Skip to content

GitLab

"tests/test_class_sh_basic.cpp" did not exist on "32d2967b0c102fdcfad9d83b4f72571e4bb0b721"
Unverified Commit a73ab0d8 authored by Illia Silin's avatar Illia Silin Committed by GitHub
Browse files

Merge pull request #10 from ROCmSoftwarePlatform/lwpck-702 

Merge down from public repo.
parents aaf4defa 55a6b4e3
Expand all Show whitespace changes
Inline Side-by-side
Showing with 838 additions and 184 deletions
.github/dependabot.yml 0 → 100644
View file @ a73ab0d8
# To get started with Dependabot version updates, you'll need to specify which
# package ecosystems to update and where the package manifests are located.
# Please see the documentation for all configuration options:
# https://docs.github.com/github/administering-a-repository/configuration-options-for-dependency-updates
version: 2
updates:
- package-ecosystem: "pip" # See documentation for possible values
directory: "/docs/.sphinx" # Location of package manifests
open-pull-requests-limit: 10
schedule:
interval: "daily"
Dockerfile
View file @ a73ab0d8
FROM ubuntu:20.04 FROM ubuntu:20.04
ARG ROCMVERSION=5.3 ARG ROCMVERSION=5.6
ARG compiler_version="release" ARG compiler_version=""
ARG compiler_commit="" ARG compiler_commit=""
RUN set -xe RUN set -xe
ARG DEB_ROCM_REPO=http://repo.radeon.com/rocm/apt/.apt_$ROCMVERSION/ ARG DEB_ROCM_REPO=http://repo.radeon.com/rocm/apt/.apt_$ROCMVERSION/
RUN useradd -rm -d /home/jenkins -s /bin/bash -u 1004 jenkins RUN useradd -rm -d /home/jenkins -s /bin/bash -u 1004 jenkins
RUN useradd -rm -d /home/manitera -s /bin/bash -u 1002 manitera
# Add rocm repository # Add rocm repository
RUN apt-get update RUN apt-get update
RUN apt-get install -y wget gnupg RUN apt-get install -y wget gnupg curl
RUN wget -qO - http://repo.radeon.com/rocm/rocm.gpg.key | apt-key add - RUN --mount=type=ssh if [ "$ROCMVERSION" != "5.6"]; then \
RUN sh -c "echo deb [arch=amd64] $DEB_ROCM_REPO ubuntu main > /etc/apt/sources.list.d/rocm.list" wget -qO - http://repo.radeon.com/rocm/rocm.gpg.key | apt-key add - && \
sh -c "echo deb [arch=amd64] $DEB_ROCM_REPO ubuntu main > /etc/apt/sources.list.d/rocm.list"; \
else sh -c "wget http://artifactory-cdn.amd.com/artifactory/list/amdgpu-deb/amd-nonfree-radeon_20.04-1_all.deb" && \
apt update && apt-get install -y ./amd-nonfree-radeon_20.04-1_all.deb && \
amdgpu-repo --amdgpu-build=1567752 --rocm-build=compute-rocm-dkms-no-npi-hipclang/11914 && \
DEBIAN_FRONTEND=noninteractive amdgpu-install -y --usecase=rocm ; \
fi
RUN wget --no-check-certificate -qO - https://apt.kitware.com/keys/kitware-archive-latest.asc 2>/dev/null | apt-key add - RUN wget --no-check-certificate -qO - https://apt.kitware.com/keys/kitware-archive-latest.asc 2>/dev/null | apt-key add -
RUN sh -c "echo deb http://mirrors.kernel.org/ubuntu focal main universe | tee -a /etc/apt/sources.list" RUN sh -c "echo deb http://mirrors.kernel.org/ubuntu focal main universe | tee -a /etc/apt/sources.list"
RUN curl -fsSL https://repo.radeon.com/rocm/rocm.gpg.key | gpg --dearmor -o /etc/apt/trusted.gpg.d/rocm-keyring.gpg
# Install dependencies # Install dependencies
RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \
apt-utils \ apt-utils \
build-essential \ build-essential \
ccache \ ccache \
cmake-data \
cmake \ cmake \
curl \
git \ git \
hip-rocclr \ hip-rocclr \
jq \ jq \
...@@ -45,6 +49,7 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow- ...@@ -45,6 +49,7 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-
rocm-device-libs \ rocm-device-libs \
rocm-cmake \ rocm-cmake \
vim \ vim \
nano \
zlib1g-dev \ zlib1g-dev \
openssh-server \ openssh-server \
clang-format-10 \ clang-format-10 \
...@@ -52,6 +57,17 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow- ...@@ -52,6 +57,17 @@ RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-
apt-get clean && \ apt-get clean && \
rm -rf /var/lib/apt/lists/* rm -rf /var/lib/apt/lists/*
#Install latest version of cmake
RUN apt purge --auto-remove -y cmake
RUN apt update
RUN apt install -y software-properties-common lsb-release
RUN apt clean all
RUN wget -O - https://apt.kitware.com/keys/kitware-archive-latest.asc 2>/dev/null | gpg --dearmor - | tee /etc/apt/trusted.gpg.d/kitware.gpg >/dev/null
RUN apt-add-repository "deb https://apt.kitware.com/ubuntu/ $(lsb_release -cs) main"
RUN apt install -y kitware-archive-keyring
RUN rm /etc/apt/trusted.gpg.d/kitware.gpg
RUN apt install -y cmake
# Setup ubsan environment to printstacktrace # Setup ubsan environment to printstacktrace
RUN ln -s /usr/bin/llvm-symbolizer-3.8 /usr/local/bin/llvm-symbolizer RUN ln -s /usr/bin/llvm-symbolizer-3.8 /usr/local/bin/llvm-symbolizer
ENV UBSAN_OPTIONS=print_stacktrace=1 ENV UBSAN_OPTIONS=print_stacktrace=1
...@@ -87,12 +103,7 @@ ENV compiler_commit=$compiler_commit ...@@ -87,12 +103,7 @@ ENV compiler_commit=$compiler_commit
RUN sh -c "echo compiler version = '$compiler_version'" RUN sh -c "echo compiler version = '$compiler_version'"
RUN sh -c "echo compiler commit = '$compiler_commit'" RUN sh -c "echo compiler commit = '$compiler_commit'"
RUN --mount=type=ssh if [ "$compiler_version" = "amd-stg-open" ]; then \ RUN --mount=type=ssh if [ "$compiler_version" = "amd-stg-open" ] && [ "$compiler_commit" = "" ]; then \
sed -i '/$HIP_CLANG_TARGET = chomp($HIP_CLANG_TARGET);/c\ chomp($HIP_CLANG_TARGET);' /opt/rocm/hip/bin/hipcc.pl && \
sed -i '/$HIP_CLANG_TARGET = chomp($HIP_CLANG_TARGET);/c\ chomp($HIP_CLANG_TARGET);' /opt/rocm/bin/hipcc.pl; \
fi
RUN --mount=type=ssh if [ "$compiler_version" != "release" ] && [ "$compiler_commit" = "" ]; then \
git clone -b "$compiler_version" https://github.com/RadeonOpenCompute/llvm-project.git && \ git clone -b "$compiler_version" https://github.com/RadeonOpenCompute/llvm-project.git && \
cd llvm-project && mkdir build && cd build && \ cd llvm-project && mkdir build && cd build && \
cmake -DCMAKE_INSTALL_PREFIX=/opt/rocm/llvm -DCMAKE_BUILD_TYPE=Release -DLLVM_ENABLE_ASSERTIONS=1 -DLLVM_TARGETS_TO_BUILD="AMDGPU;X86" -DLLVM_ENABLE_PROJECTS="clang;lld;compiler-rt" ../llvm && \ cmake -DCMAKE_INSTALL_PREFIX=/opt/rocm/llvm -DCMAKE_BUILD_TYPE=Release -DLLVM_ENABLE_ASSERTIONS=1 -DLLVM_TARGETS_TO_BUILD="AMDGPU;X86" -DLLVM_ENABLE_PROJECTS="clang;lld;compiler-rt" ../llvm && \
...@@ -100,7 +111,7 @@ RUN --mount=type=ssh if [ "$compiler_version" != "release" ] && [ "$compiler_com ...@@ -100,7 +111,7 @@ RUN --mount=type=ssh if [ "$compiler_version" != "release" ] && [ "$compiler_com
else echo "using the release compiler"; \ else echo "using the release compiler"; \
fi fi
RUN --mount=type=ssh if [ "$compiler_version" != "release" ] && [ "$compiler_commit" != "" ]; then \ RUN --mount=type=ssh if [ "$compiler_version" = "amd-stg-open" ] && [ "$compiler_commit" != "" ]; then \
git clone -b "$compiler_version" https://github.com/RadeonOpenCompute/llvm-project.git && \ git clone -b "$compiler_version" https://github.com/RadeonOpenCompute/llvm-project.git && \
cd llvm-project && git checkout "$compiler_commit" && echo "checking out commit $compiler_commit" && mkdir build && cd build && \ cd llvm-project && git checkout "$compiler_commit" && echo "checking out commit $compiler_commit" && mkdir build && cd build && \
cmake -DCMAKE_INSTALL_PREFIX=/opt/rocm/llvm -DCMAKE_BUILD_TYPE=Release -DLLVM_ENABLE_ASSERTIONS=1 -DLLVM_TARGETS_TO_BUILD="AMDGPU;X86" -DLLVM_ENABLE_PROJECTS="clang;lld;compiler-rt" ../llvm && \ cmake -DCMAKE_INSTALL_PREFIX=/opt/rocm/llvm -DCMAKE_BUILD_TYPE=Release -DLLVM_ENABLE_ASSERTIONS=1 -DLLVM_TARGETS_TO_BUILD="AMDGPU;X86" -DLLVM_ENABLE_PROJECTS="clang;lld;compiler-rt" ../llvm && \
......
Jenkinsfile
View file @ a73ab0d8
...@@ -19,6 +19,11 @@ def runShell(String command){ ...@@ -19,6 +19,11 @@ def runShell(String command){
def getDockerImageName(){ def getDockerImageName(){
def img def img
if (params.ROCMVERSION != "5.5" && params.ROCMVERSION != "5.6"){
if (params.COMPILER_VERSION == "") {
img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}"
}
else{
if (params.COMPILER_COMMIT == ""){ if (params.COMPILER_COMMIT == ""){
img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}" img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}"
} }
...@@ -26,6 +31,22 @@ def getDockerImageName(){ ...@@ -26,6 +31,22 @@ def getDockerImageName(){
def commit = "${params.COMPILER_COMMIT}"[0..6] def commit = "${params.COMPILER_COMMIT}"[0..6]
img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}_${commit}" img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}_${commit}"
} }
}
}
else{
if (params.COMPILER_VERSION == "") {
img = "${env.CK_DOCKERHUB_PRIVATE}:ck_ub20.04_rocm${params.ROCMVERSION}"
}
else{
if (params.COMPILER_COMMIT == ""){
img = "${env.CK_DOCKERHUB_PRIVATE}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}"
}
else{
def commit = "${params.COMPILER_COMMIT}"[0..6]
img = "${env.CK_DOCKERHUB_PRIVATE}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}_${commit}"
}
}
}
return img return img
} }
...@@ -49,11 +70,11 @@ def build_compiler(){ ...@@ -49,11 +70,11 @@ def build_compiler(){
compiler = '/opt/rocm/bin/hipcc' compiler = '/opt/rocm/bin/hipcc'
} }
else{ else{
if (params.COMPILER_VERSION == "release"){ if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
compiler = "/opt/rocm/llvm/bin/clang++" compiler = "/llvm-project/build/bin/clang++"
} }
else{ else{
compiler = "/llvm-project/build/bin/clang++" compiler = "/opt/rocm/llvm/bin/clang++"
} }
} }
return compiler return compiler
...@@ -232,7 +253,7 @@ def buildHipClangJob(Map conf=[:]){ ...@@ -232,7 +253,7 @@ def buildHipClangJob(Map conf=[:]){
dockerOpts = dockerOpts + " --env HSA_XNACK=1 " dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
} }
def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' " def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if (params.COMPILER_VERSION != "release"){ if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' " dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
} }
...@@ -287,7 +308,7 @@ def runCKProfiler(Map conf=[:]){ ...@@ -287,7 +308,7 @@ def runCKProfiler(Map conf=[:]){
dockerOpts = dockerOpts + " --env HSA_XNACK=1 " dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
} }
def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' " def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if (params.COMPILER_VERSION != "release"){ if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' " dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
} }
...@@ -420,7 +441,7 @@ def Build_CK(Map conf=[:]){ ...@@ -420,7 +441,7 @@ def Build_CK(Map conf=[:]){
dockerOpts = dockerOpts + " --env HSA_XNACK=1 " dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
} }
def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' " def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
if (params.COMPILER_VERSION != "release"){ if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' " dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
} }
...@@ -586,16 +607,16 @@ pipeline { ...@@ -586,16 +607,16 @@ pipeline {
description: "Force building docker image (default: false), set to true if docker image needs to be updated.") description: "Force building docker image (default: false), set to true if docker image needs to be updated.")
string( string(
name: 'ROCMVERSION', name: 'ROCMVERSION',
defaultValue: '5.4.3', defaultValue: '5.6',
description: 'Specify which ROCM version to use: 5.4.3 (default).') description: 'Specify which ROCM version to use: 5.6 (default).')
string( string(
name: 'COMPILER_VERSION', name: 'COMPILER_VERSION',
defaultValue: 'amd-stg-open', defaultValue: '',
description: 'Specify which version of compiler to use: ck-9110, release, or amd-stg-open (default).') description: 'Specify which version of compiler to use: release, amd-stg-open, or leave blank (default).')
string( string(
name: 'COMPILER_COMMIT', name: 'COMPILER_COMMIT',
defaultValue: '5541927df00eabd6a110180170eca7785d436ee3', defaultValue: '',
description: 'Specify which commit of compiler branch to use: leave empty to use the latest commit, or use 5541927df00eabd6a110180170eca7785d436ee3 (default) commit of amd-stg-open branch.') description: 'Specify which commit of compiler branch to use: leave blank to use the latest commit, or use 5541927df00eabd6a110180170eca7785d436ee3 (default) commit of amd-stg-open branch.')
string( string(
name: 'BUILD_COMPILER', name: 'BUILD_COMPILER',
defaultValue: 'hipcc', defaultValue: 'hipcc',
......
client_example/09_quantization/conv2d_fwd_bias_relu_perchannel_quantization.cpp
View file @ a73ab0d8
...@@ -73,7 +73,7 @@ int main(int argc, char* argv[]) ...@@ -73,7 +73,7 @@ int main(int argc, char* argv[])
SimpleDeviceMem in(sizeof(InDataType) * N * Hi * Wi * C); SimpleDeviceMem in(sizeof(InDataType) * N * Hi * Wi * C);
SimpleDeviceMem wei(sizeof(WeiDataType) * K * Y * X * C); SimpleDeviceMem wei(sizeof(WeiDataType) * K * Y * X * C);
SimpleDeviceMem bias(sizeof(BiasDataType) * K * Y * X * C); SimpleDeviceMem bias(sizeof(BiasDataType) * K * Y * X * C);
SimpleDeviceMem requant_scale(sizeof(RequantScaleDataType) * K * Y * X * C); SimpleDeviceMem requant_scale(sizeof(RequantScaleDataType) * K);
SimpleDeviceMem out(sizeof(OutDataType) * N * Ho * Wo * K); SimpleDeviceMem out(sizeof(OutDataType) * N * Ho * Wo * K);
using DeviceOp = ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD< using DeviceOp = ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD<
......
client_example/09_quantization/conv2d_fwd_bias_tanh_perchannel_quantization.cpp
View file @ a73ab0d8
...@@ -76,7 +76,7 @@ int main(int argc, char* argv[]) ...@@ -76,7 +76,7 @@ int main(int argc, char* argv[])
SimpleDeviceMem in(sizeof(InDataType) * N * Hi * Wi * C); SimpleDeviceMem in(sizeof(InDataType) * N * Hi * Wi * C);
SimpleDeviceMem wei(sizeof(WeiDataType) * K * Y * X * C); SimpleDeviceMem wei(sizeof(WeiDataType) * K * Y * X * C);
SimpleDeviceMem bias(sizeof(BiasDataType) * K * Y * X * C); SimpleDeviceMem bias(sizeof(BiasDataType) * K * Y * X * C);
SimpleDeviceMem requant_scale(sizeof(RequantScaleDataType) * K * Y * X * C); SimpleDeviceMem requant_scale(sizeof(RequantScaleDataType) * K);
SimpleDeviceMem out(sizeof(OutDataType) * N * Ho * Wo * K); SimpleDeviceMem out(sizeof(OutDataType) * N * Ho * Wo * K);
using DeviceOp = ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD< using DeviceOp = ck::tensor_operation::device::DeviceGroupedConvFwdMultipleD<
......
client_example/09_quantization/conv2d_fwd_perchannel_quantization.cpp
View file @ a73ab0d8
...@@ -69,7 +69,7 @@ int main(int argc, char* argv[]) ...@@ -69,7 +69,7 @@ int main(int argc, char* argv[])
SimpleDeviceMem in(sizeof(InDataType) * N * Hi * Wi * C); SimpleDeviceMem in(sizeof(InDataType) * N * Hi * Wi * C);
SimpleDeviceMem wei(sizeof(WeiDataType) * K * Y * X * C); SimpleDeviceMem wei(sizeof(WeiDataType) * K * Y * X * C);
SimpleDeviceMem requant_scale(sizeof(RequantScaleDataType) * K * Y * X * C); SimpleDeviceMem requant_scale(sizeof(RequantScaleDataType) * K);
SimpleDeviceMem out(sizeof(OutDataType) * N * Ho * Wo * K); SimpleDeviceMem out(sizeof(OutDataType) * N * Ho * Wo * K);
using DeviceOp = using DeviceOp =
......
client_example/18_groupnorm/CMakeLists.txt 0 → 100644
View file @ a73ab0d8
add_executable(client_groupnorm_swish groupnorm_swish.cpp)
target_link_libraries(client_groupnorm_swish PRIVATE composable_kernel::device_operations)
client_example/18_groupnorm/groupnorm_swish.cpp 0 → 100644
View file @ a73ab0d8
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include <iomanip>
#include <vector>
#include <iostream>
#include "ck/ck.hpp"
#include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
#include "ck/tensor_operation/gpu/device/device_normalization.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
#include "ck/library/tensor_operation_instance/gpu/normalization_swish.hpp"
using XDataType = ck::half_t;
using GammaDataType = float;
using BetaDataType = float;
using YDataType = ck::half_t;
using ComputeDataType = float;
using Swish = ck::tensor_operation::element_wise::Swish;
constexpr int Rank = 5;
constexpr int NumReduceDim = 3;
struct SimpleDeviceMem
{
SimpleDeviceMem() = delete;
SimpleDeviceMem(std::size_t mem_size) : p_mem_{}
{
(void)hipMalloc(static_cast<void**>(&p_mem_), mem_size);
}
void* GetDeviceBuffer() { return p_mem_; }
~SimpleDeviceMem() { (void)hipFree(p_mem_); }
void* p_mem_;
};
int main(int argc, char* argv[])
{
ck::index_t N = 32;
ck::index_t H = 16;
ck::index_t W = 16;
ck::index_t G = 64;
ck::index_t C = 128;
std::size_t xy_size = N * H * W * G * C;
std::size_t gamma_beta_size = G * C;
std::vector<ck::index_t> xy_strides = {H * W * G * C, W * G * C, G * C, C, 1};
std::vector<ck::index_t> gamma_beta_strides = {0, 0, 0, C, 1};
SimpleDeviceMem x_device_buf(sizeof(XDataType) * xy_size);
SimpleDeviceMem gamma_device_buf(sizeof(GammaDataType) * gamma_beta_size);
SimpleDeviceMem beta_device_buf(sizeof(BetaDataType) * gamma_beta_size);
SimpleDeviceMem y_device_buf(sizeof(YDataType) * xy_size);
using DeviceOp = ck::tensor_operation::device::DeviceNormalization<XDataType,
GammaDataType,
BetaDataType,
ComputeDataType,
YDataType,
Swish,
Rank,
NumReduceDim>;
// get device op instances
const auto op_ptrs = ck::tensor_operation::device::instance::DeviceOperationInstanceFactory<
DeviceOp>::GetInstances();
std::cout << "found " << op_ptrs.size() << " instances" << std::endl;
std::string best_op_name;
bool found = false;
int best_op_id = -1;
float best_ave_time = std::numeric_limits<float>::max();
float best_gb_per_sec = 0;
// profile device operation instances
std::cout << "Run all instances and do timing" << std::endl;
for(int i = 0; i < op_ptrs.size(); ++i)
{
auto& op_ptr = op_ptrs[i];
auto argument_ptr = op_ptr->MakeArgumentPointer({N, H, W, G, C}, // lengths
xy_strides, // xStrides
gamma_beta_strides, // gammaStrides
gamma_beta_strides, // betaStrides
xy_strides, // yStrides
{1, 2, 4}, // reduceDims
1e-6,
x_device_buf.GetDeviceBuffer(),
gamma_device_buf.GetDeviceBuffer(),
beta_device_buf.GetDeviceBuffer(),
y_device_buf.GetDeviceBuffer(),
nullptr,
nullptr,
Swish{});
auto invoker_ptr = op_ptr->MakeInvokerPointer();
std::string op_name = op_ptr->GetTypeString();
if(op_ptr->IsSupportedArgument(argument_ptr.get()))
{
float ave_time = invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, true});
std::size_t num_byte =
sizeof(XDataType) * xy_size + sizeof(GammaDataType) * gamma_beta_size +
sizeof(BetaDataType) * gamma_beta_size + sizeof(YDataType) * xy_size;
float gb_per_sec = num_byte / 1.E6 / ave_time;
std::cout << "Perf: " << std::setw(10) << ave_time << " ms, " << gb_per_sec << " GB/s, "
<< op_name << std::endl;
if(ave_time < best_ave_time)
{
found = true;
best_op_id = i;
best_op_name = op_name;
best_ave_time = ave_time;
best_gb_per_sec = gb_per_sec;
}
}
else
{
std::cout << op_name << " does not support this problem" << std::endl;
}
}
std::cout << "Best Perf: " << best_ave_time << " ms, " << best_gb_per_sec << " GB/s, "
<< best_op_name << std::endl;
// run the best intance
{
auto& op_ptr = op_ptrs[best_op_id];
std::cout << "Run the best instance without timing: " << op_ptr->GetTypeString()
<< std::endl;
auto argument_ptr = op_ptr->MakeArgumentPointer({N, H, W, G, C}, // lengths
xy_strides, // xStrides
gamma_beta_strides, // gammaStrides
gamma_beta_strides, // betaStrides
xy_strides, // yStrides
{1, 2, 4}, // reduceDims
1e-6,
x_device_buf.GetDeviceBuffer(),
gamma_device_buf.GetDeviceBuffer(),
beta_device_buf.GetDeviceBuffer(),
y_device_buf.GetDeviceBuffer(),
nullptr,
nullptr,
Swish{});
auto invoker_ptr = op_ptr->MakeInvokerPointer();
if(op_ptr->IsSupportedArgument(argument_ptr.get()))
{
invoker_ptr->Run(argument_ptr.get(), StreamConfig{nullptr, false});
}
std::cout << "Done" << std::endl;
}
return 0;
}
cmake/googletest.cmake
View file @ a73ab0d8
...@@ -21,6 +21,7 @@ list(APPEND GTEST_CMAKE_CXX_FLAGS ...@@ -21,6 +21,7 @@ list(APPEND GTEST_CMAKE_CXX_FLAGS
-Wno-comma -Wno-comma
-Wno-old-style-cast -Wno-old-style-cast
-Wno-deprecated -Wno-deprecated
-Wno-unsafe-buffer-usage
) )
message(STATUS "Suppressing googltest warnings with flags: ${GTEST_CMAKE_CXX_FLAGS}") message(STATUS "Suppressing googltest warnings with flags: ${GTEST_CMAKE_CXX_FLAGS}")
......
docs/.sphinx/requirements.in
View file @ a73ab0d8
git+https://github.com/RadeonOpenCompute/rocm-docs-core.git rocm-docs-core==0.2.0
sphinxcontrib-bibtex==2.5.0 sphinxcontrib-bibtex==2.5.0
docs/.sphinx/requirements.txt
View file @ a73ab0d8
...@@ -2,9 +2,9 @@ ...@@ -2,9 +2,9 @@
# This file is autogenerated by pip-compile with Python 3.10 # This file is autogenerated by pip-compile with Python 3.10
# by the following command: # by the following command:
# #
# pip-compile requirements.in # pip-compile .sphinx/requirements.in
# #
accessible-pygments==0.0.4 accessible-pygments==0.0.3
# via pydata-sphinx-theme # via pydata-sphinx-theme
alabaster==0.7.13 alabaster==0.7.13
# via sphinx # via sphinx
...@@ -20,7 +20,7 @@ babel==2.12.1 ...@@ -20,7 +20,7 @@ babel==2.12.1
# sphinx # sphinx
backcall==0.2.0 backcall==0.2.0
# via ipython # via ipython
beautifulsoup4==4.12.0 beautifulsoup4==4.11.2
# via pydata-sphinx-theme # via pydata-sphinx-theme
breathe==4.34.0 breathe==4.34.0
# via rocm-docs-core # via rocm-docs-core
...@@ -34,7 +34,7 @@ click==8.1.3 ...@@ -34,7 +34,7 @@ click==8.1.3
# via # via
# jupyter-cache # jupyter-cache
# sphinx-external-toc # sphinx-external-toc
comm==0.1.3 comm==0.1.2
# via ipykernel # via ipykernel
debugpy==1.6.6 debugpy==1.6.6
# via ipykernel # via ipykernel
...@@ -65,13 +65,11 @@ idna==3.4 ...@@ -65,13 +65,11 @@ idna==3.4
# via requests # via requests
imagesize==1.4.1 imagesize==1.4.1
# via sphinx # via sphinx
importlib-metadata==6.1.0 importlib-metadata==6.0.0
# via # via
# jupyter-cache # jupyter-cache
# myst-nb # myst-nb
importlib-resources==5.10.4 ipykernel==6.21.3
# via rocm-docs-core
ipykernel==6.22.0
# via myst-nb # via myst-nb
ipython==8.11.0 ipython==8.11.0
# via # via
...@@ -87,7 +85,7 @@ jsonschema==4.17.3 ...@@ -87,7 +85,7 @@ jsonschema==4.17.3
# via nbformat # via nbformat
jupyter-cache==0.5.0 jupyter-cache==0.5.0
# via myst-nb # via myst-nb
jupyter-client==8.1.0 jupyter-client==8.0.3
# via # via
# ipykernel # ipykernel
# nbclient # nbclient
...@@ -124,7 +122,7 @@ nbclient==0.5.13 ...@@ -124,7 +122,7 @@ nbclient==0.5.13
# via # via
# jupyter-cache # jupyter-cache
# myst-nb # myst-nb
nbformat==5.8.0 nbformat==5.7.3
# via # via
# jupyter-cache # jupyter-cache
# myst-nb # myst-nb
...@@ -187,7 +185,7 @@ pyyaml==6.0 ...@@ -187,7 +185,7 @@ pyyaml==6.0
# myst-parser # myst-parser
# pybtex # pybtex
# sphinx-external-toc # sphinx-external-toc
pyzmq==25.0.2 pyzmq==25.0.1
# via # via
# ipykernel # ipykernel
# jupyter-client # jupyter-client
...@@ -195,8 +193,8 @@ requests==2.28.2 ...@@ -195,8 +193,8 @@ requests==2.28.2
# via # via
# pygithub # pygithub
# sphinx # sphinx
rocm-docs-core @ git+https://github.com/RadeonOpenCompute/rocm-docs-core.git rocm-docs-core==0.2.0
# via -r requirements.in # via -r .sphinx/requirements.in
six==1.16.0 six==1.16.0
# via # via
# asttokens # asttokens
...@@ -235,9 +233,7 @@ sphinx-notfound-page==0.8.3 ...@@ -235,9 +233,7 @@ sphinx-notfound-page==0.8.3
sphinxcontrib-applehelp==1.0.4 sphinxcontrib-applehelp==1.0.4
# via sphinx # via sphinx
sphinxcontrib-bibtex==2.5.0 sphinxcontrib-bibtex==2.5.0
# via # via -r .sphinx/requirements.in
# -r requirements.in
# rocm-docs-core
sphinxcontrib-devhelp==1.0.2 sphinxcontrib-devhelp==1.0.2
# via sphinx # via sphinx
sphinxcontrib-htmlhelp==2.0.1 sphinxcontrib-htmlhelp==2.0.1
...@@ -248,7 +244,7 @@ sphinxcontrib-qthelp==1.0.3 ...@@ -248,7 +244,7 @@ sphinxcontrib-qthelp==1.0.3
# via sphinx # via sphinx
sphinxcontrib-serializinghtml==1.1.5 sphinxcontrib-serializinghtml==1.1.5
# via sphinx # via sphinx
sqlalchemy==1.4.47 sqlalchemy==1.4.46
# via jupyter-cache # via jupyter-cache
stack-data==0.6.2 stack-data==0.6.2
# via ipython # via ipython
......
example/42_groupnorm/CMakeLists.txt
View file @ a73ab0d8
add_example_executable(example_groupnorm_sigmoid_fp16 groupnorm_sigmoid_fp16.cpp) add_example_executable(example_groupnorm_sigmoid_mul_fp16 groupnorm_sigmoid_mul_fp16.cpp)
add_example_executable(example_groupnorm_swish_fp16 groupnorm_swish_fp16.cpp)
example/42_groupnorm/common.hpp 0 → 100644
View file @ a73ab0d8
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
#include <iostream>
#include <numeric>
#include <initializer_list>
#include <cstdlib>
#include <getopt.h>
#include "ck/ck.hpp"
#include "ck/utility/reduction_enums.hpp"
#include "ck/tensor_operation/gpu/device/impl/device_normalization_impl.hpp"
#include "ck/tensor_operation/gpu/device/reduction_operator_mapping.hpp"
#include "ck/library/utility/fill.hpp"
#include "ck/library/utility/check_err.hpp"
#include "ck/library/utility/device_memory.hpp"
#include "ck/library/utility/host_common_util.hpp"
#include "ck/library/utility/host_tensor.hpp"
#include "ck/library/utility/host_tensor_generator.hpp"
#include "ck/library/reference_tensor_operation/cpu/reference_groupnorm.hpp"
example/42_groupnorm/groupnorm_sigmoid_mul_fp16.cpp 0 → 100644
View file @ a73ab0d8
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include "common.hpp"
constexpr int Rank = 5;
constexpr int NumReduceDim = 3;
using XDataType = ck::half_t;
using GammaDataType = ck::half_t;
using BetaDataType = ck::half_t;
using YDataType = ck::half_t;
using ComputeDataType = float;
struct YElementOp
{
template <typename T>
__host__ __device__ void operator()(T& y, const T& x) const
{
static_assert(ck::is_same<T, float>::value || ck::is_same<T, double>::value ||
ck::is_same<T, ck::half_t>::value,
"Data type is not supported by this operation!");
T a;
ck::tensor_operation::element_wise::Sigmoid{}(a, x);
y = x * a;
};
};
using DeviceInstance =
ck::tensor_operation::device::DeviceNormalizationImpl<XDataType,
GammaDataType,
BetaDataType,
ComputeDataType,
YDataType,
YElementOp,
Rank,
NumReduceDim,
1024, // BlockSize
1, // ClusterM
1024, // ClusterK
1, // SliceM
32, // SliceK
1, // SrcVecDim (0=M, 1=K)
2, // SrcScalarPerVector
1, // GammaVecDim (0=M, 1=K)
2, // GammaScalarPerVector
1, // BetaVecDim (0=M, 1=K)
2, // BetaScalarPerVector
2>; // OutScalarPerVector
#include "run_groupnorm_example.inc"
int main(int argc, char* argv[]) { run_groupnorm_example(argc, argv); }
example/42_groupnorm/groupnorm_swish_fp16.cpp 0 → 100644
View file @ a73ab0d8
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include "common.hpp"
constexpr int Rank = 5;
constexpr int NumReduceDim = 3;
using XDataType = ck::half_t;
using GammaDataType = ck::half_t;
using BetaDataType = ck::half_t;
using YDataType = ck::half_t;
using ComputeDataType = float;
using YElementOp = ck::tensor_operation::element_wise::Swish;
using DeviceInstance =
ck::tensor_operation::device::DeviceNormalizationImpl<XDataType,
GammaDataType,
BetaDataType,
ComputeDataType,
YDataType,
YElementOp,
Rank,
NumReduceDim,
1024, // BlockSize
1, // ClusterM
1024, // ClusterK
1, // SliceM
32, // SliceK
1, // SrcVecDim (0=M, 1=K)
2, // SrcScalarPerVector
1, // GammaVecDim (0=M, 1=K)
2, // GammaScalarPerVector
1, // BetaVecDim (0=M, 1=K)
2, // BetaScalarPerVector
2>; // OutScalarPerVector
#include "run_groupnorm_example.inc"
int main(int argc, char* argv[]) { run_groupnorm_example(argc, argv); }
example/42_groupnorm/groupnorm_sigmoid_fp16.cpp → example/42_groupnorm/run_groupnorm_example.inc
View file @ a73ab0d8
// SPDX-License-Identifier: MIT // SPDX-License-Identifier: MIT
// Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved. // Copyright (c) 2018-2022, Advanced Micro Devices, Inc. All rights reserved.
#include <iostream> #pragma once
#include <numeric>
#include <initializer_list>
#include <cstdlib>
#include <getopt.h>
#include "ck/ck.hpp"
#include "ck/utility/reduction_enums.hpp"
#include "ck/tensor_operation/gpu/device/impl/device_normalization_impl.hpp"
#include "ck/tensor_operation/gpu/device/reduction_operator_mapping.hpp"
#include "ck/library/utility/fill.hpp"
#include "ck/library/utility/check_err.hpp"
#include "ck/library/utility/device_memory.hpp"
#include "ck/library/utility/host_common_util.hpp"
#include "ck/library/utility/host_tensor.hpp"
#include "ck/library/utility/host_tensor_generator.hpp"
#include "ck/library/reference_tensor_operation/cpu/reference_groupnorm.hpp"
constexpr int Rank = 5;
constexpr int NumReduceDim = 3;
using XDataType = ck::half_t;
using GammaDataType = ck::half_t;
using BetaDataType = ck::half_t;
using YDataType = ck::half_t;
using ComputeDataType = float;
struct YElementOp
{
template <typename T>
__host__ __device__ void operator()(T& y, const T& x) const
{
static_assert(ck::is_same<T, float>::value || ck::is_same<T, double>::value ||
ck::is_same<T, ck::half_t>::value,
"Data type is not supported by this operation!");
T a;
ck::tensor_operation::element_wise::Sigmoid{}(a, x);
y = x * a;
};
};
using DeviceInstance = int run_groupnorm_example(int argc, char* argv[])
ck::tensor_operation::device::DeviceNormalizationImpl<XDataType,
GammaDataType,
BetaDataType,
ComputeDataType,
YDataType,
YElementOp,
Rank,
NumReduceDim,
1024, // BlockSize
1, // ClusterM
1024, // ClusterK
1, // SliceM
32, // SliceK
1, // SrcVecDim (0=M, 1=K)
2, // SrcScalarPerVector
1, // GammaVecDim (0=M, 1=K)
2, // GammaScalarPerVector
1, // BetaVecDim (0=M, 1=K)
2, // BetaScalarPerVector
2>; // OutScalarPerVector
int main(int argc, char* argv[])
{ {
ck::index_t N = 2; ck::index_t N = 32;
ck::index_t H = 32; ck::index_t H = 16;
ck::index_t W = 32; ck::index_t W = 16;
ck::index_t G = 32; ck::index_t G = 64;
ck::index_t C = 30; ck::index_t C = 128;
if(argc == 1) if(argc == 1)
{ {
......
include/ck/ck.hpp
View file @ a73ab0d8
...@@ -172,6 +172,11 @@ ...@@ -172,6 +172,11 @@
// flag to enable (1) or disable (0) the debugging output in some kernels // flag to enable (1) or disable (0) the debugging output in some kernels
#define DEBUG_LOG 0 #define DEBUG_LOG 0
// denorm test fix, required to work around dissue
#ifndef CK_WORKAROUND_DENORM_FIX
#define CK_WORKAROUND_DENORM_FIX 0
#endif
namespace ck { namespace ck {
enum struct InMemoryDataOperationEnum enum struct InMemoryDataOperationEnum
......
include/ck/tensor_operation/gpu/element/unary_element_wise_operation.hpp
View file @ a73ab0d8
...@@ -316,8 +316,6 @@ struct Sigmoid ...@@ -316,8 +316,6 @@ struct Sigmoid
y = 1 / (ck::type_convert<T>(1) + exp(-x)); y = 1 / (ck::type_convert<T>(1) + exp(-x));
}; };
int32_t divider_ = 1;
}; };
struct TanH struct TanH
...@@ -333,6 +331,23 @@ struct TanH ...@@ -333,6 +331,23 @@ struct TanH
}; };
}; };
struct Swish
{
Swish(float beta = 1.0f) : beta_(beta) {}
template <typename T>
__host__ __device__ void operator()(T& y, const T& x) const
{
static_assert(is_same<T, float>::value || is_same<T, double>::value ||
is_same<T, ck::half_t>::value,
"Data type is not supported by this operation!");
y = x / (ck::type_convert<T>(1) + ck::math::exp(-beta_ * x));
};
float beta_ = 1.0f;
};
} // namespace element_wise } // namespace element_wise
} // namespace tensor_operation } // namespace tensor_operation
} // namespace ck } // namespace ck
include/ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_wmma_cshuffle.hpp
View file @ a73ab0d8
...@@ -505,6 +505,15 @@ struct GridwiseGemmMultipleD_k0mk1_k0nk1_mn_wmma_cshuffle ...@@ -505,6 +505,15 @@ struct GridwiseGemmMultipleD_k0mk1_k0nk1_mn_wmma_cshuffle
} }
// TODO: also check validity of all components (blockwise-copy, threadwise-copy, etc) // TODO: also check validity of all components (blockwise-copy, threadwise-copy, etc)
constexpr long_index_t TwoGB = (long_index_t{1} << 31);
if(!(a_grid_desc_k0_m_k1.GetElementSpaceSize() * sizeof(ADataType) <= TwoGB &&
b_grid_desc_k0_n_k1.GetElementSpaceSize() * sizeof(BDataType) <= TwoGB &&
e_grid_desc_m_n.GetElementSpaceSize() * sizeof(EDataType) <= TwoGB))
{
return false;
}
return true; return true;
} }
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment