davilizh
|
41c8e69a8f
|
Change the code according to Chafast's suggestions in pull request #18. 1) change if..else if..else if to switch. 2)change spaces to tab. 3) add additioanl arg check. 4) delete commented out code. 5) throw any exception to avoid code duplication.
|
8 years ago |
|
davilizh
|
73fc65daf9
|
add switch --cuda-parallel-hash to enable and disable the parallel-hash optimization
|
8 years ago |
David Li (Engrg-Hardware 1)
|
de2c1e3485
|
delete dead code resulting from launch_bound deleting
|
8 years ago |
|
David Li (Engrg-Hardware 1)
|
2f945b2216
|
1. delete launch bound in ethash_cuda_miner_kernel.cu 2. re-format the for loop in keccak.cuh
|
8 years ago |
|
David Li (Engrg-Hardware 1)
|
63ec7023b8
|
The optimized code by Nvidia Architecturer. the performance is improved from 'min/mean/max: 22369621/22579336/22719146 H/s' to 'min/mean/max: 23767722/23907532/24117248 H/s' on a flashed GTX 1060 with 2 GPCs 9 TPCs (the product chip should have 10 TPCs). Note that the code is tested on the code pulled from May-11. The current code from github cannot generate reasonable scores ('min/max/avg is 0/0/0 H/s')
|
8 years ago |
Genoil
|
9196115f23
|
mixed AMD/Nvidia mining
|
9 years ago |
|
Genoil
|
148d71a5cb
|
synchronous stratum client test (wip)
|
9 years ago |
|
Genoil
|
ed2e888bed
|
some fixes, cuda 8 rc build
|
9 years ago |
|
Genoil
|
19459aa2b2
|
on-GPU DAG generation
|
9 years ago |
|
Genoil
|
351343d365
|
tpb
|
9 years ago |
|
Genoil
|
89cedad4a6
|
working cuda kernel with shared mem
|
9 years ago |
|
Genoil
|
41043b3c04
|
shorter work mutex lock
longer work timeout
binary
|
9 years ago |
|
Genoil
|
4e2f2f37a4
|
relaxed local work size requirements
added mutex to stratum pending calls. untested
|
9 years ago |
Jan Willem Penterman
|
6b091ff10b
|
work in progress
|
9 years ago |
RoBiK
|
d8c8582f94
|
improved error handling and usage of constant memory space
|
10 years ago |
|
RoBiK
|
3aa41b10ea
|
fixed m_search_buf races, removed unnecessary cudaMemcpyAsync calls, fixed missing cudaStreamSynchronize
|
10 years ago |
|
Jan Willem Penterman
|
e46de57509
|
renamed "cu" -> "cuda" globally
|
10 years ago |
|
Jan Willem Penterman
|
30bfe53bad
|
end of day commit..nothing works yet
|
10 years ago |
