Hi all,
I've attached Emanuele's text notes for the meeting.
Cheers,
Matt
________________________________________
From: Testbed Support for GridPP member institutes <[log in to unmask]> on behalf of Alastair Dewhurst <[log in to unmask]>
Sent: 03 November 2022 11:00
To: [log in to unmask]
Subject: [External] [Indico] [Event reminder] GridPP Technical Meeting - HEPScore (11/4/22, 11:00 AM Europe/London)
This email originated outside the University. Check before clicking links or attachments.
Please note that the event "GridPP Technical Meeting - HEPScore" will start on Nov 4, 2022, 11:00 AM (Europe/London).
It will take place at Virtual Only.
You can access the full event here:
https://indico.cern.ch/e/1205334
Agenda for GridPP Technical Meeting - HEPScore
----------------------------------------------
- [11:00 AM - 11:15 AM] Introduction to HEPScore
- [11:15 AM - 11:30 AM] Tier-1 procurements with HEPScore
- [11:30 AM - 12:00 PM] HEPScore usage with GridPP - Discussion
--
Indico :: Email Notifier
https://indico.cern.ch/e/1205334
########################################################################
To unsubscribe from the TB-SUPPORT list, click the following link:
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.jiscmail.ac.uk%2Fcgi-bin%2FWA-JISC.exe%3FSUBED1%3DTB-SUPPORT%26A%3D1&data=05%7C01%7Cdoidgem%40live.lancs.ac.uk%7Cd46dc374a4d44eb7044108dabd8ab672%7C9c9bcd11977a4e9ca9a0bc734090164a%7C0%7C0%7C638030700976920591%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=aW%2BKssXNn67LjYQZGk8d3piIwnTlu2HC%2Fcil0wpEX3k%3D&reserved=0
This message was issued to members of https://eur02.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.jiscmail.ac.uk%2FTB-SUPPORT&data=05%7C01%7Cdoidgem%40live.lancs.ac.uk%7Cd46dc374a4d44eb7044108dabd8ab672%7C9c9bcd11977a4e9ca9a0bc734090164a%7C0%7C0%7C638030700976920591%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Z4tXxM93E%2FMfiSKh5iVE2LlcDYnfctAqyrVHjV0StnA%3D&reserved=0, a mailing list hosted by https://eur02.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.jiscmail.ac.uk%2F&data=05%7C01%7Cdoidgem%40live.lancs.ac.uk%7Cd46dc374a4d44eb7044108dabd8ab672%7C9c9bcd11977a4e9ca9a0bc734090164a%7C0%7C0%7C638030700976920591%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=SVqUUnZsa10k9w%2BnglolEkZ3t9I0kKnEwKiNtA3UZ30%3D&reserved=0, terms & conditions are available at https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.jiscmail.ac.uk%2Fpolicyandsecurity%2F&data=05%7C01%7Cdoidgem%40live.lancs.ac.uk%7Cd46dc374a4d44eb7044108dabd8ab672%7C9c9bcd11977a4e9ca9a0bc734090164a%7C0%7C0%7C638030700976920591%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=UPfgDX5f8NdT2ffruqgcovWcimXxfME1YuZfctkeDYo%3D&reserved=0
########################################################################
To unsubscribe from the TB-SUPPORT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=TB-SUPPORT&A=1
This message was issued to members of www.jiscmail.ac.uk/TB-SUPPORT, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
HEP-Score within the HEP-Benchmarking suite:
--------------------------------------------
HEPscore PPT: https://indico.cern.ch/event/1096028/contributions/4761493/attachments/2404581/4114747/GDB-09-03-2022-giordano.pdf
HEPScore GitHub (HEP-score & HEP-workloads): https://gitlab.cern.ch/hep-benchmarks/hep-score
https://gitlab.cern.ch/hep-benchmarks/hep-workloads
---
Let's run on a standard workernode (wn-d22-010):
ssh -A wn-d22-010
First of all, stop HTCondor:
sudo systemctl stop condor
systemctl status condor --> Active: inactive (dead) since Tue 2022-10-18 17:26:46 BST; 4s ago
Make sure all the necessary ingredients are there:
(Python 3.6 , Pip3 , Singularity)
singularity --version --> singularity version 3.8.7-1.el7
python3 --version --> Python 3.6.8
pip3 --version --> pip 21.3.1 from /home/peppo/.local/lib/python3.6/site-packages/pip (python 3.6)
git --version --> git version 1.8.3.1
Chose a folder to execute the job:
(e.g., /tmp/ or /var/lib/condor/execute/)
sudo mkdir /var/lib/condor/execute/hstest/
sudo chown peppo /var/lib/condor/execute/hstest
sudo chgrp wheel /var/lib/condor/execute/hstest
cd /var/lib/condor/execute/hstest/
Install the HEP Benchmark Suite: https://gitlab.cern.ch/hep-benchmarks/hep-benchmark-suite/
git clone https://gitlab.cern.ch/hep-benchmarks/hep-benchmark-suite.git
cd hep-benchmark-suite/
pip3 install --user .
export PATH=$PATH:~/.local/bin
cd ..
---
Alternative version:
wget https://hep-benchmarks.web.cern.ch/hep-benchmark-suite/releases/v2.1/hep-benchmark-suite-wheels-py37-v2.1.tar
tar xvf hep-benchmark-suite-wheels-py37-v2.1.tar
pip3 install suite_wheels/* --> Successfully installed PyYAML-5.4.1 hep-benchmark-suite-2.1 hep-score-1.0.0
importlib-metadata-4.5.0 pbr-5.6.0 setuptools-44.1.1 typing-extensions-3.10.0.0 zipp-3.4.1
---
Check versions:
bmkrun --version --> v2.1
hep-score --version --> hep-score 1.0.0
Install (and upgrade) the HEP-Score package:
git clone https://gitlab.cern.ch/hep-benchmarks/hep-score.git
cd hep-score
pip3 install --user . --> Successfully installed hep-score-1.4.0 pbr-5.9.0 pyyaml-6.0
hep-score -p --> hepscore_benchmark: ...
cd ..
Ok, the packages are installed and PATH has been updated.
Let's update the configuration and run it (see "How to run" on the Git page):
mkdir workdir
cd workdir/
vi hep_benchmark.yml <-- publish: false
config: ./hepscore_conf.yml
vi hepscore_conf.yml <-- reference_machine: "CPU 2x AMD EPYC 7513 32-Core Processors HT @ 2.60 GHz (x86_64)" # wn-d22
bmkrun -c hep_benchmark.yml
Or, better run with 'nohup':
nohup bmkrun -c hep_benchmark.yml &
It takes a few hours, depending on how many jobs are enabled in the config file.
When done, check the output:
ll run_2022-11-04_*/HEPSCORE/atlas-gen-bmk/atlas-gen_bmk_summary.json
Make sure you have a Grid Certificate available:
ls -latrh /home/peppo/.globus/ --> -r--r--r--. 1 peppo wheel 1.9K Jul 14 09:33 usercert.pem
-r--r--r--. 1 peppo wheel 2.0K Jul 14 09:33 userkey.pem
And publish the results (${WORKDIR}/${MYENV}/bin/python3 ${WORKDIR}/${MYENV}/lib/python3.6/site-packages/hepbenchmarksuite/plugins/send_queue.py --port=61123 --server=dashb-mb.cern.ch --topic /topic/vm.spec --key $CERTIFKEY --cert $CERT --file $RESULTS):
find ~/. -name send_queue.py --> /home/peppo/./.local/lib/python3.6/site-packages/hepbenchmarksuite/plugins/send_queue.py
export PATH=$PATH:/home/peppo/.local/lib/python3.6/site-packages/hepbenchmarksuite/plugins/
export CERTKEY=/home/peppo/.globus/userkey.pem
export CERT=/home/peppo/.globus/usercert.pem
python3 /home/peppo/.local/lib/python3.6/site-packages/hepbenchmarksuite/plugins/send_queue.py --port=61123 --server=dashb-mb.cern.ch --topic /topic/vm.spec --key $CERTKEY --cert $CERT --file run_2022-07-20_0842/bmkrun_report.json
Done. Results will be visible on the CERN Kibana dashboard:
https://es-hep-bmk.cern.ch/kibana/app/dashboards#
--- --- ---
HEP-Score as standalone packages:
---------------------------------
List of available containers (x86 & ARM): https://gitlab.cern.ch/hep-benchmarks/hep-workloads-sif/container_registry
Good ones for x86 and ARM (October 2022):
hep-benchmarks/hep-workloads-sif/hello-world-ma-bmk --> ci-v1.0_aarch64
ci-v1.0_x86_64
hep-benchmarks/hep-workloads-sif/atlas-gen_sherpa-ma-bmk --> ci-v1.0_aarch64
ci-v1.0_x86_64
hep-benchmarks/hep-workloads-sif/atlas-sim_mt-ma-bmk --> v2.0_aarch64 (latest)
v2.0_x86_6
hep-benchmarks/hep-workloads-sif/cms-reco-run3-ma-bmk --> v1.1_aarch64 (ci-v1.0, v1.0, latest)
v1.1_x86_64
hep-benchmarks/hep-workloads-sif/cms-digi-run3-ma-bmk --> v1.0_aarch64 (ci-v1.0, latest)
v1.0_x86_64
hep-benchmarks/hep-workloads-sif/cms-gen-sim-run3-ma-bmk --> v1.0_aarch64 (ci-v0.1, latest)
v1.0_x86_64
---
Run standalone container on 'x86' test node:
ssh -A amd-e22
mkdir -p /tmp/test
chmod a+rw /tmp/test
cd /tmp/test/
vi hepscore_conf.yml <-- reference_machine: "CPU AMD EPYC 7643 48-Core Processor HT @ 2.30 GHz (x86_64)"
Check that the container exists:
singularity run -B /tmp/test:/results oras://gitlab-registry.cern.ch/hep-benchmarks/hep-workloads-sif/cms-reco-run3-ma-bmk:v1.0_x86_64 -h
And run it (with nohup):
nohup singularity run -B /tmp/test:/results oras://gitlab-registry.cern.ch/hep-benchmarks/hep-workloads-sif/cms-reco-run3-ma-bmk:v1.0_x86_64 &
Check output:
cat cms-reco-run3-ma-*/cms-reco-run3-ma_summary.json
exit
Run standalone container on 'arm' test node:
ssh -A arm-e22
mkdir -p /tmp/test
chmod a+rw /tmp/test
cd /tmp/test/
vi hepscore_conf.yml <-- reference_machine: "CPU Ampere ARMv8 Neoverse-N1 80-Core Processor @ 3.00 GHz (aarch64)"
Check that the container exists:
singularity run -B /tmp/test:/results oras://gitlab-registry.cern.ch/hep-benchmarks/hep-workloads-sif/cms-reco-run3-ma-bmk:v1.0_aarch64 -h
And run it (with nohup):
nohup singularity run -B /tmp/test:/results oras://gitlab-registry.cern.ch/hep-benchmarks/hep-workloads-sif/cms-reco-run3-ma-bmk:v1.0_aarch64 &
Check output:
cat cms-reco-run3-ma-*/cms-reco-run3-ma_summary.json
exit
Done.
--- --- ---
########################################################################
To unsubscribe from the TB-SUPPORT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=TB-SUPPORT&A=1
This message was issued to members of www.jiscmail.ac.uk/TB-SUPPORT, a mailing list hosted by www.jiscmail.ac.uk, terms & conditions are available at https://www.jiscmail.ac.uk/policyandsecurity/
|