Enable sccache in the default docker and CI. (ROCm#1009)

* replace ccache with sccache, pin package versions

* put ccache back temporarily to avoid breaking other CI jobs

* add sccashe_wrapper.sh script

* fix the package version syntax

* fix the pymysql package issue

* run sccache_wrapper before build if ccache server found

* set the paths before calling the sccache_wrapper

* use /tmp instead of /usr/local for cache

* try using sccache --start-server instead of wrapper

* try using redis server with sccache

* define SCCACHE_REDIS

* add redis and ping packages, and redis port

* use the new sccache redis server

* do not use sccache with staging compiler

* fix the condition syntax

* add stunnel to redis

* add tunnel verification

* separate caches for different architectures

* fix syntax for the cache tag

* quse double brackets for conditions

* add bash line to the script

* add a switch for sccache and only use it in build stage

* run check_host function when enabling sccache

* fix the invocation tags for sccache

* fix groovy syntax

* set the invocation tag in groovy

* disable sccache in clang-format stage

* try another syntax for invocation tags

* use local sccache server if can't connect to redis

* fix script syntax

* update README

* refresh readme

* readme updates

* remove the timing and verification caveat from readme

---------

Co-authored-by: Lisa Delaney <lisa.delaney@amd.com>

Author: illsilin

CMakeLists.txt

@@ -373,9 +373,10 @@ include_directories(BEFORE
 
 SET(BUILD_DEV ON CACHE BOOL "BUILD_DEV")
 if(BUILD_DEV)
-    add_compile_options(-Werror)
-    add_compile_options(-Weverything)
+    add_compile_options(-Werror -Weverything)
 endif()
+#add flags to reduce the size of binaries
+add_compile_options(-Oz -flto=thin)
 message("CMAKE_CXX_FLAGS: ${CMAKE_CXX_FLAGS}")
 
 add_custom_target(check COMMAND ${CMAKE_CTEST_COMMAND} --output-on-failure -C ${CMAKE_CFG_INTDIR})
@@ -390,35 +391,27 @@ IF(IS_DIRECTORY "${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu
     file(READ "${PROJECT_SOURCE_DIR}/library/src/tensor_operation_instance/gpu/${subdir_path}/CMakeLists.txt" cmake_instance)
     set(add_inst 0)
     if(("${cmake_instance}" MATCHES "fp8" OR "${cmake_instance}" MATCHES "_f8") AND DTYPES MATCHES "fp8")
-        #message("fp8 instance found!")
         set(add_inst 1)
     endif()
     if(("${cmake_instance}" MATCHES "bf8" OR "${cmake_instance}" MATCHES "_b8") AND DTYPES MATCHES "bf8")
-        #message("bf8 instance found!")
         set(add_inst 1)
     endif()
     if(("${cmake_instance}" MATCHES "fp16" OR "${cmake_instance}" MATCHES "_f16") AND DTYPES MATCHES "fp16")
-        #message("fp16 instance found!")
         set(add_inst 1)
     endif()
     if(("${cmake_instance}" MATCHES "fp32" OR "${cmake_instance}" MATCHES "_f32") AND DTYPES MATCHES "fp32")
-        #message("fp32 instance found!")
         set(add_inst 1)
     endif()
     if(("${cmake_instance}" MATCHES "fp64" OR "${cmake_instance}" MATCHES "_f64") AND DTYPES MATCHES "fp64")
-        #message("fp64 instance found!")
         set(add_inst 1)
     endif()
     if(("${cmake_instance}" MATCHES "bf16" OR "${cmake_instance}" MATCHES "_b16") AND DTYPES MATCHES "bf16")
-        #message("bf16 instance found!")
         set(add_inst 1)
     endif()
     if(("${cmake_instance}" MATCHES "int8" OR "${cmake_instance}" MATCHES "_i8") AND DTYPES MATCHES "int8")
-        #message("int8 instance found!")
         set(add_inst 1)
     endif()
     if(NOT "${cmake_instance}" MATCHES "DTYPES")
-        #message("instance should be built for all types!")
         set(add_inst 1)
     endif()
     if(add_inst EQUAL 1 OR NOT DEFINED DTYPES)

Dockerfile

@@ -26,25 +26,37 @@ RUN wget -qO - http://repo.radeon.com/rocm/rocm.gpg.key | apt-key add - && \
 RUN sh -c "echo deb http://mirrors.kernel.org/ubuntu focal main universe | tee -a /etc/apt/sources.list"
 RUN amdgpu-install -y --usecase=rocm --no-dkms
 
+## Sccache binary built from source for ROCm
+ARG SCCACHE_REPO_URL=http://compute-artifactory.amd.com/artifactory/rocm-generic-experimental/rocm-sccache
+ENV SCCACHE_INSTALL_LOCATION=/usr/local/.cargo/bin
+RUN mkdir -p ${SCCACHE_INSTALL_LOCATION} && \
+curl ${SCCACHE_REPO_URL}/portable/0.2.16/sccache-0.2.16-alpha.1-rocm --output ${SCCACHE_INSTALL_LOCATION}/sccache && \
+chmod +x ${SCCACHE_INSTALL_LOCATION}/sccache
+ENV PATH=$PATH:${SCCACHE_INSTALL_LOCATION}
+
 # Install dependencies
 RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --allow-unauthenticated \
     build-essential \
-    ccache \
     cmake \
+    ccache \
     git \
     hip-rocclr \
+    iputils-ping \
     jq \
     libelf-dev \
     libncurses5-dev \
     libnuma-dev \
     libpthread-stubs0-dev \
     llvm-amdgpu \
+    net-tools \
     pkg-config \
     python \
     python3 \
     python3-dev \
     python3-pip \
+    redis \
     sshpass \
+    stunnel \
     software-properties-common \
     vim \
     nano \
@@ -62,7 +74,7 @@ RUN gunzip /usr/local/bin/ninja.gz
 RUN chmod a+x /usr/local/bin/ninja
 RUN git clone https://github.com/nico/ninjatracing.git
 # Update the cmake to the latest version
-RUN pip install --upgrade cmake
+RUN pip install --upgrade cmake==3.27.5
 
 # Setup ubsan environment to printstacktrace
 RUN ln -s /usr/bin/llvm-symbolizer-3.8 /usr/local/bin/llvm-symbolizer
@@ -77,9 +89,9 @@ ARG PREFIX=/opt/rocm
 RUN pip3 install --upgrade pip
 RUN pip3 install sqlalchemy==1.4.46
 RUN pip3 install pymysql
-RUN pip3 install pandas
+RUN pip3 install pandas==2.0.3
 RUN pip3 install setuptools-rust
-RUN pip3 install sshtunnel
+RUN pip3 install sshtunnel==0.4.0
 # Setup ubsan environment to printstacktrace
 ENV UBSAN_OPTIONS=print_stacktrace=1
 
@@ -115,6 +127,8 @@ RUN if [ "$compiler_version" = "amd-stg-open" ] && [ "$compiler_commit" != "" ];
     else echo "using the release compiler"; \
     fi
 
+#clean-up the deb package
+RUN sh -c "rm -rf amdgpu-install*"
 
 #ENV HIP_CLANG_PATH='/llvm-project/build/bin'
 #RUN sh -c "echo HIP_CLANG_PATH = '$HIP_CLANG_PATH'"

Jenkinsfile

@@ -65,10 +65,10 @@ def getDockerImageName(){
 }
 
 def check_host() {
-    if ("${env.CK_CCACHE}" != "null"){
-        def CCACHE_SERVER="${env.CK_CCACHE.split(':')[0]}"
-        echo "ccache server: ${CCACHE_SERVER}"
-        sh '''ping -c 1 -p 6379 "${CCACHE_SERVER}" | echo $? > tmp.txt'''
+    if ("${env.CK_SCCACHE}" != "null"){
+        def SCCACHE_SERVER="${env.CK_SCCACHE.split(':')[0]}"
+        echo "sccache server: ${SCCACHE_SERVER}"
+        sh '''ping -c 1 -p 6379 "${SCCACHE_SERVER}" | echo $? > tmp.txt'''
         def output = readFile(file: "tmp.txt")
         echo "tmp.txt contents: \$output"
         return (output != "0")
@@ -96,24 +96,9 @@ def build_compiler(){
 
 def getDockerImage(Map conf=[:]){
     env.DOCKER_BUILDKIT=1
-    def prefixpath = conf.get("prefixpath", "/opt/rocm") // prefix:/opt/rocm
+    def prefixpath = conf.get("prefixpath", "/opt/rocm")
     def no_cache = conf.get("no_cache", false)
     def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
-    echo "ccache server: ${env.CK_CCACHE}"
-    if(env.CK_CCACHE)
-    {
-        if(check_host())
-        {
-            echo "FOUND CCACHE SERVER: ${env.CK_CCACHE}"
-        }
-        else 
-        {
-            echo "CCACHE SERVER: ${env.CK_CCACHE} NOT FOUND, got ${check_host} response"
-        }
-        dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CK_CCACHE}' --build-arg COMPILER_LAUNCHER='ccache' "
-        env.CCACHE_DIR = """/tmp/ccache_store"""
-        env.CCACHE_SECONDARY_STORAGE="""redis://${env.CK_CCACHE}"""
-    }
     if(no_cache)
     {
         dockerArgs = dockerArgs + " --no-cache "
@@ -142,21 +127,6 @@ def buildDocker(install_prefix){
     def image_name = getDockerImageName()
     echo "Building Docker for ${image_name}"
     def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${install_prefix} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg compiler_commit='${params.COMPILER_COMMIT}' --build-arg ROCMVERSION='${params.ROCMVERSION}' "
-    echo "ccache server: ${env.CK_CCACHE}"
-    if(env.CK_CCACHE)
-    {
-        if(check_host())
-        {
-            echo "FOUND CCACHE SERVER: ${env.CK_CCACHE}"
-        }
-        else 
-        {
-            echo "CCACHE SERVER: ${env.CK_CCACHE} NOT FOUND, got ${check_host} response"
-        }
-        dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CK_CCACHE}' --build-arg COMPILER_LAUNCHER='ccache' "
-        env.CCACHE_DIR = """/tmp/ccache_store"""
-        env.CCACHE_SECONDARY_STORAGE="""redis://${env.CK_CCACHE}"""
-    }
 
     echo "Build Args: ${dockerArgs}"
     try{
@@ -219,13 +189,9 @@ def cmake_build(Map conf=[:]){
     }else{
         setup_args = " -DCMAKE_BUILD_TYPE=release" + setup_args
     }
-    if(env.CK_CCACHE)
-    {
-        setup_args = " -DCMAKE_CXX_COMPILER_LAUNCHER='ccache' -DCMAKE_C_COMPILER_LAUNCHER='ccache' " + setup_args
-    }
-    echo "ccache server: ${env.CK_CCACHE}"
 
     def pre_setup_cmd = """
+            #!/bin/bash
             echo \$HSA_ENABLE_SDMA
             ulimit -c unlimited
             rm -rf build
@@ -234,6 +200,46 @@ def cmake_build(Map conf=[:]){
             mkdir install
             cd build
         """
+    def invocation_tag=""
+    if (setup_args.contains("gfx11")){
+        invocation_tag="gfx11"
+    }
+    if (setup_args.contains("gfx10")){
+        invocation_tag="gfx10"
+    }
+    if (setup_args.contains("gfx90")){
+        invocation_tag="gfx90"
+    }
+    if (setup_args.contains("gfx94")){
+        invocation_tag="gfx94"
+    }
+    if(check_host() && params.USE_SCCACHE && "${env.CK_SCCACHE}" != "null" && "${invocation_tag}" != "") {
+        pre_setup_cmd = pre_setup_cmd + """
+            #!/bin/bash
+            export ROCM_PATH=/opt/rocm
+            export SCCACHE_ENABLED=true
+            export SCCACHE_LOG_LEVEL=debug
+            export SCCACHE_IDLE_TIMEOUT=14400
+            export COMPILERS_HASH_DIR=/tmp/.sccache
+            export SCCACHE_BIN=/usr/local/.cargo/bin/sccache
+            export SCCACHE_EXTRAFILES=/tmp/.sccache/rocm_compilers_hash_file
+            export SCCACHE_REDIS="redis://${env.CK_SCCACHE}"
+            echo "connect = ${env.CK_SCCACHE}" >> ../script/redis-cli.conf
+            export SCCACHE_C_CUSTOM_CACHE_BUSTER="${invocation_tag}"
+            echo \$SCCACHE_C_CUSTOM_CACHE_BUSTER
+            stunnel ../script/redis-cli.conf
+            (
+                set -e
+                ../script/sccache_wrapper.sh --enforce_redis
+            )
+            error_code=\$?
+            if [ \$error_code -ne 0 ]; then
+                echo "could not connect to the redis server. using sccache locally."
+                ../script/sccache_wrapper.sh
+            fi
+        """
+        setup_args = " -DCMAKE_CXX_COMPILER_LAUNCHER=sccache -DCMAKE_C_COMPILER_LAUNCHER=sccache " + setup_args
+    }
     def setup_cmd = conf.get("setup_cmd", "${cmake_envs} cmake ${setup_args}   .. ")
     // reduce parallelism when compiling, clang uses too much memory
     def nt = nthreads()
@@ -251,7 +257,7 @@ def cmake_build(Map conf=[:]){
     sh cmd
 
     // Only archive from master or develop
-    if (package_build == true && (env.BRANCH_NAME == "develop" || env.BRANCH_NAME == "master")) {
+    if (package_build == true && (env.BRANCH_NAME == "develop" || env.BRANCH_NAME == "amd-master")) {
         archiveArtifacts artifacts: "build/*.deb", allowEmptyArchive: true, fingerprint: true
     }
 }
@@ -635,7 +641,7 @@ def process_results(Map conf=[:]){
 //launch develop branch daily at 23:00 UT in FULL_QA mode and at 19:00 UT with latest staging compiler version
 CRON_SETTINGS = BRANCH_NAME == "develop" ? '''0 23 * * * % RUN_FULL_QA=true;ROCMVERSION=5.7;COMPILER_VERSION=
                                               0 21 * * * % ROCMVERSION=5.7;COMPILER_VERSION=;COMPILER_COMMIT=
-                                              0 19 * * * % BUILD_DOCKER=true;DL_KERNELS=true;COMPILER_VERSION=amd-stg-open;COMPILER_COMMIT=''' : ""
+                                              0 19 * * * % BUILD_DOCKER=true;DL_KERNELS=true;COMPILER_VERSION=amd-stg-open;COMPILER_COMMIT=;USE_SCCACHE=false''' : ""
 
 pipeline {
     agent none
@@ -682,7 +688,10 @@ pipeline {
             name: 'hipTensor_branch',
             defaultValue: 'mainline',
             description: 'Specify which branch of hipTensor to use (default: mainline)')
-
+        booleanParam(
+            name: "USE_SCCACHE",
+            defaultValue: true,
+            description: "Use the sccache for building CK (default: ON)")
     }
     environment{
         dbuser = "${dbuser}"