEL: v_lshr_i64_32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GCN-NEXT: v_mov_b32_e32 v0, v1 |
| ; GCN-NEXT: v_mov_b32_e32 v1, 0 |
| ; GCN-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10-LABEL: v_lshr_i64_32: |
| ; GFX10: ; %bb.0: |
| ; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10-NEXT: v_mov_b32_e32 v0, v1 |
| ; GFX10-NEXT: v_mov_b32_e32 v1, 0 |
| ; GFX10-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX11-LABEL: v_lshr_i64_32: |
| ; GFX11: ; %bb.0: |
| ; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX11-NEXT: v_dual_mov_b32 v0, v1 :: v_dual_mov_b32 v1, 0 |
| ; GFX11-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i64 %value, 32 |
| ret i64 %result |
| } |
| |
| define i64 @v_lshr_i64_31(i64 %value) { |
| ; GFX6-LABEL: v_lshr_i64_31: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_lshr_b64 v[0:1], v[0:1], 31 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i64_31: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b64 v[0:1], 31, v[0:1] |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i64_31: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_lshrrev_b64 v[0:1], 31, v[0:1] |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_i64_31: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[0:1], 31, v[0:1] |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i64 %value, 31 |
| ret i64 %result |
| } |
| |
| define amdgpu_ps i64 @s_lshr_i64(i64 inreg %value, i64 inreg %amount) { |
| ; GCN-LABEL: s_lshr_i64: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b64 s[0:1], s[0:1], s2 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i64: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b64 s[0:1], s[0:1], s2 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i64 %value, %amount |
| ret i64 %result |
| } |
| |
| define amdgpu_ps i64 @s_lshr_i64_63(i64 inreg %value) { |
| ; GCN-LABEL: s_lshr_i64_63: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s1, 31 |
| ; GCN-NEXT: s_mov_b32 s1, 0 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i64_63: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s1, 31 |
| ; GFX10PLUS-NEXT: s_mov_b32 s1, 0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i64 %value, 63 |
| ret i64 %result |
| } |
| |
| define amdgpu_ps i64 @s_lshr_i64_33(i64 inreg %value) { |
| ; GCN-LABEL: s_lshr_i64_33: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b32 s0, s1, 1 |
| ; GCN-NEXT: s_mov_b32 s1, 0 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i64_33: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b32 s0, s1, 1 |
| ; GFX10PLUS-NEXT: s_mov_b32 s1, 0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i64 %value, 33 |
| ret i64 %result |
| } |
| |
| define amdgpu_ps i64 @s_lshr_i64_32(i64 inreg %value) { |
| ; GCN-LABEL: s_lshr_i64_32: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_mov_b32 s0, s1 |
| ; GCN-NEXT: s_mov_b32 s1, 0 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i64_32: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_mov_b32 s0, s1 |
| ; GFX10PLUS-NEXT: s_mov_b32 s1, 0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i64 %value, 32 |
| ret i64 %result |
| } |
| |
| define amdgpu_ps i64 @s_lshr_i64_31(i64 inreg %value) { |
| ; GCN-LABEL: s_lshr_i64_31: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b64 s[0:1], s[0:1], 31 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i64_31: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b64 s[0:1], s[0:1], 31 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i64 %value, 31 |
| ret i64 %result |
| } |
| |
| define amdgpu_ps <2 x float> @lshr_i64_sv(i64 inreg %value, i64 %amount) { |
| ; GFX6-LABEL: lshr_i64_sv: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: v_lshr_b64 v[0:1], s[0:1], v0 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: lshr_i64_sv: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: v_lshrrev_b64 v[0:1], v0, s[0:1] |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: lshr_i64_sv: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: v_lshrrev_b64 v[0:1], v0, s[0:1] |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_i64_sv: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[0:1], v0, s[0:1] |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i64 %value, %amount |
| %cast = bitcast i64 %result to <2 x float> |
| ret <2 x float> %cast |
| } |
| |
| define amdgpu_ps <2 x float> @lshr_i64_vs(i64 %value, i64 inreg %amount) { |
| ; GFX6-LABEL: lshr_i64_vs: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: v_lshr_b64 v[0:1], v[0:1], s0 |
| ; GFX6-NEXT: ; return to shader part epilog |
| ; |
| ; GFX8-LABEL: lshr_i64_vs: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: v_lshrrev_b64 v[0:1], s0, v[0:1] |
| ; GFX8-NEXT: ; return to shader part epilog |
| ; |
| ; GFX9-LABEL: lshr_i64_vs: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: v_lshrrev_b64 v[0:1], s0, v[0:1] |
| ; GFX9-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: lshr_i64_vs: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[0:1], s0, v[0:1] |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i64 %value, %amount |
| %cast = bitcast i64 %result to <2 x float> |
| ret <2 x float> %cast |
| } |
| |
| define <2 x i64> @v_lshr_v2i64(<2 x i64> %value, <2 x i64> %amount) { |
| ; GFX6-LABEL: v_lshr_v2i64: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_lshr_b64 v[0:1], v[0:1], v4 |
| ; GFX6-NEXT: v_lshr_b64 v[2:3], v[2:3], v6 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_v2i64: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b64 v[0:1], v4, v[0:1] |
| ; GFX8-NEXT: v_lshrrev_b64 v[2:3], v6, v[2:3] |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_v2i64: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_lshrrev_b64 v[0:1], v4, v[0:1] |
| ; GFX9-NEXT: v_lshrrev_b64 v[2:3], v6, v[2:3] |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v2i64: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[0:1], v4, v[0:1] |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[2:3], v6, v[2:3] |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <2 x i64> %value, %amount |
| ret <2 x i64> %result |
| } |
| |
| define <2 x i64> @v_lshr_v2i64_31(<2 x i64> %value) { |
| ; GFX6-LABEL: v_lshr_v2i64_31: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_lshr_b64 v[0:1], v[0:1], 31 |
| ; GFX6-NEXT: v_lshr_b64 v[2:3], v[2:3], 31 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_v2i64_31: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_lshrrev_b64 v[0:1], 31, v[0:1] |
| ; GFX8-NEXT: v_lshrrev_b64 v[2:3], 31, v[2:3] |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_v2i64_31: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_lshrrev_b64 v[0:1], 31, v[0:1] |
| ; GFX9-NEXT: v_lshrrev_b64 v[2:3], 31, v[2:3] |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10PLUS-LABEL: v_lshr_v2i64_31: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[0:1], 31, v[0:1] |
| ; GFX10PLUS-NEXT: v_lshrrev_b64 v[2:3], 31, v[2:3] |
| ; GFX10PLUS-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr <2 x i64> %value, <i64 31, i64 31> |
| ret <2 x i64> %result |
| } |
| |
| define amdgpu_ps <2 x i64> @s_lshr_v2i64(<2 x i64> inreg %value, <2 x i64> inreg %amount) { |
| ; GCN-LABEL: s_lshr_v2i64: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_lshr_b64 s[0:1], s[0:1], s4 |
| ; GCN-NEXT: s_lshr_b64 s[2:3], s[2:3], s6 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_v2i64: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_lshr_b64 s[0:1], s[0:1], s4 |
| ; GFX10PLUS-NEXT: s_lshr_b64 s[2:3], s[2:3], s6 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr <2 x i64> %value, %amount |
| ret <2 x i64> %result |
| } |
| |
| define i65 @v_lshr_i65(i65 %value, i65 %amount) { |
| ; GFX6-LABEL: v_lshr_i65: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_mov_b32_e32 v4, 1 |
| ; GFX6-NEXT: v_mov_b32_e32 v5, 0 |
| ; GFX6-NEXT: v_and_b32_e32 v4, 1, v2 |
| ; GFX6-NEXT: v_sub_i32_e32 v8, vcc, 64, v3 |
| ; GFX6-NEXT: v_add_i32_e32 v2, vcc, 0xffffffc0, v3 |
| ; GFX6-NEXT: v_lshr_b64 v[6:7], v[0:1], v3 |
| ; GFX6-NEXT: v_lshl_b64 v[8:9], v[4:5], v8 |
| ; GFX6-NEXT: v_lshr_b64 v[10:11], v[4:5], v3 |
| ; GFX6-NEXT: v_lshr_b64 v[4:5], v[4:5], v2 |
| ; GFX6-NEXT: v_or_b32_e32 v6, v6, v8 |
| ; GFX6-NEXT: v_or_b32_e32 v7, v7, v9 |
| ; GFX6-NEXT: v_cmp_gt_u32_e32 vcc, 64, v3 |
| ; GFX6-NEXT: v_cndmask_b32_e32 v2, v4, v6, vcc |
| ; GFX6-NEXT: v_cndmask_b32_e32 v4, v5, v7, vcc |
| ; GFX6-NEXT: v_cmp_eq_u32_e64 s[4:5], 0, v3 |
| ; GFX6-NEXT: v_cndmask_b32_e64 v0, v2, v0, s[4:5] |
| ; GFX6-NEXT: v_cndmask_b32_e64 v1, v4, v1, s[4:5] |
| ; GFX6-NEXT: v_cndmask_b32_e32 v2, 0, v10, vcc |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i65: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_mov_b32_e32 v4, 1 |
| ; GFX8-NEXT: v_mov_b32_e32 v5, 0 |
| ; GFX8-NEXT: v_and_b32_e32 v4, 1, v2 |
| ; GFX8-NEXT: v_sub_u32_e32 v8, vcc, 64, v3 |
| ; GFX8-NEXT: v_add_u32_e32 v2, vcc, 0xffffffc0, v3 |
| ; GFX8-NEXT: v_lshrrev_b64 v[6:7], v3, v[0:1] |
| ; GFX8-NEXT: v_lshlrev_b64 v[8:9], v8, v[4:5] |
| ; GFX8-NEXT: v_lshrrev_b64 v[10:11], v3, v[4:5] |
| ; GFX8-NEXT: v_lshrrev_b64 v[4:5], v2, v[4:5] |
| ; GFX8-NEXT: v_or_b32_e32 v6, v6, v8 |
| ; GFX8-NEXT: v_or_b32_e32 v7, v7, v9 |
| ; GFX8-NEXT: v_cmp_gt_u32_e32 vcc, 64, v3 |
| ; GFX8-NEXT: v_cndmask_b32_e32 v2, v4, v6, vcc |
| ; GFX8-NEXT: v_cndmask_b32_e32 v4, v5, v7, vcc |
| ; GFX8-NEXT: v_cmp_eq_u32_e64 s[4:5], 0, v3 |
| ; GFX8-NEXT: v_cndmask_b32_e64 v0, v2, v0, s[4:5] |
| ; GFX8-NEXT: v_cndmask_b32_e64 v1, v4, v1, s[4:5] |
| ; GFX8-NEXT: v_cndmask_b32_e32 v2, 0, v10, vcc |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i65: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_mov_b32_e32 v4, 1 |
| ; GFX9-NEXT: v_mov_b32_e32 v5, 0 |
| ; GFX9-NEXT: v_and_b32_e32 v4, 1, v2 |
| ; GFX9-NEXT: v_sub_u32_e32 v8, 64, v3 |
| ; GFX9-NEXT: v_add_u32_e32 v2, 0xffffffc0, v3 |
| ; GFX9-NEXT: v_lshrrev_b64 v[6:7], v3, v[0:1] |
| ; GFX9-NEXT: v_lshlrev_b64 v[8:9], v8, v[4:5] |
| ; GFX9-NEXT: v_lshrrev_b64 v[10:11], v3, v[4:5] |
| ; GFX9-NEXT: v_lshrrev_b64 v[4:5], v2, v[4:5] |
| ; GFX9-NEXT: v_or_b32_e32 v6, v6, v8 |
| ; GFX9-NEXT: v_or_b32_e32 v7, v7, v9 |
| ; GFX9-NEXT: v_cmp_gt_u32_e32 vcc, 64, v3 |
| ; GFX9-NEXT: v_cndmask_b32_e32 v2, v4, v6, vcc |
| ; GFX9-NEXT: v_cndmask_b32_e32 v4, v5, v7, vcc |
| ; GFX9-NEXT: v_cmp_eq_u32_e64 s[4:5], 0, v3 |
| ; GFX9-NEXT: v_cndmask_b32_e64 v0, v2, v0, s[4:5] |
| ; GFX9-NEXT: v_cndmask_b32_e64 v1, v4, v1, s[4:5] |
| ; GFX9-NEXT: v_cndmask_b32_e32 v2, 0, v10, vcc |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10-LABEL: v_lshr_i65: |
| ; GFX10: ; %bb.0: |
| ; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10-NEXT: v_mov_b32_e32 v4, 1 |
| ; GFX10-NEXT: v_mov_b32_e32 v5, 0 |
| ; GFX10-NEXT: v_and_b32_e32 v4, 1, v2 |
| ; GFX10-NEXT: v_sub_nc_u32_e32 v2, 64, v3 |
| ; GFX10-NEXT: v_add_nc_u32_e32 v10, 0xffffffc0, v3 |
| ; GFX10-NEXT: v_lshrrev_b64 v[6:7], v3, v[0:1] |
| ; GFX10-NEXT: v_cmp_gt_u32_e32 vcc_lo, 64, v3 |
| ; GFX10-NEXT: v_cmp_eq_u32_e64 s4, 0, v3 |
| ; GFX10-NEXT: v_lshlrev_b64 v[8:9], v2, v[4:5] |
| ; GFX10-NEXT: v_lshrrev_b64 v[10:11], v10, v[4:5] |
| ; GFX10-NEXT: v_lshrrev_b64 v[4:5], v3, v[4:5] |
| ; GFX10-NEXT: v_or_b32_e32 v2, v6, v8 |
| ; GFX10-NEXT: v_or_b32_e32 v6, v7, v9 |
| ; GFX10-NEXT: v_cndmask_b32_e32 v2, v10, v2, vcc_lo |
| ; GFX10-NEXT: v_cndmask_b32_e32 v5, v11, v6, vcc_lo |
| ; GFX10-NEXT: v_cndmask_b32_e64 v0, v2, v0, s4 |
| ; GFX10-NEXT: v_cndmask_b32_e64 v1, v5, v1, s4 |
| ; GFX10-NEXT: v_cndmask_b32_e32 v2, 0, v4, vcc_lo |
| ; GFX10-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX11-LABEL: v_lshr_i65: |
| ; GFX11: ; %bb.0: |
| ; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX11-NEXT: v_cmp_gt_u32_e32 vcc_lo, 64, v3 |
| ; GFX11-NEXT: v_mov_b32_e32 v4, 1 |
| ; GFX11-NEXT: v_dual_mov_b32 v5, 0 :: v_dual_and_b32 v4, 1, v2 |
| ; GFX11-NEXT: v_sub_nc_u32_e32 v2, 64, v3 |
| ; GFX11-NEXT: v_add_nc_u32_e32 v10, 0xffffffc0, v3 |
| ; GFX11-NEXT: v_lshrrev_b64 v[6:7], v3, v[0:1] |
| ; GFX11-NEXT: v_cmp_eq_u32_e64 s0, 0, v3 |
| ; GFX11-NEXT: v_lshlrev_b64 v[8:9], v2, v[4:5] |
| ; GFX11-NEXT: v_lshrrev_b64 v[10:11], v10, v[4:5] |
| ; GFX11-NEXT: v_lshrrev_b64 v[4:5], v3, v[4:5] |
| ; GFX11-NEXT: v_or_b32_e32 v2, v6, v8 |
| ; GFX11-NEXT: v_or_b32_e32 v6, v7, v9 |
| ; GFX11-NEXT: v_cndmask_b32_e32 v2, v10, v2, vcc_lo |
| ; GFX11-NEXT: v_cndmask_b32_e32 v5, v11, v6, vcc_lo |
| ; GFX11-NEXT: v_cndmask_b32_e64 v0, v2, v0, s0 |
| ; GFX11-NEXT: v_cndmask_b32_e64 v1, v5, v1, s0 |
| ; GFX11-NEXT: v_cndmask_b32_e32 v2, 0, v4, vcc_lo |
| ; GFX11-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i65 %value, %amount |
| ret i65 %result |
| } |
| |
| define i65 @v_lshr_i65_33(i65 %value) { |
| ; GFX6-LABEL: v_lshr_i65_33: |
| ; GFX6: ; %bb.0: |
| ; GFX6-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX6-NEXT: v_mov_b32_e32 v3, v1 |
| ; GFX6-NEXT: v_mov_b32_e32 v0, 1 |
| ; GFX6-NEXT: v_mov_b32_e32 v1, 0 |
| ; GFX6-NEXT: v_and_b32_e32 v0, 1, v2 |
| ; GFX6-NEXT: v_lshl_b64 v[0:1], v[0:1], 31 |
| ; GFX6-NEXT: v_lshrrev_b32_e32 v2, 1, v3 |
| ; GFX6-NEXT: v_or_b32_e32 v0, v0, v2 |
| ; GFX6-NEXT: v_mov_b32_e32 v2, 0 |
| ; GFX6-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX8-LABEL: v_lshr_i65_33: |
| ; GFX8: ; %bb.0: |
| ; GFX8-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX8-NEXT: v_mov_b32_e32 v3, v1 |
| ; GFX8-NEXT: v_mov_b32_e32 v0, 1 |
| ; GFX8-NEXT: v_mov_b32_e32 v1, 0 |
| ; GFX8-NEXT: v_and_b32_e32 v0, 1, v2 |
| ; GFX8-NEXT: v_lshlrev_b64 v[0:1], 31, v[0:1] |
| ; GFX8-NEXT: v_lshrrev_b32_e32 v2, 1, v3 |
| ; GFX8-NEXT: v_or_b32_e32 v0, v0, v2 |
| ; GFX8-NEXT: v_mov_b32_e32 v2, 0 |
| ; GFX8-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX9-LABEL: v_lshr_i65_33: |
| ; GFX9: ; %bb.0: |
| ; GFX9-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX9-NEXT: v_mov_b32_e32 v3, v1 |
| ; GFX9-NEXT: v_mov_b32_e32 v0, 1 |
| ; GFX9-NEXT: v_mov_b32_e32 v1, 0 |
| ; GFX9-NEXT: v_and_b32_e32 v0, 1, v2 |
| ; GFX9-NEXT: v_lshlrev_b64 v[0:1], 31, v[0:1] |
| ; GFX9-NEXT: v_lshrrev_b32_e32 v2, 1, v3 |
| ; GFX9-NEXT: v_or_b32_e32 v0, v0, v2 |
| ; GFX9-NEXT: v_mov_b32_e32 v2, 0 |
| ; GFX9-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX10-LABEL: v_lshr_i65_33: |
| ; GFX10: ; %bb.0: |
| ; GFX10-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX10-NEXT: v_mov_b32_e32 v3, v1 |
| ; GFX10-NEXT: v_mov_b32_e32 v0, 1 |
| ; GFX10-NEXT: v_mov_b32_e32 v1, 0 |
| ; GFX10-NEXT: v_and_b32_e32 v0, 1, v2 |
| ; GFX10-NEXT: v_lshrrev_b32_e32 v2, 1, v3 |
| ; GFX10-NEXT: v_lshlrev_b64 v[0:1], 31, v[0:1] |
| ; GFX10-NEXT: v_or_b32_e32 v0, v0, v2 |
| ; GFX10-NEXT: v_mov_b32_e32 v2, 0 |
| ; GFX10-NEXT: s_setpc_b64 s[30:31] |
| ; |
| ; GFX11-LABEL: v_lshr_i65_33: |
| ; GFX11: ; %bb.0: |
| ; GFX11-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0) |
| ; GFX11-NEXT: v_dual_mov_b32 v3, v1 :: v_dual_mov_b32 v0, 1 |
| ; GFX11-NEXT: v_dual_mov_b32 v1, 0 :: v_dual_and_b32 v0, 1, v2 |
| ; GFX11-NEXT: v_lshrrev_b32_e32 v2, 1, v3 |
| ; GFX11-NEXT: v_lshlrev_b64 v[0:1], 31, v[0:1] |
| ; GFX11-NEXT: v_or_b32_e32 v0, v0, v2 |
| ; GFX11-NEXT: v_mov_b32_e32 v2, 0 |
| ; GFX11-NEXT: s_setpc_b64 s[30:31] |
| %result = lshr i65 %value, 33 |
| ret i65 %result |
| } |
| |
| define amdgpu_ps i65 @s_lshr_i65(i65 inreg %value, i65 inreg %amount) { |
| ; GCN-LABEL: s_lshr_i65: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_and_b64 s[4:5], s[2:3], 1 |
| ; GCN-NEXT: s_sub_i32 s10, s3, 64 |
| ; GCN-NEXT: s_sub_i32 s8, 64, s3 |
| ; GCN-NEXT: s_cmp_lt_u32 s3, 64 |
| ; GCN-NEXT: s_cselect_b32 s11, 1, 0 |
| ; GCN-NEXT: s_cmp_eq_u32 s3, 0 |
| ; GCN-NEXT: s_cselect_b32 s12, 1, 0 |
| ; GCN-NEXT: s_lshr_b64 s[6:7], s[4:5], s3 |
| ; GCN-NEXT: s_lshr_b64 s[2:3], s[0:1], s3 |
| ; GCN-NEXT: s_lshl_b64 s[8:9], s[4:5], s8 |
| ; GCN-NEXT: s_or_b64 s[2:3], s[2:3], s[8:9] |
| ; GCN-NEXT: s_lshr_b64 s[4:5], s[4:5], s10 |
| ; GCN-NEXT: s_cmp_lg_u32 s11, 0 |
| ; GCN-NEXT: s_cselect_b64 s[2:3], s[2:3], s[4:5] |
| ; GCN-NEXT: s_cmp_lg_u32 s12, 0 |
| ; GCN-NEXT: s_cselect_b64 s[0:1], s[0:1], s[2:3] |
| ; GCN-NEXT: s_cmp_lg_u32 s11, 0 |
| ; GCN-NEXT: s_cselect_b32 s2, s6, 0 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i65: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b64 s[4:5], s[2:3], 1 |
| ; GFX10PLUS-NEXT: s_sub_i32 s10, s3, 64 |
| ; GFX10PLUS-NEXT: s_sub_i32 s2, 64, s3 |
| ; GFX10PLUS-NEXT: s_cmp_lt_u32 s3, 64 |
| ; GFX10PLUS-NEXT: s_cselect_b32 s11, 1, 0 |
| ; GFX10PLUS-NEXT: s_cmp_eq_u32 s3, 0 |
| ; GFX10PLUS-NEXT: s_cselect_b32 s12, 1, 0 |
| ; GFX10PLUS-NEXT: s_lshr_b64 s[6:7], s[0:1], s3 |
| ; GFX10PLUS-NEXT: s_lshl_b64 s[8:9], s[4:5], s2 |
| ; GFX10PLUS-NEXT: s_lshr_b64 s[2:3], s[4:5], s3 |
| ; GFX10PLUS-NEXT: s_or_b64 s[6:7], s[6:7], s[8:9] |
| ; GFX10PLUS-NEXT: s_lshr_b64 s[4:5], s[4:5], s10 |
| ; GFX10PLUS-NEXT: s_cmp_lg_u32 s11, 0 |
| ; GFX10PLUS-NEXT: s_cselect_b64 s[4:5], s[6:7], s[4:5] |
| ; GFX10PLUS-NEXT: s_cmp_lg_u32 s12, 0 |
| ; GFX10PLUS-NEXT: s_cselect_b64 s[0:1], s[0:1], s[4:5] |
| ; GFX10PLUS-NEXT: s_cmp_lg_u32 s11, 0 |
| ; GFX10PLUS-NEXT: s_cselect_b32 s2, s2, 0 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i65 %value, %amount |
| ret i65 %result |
| } |
| |
| define amdgpu_ps i65 @s_lshr_i65_33(i65 inreg %value) { |
| ; GCN-LABEL: s_lshr_i65_33: |
| ; GCN: ; %bb.0: |
| ; GCN-NEXT: s_and_b64 s[2:3], s[2:3], 1 |
| ; GCN-NEXT: s_lshr_b32 s4, s1, 1 |
| ; GCN-NEXT: s_lshl_b64 s[0:1], s[2:3], 31 |
| ; GCN-NEXT: s_or_b32 s0, s0, s4 |
| ; GCN-NEXT: s_mov_b32 s2, 0 |
| ; GCN-NEXT: ; return to shader part epilog |
| ; |
| ; GFX10PLUS-LABEL: s_lshr_i65_33: |
| ; GFX10PLUS: ; %bb.0: |
| ; GFX10PLUS-NEXT: s_and_b64 s[2:3], s[2:3], 1 |
| ; GFX10PLUS-NEXT: s_lshr_b32 s4, s1, 1 |
| ; GFX10PLUS-NEXT: s_lshl_b64 s[0:1], s[2:3], 31 |
| ; GFX10PLUS-NEXT: s_mov_b32 s2, 0 |
| ; GFX10PLUS-NEXT: s_or_b32 s0, s0, s4 |
| ; GFX10PLUS-NEXT: ; return to shader part epilog |
| %result = lshr i65 %value, 33 |
| ret i65 %result |
| } |
| |
| ; FIXME: Argument lowering asserts |
| ; define <2 x i65> @v_lshr_v2i65(<2 x i65> %value, <2 x i65> %amount) { |
| ; %result = lshr <2 x i65> %value, %amount |
| ; ret <2 x i65> %result |
| ; } |
| |
| ; define amdgpu_ps <2 x i65> @s_lshr_v2i65(<2 x i65> inreg %value, <2 x i65> inreg %amount) { |
| ; %result = lshr <2 x i65> %value, %amount |
| ; ret <2 x i65> %result |
| ; } |
