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
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
|
; RUN: llc -march=amdgcn -verify-machineinstrs< %s | FileCheck --check-prefix=SI --check-prefix=CHECK %s
; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs< %s | FileCheck --check-prefix=CI --check-prefix=CHECK %s
; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs< %s | FileCheck --check-prefix=CI --check-prefix=CHECK %s
define amdgpu_kernel void @use_gep_address_space(ptr addrspace(3) %array) nounwind {
; CHECK-LABEL: {{^}}use_gep_address_space:
; CHECK: v_mov_b32_e32 [[PTR:v[0-9]+]], s{{[0-9]+}}
; CHECK: ds_write_b32 [[PTR]], v{{[0-9]+}} offset:64
%p = getelementptr [1024 x i32], ptr addrspace(3) %array, i16 0, i16 16
store i32 99, ptr addrspace(3) %p
ret void
}
; CHECK-LABEL: {{^}}use_gep_address_space_large_offset:
; The LDS offset will be 65536 bytes, which is larger than the size of LDS on
; SI, which is why it is being OR'd with the base pointer.
; SI: s_bitset1_b32
; CI: s_add_i32
; CHECK: ds_write_b32
define amdgpu_kernel void @use_gep_address_space_large_offset(ptr addrspace(3) %array) nounwind {
%p = getelementptr [1024 x i32], ptr addrspace(3) %array, i16 0, i16 16384
store i32 99, ptr addrspace(3) %p
ret void
}
; CHECK-LABEL: {{^}}gep_as_vector_v4:
; SI: s_add_i32
; SI: s_add_i32
; SI: s_add_i32
; SI: s_add_i32
; CHECK-DAG: v_mov_b32_e32 {{v[0-9]+}}, {{s[0-9]+}}
; CHECK-DAG: v_mov_b32_e32 {{v[0-9]+}}, {{s[0-9]+}}
; CHECK-DAG: v_mov_b32_e32 {{v[0-9]+}}, {{s[0-9]+}}
; CHECK-DAG: v_mov_b32_e32 {{v[0-9]+}}, {{s[0-9]+}}
; CI-DAG: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}} offset:64
; CI-DAG: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}} offset:64
; CI-DAG: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}} offset:64
; CI-DAG: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}} offset:64
; CHECK: s_endpgm
define amdgpu_kernel void @gep_as_vector_v4(<4 x ptr addrspace(3)> %array) nounwind {
%p = getelementptr [1024 x i32], <4 x ptr addrspace(3)> %array, <4 x i16> zeroinitializer, <4 x i16> <i16 16, i16 16, i16 16, i16 16>
%p0 = extractelement <4 x ptr addrspace(3)> %p, i32 0
%p1 = extractelement <4 x ptr addrspace(3)> %p, i32 1
%p2 = extractelement <4 x ptr addrspace(3)> %p, i32 2
%p3 = extractelement <4 x ptr addrspace(3)> %p, i32 3
store i32 99, ptr addrspace(3) %p0
store i32 99, ptr addrspace(3) %p1
store i32 99, ptr addrspace(3) %p2
store i32 99, ptr addrspace(3) %p3
ret void
}
; CHECK-LABEL: {{^}}gep_as_vector_v2:
; SI: s_add_i32
; SI: s_add_i32
; CHECK-DAG: v_mov_b32_e32 {{v[0-9]+}}, {{s[0-9]+}}
; CHECK-DAG: v_mov_b32_e32 {{v[0-9]+}}, {{s[0-9]+}}
; CI-DAG: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}} offset:64
; CI-DAG: ds_write_b32 v{{[0-9]+}}, v{{[0-9]+}} offset:64
; CHECK: s_endpgm
define amdgpu_kernel void @gep_as_vector_v2(<2 x ptr addrspace(3)> %array) nounwind {
%p = getelementptr [1024 x i32], <2 x ptr addrspace(3)> %array, <2 x i16> zeroinitializer, <2 x i16> <i16 16, i16 16>
%p0 = extractelement <2 x ptr addrspace(3)> %p, i32 0
%p1 = extractelement <2 x ptr addrspace(3)> %p, i32 1
store i32 99, ptr addrspace(3) %p0
store i32 99, ptr addrspace(3) %p1
ret void
}
|
