Add a new SYCL backend for supporting Intel GPUs and embedded GPUs - #197

Merged

qliu21 merged 18 commits into

CODARcode:masterfrom

JieyangChen7:mgard-x

Jul 11, 2022

Commits on Jul 5, 2022

Working on SYCL device adapter
JieyangChen7
committed
eb65dce
Add SYCL based class instantiations
JieyangChen7
committed
20db729
Remove self-defined copy constructors and destructor of SubArray to make functors copyable to SYCL devices
JieyangChen7
committed
26acd75
Remove unused files
JieyangChen7
committed
788da38
Adjust macros for SYCL backend
JieyangChen7
committed
351d7b5
Fix queue index in decompose and recompose functions
JieyangChen7
committed
dc0c224
Add missing device sync in GenerateCL, GenerateCW, Histogram & Extend atomic operations to support multiple scopes and memory spaces
JieyangChen7
committed
d81fa66
Update SYCL backend to let it query device memory size
JieyangChen7
committed
00ff9bc
Add MGARD-X source code auto generation function
JieyangChen7
committed
dc1e6e9
Refactor building units to speed up building and reduce memory footprint
JieyangChen7
committed
9aa99bf
Remove dependecy on DPCpp library & Add build scripts for JLSE machines
JieyangChen7
committed
d73c5f8
Add print of Device Name
JieyangChen7
committed
878b903
Add adaptive block size config for SYCL
JieyangChen7
committed
e44646c
Add missing sync in low level compression
JieyangChen7
committed
5e88378
Improve adaptive config selection & Fix autotuning for SYCL
JieyangChen7
committed
4bd2edb
Add atomicMax for double&float for CUDA backend
JieyangChen7
committed
877f179
Fix conflict in SubArray
JieyangChen7
committed
47e2e70

Commits on Jul 6, 2022

Turn off clang-format for certain *.in files in mgard-x
JieyangChen7
committed
6b45e5c