Skip to content

Comment on Highly efficient matrix transpose in Mojoparent

Comments

The colors are reading writing operations ;)

You have global memory and shared memory, the global is slower.

You read in rows in the global memory (faster than reading columns)

You write in columns in the shared memory (slower than in rows, but the shared memory is fast, this is the transpose operation)

You read in rows in the shared memory (very fast)

You write in rows in the global memory (faster than writing in columns)

The idea behind that tiling is to hide the slow part in a memory that is faster.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.