1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-22 18:54:02 +01:00
llvm-mirror/test/CodeGen/AMDGPU/select-constant-cttz.ll
Austin Kerbow d1f23a1772 [AMDGPU] Update subtarget features for new target ID support
Support for XNACK and SRAMECC is not static on some GPUs. We must be able
to differentiate between different scenarios for these dynamic subtarget
features.

The possible settings are:

- Unsupported: The GPU has no support for XNACK/SRAMECC.
- Any: Preference is unspecified. Use conservative settings that can run anywhere.
- Off: Request support for XNACK/SRAMECC Off
- On: Request support for XNACK/SRAMECC On

GCNSubtarget will track the four options based on the following criteria. If
the subtarget does not support XNACK/SRAMECC we say the setting is
"Unsupported". If no subtarget features for XNACK/SRAMECC are requested we
must support "Any" mode. If the subtarget features XNACK/SRAMECC exist in the
feature string when initializing the subtarget, the settings are "On/Off".

The defaults are updated to be conservatively correct, meaning if no setting
for XNACK or SRAMECC is explicitly requested, defaults will be used which
generate code that can be run anywhere. This corresponds to the "Any" setting.

Differential Revision: https://reviews.llvm.org/D85882
2021-01-26 11:25:51 -08:00

43 lines
1.7 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -march=amdgcn -verify-machineinstrs -o - %s | FileCheck -check-prefix=GCN %s
declare i32 @llvm.cttz.i32(i32, i1) nounwind readnone
declare i32 @llvm.amdgcn.sffbh.i32(i32) nounwind readnone speculatable
define amdgpu_kernel void @select_constant_cttz(i32 addrspace(1)* noalias %out, i32 addrspace(1)* nocapture readonly %arrayidx) nounwind {
; GCN-LABEL: select_constant_cttz:
; GCN: ; %bb.0:
; GCN-NEXT: s_load_dwordx2 s[2:3], s[0:1], 0xb
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_load_dword s2, s[2:3], 0x0
; GCN-NEXT: s_load_dwordx2 s[4:5], s[0:1], 0x9
; GCN-NEXT: s_mov_b32 s7, 0xf000
; GCN-NEXT: s_waitcnt lgkmcnt(0)
; GCN-NEXT: s_lshr_b32 s0, 1, s2
; GCN-NEXT: s_ff1_i32_b32 s0, s0
; GCN-NEXT: s_mov_b32 s6, -1
; GCN-NEXT: v_mov_b32_e32 v0, s0
; GCN-NEXT: v_cmp_ne_u32_e64 s[2:3], s2, 0
; GCN-NEXT: v_cndmask_b32_e64 v0, v0, -1, s[2:3]
; GCN-NEXT: v_ffbh_i32_e32 v1, v0
; GCN-NEXT: v_cmp_eq_u32_e64 s[0:1], 0, v0
; GCN-NEXT: v_sub_i32_e32 v0, vcc, 31, v1
; GCN-NEXT: s_or_b64 s[0:1], s[2:3], s[0:1]
; GCN-NEXT: v_cndmask_b32_e64 v0, v0, -1, s[0:1]
; GCN-NEXT: buffer_store_dword v0, off, s[4:7], 0
; GCN-NEXT: s_endpgm
%v = load i32, i32 addrspace(1)* %arrayidx, align 4
%sr = lshr i32 1, %v
%cmp = icmp ne i32 %v, 0
%cttz = call i32 @llvm.cttz.i32(i32 %sr, i1 true), !range !0
%sel = select i1 %cmp, i32 -1, i32 %cttz
%ffbh = call i32 @llvm.amdgcn.sffbh.i32(i32 %sel)
%sub = sub i32 31, %ffbh
%cmp2 = icmp eq i32 %sel, 0
%or = or i1 %cmp, %cmp2
%sel2 = select i1 %or, i32 -1, i32 %sub
store i32 %sel2, i32 addrspace(1)* %out
ret void
}
!0 = !{i32 0, i32 33}