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

Open

emascarenhas wants to merge 48 commits into

ccsb-scripps:developfrom

emascarenhas:develop

Commits on Feb 28, 2022

Initial dpcpp commit at a544015a5
emascarenhas
committed
Avoiding conflict with definitions for VERSION and G when using the DPCPP compiler.
emascarenhas
committed

Commits on Mar 2, 2022

After merge conflict removal and compiling, some tests are passing.
emascarenhas
committed
Allow input files like 7cpa without receptor names in the protein.maps.fld file be valid for processing.
emascarenhas
committed
Minor cleanup
emascarenhas
committed

Commits on Mar 3, 2022

Update for DPCPP and XeGPU support.
emascarenhas
committed

Commits on Apr 7, 2022

Change in Makefile.dpcpp for AOT compile and .gitignore.
emascarenhas
committed

Commits on May 4, 2022

Fix include of unnecessary sycl.hpp and dpct include files that break openCL build.
emascarenhas
committed

Commits on May 6, 2022

Use G_AD instead of G consistently.
emascarenhas
committed

Commits on Jun 15, 2022

Use float version of pow function instead of double for performance/support in HW.
emascarenhas
committed

Commits on Jul 7, 2022

Commits on Aug 12, 2022

Explicitly use local memory for barriers.
emascarenhas
committed

Commits on Sep 16, 2022

Fix compile error for CPU.
emascarenhas
committed

Commits on Sep 27, 2022

Adding alternative device code supporting sycl::native math functions.
L30nardoSV
committed
More improvements on alternative device code based on native math.
Leonardo Solis
committed
Disabling native math.
Leonardo Solis
committed

Commits on Sep 28, 2022

Commits on Sep 30, 2022

Merge pull request #2 from LeoCollab/switch-barriers
emascarenhas
authored

Commits on Oct 2, 2022

Add Docking time to OpenCL
emascarenhas
committed

Commits on Oct 5, 2022

Commits on Nov 23, 2022

Merging fixes from local for AOT compile, etc.
emascarenhas
committed

Commits on Dec 4, 2022

Commits on Dec 22, 2022

Commits on Dec 30, 2022

Adding optimization flags.
L30nardoSV
committed

Commits on Jan 6, 2023

Forcing maximum number of registers per thread with compile option: -Xcuda-ptxas --maxrregcount=64.
L30nardoSV
committed

Commits on Jan 18, 2023

Commits on Jan 23, 2023

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

Commits on Jan 25, 2023

Commits on Feb 22, 2023

Remove hard set of sub_group size for reduction kernel.
emascarenhas
committed

Commits on Apr 13, 2023

Updates to README for SYCL
emascarenhas
committed