[do not merge] work in progress: BLS12-377

[weijiekoh]

Done:

• point addition and doubling in WGSL
• point addition for two affine points (Z1 and Z2 = 1) in WGSL
• Montgomery multiplication tests (WGSL) for the 377-bit base field

Todo:

• Convert cuzk_gpu() (contains many, many smaller tasks)

[weijiekoh]

Current code is in https://github.com/td-kwj-zp2023/webgpu-msm

[TalDerei]

Benchmarks on M1 for reference:

For 2^20 inputs, the points were taking too long to load on public wifi.

[weijiekoh]

I'll start a new PR after rebasing the code in https://github.com/td-kwj-zp2023/webgpu-msm to the new structure from #110.