|
About the hardware-backends category
|
|
0
|
352
|
January 22, 2021
|
|
Set of ops for a backend to register
|
|
0
|
50
|
February 2, 2023
|
|
Using Nsight Systems to profile GPU workload
|
|
6
|
6542
|
December 26, 2022
|
|
Private use opencl device
|
|
7
|
234
|
November 11, 2022
|
|
OpenCL Backend - Important Updates
|
|
2
|
276
|
November 8, 2022
|
|
How to keep up with updates of operator signatures
|
|
6
|
162
|
November 4, 2022
|
|
What is the reason for using old ABI on Linux?
|
|
3
|
150
|
October 30, 2022
|
|
A small MPS debugging story
|
|
0
|
295
|
September 15, 2022
|
|
Weight sharing on cuda
|
|
10
|
392
|
August 3, 2022
|
|
Lazy Tensor Core
|
|
20
|
3831
|
July 12, 2022
|
|
Memory operations on a custom backend
|
|
4
|
210
|
July 5, 2022
|
|
PyTorch and TensorFloat32
|
|
5
|
3081
|
June 28, 2022
|
|
More In-Depth Details of Floating Point Precision
|
|
0
|
574
|
June 21, 2022
|
|
Keeping PyTorch's Ops Maintainable: The Jiterator
|
|
6
|
587
|
January 19, 2022
|
|
Implementing OpenCL backend for pytorch
|
|
10
|
6320
|
January 8, 2022
|
|
Automatic out-of-tree backend loading
|
|
1
|
285
|
December 13, 2021
|
|
Debugging back-propogation results
|
|
4
|
378
|
November 18, 2021
|
|
OpenCL Backend: Broadcast/Reduce Ops
|
|
3
|
483
|
November 5, 2021
|
|
Aten::abs_out called with an undefined 'out' tensor
|
|
1
|
472
|
November 4, 2021
|
|
Aten::mul.out receives tensors from mixed devices
|
|
2
|
335
|
October 16, 2021
|
|
Asynchronous Execution and Memory Management
|
|
3
|
798
|
October 11, 2021
|
|
Slides from Structured Kernel presentation
|
|
4
|
815
|
April 8, 2021
|
|
Backend Fallbacks
|
|
0
|
1192
|
April 7, 2021
|