Benson Ma

[question] Option to run asynchronous queries with the java API

2

In the HTTP API, there is the option to add "x-arango-async: store" to the headers so that a query can be run asynchronously, in the sense that a job id...

Re-organize jagged tensor tests, pt 5

3

Summary: - Re-organize jagged tensor tests, pt 5 Differential Revision: D55042916

Add arg checks to `input_combine_with_length`

12

Summary: - Add arg checks to `input_combine_with_length` Differential Revision: D54138774

[fbgemm_gpu] Enable UVM tests for ROCm 5.7+ (Rebase of PR 2290)

2

- Enable UVM Tests on ROCm 5.7+

Optimize the cache fetch for forward split, pt. 1

9

Summary: Rewrite the kernel to use cache_hit_rate enum as template argument. We first check if the cache is empty and pass that value as a template argument. Inside the first...

TBE UVM caching conflict misses benchmark

4

Differential Revision: D47819860

Optimize the cache fetch for forward split, pt. 2

2

Summary: This follows up the work on D47778422 by plumbing the `uvm_cache_stats` argument passing up to the Python API level. `local_uvm_cache_stats` is now zeroed out before the prefetch step as...

Optimize the cache fetch for forward split, pt. 1

13

Summary: - Rewrite the kernel to use `cache_hit_rate` enum as template argument Differential Revision: D47778422