| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| linux-amd64-ffmpeg-over-ip-client.zip | 2026-06-14 | 1.3 MB | |
| linux-amd64-ffmpeg-over-ip-server.zip | 2026-06-14 | 82.9 MB | |
| linux-arm64-ffmpeg-over-ip-client.zip | 2026-06-14 | 1.2 MB | |
| linux-arm64-ffmpeg-over-ip-server.zip | 2026-06-14 | 76.7 MB | |
| macos-amd64-ffmpeg-over-ip-client.zip | 2026-06-14 | 1.3 MB | |
| macos-amd64-ffmpeg-over-ip-server.zip | 2026-06-14 | 54.2 MB | |
| macos-arm64-ffmpeg-over-ip-client.zip | 2026-06-14 | 1.2 MB | |
| macos-arm64-ffmpeg-over-ip-server.zip | 2026-06-14 | 49.1 MB | |
| SHA256SUMS | 2026-06-14 | 1.0 kB | |
| windows-amd64-ffmpeg-over-ip-client.zip | 2026-06-14 | 1.3 MB | |
| windows-amd64-ffmpeg-over-ip-server.zip | 2026-06-14 | 79.2 MB | |
| README.md | 2026-06-14 | 749 Bytes | |
| v5.2.1 source code.tar.gz | 2026-06-14 | 257.2 MB | |
| v5.2.1 source code.zip | 2026-06-14 | 257.3 MB | |
| Totals: 14 Items | 862.8 MB | 0 | |
Performance
- Read-ahead, async prefetch, and range cache for tunneled fio reads (#24) —
the server-side fio layer now batches reads, prefetches the next block while
the current one is consumed, and caches recently-read ranges, cutting
request/response round trips for readers like FFmpeg's AVIO layer that pull
~32 KiB at a time. Reduces over-IP overhead on sequential workloads from
roughly 8–14× slower than local to under 2×. Tunable via the
FFOIP_READAHEAD_BYTESandFFOIP_RANGE_CACHE_BYTESenvironment variables (setFFOIP_READAHEAD_BYTES=0to disable). No protocol or configuration changes are required.
Full Changelog: https://github.com/steelbrain/ffmpeg-over-ip/compare/v5.2.0...v5.2.1