-
https://qifeng.xyz/posts/ed5d8a1b.html
本文主要介绍通过WMMA API 使用TensorCore的方法。
-
very nice work! i want to know does this support tensor core or AMD GPU?
-
Hello!
I have recently switched to a full FP16 pipeline using MMA F16F16F16F16.
In such cases, softmax becomes the bottleneck of TensorCore Active %, particularly when running in embedded GPUs suc…
-
Does Caffe support Tensorcore / Volta? There is a post on the Google Group saying its possible, but I'm not sure if it has been implemented
-
More of a feature request than a problem report and forgive my ignorance if this is irrelevant but the nvidia 20x series and the 1660ti have tensor cores which could be use when called out on the nvid…
-
Hi, thanks to your code. I am sorry to trouble you. if this code support tensorcore, so i can use my rtx2070 faster. Thanks again.
-
L4 GPU:
T4 GPU:
Other variables look fine. I ran this for the neural_gcm_dynamic_forcing_stochastic_1_4_deg mode.
-
hi i'm using your custom darknet very well thx!!
i have some question to using that for my lab
i want to use pure darknet+using tensorcore
so can i get where is chainging code for using tensorcore…
ghost updated
5 years ago
-
Hi, i want to know how to use Tensor Core PTX(like,mma.sync.aligned.m16n8k16.row.col.f32.f16.f16.f32) in Triton?
-
## TL ; DR
I'm proposing a change to the internal structure of `LinearAlgebra` that is more suitable for a to-be-developed `TensorAlgebra` library.
What it does:
- Identifies the vector space…