Changes between Version 1 and Version 2 of Fast_GEMM_Implementation_On_Cypress


Ignore:
Timestamp:
Nov 8, 2010 10:17:38 AM (14 years ago)
Author:
nakasato
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • Fast_GEMM_Implementation_On_Cypress

    v1 v2  
    2121to heavily use shared memory on GPUs,  we show that texture cache is very effective on the Cypress architecture.  
    2222 
     23== preprint == 
     24 
     25[attachment:Nakasato_PBMS2010.pdf] 
    2326 
    2427== Sample program for DGEMM == 
    … …  
    2932 * [wiki:"GEMM_Performance_Cypress"] 
    3033 * [wiki:"MatrixMultiply"] 
    31  
    32 == preprint == 
    33  
    34 [attachment:Nakasato_PBMS2010.pdf]