09f2921b5d
This patch ports the LZ4 decompression code that debuted in libpayload last year to coreboot for use in CBFS stages (upgrading the base algorithm to LZ4's dev branch to access the new in-place decompression checks). This is especially useful for pre-RAM stages in constrained SRAM-based systems, which previously could not be compressed due to the size requirements of the LZMA scratchpad and bounce buffer. The LZ4 algorithm offers a very lean decompressor function and in-place decompression support to achieve roughly the same boot speed gains (trading compression ratio for decompression time) with nearly no memory overhead. For now we only activate it for the stages that had previously not been compressed at all on non-XIP (read: non-x86) boards. In the future we may also consider replacing LZMA completely for certain boards, since which algorithm wins out on boot speed depends on board-specific parameters (architecture, processor speed, SPI transfer rate, etc.). BRANCH=None BUG=None TEST=Built and booted Oak, Jerry, Nyan and Falco. Measured boot time on Oak to be about ~20ms faster (cutting load times for affected stages almost in half). Change-Id: Iec256c0e6d585d1b69985461939884a54e3ab900 Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://review.coreboot.org/13638 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org> |
||
|---|---|---|
| .. | ||
| lib | ||
| Makefile | ||
| NEWS | ||
| README.md | ||
| lz4_Block_format.md | ||
| lz4_Frame_format.md | ||
README.md
LZ4 - Extremely fast compression
LZ4 is lossless compression algorithm, providing compression speed at 400 MB/s per core, scalable with multi-cores CPU. It features an extremely fast decoder, with speed in multiple GB/s per core, typically reaching RAM speed limits on multi-core systems.
Speed can be tuned dynamically, selecting an "acceleration" factor which trades compression ratio for more speed up. On the other end, a high compression derivative, LZ4_HC, is also provided, trading CPU time for improved compression ratio. All versions feature the same decompression speed.
LZ4 library is provided as open-source software using BSD license.s
| Branch | Status |
|---|---|
| master |   |
| dev |  |
Branch Policy:
- The "master" branch is considered stable, at all times.
- The "dev" branch is the one where all contributions must be merged before being promoted to master.
- If you plan to propose a patch, please commit into the "dev" branch, or its own feature branch. Direct commit to "master" are not permitted.
Benchmarks
The benchmark uses the Open-Source Benchmark program by m^2 (v0.14.3) compiled with GCC v4.8.2 on Linux Mint 64-bits v17. The reference system uses a Core i5-4300U @1.9GHz. Benchmark evaluates the compression of reference Silesia Corpus in single-thread mode.
| Compressor | Ratio | Compression | Decompression |
|---|---|---|---|
| memcpy | 1.000 | 4200 MB/s | 4200 MB/s |
| LZ4 fast 17 (r129) | 1.607 | 690 MB/s | 2220 MB/s |
| LZ4 default (r129) | 2.101 | 385 MB/s | 1850 MB/s |
| LZO 2.06 | 2.108 | 350 MB/s | 510 MB/s |
| QuickLZ 1.5.1.b6 | 2.238 | 320 MB/s | 380 MB/s |
| Snappy 1.1.0 | 2.091 | 250 MB/s | 960 MB/s |
| LZF v3.6 | 2.073 | 175 MB/s | 500 MB/s |
| zlib 1.2.8 -1 | 2.730 | 59 MB/s | 250 MB/s |
| LZ4 HC (r129) | 2.720 | 22 MB/s | 1830 MB/s |
| zlib 1.2.8 -6 | 3.099 | 18 MB/s | 270 MB/s |
Documentation
The raw LZ4 block compression format is detailed within lz4_Block_format.
To compress an arbitrarily long file or data stream, multiple blocks are required. Organizing these blocks and providing a common header format to handle their content is the purpose of the Frame format, defined into lz4_Frame_format. Interoperable versions of LZ4 must respect this frame format.
Other source versions
Beyond the C reference source, many contributors have created versions of lz4 in multiple languages (Java, C#, Python, Perl, Ruby, etc.). A list of known source ports is maintained on the LZ4 Homepage.