991663feff
Avoid (harmless) errors when shutting down the GPU.
2024-01-18 22:42:32 -05:00
9588ddabf6
Added the GPU versions of lj/cut/coul/cut/soft and lj/cut/coul/long/soft
2023-12-05 13:30:26 -06:00
c96ac858bf
GPU Package: Adding JIT test for OpenCL prefetch support.
2023-03-07 21:43:19 -08:00
37f22c8627
Misc Improvements to GPU Package
...
- Optimizations for molecular systems
- Improved kernel performance and greater CPU overlap
- Reduced GPU to CPU communications for discrete devices
- Switch classic Intel makefiles to use LLVM-based compilers
- Prefetch optimizations supported for OpenCL
- Optimized data repack for quaternions
2023-03-05 21:03:12 -08:00
4244d2e6cd
silence compiler warnings about unused parameters and variables
2023-01-19 08:56:54 -05:00
e5a808fb8d
apply correct platform selection for OpenCL context
2022-11-01 04:05:57 -04:00
80a141d9c8
silence compiler warnings
2022-11-01 03:38:08 -04:00
51c6eddd0d
Fix to make the property list empty for command queues when timing disabled.
2022-10-19 02:15:39 -04:00
7c9666798e
whitespace
2022-10-08 09:34:20 -04:00
7551c0a3ca
GPU Package: Documenting some additional preprocessor flags, updating oneapi Makefile.
2022-10-07 22:44:21 -07:00
a6a39d47e1
Fixing potential issues with automatic splitting of accelerators for NUMA.
2022-10-06 20:48:02 -07:00
e9f39f85d2
Fixing issue where shared main memory property only set for NVIDIA devices.
2022-10-06 20:05:33 -07:00
fb675028b9
whitespace
2022-09-29 02:42:11 -04:00
71464d8314
GPU Package: Fixing logic in OpenCL backend that could result in unnecessary device allocations.
2022-09-28 22:30:09 -07:00
6e34d21b24
GPU Package: Switching back to timer disabling with multiple MPI tasks per GPU. Logic added to prevent mem leak.
2022-09-28 21:02:16 -07:00
167abe9ce0
add preprocessor flags to select between the changed and the old code variant
2022-09-09 12:41:24 -04:00
294a1c2168
Use primary context in CUDA GPU code.
...
Since LAMMPS uses the low-level driver API of CUDA, it needs to ensure
that it is in the correct context when invoking such functions. At the
moment it creates and switches to its own context inside `UCL_Device::set`
but then assumes that the driver is still in that context for subsequent
calls into CUDA; if another part of the program uses a different context
(such as the CUDA runtime using the "primary" context) this will cause
failures inside LAMMPS.
This patch changes the context creation to instead use the primary
context for the requested device. While it's not perfect, in that it
still doesn't ensure that it's in the correct context before making
driver API calls, it at least allows it to work with libraries that use
the runtime API.
2022-09-06 09:28:51 +10:00
d8f8a3a36a
Handle inconsistent J molecules in tip4p/gpu
2022-05-28 00:24:24 +03:00
f3363070e7
remove support for CUDA toolkits before version 8 and GPUs older than Kepler
2022-04-12 15:48:16 -04:00
b36262cd11
cosmetic
2022-04-10 20:47:42 -04:00
537e3f2539
silence compiler warnings
2022-02-17 23:13:13 -05:00
e990a1cf61
remove ambiguity between "double_precision" class member variable and function
2021-10-16 21:07:04 -04:00
dd6e3c1acc
avoid variable length array and signed vs. unsigned warnings
2021-10-08 20:07:20 -04:00
9da8c932ab
make check more obvious
2021-09-25 21:33:10 -04:00
530912a930
detect double precision support according to OpenCL specs (1.2 and later)
2021-09-25 07:20:52 -04:00
cd3efc3fa8
initialize all members
2021-09-17 22:45:26 -04:00
00e396c921
move misplaced #endif and make code more readable
2021-09-16 07:33:24 -04:00
08eb3345af
silence some compiler warnings
2021-08-31 15:11:46 -04:00
49a81d6fba
silence compiler warnings
2021-08-26 06:30:16 -04:00
92b5b159e5
replace tabs and remove trailing whitespace in lib folder with updated script
2021-08-22 20:45:24 -04:00
8ea9179a00
Correct device pointer datatype in HIP UCL_Const
2021-06-02 15:46:57 -04:00
62423b0c75
Disable new GPU neighbor for HIP
2021-06-02 15:45:06 -04:00
b4a70880d9
Merge pull request #2745 from rbberger/gpu_updates
...
Updates to the GPU package
2021-04-30 19:32:44 -04:00
375f8e7054
support loading OpenCL functions up to the OpenCL 3.0 standard
2021-04-30 09:28:01 -04:00
8dfcdc4b3c
Revert "activate workaround for problems with new GPU code neighbor lists"
...
This reverts commit a28d9a6e69
2021-04-29 19:24:28 -04:00
24314b2316
Formatting
2021-04-23 16:36:01 -04:00
8ba1b59d8d
Correct fprintf statement in debug code
2021-04-23 16:34:12 -04:00
0f1f49afa7
Add more output to ocl_get_devices
2021-04-22 12:52:30 -04:00
a28d9a6e69
activate workaround for problems with new GPU code neighbor lists
2021-03-11 22:06:10 -05:00
70ca055113
Add workaround to compile with HIP, new neighbor code still needs to be ported
2021-02-23 11:28:05 -05:00
2a0b7a26c0
Add missing cq() getter
2021-02-23 11:26:07 -05:00
446a068159
Use hipGetDeviceProperties to avoid int overflows
2021-02-23 10:14:05 -05:00
1e5a73c468
silence warnings when using default OpenCL headers. Pick OpenCL v2.1 as default.
2021-02-17 17:36:35 -05:00
e7e2d2323b
Feb2021 GPU Package Update - GPU Package Files
2021-02-15 08:20:50 -08:00
f1ef7d85a8
T2345: Replace instances of NULL with nullptr
...
The following changes have been applied to src and lib folders:
regex replace: ([^"_])NULL ⇒ \1nullptr (8968 chgs in src, 1153 in lib)
Manually find/change: (void \*) nullptr ⇒ nullptr (1 case)
regex find: ".*?nullptr.*?"
Manually ~14 cases back to "NULL" in src, ~2 in lib
regex finds a few false positive where nullptr appears between two
strings in a function call
2020-09-12 09:34:38 -06:00
db47eeb2a9
Fix handling of CL_DEVICE_TYPE flags
2020-07-31 19:21:32 +03:00
4f6dd381c3
Merge branch 'master' into gpu_hip
2020-03-30 17:15:55 -04:00
5a3a5d86d4
reduce compiler warnings and avoid clash with KOKKOS
2020-03-05 14:27:33 -05:00
66c5fa2abd
Merge 'gpu_hip_port' into master
2020-01-28 20:35:08 +03:00
66a473455e
silence warnings from generic OpenCL header files (as seen on Fedora 31)
2020-01-06 17:34:26 -05:00
