AVX regular fft

[spapinistarkware]

This change is 

[spapinistarkware]

• ColumnOps #459
• Use precomputed twiddles in avx #443
• Precompute twiddles #401
• Remove FRI PolyOps dependency #399
• avx slow eval #398
• avx poly extend #397
• AVX fft eval on larger domains #396
• Reverse order of evaluations #395
• Incorporate fft in AVX backend #394
• AVX regular fft #390 👈
• Separate ifft #389
• Support ifft for itnermediate sizes #383
• AVX fft benchmark #382
• ifft_full #381
• Save twiddles #380
• ifft_lower #379
• avx ifft3 #378
• avx vecwise butterflies #377
• avx vecwise ibutterfly #368
• avx ibutterfly #367
• avx butterfly #366
• avx poly #365
• dev

This stack of pull requests is managed by Graphite. Learn more about stacking.

Join @spapinistarkware and the rest of your teammates on Graphite

[ilyalesokhin-starkware]

src/core/backend/avx512/fft/ifft.rs line 83 at r1 (raw file):

        ifft_vecwise_loop(values, twiddle_dbl, fft_layers - VECWISE_FFT_BITS, index_h);
        for layer in (VECWISE_FFT_BITS..fft_layers).step_by(3) {
            match fft_layers - layer {

are you trying to sneak this back in?

Code quote:

        for layer in (VECWISE_FFT_BITS..fft_layers).step_by(3) {
            match fft_layers - layer {

[ilyalesokhin-starkware]

Reviewed all commit messages.
Reviewable status: 0 of 2 files reviewed, 1 unresolved discussion (waiting on @spapinistarkware)

---

src/core/backend/avx512/fft/ifft.rs line 83 at r1 (raw file):

Previously, ilyalesokhin-starkware wrote…

are you trying to sneak this back in?

do you think doing a match on every layer is better?

[spapinistarkware]

Reviewable status: 0 of 2 files reviewed, 1 unresolved discussion (waiting on @ilyalesokhin-starkware)

---

src/core/backend/avx512/fft/ifft.rs line 83 at r1 (raw file):

Previously, ilyalesokhin-starkware wrote…

do you think doing a match on every layer is better?

Since I added the fft, I want it to be visible that fft is exactly the same thing in reverse order.
The while didn't really show this, since I needed to start the while from something weird like (fft_layers-4)/3*3+fft_layers

[ilyalesokhin-starkware]

src/core/backend/avx512/fft/ifft.rs line 83 at r1 (raw file):

Previously, spapinistarkware (Shahar Papini) wrote…

Since I added the fft, I want it to be visible that fft is exactly the same thing in reverse order.
The while didn't really show this, since I needed to start the while from something weird like (fft_layers-4)/3*3+fft_layers

where do you need to start from something weird?

[spapinistarkware]

Reviewable status: 0 of 2 files reviewed, 1 unresolved discussion (waiting on @ilyalesokhin-starkware)

---

src/core/backend/avx512/fft/ifft.rs line 83 at r1 (raw file):

Previously, ilyalesokhin-starkware wrote…

where do you need to start from something weird?

In the regular fft (which is starting from higher bits).
And I want it to look exactly like the reverse of this (easier to understand why they are inverse of each other)

[ilyalesokhin-starkware]

src/core/backend/avx512/fft/ifft.rs line 696 at r1 (raw file):

    #[test]
    fn test_ifft_lower_with_vecwise() {
        for log_size in 5..=11 {

what is this syntax?

Code quote:

5..=11 

[ilyalesokhin-starkware]

Reviewed 2 of 2 files at r1.
Reviewable status: all files reviewed, 2 unresolved discussions (waiting on @spapinistarkware)

[spapinistarkware]

Reviewable status: 0 of 2 files reviewed, 2 unresolved discussions (waiting on @ilyalesokhin-starkware)

---

src/core/backend/avx512/fft/ifft.rs line 696 at r1 (raw file):

Previously, ilyalesokhin-starkware wrote…

what is this syntax?

Done.

[ilyalesokhin-starkware]

Reviewed 2 of 2 files at r2, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @spapinistarkware)

[graphite-app]

Merge activity

• Mar 6, 9:56 AM EST: spapinistarkware added this pull request to the Graphite merge queue.
• Mar 6, 10:01 AM EST: spapinistarkware merged this pull request with the Graphite merge queue.