Browse by Tags

Cache Locality For Performance
22 June 09 06:14 PM | pavely | 1 comment(s)

It is a well known fact that accessing the CPU cache is faster than accessing physical RAM, but how do you ensure, or at least target the CPU cache? Does it actually gain a performance boost that is significant? Here’s a relatively simple example to test this idea: matrix multiplication. To multiply two matrices together (let’s assume for simplicity sake that these are square matrices), you need to do some summations and multiplications to get the result of a single cell in the output matrix. Here...

תגים:DEV, Performance