221
|
1 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=kaveri --amdhsa-code-object-version=2 | FileCheck --check-prefix=HSA %s
|
|
2 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=kaveri --amdhsa-code-object-version=2 -mattr=-flat-for-global | FileCheck --check-prefix=HSA-CI %s
|
|
3 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=carrizo --amdhsa-code-object-version=2 | FileCheck --check-prefix=HSA %s
|
|
4 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=carrizo --amdhsa-code-object-version=2 -mattr=-flat-for-global | FileCheck --check-prefix=HSA-VI %s
|
223
|
5 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=kaveri -filetype=obj --amdhsa-code-object-version=2 | llvm-readobj -S --sd --syms - | FileCheck --check-prefix=ELF %s
|
|
6 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=kaveri --amdhsa-code-object-version=2 | llvm-mc -filetype=obj -triple amdgcn--amdhsa -mcpu=kaveri --amdhsa-code-object-version=2 | llvm-readobj -S --sd --syms - | FileCheck %s --check-prefix=ELF
|
221
|
7 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=gfx1010 -mattr=+wavefrontsize32,-wavefrontsize64 | FileCheck --check-prefix=GFX10 --check-prefix=GFX10-W32 %s
|
|
8 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=gfx1010 -mattr=-wavefrontsize32,+wavefrontsize64 | FileCheck --check-prefix=GFX10 --check-prefix=GFX10-W64 %s
|
236
|
9 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=gfx1100 -mattr=+wavefrontsize32,-wavefrontsize64 | FileCheck --check-prefix=GFX10 --check-prefix=GFX10-W32 %s
|
|
10 ; RUN: llc < %s -mtriple=amdgcn--amdhsa -mcpu=gfx1100 -mattr=-wavefrontsize32,+wavefrontsize64 | FileCheck --check-prefix=GFX10 --check-prefix=GFX10-W64 %s
|
150
|
11
|
|
12 ; The SHT_NOTE section contains the output from the .hsa_code_object_*
|
|
13 ; directives.
|
|
14
|
|
15 ; ELF: Section {
|
|
16 ; ELF: Name: .text
|
|
17 ; ELF: Type: SHT_PROGBITS (0x1)
|
|
18 ; ELF: Flags [ (0x6)
|
|
19 ; ELF: SHF_ALLOC (0x2)
|
|
20 ; ELF: SHF_EXECINSTR (0x4)
|
|
21 ; ELF: }
|
|
22
|
|
23 ; ELF: SHT_NOTE
|
|
24 ; ELF: Flags [ (0x2)
|
|
25 ; ELF: SHF_ALLOC (0x2)
|
|
26 ; ELF: ]
|
|
27 ; ELF: SectionData (
|
|
28 ; ELF: 0000: 04000000 08000000 01000000 414D4400
|
|
29 ; ELF: 0010: 02000000 01000000 04000000 1B000000
|
|
30 ; ELF: 0020: 03000000 414D4400 04000700 07000000
|
|
31 ; ELF: 0030: 00000000 00000000 414D4400 414D4447
|
|
32 ; ELF: 0040: 50550000
|
|
33 ; ELF: )
|
|
34
|
|
35 ; ELF: Symbol {
|
|
36 ; ELF: Name: simple
|
|
37 ; ELF: Size: 288
|
|
38 ; ELF: Type: AMDGPU_HSA_KERNEL (0xA)
|
|
39 ; ELF: }
|
|
40
|
|
41 ; HSA-NOT: .AMDGPU.config
|
|
42 ; HSA: .text
|
|
43 ; HSA: .hsa_code_object_version 2,1
|
|
44 ; HSA-CI: .hsa_code_object_isa 7,0,0,"AMD","AMDGPU"
|
|
45 ; HSA-VI: .hsa_code_object_isa 8,0,1,"AMD","AMDGPU"
|
|
46
|
|
47 ; HSA-LABEL: .amdgpu_hsa_kernel simple
|
|
48 ; HSA: {{^}}simple:
|
|
49 ; HSA: .amd_kernel_code_t
|
|
50 ; HSA: enable_sgpr_private_segment_buffer = 1
|
|
51 ; HSA: enable_sgpr_kernarg_segment_ptr = 1
|
|
52
|
|
53 ; PRE-GFX10: enable_wavefront_size32 = 0
|
221
|
54 ; GFX10-W32: .amdhsa_wavefront_size32 1
|
|
55 ; GFX10-W64: .amdhsa_wavefront_size32 0
|
150
|
56
|
|
57 ; PRE-GFX10: wavefront_size = 6
|
|
58
|
|
59 ; HSA: call_convention = -1
|
|
60 ; HSA: .end_amd_kernel_code_t
|
236
|
61 ; HSA: s_load_{{dwordx2|b64}} s[{{[0-9]+:[0-9]+}}], s[4:5], 0x0
|
150
|
62
|
|
63 ; Make sure we are setting the ATC bit:
|
|
64 ; HSA-CI: s_mov_b32 s[[HI:[0-9]]], 0x100f000
|
|
65 ; On VI+ we also need to set MTYPE = 2
|
|
66 ; HSA-VI: s_mov_b32 s[[HI:[0-9]]], 0x1100f000
|
|
67 ; Make sure we generate flat store for HSA
|
|
68 ; PRE-GFX10: flat_store_dword v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}
|
236
|
69 ; GFX10: global_store_{{dword|b32}} v{{\[[0-9]+:[0-9]+\]}}, v{{[0-9]+}}, off
|
150
|
70
|
|
71 ; HSA: .Lfunc_end0:
|
|
72 ; HSA: .size simple, .Lfunc_end0-simple
|
|
73
|
|
74 define amdgpu_kernel void @simple(i32 addrspace(1)* %out) {
|
|
75 entry:
|
|
76 store i32 0, i32 addrspace(1)* %out
|
|
77 ret void
|
|
78 }
|
|
79
|
|
80 ; HSA-LABEL: .amdgpu_hsa_kernel simple_no_kernargs
|
|
81 ; HSA: enable_sgpr_kernarg_segment_ptr = 0
|
|
82 define amdgpu_kernel void @simple_no_kernargs() {
|
|
83 entry:
|
|
84 store volatile i32 0, i32 addrspace(1)* undef
|
|
85 ret void
|
|
86 }
|