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

Add support for SYCL/Data Parallel C++ (DPCPP) #183

Open
wants to merge 48 commits into
base: develop
Choose a base branch
from
Open

Add support for SYCL/Data Parallel C++ (DPCPP) #183

wants to merge 48 commits into from

Commits on Feb 28, 2022

  1. Initial dpcpp commit at a544015 

    Add dpcpp directory and associated kernel files, Makefile changes, host changes to support dpcpp for Intel XeGPUs.
    @emascarenhas
    emascarenhas committed Feb 28, 2022
    Configuration menu
    Copy the full SHA
    e1effc6 View commit details
    Browse the repository at this point in the history

  2. Avoiding conflict with definitions for VERSION and G when using the D… 

    …PCPP compiler.
    @emascarenhas
    emascarenhas committed Feb 28, 2022
    Configuration menu
    Copy the full SHA
    f1a7fac View commit details
    Browse the repository at this point in the history

Commits on Mar 2, 2022

  1. After merge conflict removal and compiling, some tests are passing.

    @emascarenhas
    emascarenhas committed Mar 2, 2022
    Configuration menu
    Copy the full SHA
    29f80ea View commit details
    Browse the repository at this point in the history

  2. Allow input files like 7cpa without receptor names in the protein.map… 

    …s.fld file be valid for processing.
    @emascarenhas
    emascarenhas committed Mar 2, 2022
    Configuration menu
    Copy the full SHA
    92a11ed View commit details
    Browse the repository at this point in the history

  3. Minor cleanup

    @emascarenhas
    emascarenhas committed Mar 2, 2022
    Configuration menu
    Copy the full SHA
    0ad002a View commit details
    Browse the repository at this point in the history

Commits on Mar 3, 2022

  1. Update for DPCPP and XeGPU support.

    @emascarenhas
    emascarenhas committed Mar 3, 2022
    Configuration menu
    Copy the full SHA
    9841c1b View commit details
    Browse the repository at this point in the history

Commits on Apr 7, 2022

  1. Change in Makefile.dpcpp for AOT compile and .gitignore.

    @emascarenhas
    emascarenhas committed Apr 7, 2022
    Configuration menu
    Copy the full SHA
    d03a6ae View commit details
    Browse the repository at this point in the history

Commits on May 4, 2022

  1. Fix include of unnecessary sycl.hpp and dpct include files that break… 

    … openCL build.
    @emascarenhas
    emascarenhas committed May 4, 2022
    Configuration menu
    Copy the full SHA
    4e889b1 View commit details
    Browse the repository at this point in the history

Commits on May 6, 2022

  1. Use G_AD instead of G consistently.

    @emascarenhas
    emascarenhas committed May 6, 2022
    Configuration menu
    Copy the full SHA
    62d742c View commit details
    Browse the repository at this point in the history

Commits on Jun 15, 2022

  1. Use float version of pow function instead of double for performance/s… 

    …upport in HW.
    @emascarenhas
    emascarenhas committed Jun 15, 2022
    Configuration menu
    Copy the full SHA
    6e2683f View commit details
    Browse the repository at this point in the history

Commits on Jul 7, 2022

  1. Use reduce_over_group instead on oneapi extension

    @emascarenhas
    emascarenhas committed Jul 7, 2022
    Configuration menu
    Copy the full SHA
    dfe6a70 View commit details
    Browse the repository at this point in the history

  2. Use XeDeviceSynchronize in DOCK_DEBUG ifdef code

    @emascarenhas
    emascarenhas committed Jul 7, 2022
    Configuration menu
    Copy the full SHA
    820d4c8 View commit details
    Browse the repository at this point in the history

  3. Add support for additional SYCL work group sizes and support to build… 

    … SYCL for NV platforms
    @emascarenhas
    emascarenhas committed Jul 7, 2022
    Configuration menu
    Copy the full SHA
    e26a528 View commit details
    Browse the repository at this point in the history

Commits on Aug 12, 2022

  1. Explicitly use local memory for barriers.

    @emascarenhas
    emascarenhas committed Aug 12, 2022
    Configuration menu
    Copy the full SHA
    0df9039 View commit details
    Browse the repository at this point in the history

Commits on Sep 16, 2022

  1. Fix compile error for CPU.

    @emascarenhas
    emascarenhas committed Sep 16, 2022
    Configuration menu
    Copy the full SHA
    9bdd335 View commit details
    Browse the repository at this point in the history

Commits on Sep 27, 2022

  1. Adding alternative device code supporting sycl::native math functions.

    @L30nardoSV
    L30nardoSV committed Sep 27, 2022
    Configuration menu
    Copy the full SHA
    06039ce View commit details
    Browse the repository at this point in the history

  2. More improvements on alternative device code based on native math.

    Leonardo Solis committed Sep 27, 2022
    Configuration menu
    Copy the full SHA
    0686611 View commit details
    Browse the repository at this point in the history

  3. Disabling native math.

    Leonardo Solis committed Sep 27, 2022
    Configuration menu
    Copy the full SHA
    0fc2e68 View commit details
    Browse the repository at this point in the history

Commits on Sep 28, 2022

  1. Controlling switching between native and non-native math from a uniqu… 

    …e place in code: common/defines.h
    Leonardo Solis committed Sep 28, 2022
    Configuration menu
    Copy the full SHA
    0b048a0 View commit details
    Browse the repository at this point in the history

  2. Defining macros (e.g., SYCL_SQRT, SYCL_SIN, etc) to use same label fo… 

    …r native and non-native calls in device code.
    Leonardo Solis committed Sep 28, 2022
    Configuration menu
    Copy the full SHA
    7269a08 View commit details
    Browse the repository at this point in the history

Commits on Sep 29, 2022

  1. Merge pull request #1 from LeoCollab/add-native-math 

    Add native math support
    @emascarenhas
    emascarenhas authored Sep 29, 2022
    Configuration menu
    Copy the full SHA
    c8b1bca View commit details
    Browse the repository at this point in the history

  2. Using explicitly local memory barriers in <dpcpp/calcMergeEneGra.dp.c… 

    …pp>.
    Leonardo Solis committed Sep 29, 2022
    Configuration menu
    Copy the full SHA
    952f524 View commit details
    Browse the repository at this point in the history

  3. Controlling switching between local and global memory space for barri… 

    …ers in device code.
    @L30nardoSV
    L30nardoSV committed Sep 29, 2022
    Configuration menu
    Copy the full SHA
    35db225 View commit details
    Browse the repository at this point in the history

  4. Adding missing native math switches.

    @L30nardoSV
    L30nardoSV committed Sep 29, 2022
    Configuration menu
    Copy the full SHA
    a4fdae1 View commit details
    Browse the repository at this point in the history

  5. Adding missing native math switches in ADAM kernel. 

    This is just for completeness. In fact, as ADAM is not supported by heuristics, and thus, not required to fully support native math.

Some sycl::pow calls were not transformed into native flavors, although doing so would have been valid.
    @L30nardoSV
    L30nardoSV committed Sep 29, 2022
    Configuration menu
    Copy the full SHA
    16de6fe View commit details
    Browse the repository at this point in the history

Commits on Sep 30, 2022

  1. Merge pull request #2 from LeoCollab/switch-barriers 

    Add switches between local and global memory for barriers + missing native
    @emascarenhas
    emascarenhas authored Sep 30, 2022
    Configuration menu
    Copy the full SHA
    4de77a4 View commit details
    Browse the repository at this point in the history

Commits on Oct 2, 2022

  1. Add Docking time to OpenCL

    @emascarenhas
    emascarenhas committed Oct 2, 2022
    Configuration menu
    Copy the full SHA
    922c188 View commit details
    Browse the repository at this point in the history

Commits on Oct 5, 2022

  1. use wg mem scope for atomics; clang Makefile upd

    @emascarenhas
    emascarenhas committed Oct 5, 2022
    Configuration menu
    Copy the full SHA
    7579e37 View commit details
    Browse the repository at this point in the history

  2. make atomics memory_scope a macro

    @emascarenhas
    emascarenhas committed Oct 5, 2022
    Configuration menu
    Copy the full SHA
    0442ddb View commit details
    Browse the repository at this point in the history

Commits on Nov 23, 2022

  1. Merging fixes from local for AOT compile, etc.

    @emascarenhas
    emascarenhas committed Nov 23, 2022
    Configuration menu
    Copy the full SHA
    6996999 View commit details
    Browse the repository at this point in the history

Commits on Dec 4, 2022

  1. Add support for H100

    @emascarenhas
    emascarenhas committed Dec 4, 2022
    Configuration menu
    Copy the full SHA
    b70da18 View commit details
    Browse the repository at this point in the history

  2. Add LDEBUG_VTUNE

    @emascarenhas
    emascarenhas committed Dec 4, 2022
    Configuration menu
    Copy the full SHA
    1689692 View commit details
    Browse the repository at this point in the history

  3. Add experiment for relaxed memory order for Atomics

    @emascarenhas
    emascarenhas committed Dec 4, 2022
    Configuration menu
    Copy the full SHA
    72e4309 View commit details
    Browse the repository at this point in the history

Commits on Dec 22, 2022

  1. Avoid making PVC default platform.

    @emascarenhas
    emascarenhas committed Dec 22, 2022
    Configuration menu
    Copy the full SHA
    cb6a186 View commit details
    Browse the repository at this point in the history

  2. Add support for ICX sycl compile AOT

    @emascarenhas
    emascarenhas committed Dec 22, 2022
    Configuration menu
    Copy the full SHA
    0089ddf View commit details
    Browse the repository at this point in the history

Commits on Dec 30, 2022

  1. Adding optimization flags.

    @L30nardoSV
    L30nardoSV committed Dec 30, 2022
    Configuration menu
    Copy the full SHA
    f82074c View commit details
    Browse the repository at this point in the history

Commits on Jan 5, 2023

  1. Replacing sycl::accessor<...target:local> with sycl::local_accessor<.… 

    …..>.
    @L30nardoSV
    L30nardoSV committed Jan 5, 2023
    Configuration menu
    Copy the full SHA
    50fabd2 View commit details
    Browse the repository at this point in the history

  2. Removing commented-out code.

    @L30nardoSV
    L30nardoSV committed Jan 5, 2023
    Configuration menu
    Copy the full SHA
    86b9a7d View commit details
    Browse the repository at this point in the history

  3. Merge pull request #1 from LeoCollab/remove-warns 

    Replacing sycl::accessor<... target:local> with sycl::local_accessor<...>
    @L30nardoSV
    L30nardoSV authored Jan 5, 2023
    Configuration menu
    Copy the full SHA
    ea3e2b8 View commit details
    Browse the repository at this point in the history

  4. Merge pull request #5 from LeoCollab/develop 

    Adding optimization flags.
    @emascarenhas
    emascarenhas authored Jan 5, 2023
    Configuration menu
    Copy the full SHA
    a00cd5a View commit details
    Browse the repository at this point in the history

Commits on Jan 6, 2023

  1. Forcing maximum number of registers per thread with compile option: -… 

    …Xcuda-ptxas --maxrregcount=64.
    @L30nardoSV
    L30nardoSV committed Jan 6, 2023
    Configuration menu
    Copy the full SHA
    9aeabe1 View commit details
    Browse the repository at this point in the history

Commits on Jan 18, 2023

  1. Merge pull request #6 from LeoCollab/fix-numreg-per-thread 

    Reducing register pressure via additional compiler option
    @emascarenhas
    emascarenhas authored Jan 18, 2023
    Configuration menu
    Copy the full SHA
    9750c01 View commit details
    Browse the repository at this point in the history

  2. Print best energy even if autostop is 0.

    @emascarenhas
    emascarenhas committed Jan 18, 2023
    Configuration menu
    Copy the full SHA
    fc8f57c View commit details
    Browse the repository at this point in the history

Commits on Jan 23, 2023

  1. Print best energy when autostop is false for cuda and opencl.

    @emascarenhas
    emascarenhas committed Jan 23, 2023
    Configuration menu
    Copy the full SHA
    bf07b90 View commit details
    Browse the repository at this point in the history

Commits on Jan 25, 2023

  1. Replacing pow() with native counterpart.

    @L30nardoSV
    L30nardoSV committed Jan 25, 2023
    Configuration menu
    Copy the full SHA
    64b52ea View commit details
    Browse the repository at this point in the history

  2. Merge pull request #7 from LeoCollab/develop 

    Replacing pow() with native counterpart
    @emascarenhas
    emascarenhas authored Jan 25, 2023
    Configuration menu
    Copy the full SHA
    56f66a9 View commit details
    Browse the repository at this point in the history

Commits on Feb 22, 2023

  1. Remove hard set of sub_group size for reduction kernel.

    @emascarenhas
    emascarenhas committed Feb 22, 2023
    Configuration menu
    Copy the full SHA
    66d75e1 View commit details
    Browse the repository at this point in the history

Commits on Apr 13, 2023

  1. Updates to README for SYCL

    @emascarenhas
    emascarenhas committed Apr 13, 2023
    Configuration menu
    Copy the full SHA
    480b4de View commit details
    Browse the repository at this point in the history