Closed cyphersnake closed 1 month ago
I'll also add a test that does all assignments to a single region, to see if it affects performance, vs. multiple regions
I ran a benchmark overnight with the following parameters:
One poseidon-hash pass is 24 rows
I wish I could say that this way our table step is 24*50, but considering that it doesn't fit into the table 2**20 / 24
runs of poseidon-hash, but only 800 (this is at k=22, at k=20 I'll find out after the end of the run), I suspect that V1 didn't help at all to equalize our situation of independent columns, so we'll see.
Issue Link / Motivation Close #249
Changes Overview In this branch I am performing a specific test. Depending on the results, this benchmark may or may not be included in the main. At the moment, this benchmark is running on our server