Support for bivariate convolution & normalization with offset (#126)

* Add bivariate-convolution
* Add pair-wise convolution + tests + benches
* Add take_cnv_pvec_[left/right] to Scratch & updated CHANGELOG.md
* cross-base2k normalization with positive offset
* clippy & fix CI doctest avx compile error
* more streamlined bounds derivation for normalization
* Working cross-base2k normalization with pos/neg offset
* Update normalization API & tests
* Add glwe tensoring test
* Add relinearization + preliminary test
* Fix GGLWEToGGSW key infos
* Add (X,Y) convolution by const (1, Y) poly
* Faster normalization test + add bench for cnv_by_const
* Update changelog

poulpy-cpu-ref/benches/convolution.rs

@@ -0,0 +1,35 @@
+use criterion::{Criterion, criterion_group, criterion_main};
+use poulpy_cpu_ref::FFT64Ref;
+use poulpy_hal::bench_suite::convolution::{
+    bench_cnv_apply_dft, bench_cnv_by_const_apply, bench_cnv_pairwise_apply_dft, bench_cnv_prepare_left, bench_cnv_prepare_right,
+};
+
+fn bench_cnv_prepare_left_cpu_ref_fft64(c: &mut Criterion) {
+    bench_cnv_prepare_left::<FFT64Ref>(c, "cpu_ref::fft64");
+}
+
+fn bench_cnv_prepare_right_cpu_ref_fft64(c: &mut Criterion) {
+    bench_cnv_prepare_right::<FFT64Ref>(c, "cpu_ref::fft64");
+}
+
+fn bench_bench_cnv_apply_dft_cpu_ref_fft64(c: &mut Criterion) {
+    bench_cnv_apply_dft::<FFT64Ref>(c, "cpu_ref::fft64");
+}
+
+fn bench_bench_bench_cnv_pairwise_apply_dft_cpu_ref_fft64(c: &mut Criterion) {
+    bench_cnv_pairwise_apply_dft::<FFT64Ref>(c, "cpu_ref::fft64");
+}
+
+fn bench_cnv_by_const_apply_cpu_ref_fft64(c: &mut Criterion) {
+    bench_cnv_by_const_apply::<FFT64Ref>(c, "cpu_ref::fft64");
+}
+
+criterion_group!(
+    benches,
+    bench_cnv_prepare_left_cpu_ref_fft64,
+    bench_cnv_prepare_right_cpu_ref_fft64,
+    bench_bench_cnv_apply_dft_cpu_ref_fft64,
+    bench_bench_bench_cnv_pairwise_apply_dft_cpu_ref_fft64,
+    bench_cnv_by_const_apply_cpu_ref_fft64,
+);
+criterion_main!(benches);

poulpy-cpu-ref/benches/fft.rs

@@ -11,10 +11,7 @@ pub fn bench_fft_ref(c: &mut Criterion) {
     fn runner(m: usize) -> impl FnMut() {
         let mut values: Vec<f64> = vec![0f64; m << 1];
         let scale: f64 = 1.0f64 / (2 * m) as f64;
-        values
-            .iter_mut()
-            .enumerate()
-            .for_each(|(i, x)| *x = (i + 1) as f64 * scale);
+        values.iter_mut().enumerate().for_each(|(i, x)| *x = (i + 1) as f64 * scale);
         let table: ReimFFTTable<f64> = ReimFFTTable::<f64>::new(m);
         move || {
             ReimFFTRef::reim_dft_execute(&table, &mut values);
@@ -39,10 +36,7 @@ pub fn bench_ifft_ref(c: &mut Criterion) {
     fn runner(m: usize) -> impl FnMut() {
         let mut values: Vec<f64> = vec![0f64; m << 1];
         let scale: f64 = 1.0f64 / (2 * m) as f64;
-        values
-            .iter_mut()
-            .enumerate()
-            .for_each(|(i, x)| *x = (i + 1) as f64 * scale);
+        values.iter_mut().enumerate().for_each(|(i, x)| *x = (i + 1) as f64 * scale);
         let table: ReimIFFTTable<f64> = ReimIFFTTable::<f64>::new(m);
         move || {
             ReimIFFTRef::reim_dft_execute(&table, &mut values);

poulpy-cpu-ref/benches/vec_znx.rs

@@ -5,7 +5,7 @@ use poulpy_hal::reference::vec_znx::{bench_vec_znx_add, bench_vec_znx_automorphi
 
 #[allow(dead_code)]
 fn bench_vec_znx_add_cpu_ref_fft64(c: &mut Criterion) {
-    bench_vec_znx_add::<FFT64Ref>(c, "cpu_spqlios::fft64");
+    bench_vec_znx_add::<FFT64Ref>(c, "cpu_ref::fft64");
 }
 
 #[allow(dead_code)]