-
Now IndexIVF.h support parallel_mode:
```
* 0 (default): parallelize over queries
* 1: parallelize over over inverted lists
* 2: parallelize over both
```
can IndexBinaryIVF su…
-
Related: #4213
We should add support for parallel model training via ray that goes beyond the currently implemented parallel fold model fitting and parallel HPO.
Given a portfolio of 100 models…
-
Hello!
First off, thanks for the lib. Really enjoying using it, and great work 👍 😄
I have a question related to guard logic execution.
Does stateless support executing guards in parallel?
…
-
I noticed after some debugging that the standard command
$ARCHDEFS/install_optrove
uses `make` internally to compile `libcutest.a`. Unfortunately, the compilation [rule](https://github.com/…
-
### Description
I started doing more intense compute work in Marimo, and noticed the kernel is locked when executing a cell.
### Suggested solution
Marimo could work on the execution graph in paral…
-
### Expected Behavior
Do some of the "configuring build" part in parallel. For example almost sure compiling the Gradle files for each separate module doesn't have cross-dependencies. The fact that I…
-
### Description
**Background**
Currently, there can be a number of drivers partialAgg(each run by a driver), but always 1 finalAgg/singleAgg; meaning finalAgg/singleAgg always uses a single core. …
-
Hi there,
I just wanted to express my appreciation for your incredible work. As I dive deeper into this field, I've encountered some limitations with my hardware resources - 2 GPUs T4, each with 16…
-
### What is the issue?
Hi!
I am trying to embed a lot of documents with `jina/jina-embeddings-v2-base-de`. I am trying to use request parallelism with `OLLAMA_NUM_PARALLEL=16`. I have a multiproce…
-
We should be able to speed up the flow (transforms/backends) by introducing parallelism.
Possible levels:
- Per Model (CoreDSL File)
- Per Set
- Per Instr
Since we currently handle transforms…