BASALT: nucleotide sequence similarity

Hello, all.

I’ve written my first CUDA project. It’s called BASALT, and it can be used to compute the similarity between nucleotide sequences.

[url=“http://coulouris.org/basalt/”]http://coulouris.org/basalt/[/url]

I’ve been testing on a 9400m (2.26ghz c2d macbook pro, osx 10.6, cuda 3.0 beta); can anyone give it a try on a non-mobile GPU?

USE_ZEROCOPY=0
BASALT_THREADS_PER_BLOCK=64
BASALT_BV_SIZE=2048
BASALT_NUM_QUERY_TILES=512
BASALT_NUM_SUBJECT_TILES=512
device 0 = GeForce 9400M
gpu processing time : 880.122986 (ms)
memory bandwidth : 4.544819 (GB/s)
cpu processing time : 2696.284912 (ms)
Test PASSED

USE_ZEROCOPY=0
BASALT_THREADS_PER_BLOCK=64
BASALT_BV_SIZE=2048
BASALT_NUM_QUERY_TILES=512
BASALT_NUM_SUBJECT_TILES=512
device 0 = GeForce 9600M GT
gpu processing time : 486.269989 (ms)
memory bandwidth : 8.225883 (GB/s)
cpu processing time : 2359.349121 (ms)
Test PASSED

USE_ZEROCOPY=0
BASALT_THREADS_PER_BLOCK=64
BASALT_BV_SIZE=2048
BASALT_NUM_QUERY_TILES=512
BASALT_NUM_SUBJECT_TILES=512
device 0 = GeForce 9800 GT
gpu processing time : 141.066162 (ms)
memory bandwidth : 28.355489 (GB/s)
cpu processing time : 3309.087891 (ms)
Test PASSED

USE_ZEROCOPY=0
BASALT_THREADS_PER_BLOCK=64
BASALT_BV_SIZE=2048
BASALT_NUM_QUERY_TILES=512
BASALT_NUM_SUBJECT_TILES=512
device 0 = Tesla C1060
gpu processing time : 59.300999 (ms)
memory bandwidth : 67.452490 (GB/s)
cpu processing time : 2752.378906 (ms)
Test PASSED