File: AMDGPUBaseInfo.cpp

package info (click to toggle)
llvm-toolchain-3.7 1%3A3.7.1-5
  • links: PTS, VCS
  • area: main
  • in suites: stretch
  • size: 345,556 kB
  • ctags: 362,199
  • sloc: cpp: 2,156,381; ansic: 458,339; objc: 91,547; python: 89,988; asm: 86,305; sh: 21,479; makefile: 6,853; perl: 5,601; ml: 5,458; pascal: 3,933; lisp: 2,429; xml: 686; cs: 239; php: 202; csh: 117
file content (60 lines) | stat: -rw-r--r-- 1,848 bytes parent folder | download
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
//===-- AMDGPUBaseInfo.cpp - AMDGPU Base encoding information--------------===//
//
//                     The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "AMDGPUBaseInfo.h"
#include "llvm/MC/SubtargetFeature.h"

#define GET_SUBTARGETINFO_ENUM
#include "AMDGPUGenSubtargetInfo.inc"
#undef GET_SUBTARGETINFO_ENUM

namespace llvm {
namespace AMDGPU {

IsaVersion getIsaVersion(const FeatureBitset &Features) {

  if (Features.test(FeatureISAVersion7_0_0))
    return {7, 0, 0};

  if (Features.test(FeatureISAVersion7_0_1))
    return {7, 0, 1};

  if (Features.test(FeatureISAVersion8_0_0))
    return {8, 0, 0};

  if (Features.test(FeatureISAVersion8_0_1))
    return {8, 0, 1};

  return {0, 0, 0};
}

void initDefaultAMDKernelCodeT(amd_kernel_code_t &Header,
                               const FeatureBitset &Features) {

  IsaVersion ISA = getIsaVersion(Features);

  memset(&Header, 0, sizeof(Header));

  Header.amd_kernel_code_version_major = 1;
  Header.amd_kernel_code_version_minor = 0;
  Header.amd_machine_kind = 1; // AMD_MACHINE_KIND_AMDGPU
  Header.amd_machine_version_major = ISA.Major;
  Header.amd_machine_version_minor = ISA.Minor;
  Header.amd_machine_version_stepping = ISA.Stepping;
  Header.kernel_code_entry_byte_offset = sizeof(Header);
  // wavefront_size is specified as a power of 2: 2^6 = 64 threads.
  Header.wavefront_size = 6;
  // These alignment values are specified in powers of two, so alignment =
  // 2^n.  The minimum alignment is 2^4 = 16.
  Header.kernarg_segment_alignment = 4;
  Header.group_segment_alignment = 4;
  Header.private_segment_alignment = 4;
}

} // End namespace AMDGPU
} // End namespace llvm