Tamas Bela Feher
Tamas Bela Feher
Thank you @sachinprasadhs for the answer. If we want to compare performance of any calculation, we need to ensure that the calculation is finished. The question of this issue: how...
/ok to test
/ok to test
I could reproduce this using `raft_ann_benchmark` (conda packages, `raft-ann-bench=24.08 cuda-version=12.2*`, Intel Xeon Silver 4210R CPU (10 cores), T4 GPU). Note that one needs to use the cpp benchmark executables to...
Hi @navneet1v, I could run CAGRA index creation using the IVF-PQ algorithm through the FAISS Python API. When using a memory mapped dataset with shape 1M x 1536, (6.144 GB),...
Thank you for posting the graphs! > Is it possible that rmm logging isn't capturing all of the GPU memory allocations? Yes, it only logs allocations done through RMM API....
Indeed, when we use IVF-PQ build method for the KNN graph, then we do a PQ quantization for the graph building, and another one for compressing the data for CAGRA...
/ok to test 91c6734
Closing this as not a bug.
Currently we have around 13 MiB reduction in `cagra_build`, which is around 33% reduction for those binaries. The binaries are still larger than I hoped them to be, investigating.