Get the latest tech news

AI Flame Graphs

We could have built a harder-to-use and higher-overhead version months ago using Intel GTPin but for widespread adoption it needs minimal overhead and ease of use so that developers don't hesitate to use this daily and to add it to deployment pipelines. You can search for any term in the flame graph: accelerator instructions, source paths, function names, etc., to quickly calculate the percentage of stacks where it is present (excluding vertical overlap) helping you prioritise performance work. But the real challenge is instruction profiling of actual GPU and AI accelerator programs -- shown as the aqua and green frames -- and correctly associating them with the CPU stacks beneath them.

Get the Android app

Or read this on Hacker News