Clive Chan cchan
cchan
/ bf16_convs.py
Created
December 28, 2022 03:52
A janky benchmark script for autocasted convolution in pytorch
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| import time | |
| import torch | |
| torch.backends.cuda.matmul.allow_tf32 = False | |
| GROUPS = 64 | |
| for k in [3, 5, 7]: | |
| x = torch.rand(16,64,1024,1024, device="cuda", dtype=torch.float32) | |
| weight = torch.rand(64,64//GROUPS,k,k, device="cuda", dtype=torch.float32) |
cchan
/ simplesim-3.0-patch
Last active
March 19, 2021 19:26
simplesim-3.0 fix for OSX, tested on Big Sur on M1. Run `make sim-tests` to check that it's working for you.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| --- syscall.c | |
| +++ syscall.c | |
| @@ -48,6 +48,39 @@ | |
| * Copyright (C) 1994-2003 by Todd M. Austin, Ph.D. and SimpleScalar, LLC. | |
| */ | |
| +#ifdef __APPLE__ | |
| + | |
| +#define __unix__ | |
| +#define unix |