Merge branch 'master' into internals/bad-check-handling

Author: Vasileios Karakasis

.pep8speaks.yml

@@ -0,0 +1,26 @@
+message:
+    opened:
+        header: "Hello @{name}, Thank you for submitting the Pull Request!"
+        footer: "Do see the ReFrame [Coding Style Guide](https://github.com/eth-cscs/reframe/wiki/coding-style-guide)"
+    updated:
+        header: "Hello @{name}, Thank you for updating!"
+        footer: "Do see the ReFrame [Coding Style Guide](https://github.com/eth-cscs/reframe/wiki/coding-style-guide)"
+    no_errors: "Cheers! There are no PEP8 issues in this Pull Request!"
+
+only_mention_files_with_errors: True
+
+scanner:
+    diff_only: True
+
+pycodestyle:
+    max-line-length: 79
+    ignore:
+        - E129
+        - E221
+        - E226
+        - E241
+        - E272
+        - E741
+        - E742
+        - E743
+        - W504

.travis.yml

@@ -1,8 +1,9 @@
+dist: xenial
 language: python
 python:
     - "3.5"
     - "3.6"
-    - "3.7-dev"
+    - "3.7"
 
 install:
     - pip install -r requirements.txt

Jenkinsfile

@@ -5,6 +5,7 @@ def loginBash = '#!/bin/bash -l'
 def bashScript = 'ci-scripts/ci-runner.bash'
 def machinesList = ['daint', 'dom', 'kesch']
 def machinesToRun = machinesList
+def runTests = true
 def uniqueID
 
 stage('Initialization') {
@@ -37,6 +38,11 @@ stage('Initialization') {
                 currentBuild.result = 'SUCCESS'
                 return
             }
+            else if (splittedComment[2] == 'none') {
+                runTests = false
+                currentBuild.result = 'SUCCESS'
+                return
+            }
 
             machinesRequested = []
             for (i = 2; i < splittedComment.size(); i++) {
@@ -66,6 +72,11 @@ stage('Initialization') {
     }
 }
 
+if (!runTests) {
+    println "Won't execute any test (${currentBuild.result}). Exiting..."
+    return
+}
+
 if (currentBuild.result != 'SUCCESS') {
     println "Initialization failed (${currentBuild.result}). Exiting..."
     return

README.md

@@ -19,9 +19,6 @@ Users can create their own test hierarchies, create test factories for generatin
 The official documentation is maintained [here](https://eth-cscs.github.io/reframe/index.html).
 It corresponds to the [latest](https://github.com/eth-cscs/reframe/releases/latest) stable release and not to the current status of the `master`.
 
-* SC18 presentation [[pdf](https://drive.google.com/file/d/1sIecW59E-AvhD-vl6c6QGXM14UKNzgo_/view?usp=sharing)]
-* SC18 Demo [[link](https://asciinema.org/a/6SQJTaRe2zrMInV92X0yb2gTh)]
-
 ### Building the documentation from master
 
 You may build the documentation of the master either with Python 2 or Python 3 (<= 3.5).
@@ -47,3 +44,21 @@ The documentation is now up on [localhost:8000](http://localhost:8000), where yo
 In the `cscs-checks/` folder, you can find realistic regression tests used for the CSCS systems that you can reuse and adapt to your system.
 Notice that these tests are published as examples and may not run as-is in your system.
 However, they can serve as a very good starting point for implementing your system tests in ReFrame.
+
+
+## Contact
+
+You can get in contact with the ReFrame community in the following ways:
+
+### Mailing list
+
+For keeping up with the latest news about ReFrame, posting questions and, generally getting in touch with other users and the developers, you may follow the mailing list: [[email protected]](mailto:[email protected]).
+
+Only subscribers may send messages to the list.
+To subscribe, please send an empty message to [[email protected]](mailto:[email protected]).
+
+For unsubscribing, you may send an empty message to [[email protected]](mailto:[email protected]).
+
+### Slack
+
+You may also reach the community through Slack at [reframetalk.slack.com](https://reframetalk.slack.com/join/signup). Currently, you may join the Slack workspace by invitation only, which you will get as soon as you subscribe to the mailing list.

ci-scripts/deploy.sh

@@ -39,7 +39,7 @@ cd $tmpdir
 git clone https://github.com/eth-cscs/reframe.git
 cd reframe
 found_version=$(./reframe.py -V)
-if [ $found_version -ne $version ]; then
+if [ $found_version != $version ]; then
     echo "$0: version mismatch: found $found_version, but required $version" >&2
     exit 1
 fi

config/cscs.py

@@ -117,7 +117,13 @@ class ReframeSettings:
                         'access': ['--partition=pn-regression'],
                         'environs': ['PrgEnv-cray', 'PrgEnv-cray-nompi',
                                      'PrgEnv-pgi', 'PrgEnv-pgi-nompi',
-                                     'PrgEnv-gnu', 'PrgEnv-gnu-nompi'],
+                                     'PrgEnv-gnu', 'PrgEnv-gnu-nompi',
+                                     'PrgEnv-cray-c2sm',
+                                     'PrgEnv-pgi-c2sm',
+                                     'PrgEnv-gnu-c2sm',
+                                     'PrgEnv-cray-c2sm-gpu',
+                                     'PrgEnv-pgi-c2sm-gpu',
+                                     'PrgEnv-gnu-c2sm-gpu'],
                         'descr': 'Kesch post-processing nodes'
                     },
 
@@ -131,8 +137,8 @@ class ReframeSettings:
                                      'PrgEnv-pgi-c2sm',
                                      'PrgEnv-gnu-c2sm',
                                      'PrgEnv-cray-c2sm-gpu',
-                                     'PrgEnv-gnu-c2sm-gpu',
-                                     'PrgEnv-cray-c2sm-gpu'],
+                                     'PrgEnv-pgi-c2sm-gpu',
+                                     'PrgEnv-gnu-c2sm-gpu'],
                         'descr': 'Kesch compute nodes',
                         'resources': {
                             '_rfm_gpu': ['--gres=gpu:{num_gpus_per_node}'],
@@ -244,36 +250,42 @@ class ReframeSettings:
                 },
                 'PrgEnv-cray-c2sm': {
                     'type': 'ProgEnvironment',
-                    'modules': ['env', 'c2sm/cray-env/base'],
+                    'modules': ['c2sm-rcm/1.00.00-kesch',
+                                'c2sm/cray-env/base'],
                 },
                 'PrgEnv-cray-c2sm-gpu': {
                     'type': 'ProgEnvironment',
-                    'modules': ['env', 'c2sm/cray-env/gpu'],
+                    'modules': ['c2sm-rcm/1.00.00-kesch',
+                                'c2sm/cray-env/gpu'],
                 },
                 'PrgEnv-pgi-c2sm': {
                     'type': 'ProgEnvironment',
-                    'modules': ['env', 'c2sm/pgi-env/base'],
+                    'modules': ['c2sm-rcm/1.00.00-kesch',
+                                'c2sm/pgi-env/base'],
                     'cc': 'mpicc',
                     'cxx': 'mpicxx',
                     'ftn': 'mpif90',
                 },
                 'PrgEnv-pgi-c2sm-gpu': {
                     'type': 'ProgEnvironment',
-                    'modules': ['env', 'c2sm/pgi-env/gpu'],
+                    'modules': ['c2sm-rcm/1.00.00-kesch',
+                                'c2sm/pgi-env/gpu'],
                     'cc': 'mpicc',
                     'cxx': 'mpicxx',
                     'ftn': 'mpif90',
                 },
                 'PrgEnv-gnu-c2sm': {
                     'type': 'ProgEnvironment',
-                    'modules': ['env', 'c2sm/gnu-env/base'],
+                    'modules': ['c2sm-rcm/1.00.00-kesch',
+                                'c2sm/gnu-env/base'],
                     'cc': 'mpicc',
                     'cxx': 'mpicxx',
                     'ftn': 'mpif90',
                 },
                 'PrgEnv-gnu-c2sm-gpu': {
                     'type': 'ProgEnvironment',
-                    'modules': ['env', 'c2sm/gnu-env/gpu'],
+                    'modules': ['c2sm-rcm/1.00.00-kesch',
+                                'c2sm/gnu-env/gpu'],
                     'cc': 'mpicc',
                     'cxx': 'mpicxx',
                     'ftn': 'mpif90',
@@ -414,7 +426,8 @@ class ReframeSettings:
                     '%(check_perf_var)s=%(check_perf_value)s|'
                     'ref=%(check_perf_ref)s '
                     '(l=%(check_perf_lower_thres)s, '
-                    'u=%(check_perf_upper_thres)s)'
+                    'u=%(check_perf_upper_thres)s)|'
+                    '%(check_perf_unit)s'
                 ),
                 'append': True
             }

cscs-checks/apps/cp2k/cp2k_check.py

@@ -100,7 +100,7 @@ def __init__(self, variant):
                     'perf': (251.8, None, 0.15)
                 },
                 'daint:gpu': {
-                    'perf': (182.3, None, 0.10)
+                    'perf': (222.6, None, 0.05)
                 },
             }
         else:
@@ -110,6 +110,6 @@ def __init__(self, variant):
                     'perf': (240.0, None, 0.05)
                 },
                 'daint:gpu': {
-                    'perf': (195.0, None, 0.10)
+                    'perf': (222.6, None, 0.05)
                 },
             }

cscs-checks/apps/tensorflow/tf_horovod_check.py

@@ -0,0 +1,54 @@
+import reframe as rfm
+import reframe.utility.sanity as sn
+
+
+@rfm.required_version('>=2.16-dev0')
+@rfm.simple_test
+class TensorFlowHorovodTest(rfm.RunOnlyRegressionTest):
+    def __init__(self):
+        super().__init__()
+        self.descr = 'Distributed training with TensorFlow and Horovod'
+        self.valid_systems = ['daint:gpu', 'dom:gpu']
+        self.valid_prog_environs = ['PrgEnv-gnu']
+        tfshortver = '1.11'
+        self.sourcesdir = 'https://github.com/tensorflow/benchmarks'
+        self.modules = ['Horovod/0.15.0-CrayGNU-18.08-tf-%s.0' % tfshortver]
+        self.reference = {
+            'dom:gpu': {
+                'throughput': (1133.6, None, 0.05, 'images/s'),
+            },
+            'daint:gpu': {
+                'throughput': (4403.0, None, 0.05, 'images/s')
+            },
+        }
+        self.perf_patterns = {
+            'throughput': sn.avg(sn.extractall(
+                r'total images/sec:\s+(?P<throughput>\S+)',
+                self.stdout, 'throughput', float))
+        }
+        self.sanity_patterns = sn.assert_found(
+            r'[\S+\s+] INFO NET\/IB : Using interface ipogif0'
+            r' for sideband communication', self.stdout)
+        self.num_tasks_per_node = 1
+        if self.current_system.name == 'dom':
+            self.num_tasks = 8
+        elif self.current_system.name == 'daint':
+            self.num_tasks = 32
+
+        self.pre_run = ['git checkout cnn_tf_v%s_compatible' % tfshortver]
+        self.variables = {
+            'NCCL_DEBUG': 'INFO',
+            'NCCL_IB_HCA': 'ipogif0',
+            'NCCL_IB_CUDA_SUPPORT': '1',
+            'OMP_NUM_THREADS': '$SLURM_CPUS_PER_TASK',
+        }
+        self.executable = ('python')
+        self.executable_opts = [
+            'scripts/tf_cnn_benchmarks/tf_cnn_benchmarks.py',
+            '--model inception3',
+            '--batch_size 64',
+            '--variable_update horovod',
+            '--log_dir ./logs',
+            '--train_dir ./checkpoints']
+        self.tags = {'production'}
+        self.maintainers = ['MS', 'RS']

cscs-checks/cuda/cuda_checks.py

@@ -23,6 +23,7 @@ def __init__(self):
         self.num_gpus_per_node = 1
         self.nvidia_sm = '60'
         if self.current_system.name == 'kesch':
+            self.exclusive_access = True
             self.nvidia_sm = '37'
 
         self.maintainers = ['AJ', 'VK']