Use tiled matrix transposition

[sbooth]

No description provided.

[github-actions]

Cpp-linter Review

Used clang-format v22.1.6

Click here for the full clang-format patch

diff --git a/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m b/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m
index 409faae..dd8320e 100644
--- a/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m
+++ b/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m
@@ -39 +39,2 @@ static BOOL readChunkID(SFBInputSource *inputSource, uint32_t *chunkID) {
-static void matrixTranspose(const unsigned char *restrict A, unsigned char *restrict B, NSInteger rows, NSInteger columns) {
+static void matrixTranspose(const unsigned char *restrict A, unsigned char *restrict B, NSInteger rows,
+                            NSInteger columns) {

Have any feedback or feature suggestions? Share it here.

[Copilot]

Pull request overview

This PR updates the DSF decoder’s channel deinterleaving step to use a tiled (blocked) matrix transpose implementation when interleaving DSF blocks into clustered frames, aiming to improve cache locality/performance during decoding.

Changes:

• Replaces the naive DSF block transpose with a tiled transpose (MATRIX_TRANPOSE_BLOCK_SIZE).
• Renames internal helper functions used for chunk ID reading and matrix transpose.
• Updates the DSF block read path to call the new transpose implementation.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

outdated suggestion

[github-actions]

Cpp-linter Review

Used clang-format v22.1.6

Click here for the full clang-format patch

diff --git a/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m b/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m
index cdd9e5b..8459255 100644
--- a/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m
+++ b/Sources/CSFBAudioEngine/Decoders/SFBDSFDecoder.m
@@ -39,2 +39 @@ static BOOL readChunkID(SFBInputSource *inputSource, uint32_t *chunkID) {
-static void matrixTranspose(const unsigned char *restrict A, unsigned char *restrict B, size_t rows,
-                            size_t columns) {
+static void matrixTranspose(const unsigned char *restrict A, unsigned char *restrict B, size_t rows, size_t columns) {

Have any feedback or feature suggestions? Share it here.

outdated suggestion