zxc(1)¶

NAME¶

zxc - High-performance asymmetric lossless compression

SYNOPSIS¶

zxc [OPTIONS] [INPUT-FILE] [OUTPUT-FILE]

DESCRIPTION¶

zxc is a command-line interface for the ZXC compression library, a high-performance lossless compression algorithm optimized for maximum decompression throughput.

zxc is designed for the “Write Once, Read Many” paradigm. It trades compression speed to generate a bitstream specifically structured to maximize decompression speed, effectively offloading complexity from the decoder to the encoder. It aims to provide very high decompression speeds across modern architectures while maintaining competitive compression ratios. ZXC is particularly suited for scenarios such as Game Assets, Firmware or App Bundles where data is compressed once on a build server and decompressed millions of times on user devices.

By default, zxc compresses a single INPUT-FILE. If no OUTPUT-FILE is provided, zxc will automatically append the .zxc extension to the input filename. If no INPUT-FILE is provided, zxc will read from standard input (stdin) and write to standard output (stdout).

STANDARD MODES¶

-z, –compress

Compress FILE. This is the default mode if no mode is specified.

-d, –decompress

Decompress FILE.

-l, –list

List archive information, including compressed size, uncompressed size, compression ratio, and checksum method.

-t, –test

Test the integrity of a compressed FILE. It decodes the file and verifies its checksum (if present) without writing any output.

-b, –bench [N]

Benchmark in-memory performance. Loads the input file entirely into RAM and measures raw algorithm throughput (default duration is 5 seconds).

OPTIONS¶

-m, –multiple

Process multiple files at once. When specified, all subsequent non-option arguments are treated as input files. For each input file, a corresponding .zxc file is created (or decompressed into its original name). Output cannot be written to standard output (stdout) when this mode is enabled.

-r, –recursive

Recursively process directories. When specified, any directory listed as an argument will be traversed, and all regular files within it will be processed (compressed or decompressed). This option implicitly enables --multiple mode.

-1..-6

Set the compression level from 1 (fastest compression) to 6 (highest density).

• -1, -2 (Fast): Optimized for real-time assets or when compression speed is a priority.
• -3 (Default): Balanced middle-ground offering efficient compression and superior ratio to fast codecs.
• -4, -5 (Compact): Better ratio than LZ4 with faster decoding than Zstd. Suited for embedded systems and firmware.
• -6 (Max): Highest ratio tier, matching LZ4-HC while keeping ZXC’s decode advantage. Best for archival and write-once / read-many workloads where compression time is amortized over many reads.

-T, –threads N

Set the number of threads to use for compression and decompression. A value of 0 means auto-detection based on the number of available CPU cores.

-B, –block-size SIZE

Set the compression block size. SIZE must be a power of two between 4K and 2M (e.g. 4K, 64K, 256K, 512K, 1M). Smaller blocks reduce memory usage and improve random-access decompression; larger blocks generally yield better compression ratios. The default is 512K, tuned for bulk/archival workloads where ratio and decompression throughput matter most. This option is only meaningful during compression; the block size is stored in the archive header and automatically used during decompression.

-C, –checksum

Enable block hashing during compression using the rapidhash algorithm. Recommended for data integrity validation. Checksum verification is automatically performed during extraction when enabled.

-N, –no-checksum

Explicitly disable checksum generation.

-S, –seekable

Append a seek table to the archive during compression. This transforms the file into a random-access format (Seekable Archive), allowing the decoder to instantly locate and decompress specific blocks in O(1) time without reading the entire file. Ideal for compressed filesystems, game assets, and log analysis.

-k, –keep

Keep the input file after compression or decompression. (Currently, the input file is preserved by default, but this flag ensures compatibility with future changes).

-f, –force

Force overwrite of the OUTPUT-FILE if it already exists.

-c, –stdout

Force writing to standard output (stdout), even if it is the console.

-v, –verbose

Enable verbose logging mode. Outputs more detailed information during operations.

-q, –quiet

Enable quiet mode, suppressing all non-error output (such as progress bars or real-time statistics).

-j, –json

Output results in JSON format. This is particularly useful for scripting, benchmarking, and the --list mode.

SPECIAL OPTIONS¶

-V, –version

Display the version of the zxc library and the compiled architecture information, then exit.

-h, –help

Display a help message and exit.

EXAMPLES¶

Compress a file: zxc data.txt

Compress a file with high density (Level 6, archival): zxc -6 data.bin

Decompress a file: zxc -d data.txt.zxc

Compress multiple files independently: zxc -m file1.txt file2.txt file3.txt

Compress all files in a directory recursively: zxc -r ./my_folder

Decompress all files in a directory recursively: zxc -d -r ./my_folder

Decompress a file to standard output: zxc -dc data.txt.zxc > data.txt

List archive information: zxc -l data.txt.zxc

Compress with a custom block size (64 KB): zxc -B 64K data.bin data.zxc

Compress with maximum block size (2 MB): zxc -6 -B 2M data.bin data.zxc

Run a benchmark for 10 seconds: zxc -b 10 data.txt

AUTHORS¶

Written by Bertrand Lebonnois & contributors.

LICENSE¶

BSD 3-Clause License.