File: data.cpp

package info (click to toggle)
llvm-toolchain-13 1%3A13.0.1-6~deb11u1
  • links: PTS, VCS
  • area: main
  • in suites: bullseye
  • size: 1,418,812 kB
  • sloc: cpp: 5,290,827; ansic: 996,570; asm: 544,593; python: 188,212; objc: 72,027; lisp: 30,291; f90: 25,395; sh: 24,900; javascript: 9,780; pascal: 9,398; perl: 7,484; ml: 5,432; awk: 3,523; makefile: 2,892; xml: 953; cs: 573; fortran: 539
file content (70 lines) | stat: -rw-r--r-- 2,378 bytes parent folder | download | duplicates (3)
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
//===--- amdgpu/impl/data.cpp ------------------------------------- C++ -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
#include "atmi_runtime.h"
#include "hsa_api.h"
#include "internal.h"
#include "machine.h"
#include "rt.h"
#include <cassert>
#include <stdio.h>
#include <string.h>
#include <vector>

using core::TaskImpl;
extern ATLMachine g_atl_machine;

namespace core {

namespace {
ATLProcessor &get_processor_by_mem_place(int DeviceId,
                                         atmi_devtype_t DeviceType) {
  switch (DeviceType) {
  case ATMI_DEVTYPE_CPU:
    return g_atl_machine.processors<ATLCPUProcessor>()[DeviceId];
  case ATMI_DEVTYPE_GPU:
    return g_atl_machine.processors<ATLGPUProcessor>()[DeviceId];
  }
}

hsa_amd_memory_pool_t get_memory_pool_by_mem_place(int DeviceId,
                                                   atmi_devtype_t DeviceType) {
  ATLProcessor &proc = get_processor_by_mem_place(DeviceId, DeviceType);
  return get_memory_pool(proc, 0 /*Memory Type (always zero) */);
}
} // namespace

hsa_status_t Runtime::DeviceMalloc(void **ptr, size_t size, int DeviceId) {
  return Runtime::Malloc(ptr, size, DeviceId, ATMI_DEVTYPE_GPU);
}

hsa_status_t Runtime::HostMalloc(void **ptr, size_t size) {
  hsa_status_t Err = Runtime::Malloc(ptr, size, 0, ATMI_DEVTYPE_CPU);
  if (Err == HSA_STATUS_SUCCESS) {
    Err = core::allow_access_to_all_gpu_agents(*ptr);
  }
  return Err;
}

hsa_status_t Runtime::Malloc(void **ptr, size_t size, int DeviceId,
                             atmi_devtype_t DeviceType) {
  hsa_amd_memory_pool_t pool =
      get_memory_pool_by_mem_place(DeviceId, DeviceType);
  hsa_status_t err = hsa_amd_memory_pool_allocate(pool, size, 0, ptr);
  DEBUG_PRINT("Malloced [%s %d] %p\n",
              DeviceType == ATMI_DEVTYPE_CPU ? "CPU" : "GPU", DeviceId, *ptr);
  return (err == HSA_STATUS_SUCCESS) ? HSA_STATUS_SUCCESS : HSA_STATUS_ERROR;
}

hsa_status_t Runtime::Memfree(void *ptr) {
  hsa_status_t err = hsa_amd_memory_pool_free(ptr);
  DEBUG_PRINT("Freed %p\n", ptr);

  return (err == HSA_STATUS_SUCCESS) ? HSA_STATUS_SUCCESS : HSA_STATUS_ERROR;
}

} // namespace core