Source
- URL: https://x.com/atorstling/status/2058449690572501039
- Author: Alexander Torstling (@atorstling)
- Posted: 2026-05-24 07:27:05
Branch
1/ @atorstling
@AgileJebrim
Hierarchical map reduce?
2/ @AgileJebrim
@atorstling
Okay. Now explain how that works.
3/ @atorstling
@AgileJebrim
Use shared memory for in-batch coordination, and global memory for intra-batch coordination. Each batch takes a set of numbers and produce a sum. Next round of batches operates on the output of the previous round. If fancy, you could encode the problem in kernels/registers.
4/ @atorstling
@AgileJebrim
Since you can use output of batches from different rounds as input, you can keep the pipeline full by starting new rounds eagerly
5/ @AgileJebrim
@atorstling
I’m not really satisfied with this answer.
6/ @atorstling
@AgileJebrim
You don’t need to be, all good :)