Changes between Version 26 and Version 27 of MatrixMultiply

Timestamp:

Jun 23, 2010 5:15:38 PM (14 years ago)

Author:

nakasato

Comment:

--

MatrixMultiply

@@ -2 +2 @@
 We have implemented single/double precision matrix multiply program for RV770/Cypress. In our implementation, we use two input streams for computing C=AB. One is transposed input matrix A (i.e. column major) and other is input matrix B in normal format (i.e. row major). Output matrix C is also row major. We adopted 8x8 block for single precision and 4x4 for double precision. Here is benchmark result for each case. Note only kernel execution time is measured.
 
-Add double-double performance. We used 2x2 block in this case. On Cypress architecture GPU, we take advantage of FMA_64 instruction.
+Add double-double precision performance. We used 2x2 block in this case. On Cypress architecture GPU, we take advantage of FMA_64 instruction.
 
 == Peformance Summary ==