Jenkinsfile 28 KB
Newer Older
JD's avatar
JD committed
1
2
3
4
5
6
7
8
9
10
11
12
13
def rocmnode(name) {
    return 'rocmtest && miopen && ' + name
}

def show_node_info() {
    sh """
        echo "NODE_NAME = \$NODE_NAME"
        lsb_release -sd
        uname -r
        ls /opt/ -la
    """
}

14
def runShell(String command){
15
    def responseCode = sh returnStatus: true, script: "${command} > tmp.txt"
16
    def output = readFile(file: "tmp.txt")
17
    echo "tmp.txt contents: $output"
18
19
20
    return (output != "")
}

21
def getDockerImageName(){
22
    def img = "${env.CK_IMAGE_URL}:composable_kernels_${params.COMPILER_VERSION}"
23
24
25
    return img
}

26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
def build_compiler(){
    def compiler
    if (params.BUILD_COMPILER == "hipcc"){
        compiler = '/opt/rocm/bin/hipcc'
    }
    else{
        if (params.COMPILER_VERSION == "release"){
            compiler = "/opt/rocm/llvm/bin/clang++"
        }
        else{
            compiler = "/llvm-project/build/bin/clang++"
        }        
    }
    return compiler
}

42
43
44
45
46
def getDockerImage(Map conf=[:]){
    env.DOCKER_BUILDKIT=1
    def prefixpath = conf.get("prefixpath", "/opt/rocm") // prefix:/opt/rocm
    def gpu_arch = conf.get("gpu_arch", "gfx908") // prebuilt dockers should have all the architectures enabled so one image can be used for all stages
    def no_cache = conf.get("no_cache", false)
Anthony Chang's avatar
Anthony Chang committed
47
48
49
    def ccache_dir_mount = conf.get("ccache_dir_mount", "/ccache_dir_host_shared")

    def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg ccache_dir_mount=${ccache_dir_mount} "
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
    if(env.CCACHE_HOST)
    {
        def check_host = sh(script:"""(printf "PING\r\n";) | nc -N ${env.CCACHE_HOST} 6379 """, returnStdout: true).trim()
        if(check_host == "+PONG")
        {
            echo "FOUND CCACHE SERVER: ${CCACHE_HOST}"
        }
        else 
        {
            echo "CCACHE SERVER: ${CCACHE_HOST} NOT FOUND, got ${check_host} response"
        }
        dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CCACHE_HOST}' --build-arg COMPILER_LAUNCHER='ccache' "
        env.CCACHE_DIR = """/tmp/ccache_store"""
        env.CCACHE_SECONDARY_STORAGE="""redis://${env.CCACHE_HOST}"""
    }
    if(no_cache)
    {
        dockerArgs = dockerArgs + " --no-cache "
    }
    echo "Docker Args: ${dockerArgs}"
    def image = getDockerImageName()
    //Check if image exists 
    def retimage
    try 
    {
        echo "Pulling down image: ${image}"
        retimage = docker.image("${image}")
        retimage.pull()
    }
    catch(Exception ex)
    {
        error "Unable to locate image: ${image}"
    }
    return [retimage, image]
}

def buildDocker(install_prefix){
    show_node_info()
    env.DOCKER_BUILDKIT=1
    checkout scm
    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}' "
    if(env.CCACHE_HOST)
    {
        def check_host = sh(script:"""(printf "PING\\r\\n";) | nc  -N ${env.CCACHE_HOST} 6379 """, returnStdout: true).trim()
        if(check_host == "+PONG")
        {
            echo "FOUND CCACHE SERVER: ${CCACHE_HOST}"
        }
        else 
        {
            echo "CCACHE SERVER: ${CCACHE_HOST} NOT FOUND, got ${check_host} response"
        }
        dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CCACHE_HOST}' --build-arg COMPILER_LAUNCHER='ccache' "
        env.CCACHE_DIR = """/tmp/ccache_store"""
        env.CCACHE_SECONDARY_STORAGE="""redis://${env.CCACHE_HOST}"""
    }

    echo "Build Args: ${dockerArgs}"
    try{
        echo "Checking for image: ${image_name}"
        sh "docker manifest inspect --insecure ${image_name}"
        echo "Image: ${image_name} found!! Skipping building image"
    }
    catch(Exception ex){
        echo "Unable to locate image: ${image_name}. Building image now"
        retimage = docker.build("${image_name}", dockerArgs + ' .')
        retimage.push()
    }
}

JD's avatar
JD committed
122
123
def cmake_build(Map conf=[:]){

124
    def compiler = build_compiler()
JD's avatar
JD committed
125
126
    def config_targets = conf.get("config_targets","check")
    def debug_flags = "-g -fno-omit-frame-pointer -fsanitize=undefined -fno-sanitize-recover=undefined " + conf.get("extradebugflags", "")
127
    def build_envs = "CTEST_PARALLEL_LEVEL=4 " + conf.get("build_env","")
JD's avatar
JD committed
128
129
    def prefixpath = conf.get("prefixpath","/opt/rocm")
    def setup_args = conf.get("setup_args","")
Anthony Chang's avatar
Anthony Chang committed
130
131
132
    def ccache_envs = "CCACHE_BASEDIR=/var/jenkins/workspace "

    setup_args = setup_args + " -DCMAKE_CXX_COMPILER_LAUNCHER=ccache"
JD's avatar
JD committed
133
134
135
136
137
138
139
140

    if (prefixpath != "/usr/local"){
        setup_args = setup_args + " -DCMAKE_PREFIX_PATH=${prefixpath} "
    }

    def build_type_debug = (conf.get("build_type",'release') == 'debug')

    //cmake_env can overwrite default CXX variables.
Anthony Chang's avatar
Anthony Chang committed
141
    def cmake_envs = "CXX=${compiler} CXXFLAGS='-Werror' " + "${ccache_envs}" + conf.get("cmake_ex_env","")
JD's avatar
JD committed
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172

    def package_build = (conf.get("package_build","") == "true")

    if (package_build == true) {
        config_targets = "package"
    }

    if(conf.get("build_install","") == "true")
    {
        config_targets = 'install ' + config_targets
        setup_args = ' -DBUILD_DEV=Off -DCMAKE_INSTALL_PREFIX=../install' + setup_args
    } else{
        setup_args = ' -DBUILD_DEV=On' + setup_args
    }

    if(build_type_debug){
        setup_args = " -DCMAKE_BUILD_TYPE=debug -DCMAKE_CXX_FLAGS_DEBUG='${debug_flags}'" + setup_args
    }else{
        setup_args = " -DCMAKE_BUILD_TYPE=release" + setup_args
    }

    def pre_setup_cmd = """
            echo \$HSA_ENABLE_SDMA
            ulimit -c unlimited
            rm -rf build
            mkdir build
            rm -rf install
            mkdir install
            cd build
        """
    def setup_cmd = conf.get("setup_cmd", "${cmake_envs} cmake ${setup_args}   .. ")
Chao Liu's avatar
Chao Liu committed
173
    // reduce parallelism when compiling, clang uses too much memory
174
    def build_cmd = conf.get("build_cmd", "${build_envs} dumb-init make  -j\$(( \$(nproc) / 2 )) ${config_targets}")
JD's avatar
JD committed
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
    def execute_cmd = conf.get("execute_cmd", "")

    def cmd = conf.get("cmd", """
            ${pre_setup_cmd}
            ${setup_cmd}
            ${build_cmd}
            ${execute_cmd}
        """)

    echo cmd
    sh cmd

    // Only archive from master or develop
    if (package_build == true && (env.BRANCH_NAME == "develop" || env.BRANCH_NAME == "master")) {
        archiveArtifacts artifacts: "build/*.deb", allowEmptyArchive: true, fingerprint: true
    }
}

def buildHipClangJob(Map conf=[:]){
        show_node_info()

        env.HSA_ENABLE_SDMA=0
        checkout scm

199
        def image = "composable_kernels_${params.COMPILER_VERSION}"
JD's avatar
JD committed
200
201
202
203
204
        def prefixpath = conf.get("prefixpath", "/opt/rocm")
        def gpu_arch = conf.get("gpu_arch", "gfx908")

        // Jenkins is complaining about the render group 
        // def dockerOpts="--device=/dev/kfd --device=/dev/dri --group-add video --group-add render --cap-add=SYS_PTRACE --security-opt seccomp=unconfined"
Anthony Chang's avatar
Anthony Chang committed
205
206
207
208
209

        def homedir = sh( script: "echo $HOME", returnStdout: true ).trim()
        def ccache_dir_mount = conf.get("ccache_dir_mount", "/ccache_dir_host_shared")

        def dockerOpts="--device=/dev/kfd --device=/dev/dri --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -v /${homedir}/.ccache/ccache:${ccache_dir_mount}"
JD's avatar
JD committed
210
        if (conf.get("enforce_xnack_on", false)) {
211
            dockerOpts = dockerOpts + " --env HSA_XNACK=1 --env GPU_ARCH='${gpu_arch}' "
JD's avatar
JD committed
212
        }
Anthony Chang's avatar
Anthony Chang committed
213
        def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' --build-arg ccache_dir_mount=${ccache_dir_mount}"
214
        if (params.COMPILER_VERSION != "release"){
215
216
            dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
        }
JD's avatar
JD committed
217
218
219
220

        def variant = env.STAGE_NAME

        def retimage
221
222

        gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
223
            try {
224
225
                //retimage = docker.build("${image}", dockerArgs + '.')
                (retimage, image) = getDockerImage(conf)
226
227
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
228
229
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
230
                            throw new Exception ("GPU not found")
231
232
233
234
                        }
                        else{
                            echo "GPU is OK"
                        }
JD's avatar
JD committed
235
236
237
                    }
                }
            }
238
239
240
241
242
243
244
245
            catch (org.jenkinsci.plugins.workflow.steps.FlowInterruptedException e){
                echo "The job was cancelled or aborted"
                throw e
            }
            catch(Exception ex) {
                retimage = docker.build("${image}", dockerArgs + " --no-cache .")
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
246
247
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo |tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
248
                            throw new Exception ("GPU not found")
249
250
251
252
                        }
                        else{
                            echo "GPU is OK"
                        }
253
                    }
254
255
                }
            }
JD's avatar
JD committed
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286

            withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') {
                timeout(time: 5, unit: 'HOURS')
                {
                    cmake_build(conf)
                }
            }
        }
        return retimage
}

def reboot(){
    build job: 'reboot-slaves', propagate: false , parameters: [string(name: 'server', value: "${env.NODE_NAME}"),]
}

def buildHipClangJobAndReboot(Map conf=[:]){
    try{
        buildHipClangJob(conf)
    }
    catch(e){
        echo "throwing error exception for the stage"
        echo 'Exception occurred: ' + e.toString()
        throw e
    }
    finally{
        if (!conf.get("no_reboot", false)) {
            reboot()
        }
    }
}

287
288
289
290
291
292
def runCKProfiler(Map conf=[:]){
        show_node_info()

        env.HSA_ENABLE_SDMA=0
        checkout scm

293
294

        def image = "composable_kernels_${params.COMPILER_VERSION}"
295
296
297
298
299
300
301
        def prefixpath = conf.get("prefixpath", "/opt/rocm")
        def gpu_arch = conf.get("gpu_arch", "gfx908")

        // Jenkins is complaining about the render group 
        // def dockerOpts="--device=/dev/kfd --device=/dev/dri --group-add video --group-add render --cap-add=SYS_PTRACE --security-opt seccomp=unconfined"
        def dockerOpts="--device=/dev/kfd --device=/dev/dri --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined"
        if (conf.get("enforce_xnack_on", false)) {
302
            dockerOpts = dockerOpts + " --env HSA_XNACK=1 --env GPU_ARCH='${gpu_arch}' "
303
        }
304
        def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='${params.COMPILER_VERSION}' "
305
        if (params.COMPILER_VERSION != "release"){
306
307
            dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
        }
308
309
310

        def variant = env.STAGE_NAME
        def retimage
311
312

        gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
313
            try {
314
315
                //retimage = docker.build("${image}", dockerArgs + '.')
                (retimage, image) = getDockerImage(conf)
316
317
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
318
319
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
320
                            throw new Exception ("GPU not found")
321
322
323
324
                        }
                        else{
                            echo "GPU is OK"
                        }
325
326
327
                    }
                }
            }
328
329
330
331
332
333
334
335
            catch (org.jenkinsci.plugins.workflow.steps.FlowInterruptedException e){
                echo "The job was cancelled or aborted"
                throw e
            }
            catch(Exception ex) {
                retimage = docker.build("${image}", dockerArgs + " --no-cache .")
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
336
337
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
338
                            throw new Exception ("GPU not found")
339
340
341
342
                        }
                        else{
                            echo "GPU is OK"
                        }
343
                    }
344
345
                }
            }
346
347

            withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') {
348
                timeout(time: 24, unit: 'HOURS')
349
350
351
                {
                    cmake_build(conf)
					dir("script"){
352
353
                        if (params.RUN_FULL_QA){
                            def qa_log = "qa_${gpu_arch}.log"
354
                            sh "./run_full_performance_tests.sh 1 QA_${params.COMPILER_VERSION} ${gpu_arch} ${env.BRANCH_NAME} ${NODE_NAME}"
355
356
357
                            archiveArtifacts "perf_gemm_${gpu_arch}.log"
                            archiveArtifacts "perf_resnet50_N256_${gpu_arch}.log"
                            archiveArtifacts "perf_resnet50_N4_${gpu_arch}.log"
358
                            archiveArtifacts "perf_batched_gemm_${gpu_arch}.log"
359
                            archiveArtifacts "perf_grouped_gemm_${gpu_arch}.log"
360
                            archiveArtifacts "perf_conv_fwd_${gpu_arch}.log"
361
                            archiveArtifacts "perf_conv_bwd_data_${gpu_arch}.log"
362
                            archiveArtifacts "perf_gemm_bilinear_${gpu_arch}.log"
363
364
365
366
367
                            archiveArtifacts "perf_reduction_${gpu_arch}.log"
                           // stash perf files to master
                            stash name: "perf_gemm_${gpu_arch}.log"
                            stash name: "perf_resnet50_N256_${gpu_arch}.log"
                            stash name: "perf_resnet50_N4_${gpu_arch}.log"
368
                            stash name: "perf_batched_gemm_${gpu_arch}.log"
369
                            stash name: "perf_grouped_gemm_${gpu_arch}.log"
370
                            stash name: "perf_conv_fwd_${gpu_arch}.log"
371
                            stash name: "perf_conv_bwd_data_${gpu_arch}.log"
372
                            stash name: "perf_gemm_bilinear_${gpu_arch}.log"
373
374
                            stash name: "perf_reduction_${gpu_arch}.log"
                            //we will process results on the master node
375
376
                        }
                        else{
377
                            sh "./run_performance_tests.sh 0 CI_${params.COMPILER_VERSION} ${gpu_arch} ${env.BRANCH_NAME} ${NODE_NAME}"
378
379
380
381
382
383
384
385
                            archiveArtifacts "perf_gemm_${gpu_arch}.log"
                            archiveArtifacts "perf_resnet50_N256_${gpu_arch}.log"
                            archiveArtifacts "perf_resnet50_N4_${gpu_arch}.log"
                            // stash perf files to master
                            stash name: "perf_gemm_${gpu_arch}.log"
                            stash name: "perf_resnet50_N256_${gpu_arch}.log"
                            stash name: "perf_resnet50_N4_${gpu_arch}.log"
                            //we will process the results on the master node
386
                        }
387

388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
					}
                }
            }
        }
        return retimage
}

def runPerfTest(Map conf=[:]){
    try{
        runCKProfiler(conf)
    }
    catch(e){
        echo "throwing error exception in performance tests"
        echo 'Exception occurred: ' + e.toString()
        throw e
    }
    finally{
        if (!conf.get("no_reboot", false)) {
            reboot()
        }
    }
}

411
412
413
def process_results(Map conf=[:]){
    env.HSA_ENABLE_SDMA=0
    checkout scm
414
    def image = "composable_kernels_${params.COMPILER_VERSION}"
415
416
417
418
419
420
    def prefixpath = "/opt/rocm"
    def gpu_arch = conf.get("gpu_arch", "gfx908")

    // Jenkins is complaining about the render group 
    def dockerOpts="--cap-add=SYS_PTRACE --security-opt seccomp=unconfined"
    if (conf.get("enforce_xnack_on", false)) {
421
        dockerOpts = dockerOpts + " --env HSA_XNACK=1 --env GPU_ARCH='${gpu_arch}' "
422
    }
423
    def dockerArgs = "--build-arg PREFIX=${prefixpath} --build-arg compiler_version='release' "
424
425
426
427
428
429

    def variant = env.STAGE_NAME
    def retimage

    gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
        try {
430
431
            //retimage = docker.build("${image}", dockerArgs + '.')
            (retimage, image) = getDockerImage(conf)
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
        }
        catch (org.jenkinsci.plugins.workflow.steps.FlowInterruptedException e){
            echo "The job was cancelled or aborted"
            throw e
        }
    }

    withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') {
        timeout(time: 1, unit: 'HOURS'){
            try{
                dir("script"){
                    if (params.RUN_FULL_QA){
                        // unstash perf files to master
                        unstash "perf_gemm_${gpu_arch}.log"
                        unstash "perf_resnet50_N256_${gpu_arch}.log"
                        unstash "perf_resnet50_N4_${gpu_arch}.log"
448
                        unstash "perf_batched_gemm_${gpu_arch}.log"
449
                        unstash "perf_grouped_gemm_${gpu_arch}.log"
450
                        unstash "perf_conv_fwd_${gpu_arch}.log"
451
                        unstash "perf_conv_bwd_data_${gpu_arch}.log"
452
                        unstash "perf_gemm_bilinear_${gpu_arch}.log"
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
                        unstash "perf_reduction_${gpu_arch}.log"
                        sh "./process_qa_data.sh ${gpu_arch}"
                    }
                    else{
                        // unstash perf files to master
                        unstash "perf_gemm_${gpu_arch}.log"
                        unstash "perf_resnet50_N256_${gpu_arch}.log"
                        unstash "perf_resnet50_N4_${gpu_arch}.log"
                        sh "./process_perf_data.sh ${gpu_arch}"
                    }
                }
            }
            catch(e){
                echo "throwing error exception while processing performance test results"
                echo 'Exception occurred: ' + e.toString()
                throw e
            }
        }
    }
}

//launch develop branch daily at 23:00 in FULL_QA mode
475
CRON_SETTINGS = BRANCH_NAME == "develop" ? '''0 23 * * * % RUN_FULL_QA=true''' : ""
476

JD's avatar
JD committed
477
478
pipeline {
    agent none
479
480
481
    triggers {
        parameterizedCron(CRON_SETTINGS)
    }
JD's avatar
JD committed
482
483
484
    options {
        parallelsAlwaysFailFast()
    }
485
    parameters {
486
487
488
489
        booleanParam(
            name: "BUILD_DOCKER",
            defaultValue: true,
            description: "Force building docker image (default: true)")
490
491
492
        string(
            name: 'COMPILER_VERSION', 
            defaultValue: 'ck-9110', 
493
            description: 'Specify which version of compiler to use: ck-9110 (default), release, or amd-stg-open.')
494
495
496
497
        string(
            name: 'BUILD_COMPILER', 
            defaultValue: 'hipcc', 
            description: 'Specify whether to build CK with hipcc (default) or with clang.')
498
499
500
501
        booleanParam(
            name: "RUN_FULL_QA",
            defaultValue: false,
            description: "Select whether to run small set of performance tests (default) or full QA")
502
503
504
505
        booleanParam(
            name: "TEST_NODE_PERFORMANCE",
            defaultValue: false,
            description: "Test the node GPU performance (default: false)")
506
507
508
509
510
511
512
513
    }
    environment{
        dbuser = "${dbuser}"
        dbpassword = "${dbpassword}"
        dbsship = "${dbsship}"
        dbsshport = "${dbsshport}"
        dbsshuser = "${dbsshuser}"
        dbsshpassword = "${dbsshpassword}"
514
        status_wrapper_creds = "${status_wrapper_creds}"
515
516
        gerrit_cred="${gerrit_cred}"
        DOCKER_BUILDKIT = "1"
517
    }
JD's avatar
JD committed
518
    stages{
519
520
521
522
523
524
525
526
527
528
529
530
531
        stage("Build Docker"){
            when {
                expression { params.BUILD_DOCKER.toBoolean() }
            }
            parallel{
                stage('Docker /opt/rocm'){
                    agent{ label rocmnode("nogpu") }
                    steps{
                        buildDocker('/opt/rocm')
                    }
                }
            }
        }
Anthony Chang's avatar
Anthony Chang committed
532
        /*
JD's avatar
JD committed
533
        stage("Static checks") {
534
535
536
537
            when {
                beforeAgent true
                expression { !params.TEST_NODE_PERFORMANCE.toBoolean() }
            }
JD's avatar
JD committed
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
            parallel{
                // enable after we move from hipcc to hip-clang
                // stage('Tidy') {
                //     agent{ label rocmnode("nogpu") }
                //     environment{
                //         // setup_cmd = "CXX='/opt/rocm/bin/hipcc' cmake -DBUILD_DEV=On .. "
                //         build_cmd = "make -j\$(nproc) -k analyze"
                //     }
                //     steps{
                //         buildHipClangJobAndReboot(build_cmd: build_cmd, no_reboot:true, prefixpath: '/opt/rocm', build_type: 'debug')
                //     }
                // }
                stage('Clang Format') {
                    agent{ label rocmnode("nogpu") }
                    environment{
JD's avatar
JD committed
553
                        execute_cmd = "find .. -iname \'*.h\' \
JD's avatar
JD committed
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
                                -o -iname \'*.hpp\' \
                                -o -iname \'*.cpp\' \
                                -o -iname \'*.h.in\' \
                                -o -iname \'*.hpp.in\' \
                                -o -iname \'*.cpp.in\' \
                                -o -iname \'*.cl\' \
                                | grep -v 'build/' \
                                | xargs -n 1 -P 1 -I{} -t sh -c \'clang-format-10 -style=file {} | diff - {}\'"
                    }
                    steps{
                        buildHipClangJobAndReboot(setup_cmd: "", build_cmd: "", execute_cmd: execute_cmd, no_reboot:true)
                    }
                }
            }
        }
Anthony Chang's avatar
Anthony Chang committed
569
        */
570
		stage("Tests")
571
        {
572
573
574
575
            when {
                beforeAgent true
                expression { !params.TEST_NODE_PERFORMANCE.toBoolean() }
            }
576
577
578
579
580
581
            parallel
            {
                stage("Run Tests: gfx908")
                {
                    agent{ label rocmnode("gfx908")}
                    environment{
582
                        setup_args = """ -D CMAKE_CXX_FLAGS=" --offload-arch=gfx908 -O3 " -DBUILD_DEV=On """
583
584
                    }
                    steps{
Anthony Chang's avatar
Anthony Chang committed
585
                        buildHipClangJobAndReboot(setup_args:setup_args, config_targets: "examples", no_reboot:true, build_type: 'Release', gpu_arch: "gfx908")
586
587
                    }
                }
JD's avatar
JD committed
588
589
                stage("Run Tests: gfx90a")
                {
590
591
592
593
                    when {
                        beforeAgent true
                        expression { params.RUN_FULL_QA.toBoolean() }
                    }
594
                    options { retry(2) }
JD's avatar
JD committed
595
596
597
598
599
                    agent{ label rocmnode("gfx90a")}
                    environment{
                        setup_args = """ -D CMAKE_CXX_FLAGS="--offload-arch=gfx90a -O3 " -DBUILD_DEV=On """
                    }
                    steps{
Anthony Chang's avatar
Anthony Chang committed
600
                        buildHipClangJobAndReboot(setup_args:setup_args, config_targets: "examples", no_reboot:true, build_type: 'Release', gpu_arch: "gfx90a")
JD's avatar
JD committed
601
602
                    }
                }
603
604
            }
        }
Anthony Chang's avatar
Anthony Chang committed
605
        /*
Chao Liu's avatar
Chao Liu committed
606
607
        stage("Client App")
        {
608
609
610
611
            when {
                beforeAgent true
                expression { !params.TEST_NODE_PERFORMANCE.toBoolean() }
            }
Chao Liu's avatar
Chao Liu committed
612
613
614
615
616
617
            parallel
            {
                stage("Run Client App")
                {
                    agent{ label rocmnode("gfx908")}
                    environment{
618
619
                        setup_args = """ -DBUILD_DEV=Off -DCMAKE_INSTALL_PREFIX=../install -D CMAKE_CXX_FLAGS="--offload-arch=gfx908 -O3 " """
                        execute_args = """ cd ../client_example && rm -rf build && mkdir build && cd build && cmake -D CMAKE_PREFIX_PATH="${env.WORKSPACE}/install;/opt/rocm" -D CMAKE_CXX_FLAGS=" --offload-arch=gfx908 -O3" -D CMAKE_CXX_COMPILER="${build_compiler()}" .. && make -j """ 
Chao Liu's avatar
Chao Liu committed
620
621
622
623
624
625
626
                    }
                    steps{
                        buildHipClangJobAndReboot(setup_args: setup_args, config_targets: "install", no_reboot:true, build_type: 'Release', execute_cmd: execute_args, prefixpath: '/usr/local')
                    }
                }
            }
        }
627
628
629
630
631
632
        stage("Performance Tests")
        {
            parallel
            {
                stage("Run ckProfiler: gfx908")
                {
633
634
                    when {
                        beforeAgent true
635
                        expression { !params.RUN_FULL_QA.toBoolean() && !params.TEST_NODE_PERFORMANCE.toBoolean() }
636
                    }
637
                    options { retry(2) }
638
639
640
                    agent{ label rocmnode("gfx908")}
                    environment{
                        setup_args = """ -D CMAKE_CXX_FLAGS="--offload-arch=gfx908 -O3 " -DBUILD_DEV=On """
641
                   }
642
                    steps{
643
644
645
646
647
                        runPerfTest(setup_args:setup_args, config_targets: "ckProfiler", no_reboot:true, build_type: 'Release', gpu_arch: "gfx908")
                    }
                }
                stage("Run ckProfiler: gfx90a")
                {
648
649
                    when {
                        beforeAgent true
650
                        expression { params.RUN_FULL_QA.toBoolean() || params.TEST_NODE_PERFORMANCE.toBoolean() }
651
                    }
652
                    options { retry(2) }
653
654
655
656
657
658
                    agent{ label rocmnode("gfx90a")}
                    environment{
                        setup_args = """ -D CMAKE_CXX_FLAGS="--offload-arch=gfx90a -O3 " -DBUILD_DEV=On """
                   }
                    steps{
                        runPerfTest(setup_args:setup_args, config_targets: "ckProfiler", no_reboot:true, build_type: 'Release', gpu_arch: "gfx90a")
659
660
661
662
                    }
                }
            }
        }
663
664
665
666
667
        stage("Process Performance Test Results")
        {
            parallel
            {
                stage("Process results for gfx908"){
668
669
                    when {
                        beforeAgent true
670
                        expression { !params.RUN_FULL_QA.toBoolean() && !params.TEST_NODE_PERFORMANCE.toBoolean() }
671
                    }
672
673
674
675
676
677
                    agent { label 'mici' }
                    steps{
                        process_results(gpu_arch: "gfx908")
                    }
                }
                stage("Process results for gfx90a"){
678
679
                    when {
                        beforeAgent true
680
                        expression { params.RUN_FULL_QA.toBoolean() || params.TEST_NODE_PERFORMANCE.toBoolean() }
681
                    }
682
683
684
685
686
687
688
                    agent { label 'mici' }
                    steps{
                        process_results(gpu_arch: "gfx90a")
                    }
                }
            }
        }
Anthony Chang's avatar
Anthony Chang committed
689
        */
690

691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
        /* enable after the cmake file supports packaging
        stage("Packages") {
            when {
                expression { params.BUILD_PACKAGES && params.TARGET_NOGPU && params.DATATYPE_NA }
            }
            parallel {
                stage("Package /opt/rocm") {
                    agent{ label rocmnode("nogpu") }
                    steps{
                        buildHipClangJobAndReboot( package_build: "true", prefixpath: '/opt/rocm', gpu_arch: "gfx906;gfx908;gfx90a")
                    }
                }
            }
        }
        */
JD's avatar
JD committed
706
    }
707
}