AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

P4 to Git Change 1177220 by emankov@em-hsa-amd on 2015/08/05 06:08:39

Browse Source 
ECR #333753 - ORCA RT/Compiler Lib/aoc2: AMD HSA Code Object Import feature (part II) - arbitrary hidden (extra) kernargs support

	Only HSAIL path is affected. It doesn't affect blit kernels.

	To use offline by aoc2:
	aoc2 -hsacodeobject=<importing_code_object_filename> -numhiddenkernargs=<num> -cl-std=CL2.0 -march=hsail(-64) -mdevice=Bonaire <source_cl_filename>

	To use online by setting env:
	AMD_DEBUG_HSA_NUM_HIDDEN_KERNARGS=<num>

	where num >= 0. If num == 0, then no additional arguments will be added on RT for every kernel. The default value is unchanged and equal to 6 for now.

	Misc:
	+ get rid of PRE & POST defines in Compiler Lib, as they started to conflict with ugl\gl\gs\hwl\ headers with the same defines.
	+ minor copy/paste eliminations & typo fixes
	+ ocltst complib tests update

	Testing: pre check-in, manually based on ocl sdk MatrixMultiplication

	Reviewers: Brian Sumner, German Andryeyev, Nikolay Haustov, Artem Tamazov

Affected files ...

... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/common/v0_8/if_acl.cpp#72 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/gpu/hsail_be.cpp#49 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/backends/gpu/metadata.cpp#8 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/include/v0_8/aclDefs.h#5 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/include/v0_8/aclEnums.h#19 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/include/v0_8/aclStructs.h#17 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/utils/bif_section_labels.hpp#21 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/utils/v0_8/libUtils.cpp#10 edit
... //depot/stg/opencl/drivers/opencl/compiler/lib/utils/v0_8/libUtils.h#20 edit
... //depot/stg/opencl/drivers/opencl/compiler/tools/aoc2/aoc2.cpp#74 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.cpp#181 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/device.hpp#249 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpukernel.cpp#291 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpukernel.hpp#113 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuprogram.cpp#199 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/gpu/gpuvirtual.cpp#369 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/hsa/hsaprogram.cpp#38 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/hsa_foundation/hsakernel.cpp#8 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/hsa_foundation/hsakernel.hpp#5 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/hsa_foundation/hsaprogram.cpp#19 edit
... //depot/stg/opencl/drivers/opencl/runtime/device/hsa_foundation/hsavirtual.cpp#43 edit
... //depot/stg/opencl/drivers/opencl/tests/ocltst/module/complib/CLAssumptionCheck.cpp#43 edit
... //depot/stg/opencl/drivers/opencl/tests/ocltst/module/complib/CLEnumCheck.cpp#44 edit
This commit is contained in:
foreman
2015-08-05 06:18:33 -04:00
parent 881eeab3b8
commit 81b331f4c5
14 changed files with 214 additions and 93 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rocclr/compiler/lib/backends/common/v0_8/if_acl.cpp
+11
View File
@@ -2641,6 +2641,17 @@ if_aclQueryInfo(aclCompiler *cl,
}
break;
}
case RT_NUM_KERNEL_HIDDEN_ARGS: {
size_t hidden_kernargs_size = sizeof(md->numHiddenKernelArgs);
if (!ptr) {
*size = hidden_kernargs_size;
success = true;
} else if (*size >= hidden_kernargs_size) {
memcpy(ptr, &md->numHiddenKernelArgs, hidden_kernargs_size);
success = true;
}
break;
}
}
return (success) ? ACL_SUCCESS : ACL_ERROR;
}
rocclr/compiler/lib/include/v0_8/aclDefs.h
+6
View File
@@ -28,4 +28,10 @@
#define BIF_API_3_0
#endif
#ifndef MAX_HIDDEN_KERNARGS_NUM
#define MAX_HIDDEN_KERNARGS_NUM 6
#else
#error "MAX_HIDDEN_KERNARGS_NUM is already defined"
#endif
#endif // _ACL_DEFS_0_8_H_
rocclr/compiler/lib/include/v0_8/aclEnums.h
+24 -23
View File
@@ -188,29 +188,30 @@ typedef enum _bif_sections_enum_0_8 {
//! An enumeration that defines what are valid queries for aclQueryInfo.
typedef enum _rt_query_types_enum_0_8 {
RT_ABI_VERSION = 0,
RT_DEVICE_NAME = 1,
RT_MEM_SIZES = 2,
RT_GPU_FUNC_CAPS = 3,
RT_GPU_FUNC_ID = 4,
RT_GPU_DEFAULT_ID = 5,
RT_WORK_GROUP_SIZE = 6,
RT_WORK_REGION_SIZE = 7,
RT_ARGUMENT_ARRAY = 8,
RT_GPU_PRINTF_ARRAY = 9,
RT_CPU_BARRIER_NAMES = 10,
RT_DEVICE_ENQUEUE = 11,
RT_KERNEL_INDEX = 12,
RT_KERNEL_NAME = 13,
RT_KERNEL_NAMES = 14,
RT_CONTAINS_LLVMIR = 15,
RT_CONTAINS_OPTIONS = 16,
RT_CONTAINS_BRIG = 17,
RT_CONTAINS_HSAIL = 18,
RT_CONTAINS_ISA = 19,
RT_CONTAINS_LOADER_MAP = 20,
RT_CONTAINS_SPIR = 21,
RT_LAST_TYPE = 22
RT_ABI_VERSION = 0,
RT_DEVICE_NAME = 1,
RT_MEM_SIZES = 2,
RT_GPU_FUNC_CAPS = 3,
RT_GPU_FUNC_ID = 4,
RT_GPU_DEFAULT_ID = 5,
RT_WORK_GROUP_SIZE = 6,
RT_WORK_REGION_SIZE = 7,
RT_ARGUMENT_ARRAY = 8,
RT_GPU_PRINTF_ARRAY = 9,
RT_CPU_BARRIER_NAMES = 10,
RT_DEVICE_ENQUEUE = 11,
RT_KERNEL_INDEX = 12,
RT_KERNEL_NAME = 13,
RT_KERNEL_NAMES = 14,
RT_CONTAINS_LLVMIR = 15,
RT_CONTAINS_OPTIONS = 16,
RT_CONTAINS_BRIG = 17,
RT_CONTAINS_HSAIL = 18,
RT_CONTAINS_ISA = 19,
RT_CONTAINS_LOADER_MAP = 20,
RT_CONTAINS_SPIR = 21,
RT_NUM_KERNEL_HIDDEN_ARGS = 22,
RT_LAST_TYPE = 23
} aclQueryType_0_8;
//! An enumeration for the various GPU capabilities
rocclr/compiler/lib/include/v0_8/aclStructs.h
+1
View File
@@ -119,6 +119,7 @@ typedef struct _acl_metadata_0_8 {
const char *deviceName; // RT_DEVICE_NAME
bool enqueue_kernel; // RT_DEVICE_ENQUEUE
uint32_t kernel_index; // RT_KERNEL_INDEX
uint32_t numHiddenKernelArgs; // RT_NUM_KERNEL_HIDDEN_ARGS
} aclMetadata_0_8;
//! An structure that holds information on the capabilities of the bif device.
rocclr/compiler/lib/utils/bif_section_labels.hpp
+4 -2
View File
@@ -6,8 +6,10 @@
#ifdef __cplusplus
extern "C" {
#endif
#define PRE 0
#define POST 1
namespace bif {
const unsigned PRE = 0;
const unsigned POST = 1;
}
typedef enum {
symOpenclCompilerOptions,
rocclr/compiler/lib/utils/v0_8/libUtils.cpp
+1 -3
View File
@@ -1,8 +1,7 @@
//
// Copyright (c) 2011 Advanced Micro Devices, Inc. All rights reserved.
//
#include "acl.h"
#include "aclTypes.h"
#include "api/v0_8/aclValidation.h"
#include "libUtils.h"
#include "bif/bifbase.hpp"
@@ -10,7 +9,6 @@
#include "utils/versions.hpp"
#include "utils/options.hpp"
#include "backends/gpu/scwrapper/devState.h"
#include <cassert>
#include <cstring>
#include "bif/bif.hpp"
extern aclBinary* constructBinary(size_t struct_version,
rocclr/compiler/lib/utils/v0_8/libUtils.h
+82 -10
View File
@@ -3,12 +3,16 @@
//
#ifndef _CL_LIB_UTILS_0_8_H_
#define _CL_LIB_UTILS_0_8_H_
#include "v0_8/aclTypes.h"
#include "acl.h"
#include <string>
#include <sstream>
#include <iterator>
#include <cstdlib>
#include <cassert>
#include "library.hpp"
#include "utils/bif_section_labels.hpp"
using namespace bif;
// Utility function to set a flag in option structure
// of the aclDevCaps.
void
@@ -150,6 +154,83 @@ aclutAlloc(const aclCompilerOptions *bin);
FreeFunc
aclutFree(const aclCompilerOptions *bin);
inline std::vector<std::string> splitSpaceSeparatedString(char *str)
{
std::string s(str);
std::stringstream ss(s);
std::istream_iterator<std::string> beg(ss), end;
std::vector<std::string> vec(beg, end);
return vec;
}
// Helper function that returns OpenCL mangled kernel name.
inline std::string
aclutOpenclMangledKernelName(const std::string& kernel_name)
{
const oclBIFSymbolStruct* sym = findBIF30SymStruct(symOpenclKernel);
assert(sym && "symbol not found");
return std::string("&") + sym->str[PRE] + kernel_name + sym->str[POST];
}
// Helper function that returns OpenCL mangled kernel metadata symbol name.
inline std::string
aclutOpenclMangledKernelMetadataName(const std::string& kernel_name)
{
const oclBIFSymbolStruct* sym = findBIF30SymStruct(symOpenclMeta);
assert(sym && "symbol not found");
return sym->str[PRE] + aclutOpenclMangledKernelName(kernel_name) + sym->str[POST];
}
#ifdef WITH_TARGET_HSAIL
// Helper function that updates metadata for all the kernels in binary;
// the updated attribute is the number of hidden kernel arguments.
inline acl_error
aclutUpdateMetadataWithHiddenKernargsNum(aclCompiler* cl, aclBinary* bin, uint32_t num) {
if (num == MAX_HIDDEN_KERNARGS_NUM) {
return ACL_SUCCESS;
}
const oclBIFSymbolStruct* sym = findBIF30SymStruct(symOpenclMeta);
assert(sym && "symbol not found");
aclSections secID = sym->sections[0];
size_t kernelNamesSize = 0;
acl_error error_code = aclQueryInfo(cl, bin, RT_KERNEL_NAMES, NULL, NULL, &kernelNamesSize);
if (error_code != ACL_SUCCESS) {
return error_code;
}
char* kernelNames = new char[kernelNamesSize];
error_code = aclQueryInfo(cl, bin, RT_KERNEL_NAMES, NULL, kernelNames, &kernelNamesSize);
if (error_code != ACL_SUCCESS) {
delete kernelNames;
return error_code;
}
std::vector<std::string> vKernels = splitSpaceSeparatedString(kernelNames);
delete kernelNames;
size_t roSize = 0;
for (auto it = vKernels.begin(); it != vKernels.end(); ++it) {
std::string symbol = aclutOpenclMangledKernelMetadataName(*it);
void* roSec = const_cast<void*>(aclExtractSymbol(cl, bin, &roSize, secID, symbol.c_str(), &error_code));
if (error_code != ACL_SUCCESS) {
return error_code;
}
if (!roSec || roSize == 0) {
error_code = ACL_ELF_ERROR;
return error_code;
}
aclMetadata *md = reinterpret_cast<aclMetadata*>(roSec);
md->numHiddenKernelArgs = num;
error_code = aclRemoveSymbol(cl, bin, secID, symbol.c_str());
if (error_code != ACL_SUCCESS) {
return error_code;
}
error_code = aclInsertSymbol(cl, bin, md, roSize, secID, symbol.c_str());
if (error_code != ACL_SUCCESS) {
return error_code;
}
}
return error_code;
}
#endif
inline bool is64BitTarget(const aclTargetInfo& target)
{
return (target.arch_id == aclX64 ||
@@ -184,15 +265,6 @@ enum scId {
SC_LAST,
};
inline std::vector<std::string> splitSpaceSeparatedString(char *str)
{
std::string s(str);
std::stringstream ss(s);
std::istream_iterator<std::string> beg(ss), end;
std::vector<std::string> vec(beg, end);
return vec;
}
// Helper function that allocates an aligned memory.
inline void*
alignedMalloc(size_t size, size_t alignment)