Changes between Version 9 and Version 10 of MatrixMultiply


Ignore:
Timestamp:
Jun 7, 2010 10:11:16 PM (14 years ago)
Author:
nakasato
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • MatrixMultiply

    v9 v10  
    11= Matrix Multiply on GPU = 
    2 == Our Results == 
     2We have implemented single/double precision matrix multiply program for RV770/Cypress. In our implementation, we use two input streams. One is transposed input matrix A and other is input matrix B in normal format. Output matrix C is also not transposed. We adopted 8x8 block for single precision and 4x4 for double precision. Here is benchmark result: 
    33 
     4== Single precision == 
     5[[Image(SMM.png)]] 
    46 
    5  
    6  
    7  
    8  
    9  
    10 [[Image(MM1.png)]] 
     7== Double precision == 
     8[[Image(DMM.png)]] 
    119 
    1210= Useful forum discussions = 
    … …  
    2119 
    2220Meta-programing works in reality. 
     21 
     22 
     23[[Image(MM1.png)]]