AI/rocm-systems
2
0
Forka 0
Codice Problemi Pull Request Actions Pacchetti Progetti Rilasci Wiki Attività

Merge 'master' into 'amd-master'

Sfoglia il codice sorgente 
Change-Id: I1dd6af19c418181b78b67f4b47beaf82b19170fc


[ROCm/hip commit: ec8df015e0]
This commit is contained in:
Jenkins
2019-08-21 15:50:56 -04:00
parent 78cc0f14b3 7dae0e25d1
commit e61530e8fe
28 ha cambiato i file con 288 aggiunte e 167 eliminazioni
Scarica il file Patch Scarica il file Diff Expand all files Collapse all files
projects/hip/CMakeLists.txt
+20 -3
Vedi File
@@ -44,12 +44,25 @@ execute_process(COMMAND git rev-parse --short HEAD
OUTPUT_VARIABLE HIP_VERSION_GITHASH
OUTPUT_STRIP_TRAILING_WHITESPACE)
# get commit count
execute_process(COMMAND git rev-list --count HEAD
WORKING_DIRECTORY ${PROJECT_SOURCE_DIR}
OUTPUT_VARIABLE HIP_VERSION_GITCOUNT
OUTPUT_STRIP_TRAILING_WHITESPACE)
set(HIP_VERSION_PATCH ${HIP_VERSION_GITDATE}-${HIP_VERSION_GITHASH})
set(HIP_VERSION ${HIP_VERSION_MAJOR}.${HIP_VERSION_MINOR}.${HIP_VERSION_PATCH})
add_to_config(_versionInfo HIP_VERSION_MAJOR)
add_to_config(_versionInfo HIP_VERSION_MINOR)
add_to_config(_versionInfo HIP_VERSION_PATCH)
# overwrite HIP_VERSION_PATCH for packaging
if(DEFINED ENV{ROCM_BUILD_ID})
set(HIP_VERSION_PATCH ${HIP_VERSION_GITDATE}.${HIP_VERSION_GITCOUNT}-$ENV{ROCM_BUILD_ID}-${HIP_VERSION_GITHASH})
else()
set(HIP_VERSION_PATCH ${HIP_VERSION_GITDATE}.${HIP_VERSION_GITCOUNT}-${HIP_VERSION_GITHASH})
endif()
set(HIP_VERSION ${HIP_VERSION_MAJOR}.${HIP_VERSION_MINOR}.${HIP_VERSION_PATCH})
#############################
# Configure variables
#############################
@@ -424,7 +437,7 @@ add_custom_target(pkg_hip_hcc COMMAND ${CMAKE_COMMAND} .
COMMAND cp *.rpm ${PROJECT_BINARY_DIR}
COMMAND cp *.tar.gz ${PROJECT_BINARY_DIR}
WORKING_DIRECTORY ${BUILD_DIR}
DEPENDS hip_hcc hip_hcc_static)
DEPENDS hip_hcc hip_hcc_static hiprtc)
# Package: hip_nvcc
set(BUILD_DIR ${CMAKE_CURRENT_BINARY_DIR}/packages/hip_nvcc)
@@ -464,7 +477,11 @@ if(POLICY CMP0037)
cmake_policy(PUSH)
cmake_policy(SET CMP0037 OLD)
endif()
add_custom_target(package DEPENDS pkg_hip_base pkg_hip_hcc pkg_hip_nvcc pkg_hip_doc pkg_hip_samples)
add_custom_target(package
COMMAND rename -v "'s/([a-z0-9_.\-]).deb/$1-amd64.deb/'" *.deb
COMMAND rename -v "'s/([a-z0-9_.\-]).rpm/$1.x86_64.rpm/'" *.rpm
WORKING_DIRECTORY ${PROJECT_BINARY_DIR}
DEPENDS pkg_hip_base pkg_hip_hcc pkg_hip_nvcc pkg_hip_doc pkg_hip_samples)
if(POLICY CMP0037)
cmake_policy(POP)
endif()
projects/hip/Jenkinsfile
esterno
+4 -4
Vedi File
@@ -295,13 +295,13 @@ def docker_upload_dockerhub( String local_org, String image_name, String remote_
String build_config = 'Release'
String job_name = env.JOB_NAME.toLowerCase( )
// The following launches 3 builds in parallel: rocm-head, rocm-2.6.x and cuda-10.x
parallel rocm_2_6:
// The following launches 3 builds in parallel: rocm-head, rocm-2.7.x and cuda-10.x
parallel rocm_2_7:
{
node('hip-rocm')
{
String hcc_ver = 'rocm-2.6.x'
String from_image = 'ci_test_nodes/rocm-2.6.x/ubuntu-16.04:latest'
String hcc_ver = 'rocm-2.7.x'
String from_image = 'ci_test_nodes/rocm-2.7.x/ubuntu-16.04:latest'
String inside_args = '--device=/dev/kfd --device=/dev/dri --group-add=video'
// Checkout source code, dependencies and version files
projects/hip/bin/hipcc
+2 -2
Vedi File
@@ -107,7 +107,7 @@ $HIP_COMPILER= $hipConfig{'HIP_COMPILER'};
$HIP_RUNTIME= $hipConfig{'HIP_RUNTIME'};
# If using VDI runtime, need to find HIP_VDI_HOME
if ($HIP_RUNTIME eq "VDI" and !defined $HIP_VDI_HOME) {
if (defined $HIP_RUNTIME and $HIP_RUNTIME eq "VDI" and !defined $HIP_VDI_HOME) {
my $hipcc_dir = dirname($0);
if (-e "$hipcc_dir/../lib/bitcode") {
$HIP_VDI_HOME = abs_path($hipcc_dir . "/..");
@@ -133,7 +133,7 @@ if (defined $HIP_VDI_HOME) {
}
}
if ($HIP_COMPILER eq "clang") {
if (defined $HIP_COMPILER and $HIP_COMPILER eq "clang") {
$HIP_PLATFORM = "clang";
if (!defined $HIP_CLANG_PATH) {
$HIP_CLANG_PATH = "/opt/rocm/llvm/bin";
projects/hip/docs/markdown/CUBLAS_API_supported_by_HIP.md
+5
Vedi File
@@ -4,6 +4,11 @@
| **type** | **CUDA** |**CUDA version\***| **HIP** |**HIP value** (if differs) |
|-------------:|---------------------------------------------------------------|:----------------:|------------------------------------------------------------|---------------------------|
| define |`CUBLAS_VER_MAJOR` | 10.1 Update 2 | |
| define |`CUBLAS_VER_MINOR` | 10.1 Update 2 | |
| define |`CUBLAS_VER_PATCH` | 10.1 Update 2 | |
| define |`CUBLAS_VER_BUILD` | 10.1 Update 2 | |
| define |`CUBLAS_VERSION` | 10.1 Update 2 | |
| enum |***`cublasStatus`*** | |***`hipblasStatus_t`*** |
| enum |***`cublasStatus_t`*** | |***`hipblasStatus_t`*** |
| 0 |*`CUBLAS_STATUS_SUCCESS`* | |*`HIPBLAS_STATUS_SUCCESS`* |
projects/hip/docs/markdown/CURAND_API_supported_by_HIP.md
+114 -109
Vedi File
@@ -2,115 +2,120 @@
## **1. CURAND Data types**
| **type** | **CUDA** | **HIP** | **HIP value** (if differs) |
|-------------:|---------------------------------------------------------------|------------------------------------------------------------|----------------------------|
| enum |***`curandStatus`*** |***`hiprandStatus`*** |
| enum |***`curandStatus_t`*** |***`hiprandStatus_t`*** |
| 0 |*`CURAND_STATUS_SUCCESS`* |*`HIPRAND_STATUS_SUCCESS`* |
| 100 |*`CURAND_STATUS_VERSION_MISMATCH`* |*`HIPRAND_STATUS_VERSION_MISMATCH`* |
| 101 |*`CURAND_STATUS_NOT_INITIALIZED`* |*`HIPRAND_STATUS_NOT_INITIALIZED`* |
| 102 |*`CURAND_STATUS_ALLOCATION_FAILED`* |*`HIPRAND_STATUS_ALLOCATION_FAILED`* |
| 103 |*`CURAND_STATUS_TYPE_ERROR`* |*`HIPRAND_STATUS_TYPE_ERROR`* |
| 104 |*`CURAND_STATUS_OUT_OF_RANGE`* |*`HIPRAND_STATUS_OUT_OF_RANGE`* |
| 105 |*`CURAND_STATUS_LENGTH_NOT_MULTIPLE`* |*`HIPRAND_STATUS_LENGTH_NOT_MULTIPLE`* |
| 106 |*`CURAND_STATUS_DOUBLE_PRECISION_REQUIRED`* |*`HIPRAND_STATUS_DOUBLE_PRECISION_REQUIRED`* |
| 201 |*`CURAND_STATUS_LAUNCH_FAILURE`* |*`HIPRAND_STATUS_LAUNCH_FAILURE`* |
| 202 |*`CURAND_STATUS_PREEXISTING_FAILURE`* |*`HIPRAND_STATUS_PREEXISTING_FAILURE`* |
| 203 |*`CURAND_STATUS_INITIALIZATION_FAILED`* |*`HIPRAND_STATUS_INITIALIZATION_FAILED`* |
| 204 |*`CURAND_STATUS_ARCH_MISMATCH`* |*`HIPRAND_STATUS_ARCH_MISMATCH`* |
| 999 |*`CURAND_STATUS_INTERNAL_ERROR`* |*`HIPRAND_STATUS_INTERNAL_ERROR`* |
| enum |***`curandRngType`*** |***`hiprandRngType`*** |
| enum |***`curandRngType_t`*** |***`hiprandRngType_t`*** |
| 0 |*`CURAND_RNG_TEST`* |*`HIPRAND_RNG_TEST`* |
| 100 |*`CURAND_RNG_PSEUDO_DEFAULT`* |*`HIPRAND_RNG_PSEUDO_DEFAULT`* | 400 |
| 101 |*`CURAND_RNG_PSEUDO_XORWOW`* |*`HIPRAND_RNG_PSEUDO_XORWOW`* | 401 |
| 121 |*`CURAND_RNG_PSEUDO_MRG32K3A`* |*`HIPRAND_RNG_PSEUDO_MRG32K3A`* | 402 |
| 141 |*`CURAND_RNG_PSEUDO_MTGP32`* |*`HIPRAND_RNG_PSEUDO_MTGP32`* | 403 |
| 142 |*`CURAND_RNG_PSEUDO_MT19937`* |*`HIPRAND_RNG_PSEUDO_MT19937`* | 404 |
| 161 |*`CURAND_RNG_PSEUDO_PHILOX4_32_10`* |*`HIPRAND_RNG_PSEUDO_PHILOX4_32_10`* | 405 |
| 200 |*`CURAND_RNG_QUASI_DEFAULT`* |*`HIPRAND_RNG_QUASI_DEFAULT`* | 500 |
| 201 |*`CURAND_RNG_QUASI_SOBOL32`* |*`HIPRAND_RNG_QUASI_SOBOL32`* | 501 |
| 202 |*`CURAND_RNG_QUASI_SCRAMBLED_SOBOL32`* |*`HIPRAND_RNG_QUASI_SCRAMBLED_SOBOL32`* | 502 |
| 203 |*`CURAND_RNG_QUASI_SOBOL64`* |*`HIPRAND_RNG_QUASI_SOBOL64`* | 503 |
| 204 |*`CURAND_RNG_QUASI_SCRAMBLED_SOBOL64`* |*`HIPRAND_RNG_QUASI_SCRAMBLED_SOBOL64`* | 504 |
| enum |***`curandOrdering`*** | |
| enum |***`curandOrdering_t`*** | |
| 100 |*`CURAND_ORDERING_PSEUDO_BEST`* | |
| 101 |*`CURAND_ORDERING_PSEUDO_DEFAULT`* | |
| 102 |*`CURAND_ORDERING_PSEUDO_SEEDED`* | |
| 201 |*`CURAND_ORDERING_QUASI_DEFAULT`* | |
| enum |***`curandDirectionVectorSet`*** | |
| enum |***`curandDirectionVectorSet_t`*** | |
| 101 |*`CURAND_DIRECTION_VECTORS_32_JOEKUO6`* | |
| 102 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_32_JOEKUO6`* | |
| 103 |*`CURAND_DIRECTION_VECTORS_64_JOEKUO6`* | |
| 104 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_64_JOEKUO6`* | |
| uint | `curandDirectionVectors32_t` | `hiprandDirectionVectors32_t` |
| uint | `curandDirectionVectors64_t` | |
| struct | `curandGenerator_st` | `hiprandGenerator_st` |
| struct* | `curandGenerator_t` | `hiprandGenerator_t` |
| enum |***`curandOrdering`*** | |
| enum |***`curandOrdering_t`*** | |
| 100 |*`CURAND_ORDERING_PSEUDO_BEST`* | |
| 101 |*`CURAND_ORDERING_PSEUDO_DEFAULT`* | |
| 102 |*`CURAND_ORDERING_PSEUDO_SEEDED`* | |
| 201 |*`CURAND_ORDERING_QUASI_DEFAULT`* | |
| enum |***`curandDirectionVectorSet`*** | |
| enum |***`curandDirectionVectorSet_t`*** | |
| 101 |*`CURAND_DIRECTION_VECTORS_32_JOEKUO6`* | |
| 102 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_32_JOEKUO6`* | |
| 103 |*`CURAND_DIRECTION_VECTORS_64_JOEKUO6`* | |
| 104 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_64_JOEKUO6`* | |
| uint | `curandDirectionVectors32_t` | `hiprandDirectionVectors32_t` |
| uint | `curandDirectionVectors64_t` | |
| double | `curandDistribution_st` | |
| double | `curandHistogramM2V_st` | |
| double* | `curandDistribution_t` | |
| double* | `curandHistogramM2V_t` | |
| struct | `curandDistributionShift_st` | |
| struct* | `curandDistributionShift_t` | |
| struct | `curandDistributionM2Shift_st` | |
| struct* | `curandDistributionM2Shift_t` | |
| struct | `curandHistogramM2_st` | |
| struct* | `curandHistogramM2_t` | |
| uint | `curandHistogramM2K_st` | |
| uint* | `curandHistogramM2K_t` | |
| struct | `curandDiscreteDistribution_st` | `hiprandDiscreteDistribution_st` |
| struct* | `curandDiscreteDistribution_t` | `hiprandDiscreteDistribution_t` |
| enum |***`curandMethod`*** | |
| enum |***`curandMethod_t`*** | |
| 0 |*`CURAND_CHOOSE_BEST`* | |
| 1 |*`CURAND_ITR`* | |
| 2 |*`CURAND_KNUTH`* | |
| 3 |*`CURAND_HITR`* | |
| 4 |*`CURAND_M1`* | |
| 5 |*`CURAND_M2`* | |
| 6 |*`CURAND_BINARY_SEARCH`* | |
| 7 |*`CURAND_DISCRETE_GAUSS`* | |
| 8 |*`CURAND_REJECTION`* | |
| 9 |*`CURAND_DEVICE_API`* | |
| 10 |*`CURAND_FAST_REJECTION`* | |
| 11 |*`CURAND_3RD`* | |
| 12 |*`CURAND_DEFINITION`* | |
| 13 |*`CURAND_POISSON`* | |
| struct | `curandStateMtgp32` | `hiprandStateMtgp32` |
| typedef | `curandStateMtgp32_t` | `hiprandStateMtgp32_t` |
| struct | `curandStateScrambledSobol64` | |
| typedef | `curandStateScrambledSobol64_t` | |
| struct | `curandStateSobol64` | |
| typedef | `curandStateSobol64_t` | |
| struct | `curandStateScrambledSobol32` | |
| typedef | `curandStateScrambledSobol32_t` | |
| struct | `curandStateSobol32` | `hiprandStateSobol32` |
| typedef | `curandStateSobol32_t` | `hiprandStateSobol32_t` |
| struct | `curandStateMRG32k3a` | `hiprandStateMRG32k3a` |
| typedef | `curandStateMRG32k3a_t` | `hiprandStateMRG32k3a_t` |
| struct | `curandStatePhilox4_32_10` | `hiprandStatePhilox4_32_10` |
| typedef | `curandStatePhilox4_32_10_t` | `hiprandStatePhilox4_32_10_t` |
| struct | `curandStateXORWOW` | `hiprandStateXORWOW` |
| typedef | `curandStateXORWOW_t` | `hiprandStateXORWOW_t` |
| struct | `curandState` | `hiprandState` |
| typedef | `curandState_t` | `hiprandState_t` |
| **type** | **CUDA** |**CUDA version\***| **HIP** | **HIP value** (if differs) |
|-------------:|---------------------------------------------------------------|:----------------:|------------------------------------------------------------|----------------------------|
| define |`CURAND_VER_MAJOR` | 10.1 Update 2 | |
| define |`CURAND_VER_MINOR` | 10.1 Update 2 | |
| define |`CURAND_VER_PATCH` | 10.1 Update 2 | |
| define |`CURAND_VER_BUILD` | 10.1 Update 2 | |
| define |`CURAND_VERSION` | 10.1 Update 2 | |
| enum |***`curandStatus`*** | |***`hiprandStatus`*** |
| enum |***`curandStatus_t`*** | |***`hiprandStatus_t`*** |
| 0 |*`CURAND_STATUS_SUCCESS`* | |*`HIPRAND_STATUS_SUCCESS`* |
| 100 |*`CURAND_STATUS_VERSION_MISMATCH`* | |*`HIPRAND_STATUS_VERSION_MISMATCH`* |
| 101 |*`CURAND_STATUS_NOT_INITIALIZED`* | |*`HIPRAND_STATUS_NOT_INITIALIZED`* |
| 102 |*`CURAND_STATUS_ALLOCATION_FAILED`* | |*`HIPRAND_STATUS_ALLOCATION_FAILED`* |
| 103 |*`CURAND_STATUS_TYPE_ERROR`* | |*`HIPRAND_STATUS_TYPE_ERROR`* |
| 104 |*`CURAND_STATUS_OUT_OF_RANGE`* | |*`HIPRAND_STATUS_OUT_OF_RANGE`* |
| 105 |*`CURAND_STATUS_LENGTH_NOT_MULTIPLE`* | |*`HIPRAND_STATUS_LENGTH_NOT_MULTIPLE`* |
| 106 |*`CURAND_STATUS_DOUBLE_PRECISION_REQUIRED`* | |*`HIPRAND_STATUS_DOUBLE_PRECISION_REQUIRED`* |
| 201 |*`CURAND_STATUS_LAUNCH_FAILURE`* | |*`HIPRAND_STATUS_LAUNCH_FAILURE`* |
| 202 |*`CURAND_STATUS_PREEXISTING_FAILURE`* | |*`HIPRAND_STATUS_PREEXISTING_FAILURE`* |
| 203 |*`CURAND_STATUS_INITIALIZATION_FAILED`* | |*`HIPRAND_STATUS_INITIALIZATION_FAILED`* |
| 204 |*`CURAND_STATUS_ARCH_MISMATCH`* | |*`HIPRAND_STATUS_ARCH_MISMATCH`* |
| 999 |*`CURAND_STATUS_INTERNAL_ERROR`* | |*`HIPRAND_STATUS_INTERNAL_ERROR`* |
| enum |***`curandRngType`*** | |***`hiprandRngType`*** |
| enum |***`curandRngType_t`*** | |***`hiprandRngType_t`*** |
| 0 |*`CURAND_RNG_TEST`* | |*`HIPRAND_RNG_TEST`* |
| 100 |*`CURAND_RNG_PSEUDO_DEFAULT`* | |*`HIPRAND_RNG_PSEUDO_DEFAULT`* | 400 |
| 101 |*`CURAND_RNG_PSEUDO_XORWOW`* | |*`HIPRAND_RNG_PSEUDO_XORWOW`* | 401 |
| 121 |*`CURAND_RNG_PSEUDO_MRG32K3A`* | |*`HIPRAND_RNG_PSEUDO_MRG32K3A`* | 402 |
| 141 |*`CURAND_RNG_PSEUDO_MTGP32`* | |*`HIPRAND_RNG_PSEUDO_MTGP32`* | 403 |
| 142 |*`CURAND_RNG_PSEUDO_MT19937`* | |*`HIPRAND_RNG_PSEUDO_MT19937`* | 404 |
| 161 |*`CURAND_RNG_PSEUDO_PHILOX4_32_10`* | |*`HIPRAND_RNG_PSEUDO_PHILOX4_32_10`* | 405 |
| 200 |*`CURAND_RNG_QUASI_DEFAULT`* | |*`HIPRAND_RNG_QUASI_DEFAULT`* | 500 |
| 201 |*`CURAND_RNG_QUASI_SOBOL32`* | |*`HIPRAND_RNG_QUASI_SOBOL32`* | 501 |
| 202 |*`CURAND_RNG_QUASI_SCRAMBLED_SOBOL32`* | |*`HIPRAND_RNG_QUASI_SCRAMBLED_SOBOL32`* | 502 |
| 203 |*`CURAND_RNG_QUASI_SOBOL64`* | |*`HIPRAND_RNG_QUASI_SOBOL64`* | 503 |
| 204 |*`CURAND_RNG_QUASI_SCRAMBLED_SOBOL64`* | |*`HIPRAND_RNG_QUASI_SCRAMBLED_SOBOL64`* | 504 |
| enum |***`curandOrdering`*** | | |
| enum |***`curandOrdering_t`*** | | |
| 100 |*`CURAND_ORDERING_PSEUDO_BEST`* | | |
| 101 |*`CURAND_ORDERING_PSEUDO_DEFAULT`* | | |
| 102 |*`CURAND_ORDERING_PSEUDO_SEEDED`* | | |
| 201 |*`CURAND_ORDERING_QUASI_DEFAULT`* | | |
| enum |***`curandDirectionVectorSet`*** | | |
| enum |***`curandDirectionVectorSet_t`*** | | |
| 101 |*`CURAND_DIRECTION_VECTORS_32_JOEKUO6`* | | |
| 102 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_32_JOEKUO6`* | | |
| 103 |*`CURAND_DIRECTION_VECTORS_64_JOEKUO6`* | | |
| 104 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_64_JOEKUO6`* | | |
| uint | `curandDirectionVectors32_t` | | `hiprandDirectionVectors32_t` |
| uint | `curandDirectionVectors64_t` | | |
| struct | `curandGenerator_st` | | `hiprandGenerator_st` |
| struct* | `curandGenerator_t` | | `hiprandGenerator_t` |
| enum |***`curandOrdering`*** | | |
| enum |***`curandOrdering_t`*** | | |
| 100 |*`CURAND_ORDERING_PSEUDO_BEST`* | | |
| 101 |*`CURAND_ORDERING_PSEUDO_DEFAULT`* | | |
| 102 |*`CURAND_ORDERING_PSEUDO_SEEDED`* | | |
| 201 |*`CURAND_ORDERING_QUASI_DEFAULT`* | | |
| enum |***`curandDirectionVectorSet`*** | | |
| enum |***`curandDirectionVectorSet_t`*** | | |
| 101 |*`CURAND_DIRECTION_VECTORS_32_JOEKUO6`* | | |
| 102 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_32_JOEKUO6`* | | |
| 103 |*`CURAND_DIRECTION_VECTORS_64_JOEKUO6`* | | |
| 104 |*`CURAND_SCRAMBLED_DIRECTION_VECTORS_64_JOEKUO6`* | | |
| uint | `curandDirectionVectors32_t` | | `hiprandDirectionVectors32_t` |
| uint | `curandDirectionVectors64_t` | | |
| double | `curandDistribution_st` | | |
| double | `curandHistogramM2V_st` | | |
| double* | `curandDistribution_t` | | |
| double* | `curandHistogramM2V_t` | | |
| struct | `curandDistributionShift_st` | | |
| struct* | `curandDistributionShift_t` | | |
| struct | `curandDistributionM2Shift_st` | | |
| struct* | `curandDistributionM2Shift_t` | | |
| struct | `curandHistogramM2_st` | | |
| struct* | `curandHistogramM2_t` | | |
| uint | `curandHistogramM2K_st` | | |
| uint* | `curandHistogramM2K_t` | | |
| struct | `curandDiscreteDistribution_st` | | `hiprandDiscreteDistribution_st` |
| struct* | `curandDiscreteDistribution_t` | | `hiprandDiscreteDistribution_t` |
| enum |***`curandMethod`*** | | |
| enum |***`curandMethod_t`*** | | |
| 0 |*`CURAND_CHOOSE_BEST`* | | |
| 1 |*`CURAND_ITR`* | | |
| 2 |*`CURAND_KNUTH`* | | |
| 3 |*`CURAND_HITR`* | | |
| 4 |*`CURAND_M1`* | | |
| 5 |*`CURAND_M2`* | | |
| 6 |*`CURAND_BINARY_SEARCH`* | | |
| 7 |*`CURAND_DISCRETE_GAUSS`* | | |
| 8 |*`CURAND_REJECTION`* | | |
| 9 |*`CURAND_DEVICE_API`* | | |
| 10 |*`CURAND_FAST_REJECTION`* | | |
| 11 |*`CURAND_3RD`* | | |
| 12 |*`CURAND_DEFINITION`* | | |
| 13 |*`CURAND_POISSON`* | | |
| struct | `curandStateMtgp32` | | `hiprandStateMtgp32` |
| typedef | `curandStateMtgp32_t` | | `hiprandStateMtgp32_t` |
| struct | `curandStateScrambledSobol64` | | |
| typedef | `curandStateScrambledSobol64_t` | | |
| struct | `curandStateSobol64` | | |
| typedef | `curandStateSobol64_t` | | |
| struct | `curandStateScrambledSobol32` | | |
| typedef | `curandStateScrambledSobol32_t` | | |
| struct | `curandStateSobol32` | | `hiprandStateSobol32` |
| typedef | `curandStateSobol32_t` | | `hiprandStateSobol32_t` |
| struct | `curandStateMRG32k3a` | | `hiprandStateMRG32k3a` |
| typedef | `curandStateMRG32k3a_t` | | `hiprandStateMRG32k3a_t` |
| struct | `curandStatePhilox4_32_10` | | `hiprandStatePhilox4_32_10` |
| typedef | `curandStatePhilox4_32_10_t` | | `hiprandStatePhilox4_32_10_t` |
| struct | `curandStateXORWOW` | | `hiprandStateXORWOW` |
| typedef | `curandStateXORWOW_t` | | `hiprandStateXORWOW_t` |
| struct | `curandState` | | `hiprandState` |
| typedef | `curandState_t` | | `hiprandState_t` |
## **2. Host API Functions**
projects/hip/docs/markdown/CUSPARSE_API_supported_by_HIP.md
+5
Vedi File
@@ -4,6 +4,11 @@
| **type** | **CUDA** |**CUDA version\***| **HIP** |
|-------------:|---------------------------------------------------------------|:-----------------|------------------------------------------------------------|
| define |`CUSPARSE_VER_MAJOR` | 10.1 Update 2 | |
| define |`CUSPARSE_VER_MINOR` | 10.1 Update 2 | |
| define |`CUSPARSE_VER_PATCH` | 10.1 Update 2 | |
| define |`CUSPARSE_VER_BUILD` | 10.1 Update 2 | |
| define |`CUSPARSE_VERSION` | 10.1 Update 2 | |
| enum |***`cusparseAction_t`*** | |***`hipsparseAction_t`*** |
| 0 |*`CUSPARSE_ACTION_SYMBOLIC`* | |*`HIPSPARSE_ACTION_SYMBOLIC`* |
| 1 |*`CUSPARSE_ACTION_NUMERIC`* | |*`HIPSPARSE_ACTION_NUMERIC`* |
projects/hip/hipify-clang/src/CUDA2HIP_BLAS_API_types.cpp
+7
Vedi File
@@ -24,6 +24,13 @@ THE SOFTWARE.
// Map of all functions
const std::map<llvm::StringRef, hipCounter> CUDA_BLAS_TYPE_NAME_MAP{
// Blas defines
{"CUBLAS_VER_MAJOR", {"HIPBLAS_VER_MAJOR", "", CONV_DEFINE, API_BLAS, HIP_UNSUPPORTED}},
{"CUBLAS_VER_MINOR", {"HIPBLAS_VER_MINOR", "", CONV_DEFINE, API_BLAS, HIP_UNSUPPORTED}},
{"CUBLAS_VER_PATCH", {"HIPBLAS_VER_PATCH", "", CONV_DEFINE, API_BLAS, HIP_UNSUPPORTED}},
{"CUBLAS_VER_BUILD", {"HIPBLAS_VER_BUILD", "", CONV_DEFINE, API_BLAS, HIP_UNSUPPORTED}},
{"CUBLAS_VERSION", {"HIPBLAS_VERSION", "", CONV_DEFINE, API_BLAS, HIP_UNSUPPORTED}},
// Blas operations
{"cublasOperation_t", {"hipblasOperation_t", "rocblas_operation", CONV_TYPE, API_BLAS}},
{"CUBLAS_OP_N", {"HIPBLAS_OP_N", "rocblas_operation_none", CONV_NUMERIC_LITERAL, API_BLAS}},
projects/hip/hipify-clang/src/CUDA2HIP_RAND_API_types.cpp
+10 -1
Vedi File
@@ -24,6 +24,14 @@ THE SOFTWARE.
// Map of all functions
const std::map<llvm::StringRef, hipCounter> CUDA_RAND_TYPE_NAME_MAP{
// RAND defines
{"CURAND_VER_MAJOR", {"HIPRAND_VER_MAJOR", "", CONV_DEFINE, API_RAND, HIP_UNSUPPORTED}},
{"CURAND_VER_MINOR", {"HIPRAND_VER_MINOR", "", CONV_DEFINE, API_RAND, HIP_UNSUPPORTED}},
{"CURAND_VER_PATCH", {"HIPRAND_VER_PATCH", "", CONV_DEFINE, API_RAND, HIP_UNSUPPORTED}},
{"CURAND_VER_BUILD", {"HIPRAND_VER_BUILD", "", CONV_DEFINE, API_RAND, HIP_UNSUPPORTED}},
{"CURAND_VERSION", {"HIPRAND_VERSION", "", CONV_DEFINE, API_RAND, HIP_UNSUPPORTED}},
// RAND Host types
{"curandStatus", {"hiprandStatus_t", "", CONV_TYPE, API_RAND}},
{"curandStatus_t", {"hiprandStatus_t", "", CONV_TYPE, API_RAND}},
{"curandRngType", {"hiprandRngType_t", "", CONV_TYPE, API_RAND}},
@@ -52,7 +60,8 @@ const std::map<llvm::StringRef, hipCounter> CUDA_RAND_TYPE_NAME_MAP{
{"curandMethod_t", {"hiprandMethod_t", "", CONV_TYPE, API_RAND, HIP_UNSUPPORTED}},
{"curandDirectionVectors32_t", {"hiprandDirectionVectors32_t", "", CONV_TYPE, API_RAND}},
{"curandDirectionVectors64_t", {"hiprandDirectionVectors64_t", "", CONV_TYPE, API_RAND, HIP_UNSUPPORTED}},
// cuRAND types for Device functions
// RAND types for Device functions
{"curandStateMtgp32", {"hiprandStateMtgp32", "", CONV_TYPE, API_RAND}},
{"curandStateMtgp32_t", {"hiprandStateMtgp32_t", "", CONV_TYPE, API_RAND}},
{"curandStateScrambledSobol64", {"hiprandStateScrambledSobol64", "", CONV_TYPE, API_RAND, HIP_UNSUPPORTED}},
projects/hip/hipify-clang/src/CUDA2HIP_SPARSE_API_types.cpp
+6
Vedi File
@@ -175,4 +175,10 @@ const std::map<llvm::StringRef, hipCounter> CUDA_SPARSE_TYPE_NAME_MAP{
{"CUSPARSE_INDEX_32I", {"HIPSPARSE_INDEX_32I", "", CONV_NUMERIC_LITERAL, API_SPARSE, HIP_UNSUPPORTED}},
{"CUSPARSE_INDEX_64I", {"HIPSPARSE_INDEX_64I", "", CONV_NUMERIC_LITERAL, API_SPARSE, HIP_UNSUPPORTED}},
// 3. Defines
{"CUSPARSE_VER_MAJOR", {"HIPSPARSE_VER_MAJOR", "", CONV_DEFINE, API_SPARSE, HIP_UNSUPPORTED}},
{"CUSPARSE_VER_MINOR", {"HIPSPARSE_VER_MINOR", "", CONV_DEFINE, API_SPARSE, HIP_UNSUPPORTED}},
{"CUSPARSE_VER_PATCH", {"HIPSPARSE_VER_PATCH", "", CONV_DEFINE, API_SPARSE, HIP_UNSUPPORTED}},
{"CUSPARSE_VER_BUILD", {"HIPSPARSE_VER_BUILD", "", CONV_DEFINE, API_SPARSE, HIP_UNSUPPORTED}},
{"CUSPARSE_VERSION", {"HIPSPARSE_VERSION", "", CONV_DEFINE, API_SPARSE, HIP_UNSUPPORTED}},
};
projects/hip/include/hip/hcc_detail/device_functions.h
+1 -1
Vedi File
@@ -50,7 +50,7 @@ __device__ static inline int __clz(int input) {
}
__device__ static inline int __clzll(long long int input) {
return __ockl_clz_u64((ulong)input);
return __ockl_clz_u64((ullong)input);
}
__device__ static inline unsigned int __ffs(unsigned int input) {
projects/hip/include/hip/hcc_detail/device_library_decls.h
+2 -1
Vedi File
@@ -34,6 +34,7 @@ typedef unsigned char uchar;
typedef unsigned short ushort;
typedef unsigned int uint;
typedef unsigned long ulong;
typedef unsigned long long ullong;
extern "C" __device__ __attribute__((const)) bool __ockl_wfany_i32(int);
extern "C" __device__ __attribute__((const)) bool __ockl_wfall_i32(int);
@@ -48,7 +49,7 @@ extern "C" __device__ __attribute__((const)) uint __ockl_sad_u32(uint, uint, uin
extern "C" __device__ __attribute__((const)) uchar __ockl_clz_u8(uchar);
extern "C" __device__ __attribute__((const)) ushort __ockl_clz_u16(ushort);
extern "C" __device__ __attribute__((const)) uint __ockl_clz_u32(uint);
extern "C" __device__ __attribute__((const)) ulong __ockl_clz_u64(ulong);
extern "C" __device__ __attribute__((const)) ullong __ockl_clz_u64(ullong);
extern "C" __device__ __attribute__((const)) float __ocml_floor_f32(float);
extern "C" __device__ __attribute__((const)) float __ocml_rint_f32(float);
projects/hip/include/hip/hcc_detail/hip_fp16.h
+2 -3
Vedi File
@@ -106,7 +106,7 @@ THE SOFTWARE.
#if !defined(__HIP_NO_HALF_CONVERSIONS__)
template<
typename T, Enable_if_t<std::is_integral<T>{}>* = nullptr>
__device__
__host__ __device__
__half(T x) : data{static_cast<_Float16>(x)} {}
#endif
@@ -233,11 +233,10 @@ THE SOFTWARE.
return __half_raw{data};
}
// ACCESSORS - DEVICE ONLY
#if !defined(__HIP_NO_HALF_CONVERSIONS__)
template<
typename T, Enable_if_t<std::is_integral<T>{}>* = nullptr>
__device__
__host__ __device__
operator T() const { return data; }
#endif
projects/hip/include/hip/hcc_detail/hip_runtime.h
+19
Vedi File
@@ -462,6 +462,25 @@ hc_get_workitem_absolute_id(int dim)
#undef __CUDA__
#pragma pop_macro("__CUDA__")
#if __HIP_VDI__
hipError_t hipExtModuleLaunchKernel(hipFunction_t f, uint32_t globalWorkSizeX,
uint32_t globalWorkSizeY, uint32_t globalWorkSizeZ,
uint32_t localWorkSizeX, uint32_t localWorkSizeY,
uint32_t localWorkSizeZ, size_t sharedMemBytes,
hipStream_t hStream, void** kernelParams, void** extra,
hipEvent_t startEvent = nullptr,
hipEvent_t stopEvent = nullptr,
uint32_t flags = 0);
hipError_t hipHccModuleLaunchKernel(hipFunction_t f, uint32_t globalWorkSizeX,
uint32_t globalWorkSizeY, uint32_t globalWorkSizeZ,
uint32_t localWorkSizeX, uint32_t localWorkSizeY,
uint32_t localWorkSizeZ, size_t sharedMemBytes,
hipStream_t hStream, void** kernelParams, void** extra,
hipEvent_t startEvent = nullptr,
hipEvent_t stopEvent = nullptr)
__attribute__((deprecated("use hipExtModuleLaunchKernel instead")));
#endif // __HIP_VDI__
#endif // defined(__clang__) && defined(__HIP__)
#include <hip/hcc_detail/hip_memory.h>
projects/hip/include/hip/hcc_detail/hip_runtime_api.h
+13
Vedi File
@@ -2022,6 +2022,19 @@ hipError_t hipMemcpyHtoA(hipArray* dstArray, size_t dstOffset, const void* srcHo
*/
hipError_t hipMemcpy3D(const struct hipMemcpy3DParms* p);
/**
* @brief Copies data between host and device asynchronously.
*
* @param[in] p 3D memory copy parameters
* @param[in] stream Stream to use
* @return #hipSuccess, #hipErrorInvalidValue, #hipErrorInvalidPitchValue,
* #hipErrorInvalidDevicePointer, #hipErrorInvalidMemcpyDirection
*
* @see hipMemcpy, hipMemcpy2DToArray, hipMemcpy2D, hipMemcpyFromArray, hipMemcpyToSymbol,
* hipMemcpyAsync
*/
hipError_t hipMemcpy3DAsync(const struct hipMemcpy3DParms* p, hipStream_t stream __dparm(0));
// doxygen end Memory
/**
* @}
projects/hip/include/hip/hcc_detail/math_functions.h
+7
Vedi File
@@ -29,7 +29,14 @@ THE SOFTWARE.
#include <hip/hcc_detail/host_defines.h>
#include <algorithm>
// assert.h is only for the host version of assert.
// The device version of assert is implemented in hip/hcc_detail/hip_runtime.h.
// Users should include hip_runtime.h for the device version of assert.
#if !__HIP_DEVICE_COMPILE__
#include <assert.h>
#endif
#include <limits.h>
#include <limits>
#include <stdint.h>
projects/hip/include/hip/nvcc_detail/hip_runtime_api.h
+11 -6
Vedi File
@@ -67,12 +67,12 @@ typedef enum hipMemcpyKind {
#define hipReadModeElementType cudaReadModeElementType
#define hipReadModeNormalizedFloat cudaReadModeNormalizedFloat
typedef enum hipChannelFormatKind {
hipChannelFormatKindSigned = 0,
hipChannelFormatKindUnsigned = 1,
hipChannelFormatKindFloat = 2,
hipChannelFormatKindNone = 3
} hipChannelFormatKind;
// hipChannelFormatKind
#define hipChannelFormatKind cudaChannelFormatKind
#define hipChannelFormatKindSigned cudaChannelFormatKindSigned
#define hipChannelFormatKindUnsigned cudaChannelFormatKindUnsigned
#define hipChannelFormatKindFloat cudaChannelFormatKindFloat
#define hipChannelFormatKindNone cudaChannelFormatKindNone
#define hipSurfaceBoundaryMode cudaSurfaceBoundaryMode
#define hipBoundaryModeZero cudaBoundaryModeZero
@@ -607,6 +607,11 @@ inline static hipError_t hipMemcpy3D(const struct hipMemcpy3DParms *p)
return hipCUDAErrorTohipError(cudaMemcpy3D(p));
}
inline static hipError_t hipMemcpy3DAsync(const struct hipMemcpy3DParms *p, hipStream_t stream)
{
return hipCUDAErrorTohipError(cudaMemcpy3DAsync(p, stream));
}
inline static hipError_t hipMemcpy2DAsync(void* dst, size_t dpitch, const void* src, size_t spitch,
size_t width, size_t height, hipMemcpyKind kind,
hipStream_t stream) {
projects/hip/packaging/hip_doc.txt
+2 -1
Vedi File
@@ -35,7 +35,8 @@ set(CPACK_DEBIAN_PACKAGE_DEPENDS "hip_base (= ${CPACK_PACKAGE_VERSION})")
set(CPACK_BINARY_RPM "ON")
set(CPACK_RPM_PACKAGE_ARCHITECTURE "${CMAKE_SYSTEM_PROCESSOR}")
set(CPACK_RPM_PACKAGE_AUTOREQPROV " no")
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}")
string(REPLACE "-" "_" HIP_BASE_VERSION ${CPACK_PACKAGE_VERSION})
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${HIP_BASE_VERSION}")
set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt")
set(CPACK_SOURCE_GENERATOR "TGZ")
include(CPack)
projects/hip/packaging/hip_hcc.txt
+3 -2
Vedi File
@@ -42,10 +42,11 @@ set(CPACK_RPM_PACKAGE_ARCHITECTURE "${CMAKE_SYSTEM_PROCESSOR}")
set(CPACK_RPM_POST_INSTALL_SCRIPT_FILE "${PROJECT_BINARY_DIR}/postinst")
set(CPACK_RPM_PRE_UNINSTALL_SCRIPT_FILE "${PROJECT_BINARY_DIR}/prerm")
set(CPACK_RPM_PACKAGE_AUTOREQPROV " no")
string(REPLACE "-" "_" HIP_BASE_VERSION ${CPACK_PACKAGE_VERSION})
if(@COMPILE_HIP_ATP_MARKER@)
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}, ${HCC_PACKAGE_NAME} = @HCC_PACKAGE_VERSION@, rocm-profiler, comgr >= 1.1")
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${HIP_BASE_VERSION}, ${HCC_PACKAGE_NAME} = @HCC_PACKAGE_VERSION@, rocm-profiler, comgr >= 1.1")
else()
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}, ${HCC_PACKAGE_NAME} = @HCC_PACKAGE_VERSION@, comgr >= 1.1")
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${HIP_BASE_VERSION}, ${HCC_PACKAGE_NAME} = @HCC_PACKAGE_VERSION@, comgr >= 1.1")
endif()
set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt")
set(CPACK_SOURCE_GENERATOR "TGZ")
projects/hip/packaging/hip_nvcc.txt
+2 -1
Vedi File
@@ -24,7 +24,8 @@ set(CPACK_RPM_PACKAGE_ARCHITECTURE "${CMAKE_SYSTEM_PROCESSOR}")
#set(CPACK_RPM_POST_INSTALL_SCRIPT_FILE "${PROJECT_BINARY_DIR}/postinst")
#set(CPACK_RPM_PRE_UNINSTALL_SCRIPT_FILE "${PROJECT_BINARY_DIR}/prerm")
set(CPACK_RPM_PACKAGE_AUTOREQPROV " no")
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}, cuda >= 7.5")
string(REPLACE "-" "_" HIP_BASE_VERSION ${CPACK_PACKAGE_VERSION})
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${HIP_BASE_VERSION}, cuda >= 7.5")
set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt")
set(CPACK_SOURCE_GENERATOR "TGZ")
include(CPack)
projects/hip/packaging/hip_samples.txt
+2 -1
Vedi File
@@ -23,7 +23,8 @@ set(CPACK_DEBIAN_PACKAGE_DEPENDS "hip_base (= ${CPACK_PACKAGE_VERSION})")
set(CPACK_BINARY_RPM "ON")
set(CPACK_RPM_PACKAGE_ARCHITECTURE "${CMAKE_SYSTEM_PROCESSOR}")
set(CPACK_RPM_PACKAGE_AUTOREQPROV " no")
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${CPACK_PACKAGE_VERSION}")
string(REPLACE "-" "_" HIP_BASE_VERSION ${CPACK_PACKAGE_VERSION})
set(CPACK_RPM_PACKAGE_REQUIRES "hip_base = ${HIP_BASE_VERSION}")
set(CPACK_RPM_EXCLUDE_FROM_AUTO_FILELIST_ADDITION "/opt")
set(CPACK_SOURCE_GENERATOR "TGZ")
include(CPack)
projects/hip/samples/1_Utils/hipBusBandwidth/Makefile
+1 -1
Vedi File
@@ -5,7 +5,7 @@ endif
HIPCC=$(HIP_PATH)/bin/hipcc
EXE=hipBusBandwidth
CXXFLAGS = -O3 -g
CXXFLAGS = -O3
all: install
projects/hip/src/hip_clang.cpp
+7 -1
Vedi File
@@ -29,6 +29,10 @@ THE SOFTWARE.
#include "hip_fatbin.h"
#include "trace_helper.h"
#ifdef __GNUC__
#pragma GCC visibility push (default)
#endif
extern "C" std::vector<hipModule_t>*
__hipRegisterFatBinary(const void* data)
{
@@ -249,4 +253,6 @@ hipError_t hipLaunchByPtr(const void *hostFunction)
return ihipLogStatus(e);
}
#ifdef __GNUC__
#pragma GCC visibility pop
#endif
projects/hip/src/hip_memory.cpp
+32 -6
Vedi File
@@ -1387,10 +1387,9 @@ hipError_t hipMemcpyAtoH(void* dst, hipArray* srcArray, size_t srcOffset, size_t
return ihipLogStatus(e);
}
hipError_t hipMemcpy3D(const struct hipMemcpy3DParms* p) {
HIP_INIT_SPECIAL_API(hipMemcpy3D, (TRACE_MCMD), p);
hipError_t ihipMemcpy3D(const struct hipMemcpy3DParms* p, hipStream_t stream, bool isAsync) {
hipError_t e = hipSuccess;
if (p) {
if(p) {
size_t byteSize;
size_t depth;
size_t height;
@@ -1448,11 +1447,14 @@ hipError_t hipMemcpy3D(const struct hipMemcpy3DParms* p) {
ySize = p->srcPtr.ysize;
dstPitch = p->dstPtr.pitch;
}
hipStream_t stream = ihipSyncAndResolveStream(hipStreamNull);
stream = ihipSyncAndResolveStream(stream);
hc::completion_future marker;
try {
if((widthInBytes == dstPitch) && (widthInBytes == srcPitch)) {
stream->locked_copySync((void*)dstPtr, (void*)srcPtr, widthInBytes*height*depth, p->kind, false);
if(isAsync)
stream->locked_copyAsync((void*)dstPtr, (void*)srcPtr, widthInBytes*height*depth, p->kind);
else
stream->locked_copySync((void*)dstPtr, (void*)srcPtr, widthInBytes*height*depth, p->kind, false);
} else {
for (int i = 0; i < depth; i++) {
for (int j = 0; j < height; j++) {
@@ -1461,7 +1463,10 @@ hipError_t hipMemcpy3D(const struct hipMemcpy3DParms* p) {
(unsigned char*)srcPtr + i * ySize * srcPitch + j * srcPitch;
unsigned char* dst =
(unsigned char*)dstPtr + i * height * dstPitch + j * dstPitch;
stream->locked_copySync(dst, src, widthInBytes, p->kind);
if(isAsync)
stream->locked_copyAsync(dst, src, widthInBytes, p->kind);
else
stream->locked_copySync(dst, src, widthInBytes, p->kind);
}
}
}
@@ -1471,6 +1476,20 @@ hipError_t hipMemcpy3D(const struct hipMemcpy3DParms* p) {
} else {
e = hipErrorInvalidValue;
}
return e;
}
hipError_t hipMemcpy3D(const struct hipMemcpy3DParms* p) {
HIP_INIT_SPECIAL_API(hipMemcpy3D, (TRACE_MCMD), p);
hipError_t e = hipSuccess;
e = ihipMemcpy3D(p, hipStreamNull, false);
return ihipLogStatus(e);
}
hipError_t hipMemcpy3DAsync(const struct hipMemcpy3DParms* p, hipStream_t stream) {
HIP_INIT_SPECIAL_API(hipMemcpy3DAsync, (TRACE_MCMD), p, stream);
hipError_t e = hipSuccess;
e = ihipMemcpy3D(p, stream, true);
return ihipLogStatus(e);
}
@@ -1538,6 +1557,13 @@ hipError_t ihipMemPtrGetInfo(void* ptr, size_t* size) {
template <typename T>
void ihipMemsetKernel(hipStream_t stream, T* ptr, T val, size_t count) {
// Just Use count, instead of dividing by 4, the calling API already does it
if (sizeof(T) == sizeof(uint32_t) && (count % sizeof(uint32_t) == 0) &&
!hsa_amd_memory_fill(ptr, reinterpret_cast<const std::uint32_t&>(val), count)) {
// Only return if the execution completes without error
// if error occured, try the normal version
return;
}
static constexpr uint32_t block_dim = 256;
const uint32_t grid_dim = clamp_integer<size_t>(count / block_dim, 1, UINT32_MAX);
projects/hip/tests/src/hipEnvVar.cpp
+1 -8
Vedi File
@@ -31,14 +31,7 @@ THE SOFTWARE.
#include "clara/clara.hpp"
#include <string>
#include "hip/hip_runtime.h"
#ifdef _WIN64
#include <tchar.h>
#endif
#ifdef _WIN64
#define setenv(x,y,z) _putenv_s(x,y)
#endif
#include "test_common.h"
using namespace std;
using namespace clara;
projects/hip/tests/src/hipEnvVarDriver.cpp
+1 -7
Vedi File
@@ -30,13 +30,7 @@ OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWA
#include "hip/hip_runtime.h"
#include <chrono>
#include <thread>
#ifdef _WIN64
#define popen(x,y) _popen(x,y)
#define pclose(x) _pclose(x)
#define setenv(x,y,z) _putenv_s(x,y)
#define unsetenv(x) _putenv_s(x,"")
#endif
#include "test_common.h"
using namespace std;
projects/hip/tests/src/runtimeApi/memory/hipMemcpy_simple.cpp
-3
Vedi File
@@ -29,9 +29,6 @@ THE SOFTWARE.
#include "hip/hip_runtime.h"
#include "test_common.h"
#ifdef _WIN64
#define aligned_alloc _aligned_malloc
#endif
bool p_async = false;
projects/hip/tests/src/runtimeApi/memory/hipMemoryAllocateCoherentDriver.cpp
+1 -6
Vedi File
@@ -29,12 +29,7 @@ OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWA
#include <assert.h>
#include <string>
#include "hip/hip_runtime.h"
#ifdef _WIN64
#define popen(x,y) _popen(x,y)
#define pclose(x) _pclose(x)
#define setenv(x,y,z) _putenv_s(x,y)
#endif
#include "test_common.h"
using namespace std;
projects/hip/tests/src/test_common.h
+8
Vedi File
@@ -95,6 +95,14 @@ THE SOFTWARE.
} \
}
#ifdef _WIN64
#include <tchar.h>
#define aligned_alloc _aligned_malloc
#define popen(x,y) _popen(x,y)
#define pclose(x) _pclose(x)
#define setenv(x,y,z) _putenv_s(x,y)
#endif
// standard command-line variables:
extern size_t N;
extern char memsetval;