To get this branch, use:
bzr branch
http://suren.me/webbzr/normxcorr/trunk
- Committer: Suren A. Chilingaryan
- Date: 2009-12-10 16:34:35 UTC
- Revision ID: csa@dside.dyndns.org-20091210163435-43ccu65nj1qi96d2
Compile with --use_fast_math option of nvcc
| Filename | Latest Rev | Last Changed | Committer | Comment | Size | ||
|---|---|---|---|---|---|---|---|
 ..
|
|||||||
INFO |
7 | 14 years ago | Suren A. Chilingaryan | FindPeak optimization | 101 bytes |
|
|
|
local_sum.cu |
16 | 14 years ago | Suren A. Chilingaryan | Optimize FFT size | 3.2 KB |
|
|
|
local_sum.h |
15 | 14 years ago | Suren A. Chilingaryan | First attempt with CUDA streams | 354 bytes |
|
|
|
local_sum_kernel.cu |
9 | 14 years ago | Suren A. Chilingaryan | Avoid bank conflicts in local sum kernels | 1.5 KB |
|
|
|
Makefile |
10 | 14 years ago | Suren A. Chilingaryan | Block computational kernels | 1.6 KB |
|
|
|
normxcorr_hw.cu |
17 | 14 years ago | Suren A. Chilingaryan | Precompute if side and base blocks amount is power | 33.9 KB |
|
|
|
normxcorr_hw.h |
17 | 14 years ago | Suren A. Chilingaryan | Precompute if side and base blocks amount is power | 4 KB |
|
|
|
normxcorr_hw_kernel.cu |
18 | 14 years ago | Suren A. Chilingaryan | Compile with --use_fast_math option of nvcc | 9.6 KB |
|
|
|
normxcorr_hw_msg.h |
1 | 15 years ago | Suren A. Chilingaryan | Initial import | 439 bytes |
|
|
nvmex |
1 | 15 years ago | Suren A. Chilingaryan | Initial import | 44.9 KB |
|
|
|
nvopts.sh |
18 | 14 years ago | Suren A. Chilingaryan | Compile with --use_fast_math option of nvcc | 7.8 KB |
|
|