-
The anatomy of AVX-512 intrinsics
-
Making CPUs adequate at linear algebra
-
When you should not use a GPU to run a machine learning model
-
Mulling over matrix multiplications in CUDA
-
Birdbrain batching and tips on never using it
-
Batching arises naturally from the fundamental design of neural networks
-
Counting combinations
-
Fortran out here impressing
-
Immutability at home
-
If you work hard enough, you can be mediocre at anything
-
Reversals & Displeasure
-
Visual intuition for Fourier transforms in two dimensions