Asynchronous Versus Synchronous Execution PyTorch Developer podcast

Artwork

Tech PyTorch Edward Yang Team PyTorch Deep Learning Machine Learning

Innhold levert av PyTorch, Edward Yang, and Team PyTorch. Alt podcastinnhold, inkludert episoder, grafikk og podcastbeskrivelser, lastes opp og leveres direkte av PyTorch, Edward Yang, and Team PyTorch eller deres podcastplattformpartner. Hvis du tror at noen bruker det opphavsrettsbeskyttede verket ditt uten din tillatelse, kan du følge prosessen skissert her https://no.player.fm/legal.

PyTorch Developer Podcast « »
Asynchronous versus synchronous execution

3y ago 15:03

Del

MP3•Episoder hjem

Innhold levert av PyTorch, Edward Yang, and Team PyTorch. Alt podcastinnhold, inkludert episoder, grafikk og podcastbeskrivelser, lastes opp og leveres direkte av PyTorch, Edward Yang, and Team PyTorch eller deres podcastplattformpartner. Hvis du tror at noen bruker det opphavsrettsbeskyttede verket ditt uten din tillatelse, kan du følge prosessen skissert her https://no.player.fm/legal.

CUDA is asynchronous, CPU is synchronous. Making them play well together can be one of the more thorny and easy to get wrong aspects of the PyTorch API. I talk about why non_blocking is difficult to use correctly, a hypothetical "asynchronous CPU" device which would help smooth over some of the API problems and also why it used to be difficult to implement async CPU (but it's not hard anymore!) At the end, I also briefly talk about how async/sync impedance can also show up in unusual places, namely the CUDA caching allocator.

Further reading.

CUDA semantics which discuss non_blocking somewhat https://pytorch.org/docs/stable/notes/cuda.html
Issue requesting async cpu https://github.com/pytorch/pytorch/issues/44343

… continue reading

82 episoder

#Tech #PyTorch #Edward Yang #Team PyTorch #Deep Learning #Machine Learning

Artwork

Asynchronous versus synchronous execution

PyTorch Developer Podcast

33 subscribers

published 3y ago

Del

MP3•Episoder hjem

Innhold levert av PyTorch, Edward Yang, and Team PyTorch. Alt podcastinnhold, inkludert episoder, grafikk og podcastbeskrivelser, lastes opp og leveres direkte av PyTorch, Edward Yang, and Team PyTorch eller deres podcastplattformpartner. Hvis du tror at noen bruker det opphavsrettsbeskyttede verket ditt uten din tillatelse, kan du følge prosessen skissert her https://no.player.fm/legal.

CUDA is asynchronous, CPU is synchronous. Making them play well together can be one of the more thorny and easy to get wrong aspects of the PyTorch API. I talk about why non_blocking is difficult to use correctly, a hypothetical "asynchronous CPU" device which would help smooth over some of the API problems and also why it used to be difficult to implement async CPU (but it's not hard anymore!) At the end, I also briefly talk about how async/sync impedance can also show up in unusual places, namely the CUDA caching allocator.

Further reading.

CUDA semantics which discuss non_blocking somewhat https://pytorch.org/docs/stable/notes/cuda.html
Issue requesting async cpu https://github.com/pytorch/pytorch/issues/44343

… continue reading

82 episoder

#Tech #PyTorch #Edward Yang #Team PyTorch #Deep Learning #Machine Learning

Alle episoder

×

Velkommen til Player FM!

Player FM scanner netter for høykvalitets podcaster som du kan nyte nå. Det er den beste podcastappen og fungerer på Android, iPhone og internett. Registrer deg for å synkronisere abonnement på flere enheter.

Lytt til 500+ tema