-
I would like to profile the time spent to move from epoch to another while training a net but didn't understand how to enable to profiler to do so. If I got it right, the `profile_batch` parameter onl…
-
For CUDA it's possible we can track allocations by intercepting `cudaMalloc()` and friends. This would need to be tracked and reported separately than CPU memory.
https://docs.nvidia.com/cuda/cuda-…
-
Add a class which would manage time takers and draw them like this:
```
/ Server step
| / Environment step
| | + Modify in active blocks, took X ms
| | + Manage active blocks, took X ms
| \ Environme…
xyzz updated
10 years ago
-
> #### Description:
Opis had the ability to profile packet flow between the server and client (in terms of raw bytes/sec), and give a breakdown by mod. I know that network bandwidth is not somethin…
-
**Environment**
```
rkt Version: 1.13.0+gitc4b224a
appc Version: 0.8.6
Go Version: go1.6
Go OS/Arch: linux/amd64
Features: +TPM +SDJOURNAL
--
Linux 4.4.0-31-generic x86_64
--
NAME="Ubuntu"
VERSION="1…
-
The numpy temporal calculator is slow.
Here is the profiling for a the top htm layer performing temporal poooling with the following configuration parameters for the HTM network;
```
testParameters…
-
```
Tests in the test/stress directory should be augumented and a profiling
report should be generated to help addressing improvement areas.
Ideally profiling should be done on many different operati…
-
**Please describe the problem you are trying to solve.**
Given that the reported runtime measures are in `us` by default, it might be hard to spot the most expensive events at a first glance.
**D…
-
### Website or app
N/A
### Repro steps
When profiling a complex, proprietary app, I regularly get an error message, "Message length exceeded maximum allowed length."
Steps to reproduce:
1. Wi…
-
CNN or Preprocessor could offer a way to profile the speed of
- data access/loading
- sample preprocessing
- parallelized preprocessing
- forward pass
and log information to wandb