Jack Grigg

And some more rough benchmarks: ## AMD Ryzen 9 5950X ### Serial: 17.2% decrease ``` Variable-time aA+bB, A variable, B fixed time: [34.671 µs 34.687 µs 34.704 µs] Found 3...

I'm going to rebase this onto `main` shortly, and if I have time I'll address the other two (non-blocking) issues raised above.

Rebased on `main` and fixed the resulting merge conflicts.

[Force-pushed](https://github.com/dalek-cryptography/curve25519-dalek/compare/219ceff01f76db4f748d8b0ff869c38da02bd1e4..b13b3a65a2ef67f76b98a7b2653365cab4d7f864) to fix post-rebase bugs and get CI passing.

[Force-pushed](https://github.com/dalek-cryptography/curve25519-dalek/compare/b13b3a65a2ef67f76b98a7b2653365cab4d7f864..20e355e578da92dddfcd496a2f559344ad023b5e) to add changelog entries and fix documentation.

[Force-pushed](https://github.com/dalek-cryptography/curve25519-dalek/compare/20e355e578da92dddfcd496a2f559344ad023b5e..fd8952c07717cac0eca591df6d5a3bd16537455e) to move the new generated serial tables into separate submodules, and added cfg-flagged tests to generate them, and a CI job that verifies them. If this works, I'll attempt...

[Force-pushed](https://github.com/dalek-cryptography/curve25519-dalek/compare/fd8952c07717cac0eca591df6d5a3bd16537455e..5e03d5cb9d26764637e9881a1728548ec7f510a0) to fix the Fiat backends, and adjust the new CI check to fail if the table generators do nothing (as they generate output that is incorrectly formatted, and thus...

[Force-pushed](https://github.com/dalek-cryptography/curve25519-dalek/compare/5e03d5cb9d26764637e9881a1728548ec7f510a0..c96c810c02c6d7c7277995d42d51f1348bace6a1) to implement a similar kind of generator approach for the AVX2 vector table. It doesn't currently work because the `Debug` impl for `u32x8` doesn't print out values that are...

[Force-pushed](https://github.com/dalek-cryptography/curve25519-dalek/compare/c96c810c02c6d7c7277995d42d51f1348bace6a1..a77e13be04ef65225d994376fb1bd254fea05f39) to fix the AVX2 table generator. The generated constant is concretely different from before (I presume something changed about the wNAF implementation in the intervening four years), but tests...

[Force-pushed](https://github.com/dalek-cryptography/curve25519-dalek/compare/a77e13be04ef65225d994376fb1bd254fea05f39..01a9e9e7a6ffa813977f3a61b898b2ec407d7a17) to implement a generator for the IFMA vector table, based on the working AVX2 generator. It should work, but I don't have the hardware to run it, and so...