*   Python 3.3+ support via changes to python codebase and ability to specify
    python version via ./configure.

*   Some improvements to GPU performance and memory usage:
    [convnet benchmarks](https://github.com/soumith/convnet-benchmarks/issues/66)
    roughly equivalent with native cudnn v2 performance. Improvements mostly due
    to moving to 32-bit indices, faster shuffling kernels. More improvements to
    come in later releases.

### Migration

func BenchmarkCAST(b *testing.B) { // IG 10.3.A says "ML-DSA digital signature generation CASTs should cover // all applicable rejection sampling loop paths". For ML-DSA-44, there are // four paths. For ML-DSA-65 and ML-DSA-87, only three. This benchmark helps // us figure out which is faster: four rejections of ML-DSA-44, or three of // ML-DSA-65. (It's the former, but only barely.) b.Run("ML-DSA-44", func(b *testing.B) { // Same as TestACVPRejectionKAT/Test/Path/ML-DSA-44/1. seed := fromHex("5...

B(&x, B, &y) q1.ScalarBaseMult(&x) q2.ScalarBaseMult(&y) check.Add(&q1, &q2) checkOnCurve(t, &p, &check, &q1, &q2) return p.Equal(&check) == 1 } if err := quick.Check(varTimeDoubleBaseMul, quickCheckConfig(32)); err != nil { t.Error(err) } } // Benchmarks. func BenchmarkScalarBaseM(b *testing.B) { var p Point for i := 0; i < b.N; i++ { p.ScalarBaseMult(dalekScalar) } } func BenchmarkScalarMult(b *testing.B) { var p Point for i := 0; i < b.N; i++ { p.ScalarMult(dalekScalar, B) } } func BenchmarkVarTimeDoub(b...