Speeding up lilcom decompression

[pzelasko]

As lilcom is often used through Lhotse, several Lhotse users have reported that the decompression is computationally expensive; some of my colleagues performed dataloader profiling and found it takes 50% of the time in their data pipeline. I am not familiar with the details of lilcom’s algorithm, but perhaps it could be optimized somehow?

[danpovey]

Hm, it may be hard to optimize much. Might be worth investigating whether compiling locally, instead of using pip, helps (not sure if pip is installing binaries, or if it gets installed locally via compilation when you use pip).

…

On Sat, Nov 27, 2021 at 11:36 AM Piotr Żelasko ***@***.***> wrote: As lilcom is often used through Lhotse, several Lhotse users have reported that the decompression is computationally expensive; some of my colleagues performed dataloader profiling and found it takes 50% of the time in their data pipeline. I am not familiar with the details of lilcom’s algorithm, but perhaps it could be optimized somehow? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#33>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAZFLOZVEAF5HL6P3QMOFHTUOBG5TANCNFSM5I3W3OUQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[pzelasko]

Thanks, we'll check it out.