Internal error using MPSCNNConvolution

Graphics & Games General
s1ddok OP
Created Jan ’23
Replies 2
Boosts 0
Views 743
Participants 3

Hello!

I'm a long-standing user of MPSCNN framework, it usually works fine but during the implementation of one my recent networks I started to get this errors:


2023-01-06 00:17:46.017908+0600 -[44642:879994] [GPUDebug] Invalid device load executing kernel function "cnnConvWinograd_8x8_3x3_32x32_256" encoder: "", dispatch: 0, at offset 120```

Also, strangely networks produces different results over the same inputs across multiple runs. I assume there are some race conditions inside which causes that. Is it possible to somehow enforce MPSCNN to use some other implementation of convolution? Or I'm stuck with it forever?
Replies  2
Boosts  0
Views  743
Participants  3
Frameworks Engineer OP
Apple
Jan ’23

@s1ddok, can you provide more information about :

  1. What platform you are seeing this error ?
  2. What input / output sizes and data types you are seeing this error.

It will be great if you can provide a small repro case. Also if you can file an issue using Feedback Assistant, it will be great.

0
Frameworks Engineer OP
Apple
Jan ’23

I got the Feedback Assistant report you filed. Can you please tell us what OS and device its happening. Thanks.

0
Internal error using MPSCNNConvolution
First post date Last post date
 
 
Q