Sergey "Shnatsel" DavidoffandClaude c74f23784c Vectorize byte-level horizontal differencing in fp_predict ...

Process the difference step in reverse 16-byte chunks, reading both
curr and prev into local arrays before writing. This makes all 16
subtractions independent, enabling autovectorization (confirmed with
SSE2: movdqu + psubb).

Co-Authored-By: Claude <noreply@anthropic.com>

2026-02-26 15:45:21 +00:00

..

decoder

Skip endianess swap for SGI compression

2026-02-14 18:04:07 +01:00

encoder

Vectorize byte-level horizontal differencing in fp_predict

2026-02-26 15:45:21 +00:00

bytecast.rs

Mark f16 bytecast as expected unused

2025-07-14 19:25:12 +02:00

directory.rs

Build Directory FromIterator

2025-11-12 01:21:53 +01:00

error.rs

Improve unexpected tile count error

2026-02-14 06:55:29 +01:00

lib.rs

Public ColorType::{num_samples, bit_depth}

2026-01-24 05:24:19 +01:00

tags.rs

Implement floating-point predictor (Predictor 3) encoding

2026-02-25 23:05:35 +00:00