1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-26 04:32:44 +01:00
llvm-mirror/test/CodeGen/AMDGPU/nand.ll
Graham Sellers 95cb757dac [AMDGPU] Add and update scalar instructions
This patch adds support for S_ANDN2, S_ORN2 32-bit and 64-bit instructions and adds splits to move them to the vector unit (for which there is no equivalent instruction). It modifies the way that the more complex scalar instructions are lowered to vector instructions by first breaking them down to sequences of simpler scalar instructions which are then lowered through the existing code paths. The pattern for S_XNOR has also been updated to apply inversion to one input rather than the output of the XOR as the result is equivalent and may allow leaving the NOT instruction on the scalar unit.

A new tests for NAND, NOR, ANDN2 and ORN2 have been added, and existing tests now hit the new instructions (and have been modified accordingly).

Differential: https://reviews.llvm.org/D54714
llvm-svn: 347877
2018-11-29 16:05:38 +00:00

84 lines
2.3 KiB
LLVM

; RUN: llc -march=amdgcn -mcpu=gfx600 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX600 %s
; RUN: llc -march=amdgcn -mcpu=gfx700 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX700 %s
; RUN: llc -march=amdgcn -mcpu=gfx801 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX801 %s
; RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck --check-prefix=GCN --check-prefix=GFX900 %s
; GCN-LABEL: {{^}}scalar_nand_i32_one_use
; GCN: s_nand_b32
define amdgpu_kernel void @scalar_nand_i32_one_use(
i32 addrspace(1)* %r0, i32 %a, i32 %b) {
entry:
%and = and i32 %a, %b
%r0.val = xor i32 %and, -1
store i32 %r0.val, i32 addrspace(1)* %r0
ret void
}
; GCN-LABEL: {{^}}scalar_nand_i32_mul_use
; GCN-NOT: s_nand_b32
; GCN: s_and_b32
; GCN: s_not_b32
; GCN: s_add_i32
define amdgpu_kernel void @scalar_nand_i32_mul_use(
i32 addrspace(1)* %r0, i32 addrspace(1)* %r1, i32 %a, i32 %b) {
entry:
%and = and i32 %a, %b
%r0.val = xor i32 %and, -1
%r1.val = add i32 %and, %a
store i32 %r0.val, i32 addrspace(1)* %r0
store i32 %r1.val, i32 addrspace(1)* %r1
ret void
}
; GCN-LABEL: {{^}}scalar_nand_i64_one_use
; GCN: s_nand_b64
define amdgpu_kernel void @scalar_nand_i64_one_use(
i64 addrspace(1)* %r0, i64 %a, i64 %b) {
entry:
%and = and i64 %a, %b
%r0.val = xor i64 %and, -1
store i64 %r0.val, i64 addrspace(1)* %r0
ret void
}
; GCN-LABEL: {{^}}scalar_nand_i64_mul_use
; GCN-NOT: s_nand_b64
; GCN: s_and_b64
; GCN: s_not_b64
; GCN: s_add_u32
; GCN: s_addc_u32
define amdgpu_kernel void @scalar_nand_i64_mul_use(
i64 addrspace(1)* %r0, i64 addrspace(1)* %r1, i64 %a, i64 %b) {
entry:
%and = and i64 %a, %b
%r0.val = xor i64 %and, -1
%r1.val = add i64 %and, %a
store i64 %r0.val, i64 addrspace(1)* %r0
store i64 %r1.val, i64 addrspace(1)* %r1
ret void
}
; GCN-LABEL: {{^}}vector_nand_i32_one_use
; GCN-NOT: s_nand_b32
; GCN: v_and_b32
; GCN: v_not_b32
define i32 @vector_nand_i32_one_use(i32 %a, i32 %b) {
entry:
%and = and i32 %a, %b
%r = xor i32 %and, -1
ret i32 %r
}
; GCN-LABEL: {{^}}vector_nand_i64_one_use
; GCN-NOT: s_nand_b64
; GCN: v_and_b32
; GCN: v_and_b32
; GCN: v_not_b32
; GCN: v_not_b32
define i64 @vector_nand_i64_one_use(i64 %a, i64 %b) {
entry:
%and = and i64 %a, %b
%r = xor i64 %and, -1
ret i64 %r
}