| Age | Commit message (Expand) | Author |
|---|
| 2024-04-16 | refmvs: Fix buffer overread in save_tmvs() asmHEADmastermain | Henrik Gramner |
| 2024-03-23 | Upgrade libdav1d to 1.4.1 | Harish Mahendrakar |
| 2024-03-09 | NEWS: Forgotten intro sentence | Jean-Baptiste Kempf |
| 2024-03-08 | Update to 1.4.1 | Jean-Baptiste Kempf |
| 2024-03-08 | Update THANKS.md | Matthias Dressel |
| 2024-03-08 | arm32: Fix right shifts in the 16bpc iwht implementation | Martin Storsjö |
| 2024-03-08 | arm32/msac: Trim C functions, saves 1024 bytes | Nathan E. Egge |
| 2024-03-08 | arm64/msac: Trim C functions, saves 1392 bytes | Nathan E. Egge |
| 2024-03-08 | arm: Use -fno-align-functions when building | Nathan E. Egge |
| 2024-03-08 | arm32/itx: Trim dav1d_inv_wht4_1d_c, saves 68 bytes | Nathan E. Egge |
| 2024-03-08 | arm64/itx: Trim dav1d_inv_wht4_1d_c, saves 92 bytes | Nathan E. Egge |
| 2024-03-08 | arm32/itx16: Add 4x4 12bpc NEON wht_wht transform | Nathan E. Egge |
| 2024-03-08 | arm64/itx16: Add 4x4 12bpc NEON wht_wht transform | Nathan E. Egge |
| 2024-03-07 | x86: Fix out-of-bounds read in 8bpc SSE2/SSSE3 wiener_filter | Henrik Gramner |
| 2024-03-05 | AArch64: Specialise HBD Neon convolutions for 6-tap filters | Arpad Panyik |
| 2024-03-05 | AArch64: Optimize 6-tap SBD HV Neon convolution | Arpad Panyik |
| 2024-03-04 | checkasm: aarch64: Print the SVE vector length, if available | Martin Storsjö |
| 2024-03-04 | aarch64: Check for assembler support for various aarch64 extensions | Martin Storsjö |
| 2024-02-29 | checkasm: Add --list-cpuflags option | Henrik Gramner |
| 2024-02-28 | ci: Add an aarch64 cross compile CI job with a recent Clang | Martin Storsjö |
| 2024-02-28 | ci: Test aarch64 with QEMU, with varying SVE vector lengths | Martin Storsjö |
| 2024-02-28 | ci: Bump to the latest dav1d-debian-unstable image | Martin Storsjö |
| 2024-02-28 | Extend Arm and AArch64 run-time CPU feature detection | Arpad Panyik |
| 2024-02-27 | riscv64/itx: Add 16x16 8bpc eob test | Nathan E. Egge |
| 2024-02-27 | riscv64/itx: Add 8x16 8bpc eob test | Nathan E. Egge |
| 2024-02-27 | riscv64/itx: Add 4x16 8bpc eob test | Nathan E. Egge |
| 2024-02-26 | riscv/checkasm: Print the RVV vector length, if available | Nathan E. Egge |
| 2024-02-26 | arm/msac: Enable NEON optimizations on more platforms | Nathan E. Egge |
| 2024-02-22 | CI: Add riscv64 clang build | Matthias Dressel |
| 2024-02-22 | CI: Update image | Matthias Dressel |
| 2024-02-22 | gcovr: Fix config file | Matthias Dressel |
| 2024-02-22 | riscv64/itx: Fix build issues with clang | Nathan E. Egge |
| 2024-02-22 | x86inc: Fix warnings with old nasm versions | Henrik Gramner |
| 2024-02-22 | AArch64: Enable benchmarks for 8-tap sharp filters | Arpad Panyik |
| 2024-02-22 | AArch64: Specialise Neon convolutions for 6-tap filters | Arpad Panyik |
| 2024-02-21 | ppc: Factor out the shift computation | Luca Barbato |
| 2024-02-21 | ppc: Use vec_madd instead of shifts | Luca Barbato |
| 2024-02-21 | ppc: Unify cdef_directions | Luca Barbato |
| 2024-02-21 | ppc: Split the cdef_filter in pri/sec as done for the other arches | Luca Barbato |
| 2024-02-21 | arm64/itx16: Reuse horz_16x4 epilog, saves 96 bytes | Nathan E. Egge |
| 2024-02-21 | arm32/itx16: Reuse horz_16x2 epilog, saves 24 bytes | Nathan E. Egge |
| 2024-02-21 | riscv64/itx: Reuse horz_16x8 epilog, saves 94 bytes | Nathan E. Egge |
| 2024-02-21 | arm64/itx: Reuse horz_16x8 epilog, saves 512 bytes | Nathan E. Egge |
| 2024-02-21 | arm32/itx: Reuse horz_16x4 epilog, saves 336 bytes | Nathan E. Egge |
| 2024-02-21 | riscv64/itx: Reuse 16x8 epilog, saves 706 bytes | Nathan E. Egge |
| 2024-02-21 | riscv64/itx: Reuse 8x16 epilog, saves 24 bytes | Nathan E. Egge |
| 2024-02-21 | riscv64/itx: Tail call vert_8x16, saves 1086 bytes | Nathan E. Egge |
| 2024-02-21 | riscv64/itx: Reuse 16x4 epilog, saves 354 bytes | Nathan E. Egge |
| 2024-02-21 | riscv64/itx: Reuse 4x16 epilog, saves 642 bytes | Nathan E. Egge |
| 2024-02-21 | riscv64/itx: Fix unrolled .irp loops, saves 12 bytes | Nathan E. Egge |
