Jenkinsfile 30.2 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
17
18
19
    def output = readFile(file: "tmp.txt")
    return (output != "")
}

20
def getDockerImageName(){
21
    def img
22
    if (params.ROCMVERSION != "5.6"){
23
24
       if (params.COMPILER_VERSION == "") {
           img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}"
25
26
       }
       else{
27
28
29
30
31
32
33
          if (params.COMPILER_COMMIT == ""){
             img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}"
          }
          else{
             def commit = "${params.COMPILER_COMMIT}"[0..6]
             img = "${env.CK_DOCKERHUB}:ck_ub20.04_rocm${params.ROCMVERSION}_${params.COMPILER_VERSION}_${commit}"
          }
34
       }
35
36
    }
    else{
37
38
       if (params.COMPILER_VERSION == "") {
           img = "${env.CK_DOCKERHUB_PRIVATE}:ck_ub20.04_rocm${params.ROCMVERSION}"
39
40
       }
       else{
41
42
43
44
45
46
47
          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}"
          }
48
       }
49
    }
50
51
52
    return img
}

53
54
55
56
57
58
59
60
61
62
63
64
65
66
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'''
        def output = readFile(file: "tmp.txt")
        echo "tmp.txt contents: \$output"
        return (output != "0")
    }
    else{
        return 1
    }
}

67
68
69
70
71
72
def build_compiler(){
    def compiler
    if (params.BUILD_COMPILER == "hipcc"){
        compiler = '/opt/rocm/bin/hipcc'
    }
    else{
73
74
        if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
            compiler = "/llvm-project/build/bin/clang++"
75
76
        }
        else{
77
            compiler = "/opt/rocm/llvm/bin/clang++"
78
79
80
81
82
        }        
    }
    return compiler
}

83
84
85
86
def getDockerImage(Map conf=[:]){
    env.DOCKER_BUILDKIT=1
    def prefixpath = conf.get("prefixpath", "/opt/rocm") // prefix:/opt/rocm
    def no_cache = conf.get("no_cache", false)
87
88
89
    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)
90
    {
91
        if(check_host())
92
        {
93
            echo "FOUND CCACHE SERVER: ${env.CK_CCACHE}"
94
95
96
        }
        else 
        {
97
            echo "CCACHE SERVER: ${env.CK_CCACHE} NOT FOUND, got ${check_host} response"
98
        }
99
        dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CK_CCACHE}' --build-arg COMPILER_LAUNCHER='ccache' "
100
        env.CCACHE_DIR = """/tmp/ccache_store"""
101
        env.CCACHE_SECONDARY_STORAGE="""redis://${env.CK_CCACHE}"""
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
    }
    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 ${image_name}"
130
131
132
    def dockerArgs = "--build-arg BUILDKIT_INLINE_CACHE=1 --build-arg PREFIX=${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)
133
    {
134
        if(check_host())
135
        {
136
            echo "FOUND CCACHE SERVER: ${env.CK_CCACHE}"
137
138
139
        }
        else 
        {
140
            echo "CCACHE SERVER: ${env.CK_CCACHE} NOT FOUND, got ${check_host} response"
141
        }
142
        dockerArgs = dockerArgs + " --build-arg CCACHE_SECONDARY_STORAGE='redis://${env.CK_CCACHE}' --build-arg COMPILER_LAUNCHER='ccache' "
143
        env.CCACHE_DIR = """/tmp/ccache_store"""
144
        env.CCACHE_SECONDARY_STORAGE="""redis://${env.CK_CCACHE}"""
145
146
147
148
    }

    echo "Build Args: ${dockerArgs}"
    try{
149
150
151
152
153
154
155
156
157
158
159
        if(params.BUILD_DOCKER){
            //force building the new docker if that parameter is true
            echo "Building image: ${image_name}"
            retimage = docker.build("${image_name}", dockerArgs + ' .')
            retimage.push()
        }
        else{
            echo "Checking for image: ${image_name}"
            sh "docker manifest inspect --insecure ${image_name}"
            echo "Image: ${image_name} found!! Skipping building image"
        }
160
161
162
163
164
165
166
167
    }
    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
168
169
def cmake_build(Map conf=[:]){

170
    def compiler = build_compiler()
JD's avatar
JD committed
171
172
    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", "")
173
    def build_envs = "CTEST_PARALLEL_LEVEL=4 " + conf.get("build_env","")
JD's avatar
JD committed
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
    def prefixpath = conf.get("prefixpath","/opt/rocm")
    def setup_args = conf.get("setup_args","")

    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.
    def cmake_envs = "CXX=${compiler} CXXFLAGS='-Werror' " + conf.get("cmake_ex_env","")

    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
195
        setup_args = ' -DBUILD_DEV=On -DCMAKE_INSTALL_PREFIX=../install' + setup_args
JD's avatar
JD committed
196
197
198
199
200
201
202
203
204
    } 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
    }
205
    if(env.CK_CCACHE)
206
207
208
    {
        setup_args = " -DCMAKE_CXX_COMPILER_LAUNCHER='ccache' -DCMAKE_C_COMPILER_LAUNCHER='ccache' " + setup_args
    }
209
    echo "ccache server: ${env.CK_CCACHE}"
JD's avatar
JD committed
210
211
212
213
214
215
216
217
218
219
220

    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 ${setup_args}   .. ")
Chao Liu's avatar
Chao Liu committed
221
    // reduce parallelism when compiling, clang uses too much memory
222
    def build_cmd = conf.get("build_cmd", "${build_envs} dumb-init make  -j\$(( \$(nproc) / 2 )) ${config_targets}")
JD's avatar
JD committed
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
    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

247
        def image = getDockerImageName() 
JD's avatar
JD committed
248
249
250
        def prefixpath = conf.get("prefixpath", "/opt/rocm")

        // Jenkins is complaining about the render group 
251
        def dockerOpts="--device=/dev/kfd --device=/dev/dri --group-add video --group-add render --cap-add=SYS_PTRACE --security-opt seccomp=unconfined"
JD's avatar
JD committed
252
        if (conf.get("enforce_xnack_on", false)) {
253
            dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
JD's avatar
JD committed
254
        }
255
        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}' "
256
        if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
257
258
            dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
        }
JD's avatar
JD committed
259
260
261
262

        def variant = env.STAGE_NAME

        def retimage
263
        (retimage, image) = getDockerImage(conf)
264
265

        gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
JD's avatar
JD committed
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
            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()
        }
    }
}

296
297
298
299
300
301
def runCKProfiler(Map conf=[:]){
        show_node_info()

        env.HSA_ENABLE_SDMA=0
        checkout scm

302
        def image = getDockerImageName()
303
304
305
        def prefixpath = conf.get("prefixpath", "/opt/rocm")

        // Jenkins is complaining about the render group 
306
        def dockerOpts="--device=/dev/kfd --device=/dev/dri --group-add video --group-add render --cap-add=SYS_PTRACE --security-opt seccomp=unconfined"
307
        if (conf.get("enforce_xnack_on", false)) {
308
            dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
309
        }
310
        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}' "
311
        if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
312
313
            dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
        }
314
315
316

        def variant = env.STAGE_NAME
        def retimage
317
318

        gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
319
            try {
320
                (retimage, image) = getDockerImage(conf)
321
322
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
323
324
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
325
                            throw new Exception ("GPU not found")
326
327
328
329
                        }
                        else{
                            echo "GPU is OK"
                        }
330
331
332
                    }
                }
            }
333
334
335
336
337
338
339
340
            catch (org.jenkinsci.plugins.workflow.steps.FlowInterruptedException e){
                echo "The job was cancelled or aborted"
                throw e
            }
            catch(Exception ex) {
                retimage = docker.build("${image}", dockerArgs + " --no-cache .")
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
341
342
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
343
                            throw new Exception ("GPU not found")
344
345
346
347
                        }
                        else{
                            echo "GPU is OK"
                        }
348
                    }
349
350
                }
            }
351
352

            withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') {
353
                timeout(time: 24, unit: 'HOURS')
354
                {
355
356
357
358
359
360
361
362
363
364
365
                    //cmake_build(conf)
                    //instead of building, just unstash the ckProfiler and install it
                    sh """
                        rm -rf build
                        mkdir build
                    """
                    dir("build"){
                        unstash 'ckProfiler.tar.gz'
                        sh 'tar -xvf ckProfiler.tar.gz'
                    }

366
					dir("script"){
367
                        if (params.RUN_FULL_QA){
368
369
370
371
372
373
374
375
376
377
378
379
380
                            sh "./run_full_performance_tests.sh 1 QA_${params.COMPILER_VERSION} ${env.BRANCH_NAME} ${NODE_NAME}"
                            archiveArtifacts "perf_gemm.log"
                            archiveArtifacts "perf_resnet50_N256.log"
                            archiveArtifacts "perf_resnet50_N4.log"
                            archiveArtifacts "perf_batched_gemm.log"
                            archiveArtifacts "perf_grouped_gemm.log"
                            archiveArtifacts "perf_conv_fwd.log"
                            archiveArtifacts "perf_conv_bwd_data.log"
                            archiveArtifacts "perf_gemm_bilinear.log"
                            archiveArtifacts "perf_reduction.log"
                            archiveArtifacts "perf_splitK_gemm_verify.log"
                            archiveArtifacts "perf_splitK_gemm.log"
                            archiveArtifacts "perf_onnx_gemm.log"
381
                           // stash perf files to master
382
383
384
385
386
387
388
389
390
391
392
                            stash name: "perf_gemm.log"
                            stash name: "perf_resnet50_N256.log"
                            stash name: "perf_resnet50_N4.log"
                            stash name: "perf_batched_gemm.log"
                            stash name: "perf_grouped_gemm.log"
                            stash name: "perf_conv_fwd.log"
                            stash name: "perf_conv_bwd_data.log"
                            stash name: "perf_gemm_bilinear.log"
                            stash name: "perf_reduction.log"
                            stash name: "perf_splitK_gemm.log"
                            stash name: "perf_onnx_gemm.log"
393
                            //we will process results on the master node
394
395
                        }
                        else{
396
397
398
399
                            sh "./run_performance_tests.sh 0 CI_${params.COMPILER_VERSION} ${env.BRANCH_NAME} ${NODE_NAME}"
                            archiveArtifacts "perf_gemm.log"
                            archiveArtifacts "perf_resnet50_N256.log"
                            archiveArtifacts "perf_resnet50_N4.log"
400
                            // stash perf files to master
401
402
403
                            stash name: "perf_gemm.log"
                            stash name: "perf_resnet50_N256.log"
                            stash name: "perf_resnet50_N4.log"
404
                            //we will process the results on the master node
405
                        }
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
					}
                }
            }
        }
        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()
        }
    }
}

429
430
431
432
433
434
435
436
437
438
439
440
441
442
def Build_CK(Map conf=[:]){
        show_node_info()

        env.HSA_ENABLE_SDMA=0
        checkout scm

        def image = getDockerImageName() 
        def prefixpath = conf.get("prefixpath", "/opt/rocm")

        // 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"
        if (conf.get("enforce_xnack_on", false)) {
            dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
        }
443
        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}' "
444
        if (params.COMPILER_VERSION == "amd-stg-open" || params.COMPILER_COMMIT != ""){
445
446
447
448
449
            dockerOpts = dockerOpts + " --env HIP_CLANG_PATH='/llvm-project/build/bin' "
        }

        def variant = env.STAGE_NAME
        def retimage
Illia Silin's avatar
Illia Silin committed
450
        def navi_node = 0
451
452
453
454
455
456
457
458
459
460
461
462
463

        gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
            try {
                (retimage, image) = getDockerImage(conf)
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo | tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
                            throw new Exception ("GPU not found")
                        }
                        else{
                            echo "GPU is OK"
                        }
Illia Silin's avatar
Illia Silin committed
464
465
466
                        if ( runShell('grep -n "gfx1030" clinfo.log') ){
                            navi_node = 1
                        }
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
                    }
                }
            }
            catch (org.jenkinsci.plugins.workflow.steps.FlowInterruptedException e){
                echo "The job was cancelled or aborted"
                throw e
            }
            catch(Exception ex) {
                retimage = docker.build("${image}", dockerArgs + " --no-cache .")
                withDockerContainer(image: image, args: dockerOpts) {
                    timeout(time: 5, unit: 'MINUTES'){
                        sh 'PATH="/opt/rocm/opencl/bin:/opt/rocm/opencl/bin/x86_64:$PATH" clinfo |tee clinfo.log'
                        if ( runShell('grep -n "Number of devices:.*. 0" clinfo.log') ){
                            throw new Exception ("GPU not found")
                        }
                        else{
                            echo "GPU is OK"
                        }
Illia Silin's avatar
Illia Silin committed
485
486
487
                        if ( runShell('grep -n "gfx1030" clinfo.log') ){
                            navi_node = 1
                        }
488
489
490
491
492
493
494
495
                    }
                }
            }
            withDockerContainer(image: image, args: dockerOpts + ' -v=/var/jenkins/:/var/jenkins') {
                timeout(time: 24, unit: 'HOURS')
                {
                    cmake_build(conf)
                    dir("build"){
Illia Silin's avatar
Illia Silin committed
496
497
498
499
500
501
502
                        if (navi_node == 0 ){
                           //run tests and examples on all nodes except Navi
                           sh 'make -j check'
                           //we only need the ckProfiler to run the performance tests, so we pack and stash it
                           sh 'tar -zcvf ckProfiler.tar.gz bin/ckProfiler'
                           stash "ckProfiler.tar.gz"
                        }
503
504
505
506
507
                        if (params.RUN_FULL_QA){
                           // build deb packages
                           sh 'make -j package'
                           archiveArtifacts artifacts: 'composablekernel-ckprofiler_*.deb'
                           archiveArtifacts artifacts: 'composablekernel-tests_*.deb'
Illia Silin's avatar
Illia Silin committed
508
509
                           sh 'mv composablekernel-ckprofiler_*.deb ckprofiler_0.2.0_amd64.deb'
                           stash "ckprofiler_0.2.0_amd64.deb"
510
                        }
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
                    }
                }
            }
        }
        return retimage
}

def Build_CK_and_Reboot(Map conf=[:]){
    try{
        Build_CK(conf)
    }
    catch(e){
        echo "throwing error exception while building CK"
        echo 'Exception occurred: ' + e.toString()
        throw e
    }
    finally{
        if (!conf.get("no_reboot", false)) {
            reboot()
        }
    }
}

534
535
536
def process_results(Map conf=[:]){
    env.HSA_ENABLE_SDMA=0
    checkout scm
537
    def image = getDockerImageName() 
538
539
540
541
542
    def prefixpath = "/opt/rocm"

    // Jenkins is complaining about the render group 
    def dockerOpts="--cap-add=SYS_PTRACE --security-opt seccomp=unconfined"
    if (conf.get("enforce_xnack_on", false)) {
543
        dockerOpts = dockerOpts + " --env HSA_XNACK=1 "
544
545
546
547
548
549
550
    }

    def variant = env.STAGE_NAME
    def retimage

    gitStatusWrapper(credentialsId: "${status_wrapper_creds}", gitHubContext: "Jenkins - ${variant}", account: 'ROCmSoftwarePlatform', repo: 'composable_kernel') {
        try {
551
            (retimage, image) = getDockerImage(conf)
552
553
554
555
556
557
558
559
560
561
562
563
564
        }
        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
565
566
567
568
569
570
571
572
573
574
575
576
                        unstash "perf_gemm.log"
                        unstash "perf_resnet50_N256.log"
                        unstash "perf_resnet50_N4.log"
                        unstash "perf_batched_gemm.log"
                        unstash "perf_grouped_gemm.log"
                        unstash "perf_conv_fwd.log"
                        unstash "perf_conv_bwd_data.log"
                        unstash "perf_gemm_bilinear.log"
                        unstash "perf_reduction.log"
                        unstash "perf_splitK_gemm.log"
                        unstash "perf_onnx_gemm.log"
                        sh "./process_qa_data.sh"
Illia Silin's avatar
Illia Silin committed
577
578
                        unstash "ckprofiler_0.2.0_amd64.deb"
                        sh "sshpass -p ${env.ck_deb_pw} scp -o StrictHostKeyChecking=no ckprofiler_0.2.0_amd64.deb ${env.ck_deb_user}@${env.ck_deb_ip}:/var/www/html/composable_kernel/"
579
580
581
                    }
                    else{
                        // unstash perf files to master
582
583
584
585
                        unstash "perf_gemm.log"
                        unstash "perf_resnet50_N256.log"
                        unstash "perf_resnet50_N4.log"
                        sh "./process_perf_data.sh"
586
587
588
589
590
591
592
593
594
595
596
597
                    }
                }
            }
            catch(e){
                echo "throwing error exception while processing performance test results"
                echo 'Exception occurred: ' + e.toString()
                throw e
            }
        }
    }
}

598
//launch develop branch daily at 23:00 UT in FULL_QA mode and at 19:00 UT with latest staging compiler version
Illia Silin's avatar
Illia Silin committed
599
CRON_SETTINGS = BRANCH_NAME == "develop" ? '''0 23 * * * % RUN_FULL_QA=true
600
                                              0 21 * * * % ROCMVERSION=5.5;COMPILER_VERSION=release;COMPILER_COMMIT=
601
                                              0 19 * * * % BUILD_DOCKER=true;COMPILER_VERSION=amd-stg-open;COMPILER_COMMIT=''' : ""
602

JD's avatar
JD committed
603
604
pipeline {
    agent none
605
606
607
    triggers {
        parameterizedCron(CRON_SETTINGS)
    }
JD's avatar
JD committed
608
609
610
    options {
        parallelsAlwaysFailFast()
    }
611
    parameters {
612
613
        booleanParam(
            name: "BUILD_DOCKER",
614
            defaultValue: false,
615
616
617
            description: "Force building docker image (default: false), set to true if docker image needs to be updated.")
        string(
            name: 'ROCMVERSION', 
618
619
            defaultValue: '5.6', 
            description: 'Specify which ROCM version to use: 5.6 (default).')
620
621
        string(
            name: 'COMPILER_VERSION', 
622
623
            defaultValue: '', 
            description: 'Specify which version of compiler to use: release, amd-stg-open, or leave blank (default).')
624
625
        string(
            name: 'COMPILER_COMMIT', 
626
627
            defaultValue: '', 
            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.')
628
629
630
631
        string(
            name: 'BUILD_COMPILER', 
            defaultValue: 'hipcc', 
            description: 'Specify whether to build CK with hipcc (default) or with clang.')
632
633
634
635
        booleanParam(
            name: "RUN_FULL_QA",
            defaultValue: false,
            description: "Select whether to run small set of performance tests (default) or full QA")
636
637
638
639
640
641
642
643
    }
    environment{
        dbuser = "${dbuser}"
        dbpassword = "${dbpassword}"
        dbsship = "${dbsship}"
        dbsshport = "${dbsshport}"
        dbsshuser = "${dbsshuser}"
        dbsshpassword = "${dbsshpassword}"
644
        status_wrapper_creds = "${status_wrapper_creds}"
645
646
        gerrit_cred="${gerrit_cred}"
        DOCKER_BUILDKIT = "1"
647
    }
JD's avatar
JD committed
648
    stages{
649
        stage("Build Docker"){
650
651
652
653
            //when {
            //    beforeAgent true
            //    expression { params.BUILD_DOCKER.toBoolean() }
            //}
654
655
656
657
658
659
660
661
662
            parallel{
                stage('Docker /opt/rocm'){
                    agent{ label rocmnode("nogpu") }
                    steps{
                        buildDocker('/opt/rocm')
                    }
                }
            }
        }
JD's avatar
JD committed
663
664
665
666
667
        stage("Static checks") {
            parallel{
                stage('Clang Format') {
                    agent{ label rocmnode("nogpu") }
                    environment{
Illia Silin's avatar
Illia Silin committed
668
669
670
                        execute_cmd = "find .. -not -path \'*.git*\' -iname \'*.h\' \
                                -o -not -path \'*.git*\' -iname \'*.hpp\' \
                                -o -not -path \'*.git*\' -iname \'*.cpp\' \
JD's avatar
JD committed
671
672
673
674
675
676
677
678
679
680
681
682
683
                                -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)
                    }
                }
            }
        }
684
685
    
		stage("Build CK and run Tests")
686
687
688
        {
            parallel
            {
Illia Silin's avatar
Illia Silin committed
689
                stage("Build CK and run Tests on MI100/MI200")
690
                {
691
                    agent{ label rocmnode("gfx908 || gfx90a") }
692
                    environment{
Illia Silin's avatar
Illia Silin committed
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
                        setup_args = """ -DCMAKE_INSTALL_PREFIX=../install -DGPU_TARGETS="gfx908;gfx90a" """
                        execute_args = """ cd ../client_example && rm -rf build && mkdir build && cd build && cmake -D CMAKE_PREFIX_PATH="${env.WORKSPACE}/install;/opt/rocm" -DGPU_TARGETS="gfx908,gfx90a" -D CMAKE_CXX_COMPILER="${build_compiler()}" .. && make -j """ 
                    }
                    steps{
                        Build_CK_and_Reboot(setup_args: setup_args, config_targets: "install", no_reboot:true, build_type: 'Release', execute_cmd: execute_args, prefixpath: '/usr/local')
                    }
                }
                stage("Build CK and run Tests on Navi")
                {
                    when {
                        beforeAgent true
                        expression { !params.RUN_FULL_QA.toBoolean() }
                    }
                    agent{ label rocmnode("navi21") }
                    environment{
708
                        setup_args = """ -DCMAKE_INSTALL_PREFIX=../install """ 
709
                        execute_args = """ cd ../client_example && rm -rf build && mkdir build && cd build && cmake -D CMAKE_PREFIX_PATH="${env.WORKSPACE}/install;/opt/rocm" -DGPU_TARGETS="gfx1030;gfx1100;gfx1101;gfx1102" -D CMAKE_CXX_COMPILER="${build_compiler()}" .. && make -j """
Illia Silin's avatar
Illia Silin committed
710

711
712
                    }
                    steps{
713
                        Build_CK_and_Reboot(setup_args: setup_args, config_targets: "install", no_reboot:true, build_type: 'Release', execute_cmd: execute_args, prefixpath: '/usr/local')
JD's avatar
JD committed
714
715
                    }
                }
716
717
            }
        }
718

719
720
721
722
        stage("Performance Tests")
        {
            parallel
            {
Illia Silin's avatar
Illia Silin committed
723
                stage("Run ckProfiler: gfx90*")
724
                {
725
726
                    when {
                        beforeAgent true
727
                        expression { !params.RUN_FULL_QA.toBoolean() }
728
                    }
729
                    options { retry(2) }
730
                    agent{ label rocmnode("gfx908 || gfx90a")}
731
                    environment{
Illia Silin's avatar
Illia Silin committed
732
                        setup_args = """ -DGPU_TARGETS="gfx908;gfx90a" -DBUILD_DEV=On """
733
                   }
734
                    steps{
735
                        runPerfTest(setup_args:setup_args, config_targets: "ckProfiler", no_reboot:true, build_type: 'Release')
736
737
738
739
                    }
                }
                stage("Run ckProfiler: gfx90a")
                {
740
741
                    when {
                        beforeAgent true
742
                        expression { params.RUN_FULL_QA.toBoolean() }
743
                    }
744
                    options { retry(2) }
745
746
                    agent{ label rocmnode("gfx90a")}
                    environment{
Illia Silin's avatar
Illia Silin committed
747
                        setup_args = """ -DGPU_TARGETS="gfx90a" -DBUILD_DEV=On """
748
                    }
749
                    steps{
750
                        runPerfTest(setup_args:setup_args, config_targets: "ckProfiler", no_reboot:true, build_type: 'Release')
751
752
753
754
                    }
                }
            }
        }
755
756
757
758
        stage("Process Performance Test Results")
        {
            parallel
            {
759
                stage("Process results"){
760
761
                    agent { label 'mici' }
                    steps{
762
                        process_results()
763
764
765
766
                    }
                }
            }
        }
JD's avatar
JD committed
767
    }
768
}