-
Changed the code a bit so the GPU would like it just a tiny bit more:
Changed all possible data types to unsigned, I thought I saw a tiny improvement on d_maxOnes, but this should be analyzed more; m…
-
### Describe you feature request
**Is your feature request related to a problem? Please describe.**
Reloading container only to toggle XDebug is rather expensive because it involves restarting whole…
-
We don't need a state commitment and it's slow so let's mock it and re-execute transactions as needed (we are already doing this for tracing after the latest go-ethereum upgrade). The only trie that i…
-
# Links to Optimizations Resources
- [Build for Billions](https://developer.android.com/docs/quality-guidelines/build-for-billions)
- [General information on Optimize for Android (Go edition) ](ht…
-
Issue tracking the difference between the expected bfp8_b AllGather perf in Llama (24 µs) and the values measured with tracing (32 µs)
-
Currently the compiler generates not optimal output, which can lead to troubles due to limits of Housing.
-
Look into [tearing](https://wiki.hyprland.org/Configuring/Tearing/) and [Pipewire latency](https://nixos.wiki/wiki/PipeWire)
-
This is just a list of potential optimizations that we think could work.
**Maintained list of optimizations**
- [x] We currently use random numbers (uuid) for our caching system, my understandi…
-
- [ ] Require only one recursive loop to do all parsing of classes and formatting of children
- [ ] Connect images to elementInfo class instead of finding it in array
-
### Description
As the project develops, many of our tools work with lists of Audio objects with the goal to be that they can be optimized into Pydra workflows and have easy to use pipelines, especia…