Remove artificial bottleneck in load/store queue #12
Comments
JosseVanDelm
pushed a commit
to JosseVanDelm/snitch_cluster
that referenced
this issue
Dec 11, 2023
* hw: Add HWPE MAC ctrl code * hw: Add HWPE MAC reqrsp code * hw: Add HWPE MAC top code * bender: add HWPE mac files to bender list * lint: Exclude HWPE MAC files from license checker * hw: make sure the module name matches the fine name * bender: change the file name of hwpe mac * hw: rm the hwpe mac file with wrong filename * bender: change the order of hwpe mac in bender list * hw: edit comments * hw: Clarify comments about the data addressing --------- Co-authored-by: rgantonio <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Currently, the load-store queue is quite limited in Snitch and since the addition of store response handling becomes easily full. The idea would be to compress back-to-back stores as we are not interested in the actual value but just that we have an outstanding store.
I think something like a
compressable_fifowould be a good start where element of the same time pushed back-to-back could increment a counter instead of occupying an actual queue item. Critical paths need to be checked as well.The text was updated successfully, but these errors were encountered: