Is there any way to optimize PC1?

[nujz]

Is there any way to optimize PC1?

[hail100]

https://github.com/storswiftlabs/rust-fil-proofs-mpool/blob/master/storage-proofs-porep/src/stacked/vanilla/create_label/mem_pool_arch.md

[hail100]

hi @vmx, I found PC1 create_base_merkle_tree hash() use of cpu will affect SDR bind cores, so 32G PC1 tasks will run for one more hour.

[vmx]

@hail100: yes only the layer generation is bound to cores. The tree building of create_base_merkle_tree isn't bound to any specific cores.

[hail100]

Maybe need other project support. https://github.com/filecoin-project/merkletree/blob/master/src/hash.rs

[hail100]

https://github.com/supranational/filecoin_pc1