Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

OpenCL verbose output and documentation, improved auto-tuning scripts, minor fixes after #419 #425

Merged
merged 23 commits into from
Feb 8, 2021
Merged

OpenCL verbose output and documentation, improved auto-tuning scripts, minor fixes after #419 #425

merged 23 commits into from
Feb 8, 2021

Commits on Feb 4, 2021

  1. OpenCL-BE/LIBSMM: verbose output and documentation. Improved auto-tun… 

    …ing scripts. Minor fixes after cp2k#419.

* Introduced (runtime-)verbosity level. Print device name (non-zero verbosity).
* Fixed issue (cp2k#419 (comment)).
* Renamed ACC_OPENCL_VERBOSE to ACC_OPENCL_DEBUG.
* ACC benchmark drivers: inform if no device was found.
* Improved documentation and documented ACC_OPENCL_VERBOSE.
* Introduced verbose output (time needed for kernel compilation, etc).
* tune_multiply.py: option to only rely on primary objective.
* tune_multiply.py: catch CTRL-C and save configuration.
* tune_multiply.sh: relay result code of failing script.
* tune_multiply.sh: continuation with wrapper script.
    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    1ce2427 View commit details
    Browse the repository at this point in the history

  2. Fixed Makefile used to build acc_bench_trans/acc_bench_smm with CUDA … 

    …(accommodate changes from cp2k#419).
    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    69a0e85 View commit details
    Browse the repository at this point in the history

  3. Disabled ACC_OPENCL_THREADLOCAL_CONTEXT since DBCSR calls init/finali… 

    …ze in a parallel region which makes this code ineffective.
    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    3bf854c View commit details
    Browse the repository at this point in the history

  4. Updated LIBXSMM prior to v1.17.

    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    f14eee4 View commit details
    Browse the repository at this point in the history

  5. Attempt to runtime-test OpenCL BE/LIBSMM.

    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    9598017 View commit details
    Browse the repository at this point in the history

  6. Reduced console output to potentially improve runtime of (CI-)tests.

    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    6721baa View commit details
    Browse the repository at this point in the history

  7. Increased timeout from 15m to 20m.

    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    ee12e07 View commit details
    Browse the repository at this point in the history

  8. Fetch all commits before referring to some SHA.

    @hfp
    hfp committed Feb 4, 2021
    1 Configuration menu
    Copy the full SHA
    8baf7ae View commit details
    Browse the repository at this point in the history

  9. Revert "Attempt to runtime-test OpenCL BE/LIBSMM." 

    This reverts commit 9598017.
    @hfp
    hfp committed Feb 4, 2021
    Configuration menu
    Copy the full SHA
    2b9335f View commit details
    Browse the repository at this point in the history

Commits on Feb 5, 2021

  1. Revised enabling ACC_OPENCL_THREADLOCAL_CONTEXT.

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    20f9d25 View commit details
    Browse the repository at this point in the history

  2. Repeated note about combining auto-tuned parameters for SP and DP in … 

    …one application.
    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    a26e779 View commit details
    Browse the repository at this point in the history

  3. Only print device name if the device changed (and avoid duplicated ve… 

    …rbose output).
    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    cb91474 View commit details
    Browse the repository at this point in the history

  4. Removed tabs from source file (minor/unrelated change).

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    9221d1b View commit details
    Browse the repository at this point in the history

  5. More prefixes in follow-up of cp2k#419 (c_dbcsr_).

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    8ac6f1d View commit details
    Browse the repository at this point in the history

  6. Supply platform when forming context.

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    b58a37b View commit details
    Browse the repository at this point in the history

  7. Code cleanup.

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    6f3c910 View commit details
    Browse the repository at this point in the history

  8. Try to avoid MPS issue (temporarily) testing with only one rank. Sync… 

    … up with cp2k#428.
    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    b5cb129 View commit details
    Browse the repository at this point in the history

  9. Enabled OpenCL based runtime tests.

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    367a117 View commit details
    Browse the repository at this point in the history

  10. Fixed CI-scripts.

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    6c9f84c View commit details
    Browse the repository at this point in the history

  11. Fixed another variable which was left unbound (CI-script).

    @hfp
    hfp committed Feb 5, 2021
    Configuration menu
    Copy the full SHA
    9ff03d9 View commit details
    Browse the repository at this point in the history

Commits on Feb 8, 2021

  1. Incorporated cp2k#428.

    @hfp
    hfp committed Feb 8, 2021
    Configuration menu
    Copy the full SHA
    d17f03f View commit details
    Browse the repository at this point in the history

  2. Merge branch 'develop' of https://github.com/cp2k/dbcsr into oclverbose 

    # Conflicts:
#	.ci/daint.cscs.ch/ocl.build.sh
#	.ci/daint.cscs.ch/ocl.test.sh
    @hfp
    hfp committed Feb 8, 2021
    Configuration menu
    Copy the full SHA
    1265b26 View commit details
    Browse the repository at this point in the history

  3. Warn about potentially exclusive device-mode.

    @hfp
    hfp committed Feb 8, 2021
    Configuration menu
    Copy the full SHA
    2e57682 View commit details
    Browse the repository at this point in the history