-
I would like to see memory bandwidth metrics from Cadvisor. There are specific metrics named "container_memory_bandwidth_bytes".
I know they are under resctrl. However, I can not enable the cadvis…
-
Thanks for the great work! Would you mind share the system requirement to run inference? Can I run it on free google colab T4 gpu with 15G GPU RAM?
-
It would be great if the syspapi sampler supported collection of UNCORE counters.
Access to UNCORE counters on Haswell and KNL will allow for collection of memory bandwidth.
These counters are a…
-
Hi,
When I use the [benchmark script](https://www.tensorflow.org/lite/performance/measurement#native_benchmark_binary) and [benchmark apk](https://www.tensorflow.org/lite/performance/measurement#andr…
-
## Describe the bug
Setting up Talos v1.8.3 by [the refined terraform script](https://github.com/longhorn/longhorn-tests/pull/2172) and installing Longhorn. Test case `test_node_metrics` fails beca…
-
I haven't looked through the code, but I copied the bandwidth and OpenSBLI performance numbers into LibreOffice, made a scatter plot out of it, and the R^2 is coming out as 0.49 (=> R = 0.7), which is…
-
This is meant to exaplain the main concepts of a GPU architecture that are relevant to GPU programming.
Base material on `L11-Accelerated_Architectures.pptx`.
For instance explaining:
- presence of …
-
### Motivation
최근 DNN 모델은 더욱 큰 메모리 공간을 요구한다. 하지만 GPU, TPU와 같은 전통적인 HBM 기반 DNN training platform은 불충분한 용량을 제공할 뿐 아니라, 추가적인 비용과 memory bandwidth 측면에서의 under-utilization을 유발한다.
### Proposal
저자들은 비용 …
-
I am trying to profile my application on AMD Ryzen Threadripper 3960X (Zen2 arch). Specifically, I am trying to measure memory bandwidth and I get all 0's in the output. The OS I am using is Ubuntu 18…
-
### Feature Description
## GPUDirect Storage
[Nvidia GPUDirect Storage](https://developer.nvidia.com/blog/gpudirect-storage/) is a technology introduced by Nvidia that enables GPUs to read files d…