Can prove key deserialization be optimized?

[p0vidl0]

Thanks for great library. For a custom use-case, serialized prove key have size about 29332912 bytes. The key deserialization takes 26 seconds and proof generation takes less than 2 seconds.

Please advise, is it possible in theory to optimize the deserialization process to make overall process faster?

[Schaeff]

Hi, thanks for flagging. We would welcome contributions on this.

[dark64]

Thanks for great library. For a custom use-case, serialized prove key have size about 29332912 bytes. The key deserialization takes 26 seconds and proof generation takes less than 2 seconds.

Please advise, is it possible in theory to optimize the deserialization process to make overall process faster?

I'm closing this for now as I could not replicate the same issue with the latest version. My assumption is this issue is resolved since 0.8.1, the issue was related to a call to the deserialization function that had validity checks which were time-consuming.