File: wave_dispatch_regs.ll

package info (click to toggle)
llvm-toolchain-14 1%3A14.0.6-12
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 1,496,180 kB
  • sloc: cpp: 5,593,972; ansic: 986,872; asm: 585,869; python: 184,223; objc: 72,530; lisp: 31,119; f90: 27,793; javascript: 9,780; pascal: 9,762; sh: 9,482; perl: 7,468; ml: 5,432; awk: 3,523; makefile: 2,538; xml: 953; cs: 573; fortran: 567
file content (34 lines) | stat: -rw-r--r-- 1,758 bytes parent folder | download | duplicates (12)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI -enable-var-scope %s
; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI -enable-var-scope %s
; RUN: llc -mtriple=amdgcn--amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -enable-var-scope %s

; This compute shader has input args that claim that it has 17 sgprs and 5 vgprs
; in wave dispatch. Ensure that the sgpr and vgpr counts in COMPUTE_PGM_RSRC1
; are set to reflect that, even though the registers are not used in the shader.

; GCN-LABEL: {{^}}_amdgpu_cs_main:
; GCN:         .amdgpu_pal_metadata
; GCN-NEXT: ---
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT:   - .hardware_stages:
; GCN-NEXT:       .cs:
; GCN-NEXT:         .entry_point:    _amdgpu_cs_main
; GCN-NEXT:         .scratch_memory_size: 0
; SI-NEXT:          .sgpr_count:     0x11
; VI-NEXT:          .sgpr_count:     0x60
; GFX9-NEXT:        .sgpr_count:     0x11
; SI-NEXT:          .vgpr_count:     0x5
; VI-NEXT:          .vgpr_count:     0x5
; GFX9-NEXT:        .vgpr_count:     0x5
; GCN-NEXT:     .registers:
; SI-NEXT:        0x2e12 (COMPUTE_PGM_RSRC1): 0x{{[0-9a-f]*}}81
; VI-NEXT:        0x2e12 (COMPUTE_PGM_RSRC1): 0x{{[0-9a-f]*}}c1
; GFX9-NEXT:      0x2e12 (COMPUTE_PGM_RSRC1): 0x{{[0-9a-f]*}}81
; GCN-NEXT:       0x2e13 (COMPUTE_PGM_RSRC2): 0
; GCN-NEXT: ...
; GCN-NEXT:         .end_amdgpu_pal_metadata

define dllexport amdgpu_cs void @_amdgpu_cs_main(i32 inreg, i32 inreg, <2 x i32> inreg, i32 inreg, i32 inreg, i32 inreg, i32 inreg, i32 inreg, i32 inreg, i32 inreg, i32 inreg, i32 inreg, <3 x i32> inreg, i32 inreg, <5 x i32>) {
.entry:
  ret void
}