-
I think at some point @jimi-c had mentioned something about lookups being potentially faster/a workaround to full module invocations, so doing something using `lookup('k8s'...` to get info instead of …
-
Might look really nice to have an actual 3d view (e.g. with height actually going up in height).
Concern - might disrupt art style vs in-game.
-
```
Time measurement for the reading from and writing to MSR registers
```
Original issue reported on code.google.com by `Thomas.R...@googlemail.com` on 6 Mar 2014 at 12:54
-
-
To help with the performance issues caused by the mod profiler overhead.
Note:
- Must establish rules for what methods will be inlined, when.
-
Currently arrays have been largely pre-populated because this gives us noticable iteration speed ups. What has not been considered by this is the memory overhead of this. A tradeoff will always need t…
-
### 🐛 Describe the bug
The bug is that using the pool from:
```
import torch.multiprocessing as mp # use spawn context
mp = mp.get_context('spawn')
```
**incurs ENORMOUS latency (x10 what norm…
-
Any function passed into `new Benchmark(options.setup)` will result in a considerable drop in tested function performance. It appears to be a fixed overhead unrelated to what is actually done in the s…
-
Sorry to bother you again.
In my research area, each stage is scheduled to be placed on some VM node. If its child stages are placed on different VM nodes, cross-node data transmission overhead sho…
-
## 🐛 Bug
Hi, I am new to PyTorch so apologies if I am missing something obvious! My use case involves narrow models with small input dimension (