Skip to content

pra-dan/cosim-using-cuda

Repository files navigation

Notes:

The notes can be found in OneNote > Prashant's Notebook > CUDA. Also attached the notes' pdf here.

Results

On CPU (Intel XEON) - CPU 083 - 7 milliseconds

Generating random vectors...
Computing cosine similarities...
Time taken for cosine similarity computation: 7 milliseconds
Total time taken in main() for cosine similarity computation: 7 milliseconds

Sample similarities (first 5 reference vectors x first 5 test vectors):
-0.0180889      0.00903661      -0.0433678      -0.00477155     -0.0186576
-0.0826939      -0.00924551     -0.0568913      0.0429572       -0.0369354
-0.00529327     -0.00478061     -0.0728236      -0.00633934     -0.0718395
0.0502729       -0.0544535      -0.0537081      -0.0136128      -0.03955
0.002518        0.0559863       0.0574715       0.00593543      0.0264022

Ver 1: Plain - On GPU (4070Ti) - Same CPU

Generating random vectors...
Computing cosine similarities...
Time taken for cosine similarity computation: 4 milliseconds
Total time taken in main() for cosine similarity computation: 94 milliseconds

Sample similarities (first 5 reference vectors x first 5 test vectors):
0.00554452      0.0119299       -0.011022       -0.0284682      -0.00162509
-0.072349       -0.0583075      0.022736        0.024454        -0.116779
0.0469669       0.0470417       0.0816566       0.00964433      -0.00941299
-0.0248692      -0.0126302      -0.0253003      -0.0329331      -0.0729495
0.0489059       0.0503447       0.0296416       -0.0573841      0.0926538

Ver 2: Optimized

Generating random vectors...
Computing cosine similarities...
Time taken for cosine similarity computation: 0 milliseconds
Total time taken in main() for cosine similarity computation: 1 milliseconds

Sample similarities (first 5 reference vectors x first 5 test vectors):
0.0338589       0.0338589       0.0338589       0.0338589       0.0338589
0.0604582       0.0604582       0.0604582       0.0604582       0.0604582
0.020267        0.020267        0.020267        0.020267        0.020267
0.0425228       0.0425228       0.0425228       0.0425228       0.0425228
0.183263        0.183263        0.183263        0.183263        0.183263

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published