Example: using BenchmarkTools using Flux conv = Conv((7,7), 3 => 64; stride=2, pad=3) dummy = randn(Float32, 224, 224, 3, 2); conv(dummy) # ignore compile time @benchmark conv(dummy) #= BenchmarkTools.Trial: memory estimate: 19.29 MiB allocs estimate: 60 -------------- minimum time: …

`Conv` is 2x slow than pytorch `Conv` on cpu

DrChainsaw May 21, 2020, 12:37pm 7

Hmm, I just saw that the benchmark does not even use the GPU, so there is no CUDNN involved. Sorry for the noise.

Could it be an MKL vs OpenBLAS thing then?

Topic		Replies	Views
Flux vs pytorch cpu performance Machine Learning first-steps , flux	59	9506	October 2, 2020
Slow LSTM on GPU in Flux Machine Learning gpu , flux , pytorch	21	2215	February 15, 2024
Flux running slow? Machine Learning	16	2819	August 19, 2021
Understanding the need for Torch.jl? Machine Learning	23	6271	November 4, 2024
Flux convolutional layer not type stable Machine Learning performance , flux	0	624	May 14, 2020