apunkt/cuGenOpt
1
0
Fork 0
mirror of https://github.com/L-yang-yang/cugenopt.git synced 2026-04-24 12:06:22 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions
18 commits 1 branch 0 tags 528 KiB
Commit graph

18 commits

This branch
This branch
All branches
Author SHA1 Message Date
L-yang-yang
714a2ee23b feat(multi-gpu): implement inject distribution modes (OneIsland/HalfIslands/AllIslands) 
inject_check_kernel now respects MultiGpuInjectMode from SolverConfig instead of
hardcoding OneIsland. HalfIslands uses LCG-based random island selection.
Also fixes stale write_async calls in test_multi_gpu_b3.cu.

Verified on 2×V100S: all 5 B3 tests pass, e5 (12 problem types) all optimal.
 2026-03-30 20:42:32 +08:00
L-yang-yang
93fda8d900 update readme 2026-03-30 12:04:55 +08:00
LanYangYang
6f9815ecb7
Update README.md 2026-03-25 14:21:25 +08:00
LanYangYang
9a5d855e5b
Update README.md 2026-03-25 14:20:22 +08:00
L-yang-yang
be17ab6e5a add a logo — now it looks like we know what we're doing 😄 2026-03-25 12:20:25 +08:00
L-yang-yang
f19f14b716 add a logo — now it looks like we know what we're doing 😄 2026-03-25 12:19:49 +08:00
L-yang-yang
f0d0969b39 add a logo — now it looks like we know what we're doing 😄 2026-03-25 12:18:59 +08:00
L-yang-yang
292d2ee570 add a logo — now it looks like we know what we're doing 😄 2026-03-25 12:17:41 +08:00
L-yang-yang
98bcc7296d update logo hhh 2026-03-25 12:10:39 +08:00
L-yang-yang
c23e88b6a9 Merge remote-tracking branch 'origin/main' 
# Conflicts:
#	README.md
 2026-03-25 11:53:56 +08:00
L-yang-yang
a848730459 fix: harden CUDA safety checks and translate comments to English 
Safety fixes (4 critical, 4 warning) from code review:

- qap.cuh: fix clone_to_device cross-device D2H by retaining host matrices
- types.cuh: add CUDA_CHECK to InjectBuffer, track owner_gpu for safe destroy
- types.cuh: add bounds check on lexicographic priority index
- solver.cuh: cap migrate_kernel islands to MAX_ISLANDS=64 to prevent stack overflow
- multi_gpu_solver.cuh: guard against 0 GPUs, propagate stop_reason from best GPU
- types.cuh: warn on SeqRegistry overflow
- solver.cuh: warn when constraint_directed/phased_search disabled without AOS

Translate all Chinese comments to English across 25+ source files
(core/*.cuh, problems/*.cuh, Makefile, multi-GPU tests).

Verified on V100S×2 (sm_70, CUDA 12.8): e5 (12 problem types, all optimal),
e13 (multi-objective + multi-GPU, 9 configs, all passed).
 2026-03-25 11:52:50 +08:00
LanYangYang
b128a526d3
Update README.md 2026-03-25 01:53:47 +08:00
L-yang-yang
ab278d0e82 Merge remote-tracking branch 'origin/main' 2026-03-23 15:48:54 +08:00
L-yang-yang
35199b07ad update readme about install 2026-03-23 15:48:45 +08:00
青青草原懒羊羊
3fbb1cdb50
Update README.md 2026-03-22 11:42:41 +08:00
L-yang-yang
339b14291e update arvix url 2026-03-20 12:55:02 +08:00
L-yang-yang
6e95f60649 update arvix url 2026-03-20 12:48:16 +08:00
L-yang-yang
fc5a0ff4af Initial commit: cuGenOpt GPU optimization solver 2026-03-20 00:33:45 +08:00