CheatPress

zstd Fast compression

Version 1.5.7 · source

zstd, Postcard, light preview Light · Postcard 102 × 152 mm · PDF zstd, Postcard, dark preview Dark · Postcard 102 × 152 mm · PDF zstd, Postcard, ink saver preview Ink saver · Postcard 102 × 152 mm · PDF
zstd, A5, light preview Light · A5 148 × 210 mm · PDF zstd, A5, dark preview Dark · A5 148 × 210 mm · PDF zstd, A5, ink saver preview Ink saver · A5 148 × 210 mm · PDF
zstd, A4, light preview Light · A4 210 × 297 mm · PDF zstd, A4, dark preview Dark · A4 210 × 297 mm · PDF zstd, A4, ink saver preview Ink saver · A4 210 × 297 mm · PDF
zstd, A4 landscape, light preview Light · A4 landscape 297 × 210 mm · PDF zstd, A4 landscape, dark preview Dark · A4 landscape 297 × 210 mm · PDF zstd, A4 landscape, ink saver preview Ink saver · A4 landscape 297 × 210 mm · PDF
zstd, A3, light preview Light · A3 297 × 420 mm · PDF zstd, A3, dark preview Dark · A3 297 × 420 mm · PDF zstd, A3, ink saver preview Ink saver · A3 297 × 420 mm · PDF
zstd, A3 landscape, light preview Light · A3 landscape 420 × 297 mm · PDF zstd, A3 landscape, dark preview Dark · A3 landscape 420 × 297 mm · PDF zstd, A3 landscape, ink saver preview Ink saver · A3 landscape 420 × 297 mm · PDF
zstd, Letter, light preview Light · Letter 8.5 × 11 in · PDF zstd, Letter, dark preview Dark · Letter 8.5 × 11 in · PDF zstd, Letter, ink saver preview Ink saver · Letter 8.5 × 11 in · PDF
zstd, Tabloid, light preview Light · Tabloid 11 × 17 in · PDF zstd, Tabloid, dark preview Dark · Tabloid 11 × 17 in · PDF zstd, Tabloid, ink saver preview Ink saver · Tabloid 11 × 17 in · PDF

Start here

zstd <file>
Compress to <file>.zst
zstd -d <file>.zst
Decompress
zstd -19 <file>
Smaller, slower
zstd -T0 <file>
Use all cores
zstd -l <file>.zst
Show sizes and ratio
zstd -dc <file>.zst
Print to stdout
tar --zstd -cf <a>.tar.zst <dir>
Archive a directory
zstd --rm <file>
Delete original after

Concepts

  1. --fast=N fastest, largest
  2. -3 default balance
  3. --ultra -22 smallest, slow, lots of RAM
Level versus speed

Compress

zstd <file>
Compress, keep original (<file>.zst)
zstd <file> -o <out>
Choose the output name
zstd -<n> <file>
Level 1-19 (default 3)
zstd -c <file> > <out>
Write to stdout, keep input
zstd --rm <file>
Remove input after success
zstd -f <file>
Overwrite existing output
zstd -r <dir>
Compress each file in a directory
Makes one .zst per file, not one archive. Use tar for a single archive.
zstd <a> <b> <c>
Compress several files at once
zstd --fast=<n> <file>
Faster than level 1, lower ratio
zstd --ultra -22 <file>
Levels 20-22, heavy memory use
Decompression then needs lots of memory too, so check the target machine.
zstd --adapt <file>
Auto-tune level to I/O speed
zstd --exclude-compressed -r <dir>
Skip already compressed files
zstd --no-check <file>
Skip the XXH64 checksum
zstd --format=gzip <file>
Write gzip, xz, lzma or lz4 instead

Decompress

zstd -d <file>.zst
Decompress, keep the .zst
unzstd <file>.zst
Same as zstd -d
zstd -dc <file>.zst
Decompress to stdout
zstdcat <file>.zst
Print a compressed file
zstdless <file>.zst
Page a compressed file
zstdgrep <pattern> <file>.zst
Search inside a compressed file
zstd -d <file>.zst -o <out>
Choose the output name
zstd -df <file>.zst
Overwrite existing output
zstd -d -M<mb> <file>.zst
Cap decompression memory
zstd --test <file>.zst
Verify integrity only
zstd -d --long=<n> <file>.zst
Needed if made with a large --long window
Windows above 27 bits must be allowed explicitly on decompression.

Inspect

zstd -l <file>.zst
Frames, sizes, ratio, checksum
zstd -lv <file>.zst
Verbose: window size, dictionary ID
zstd -b<n> <file>
Benchmark level n on a file
zstd -b1 -e19 <file>
Benchmark a range of levels
zstd -v <file>
Verbose output

Speed and memory

zstd -T0 <file>
Use all physical cores
zstd -T<n> <file>
Use n threads
zstd --auto-threads=logical -T0 <file>
Count logical cores for -T0
zstd --long <file>
Long-range matching for big, repetitive data
zstd --long=<n> <file>
Window of 2^n bytes (default 27)
zstd --rsyncable <file>
Output that rsync can delta well
zstd --single-thread <file>
One thread for I/O and compression
zstd -B<size> <file>
Set job size per thread
zstd --size-hint=<bytes>
Tune for piped input of known size

Pipes and tar

tar --zstd -cf <a>.tar.zst <dir>
Create a tar.zst
tar --zstd -xf <a>.tar.zst
Extract a tar.zst
tar -tf <a>.tar.zst
List contents
Modern GNU tar detects the zstd format on read, so --zstd is only needed when creating.
tar -I 'zstd -T0 -19' -cf <a>.tar.zst <dir>
Pick level and threads for tar
tar -cf - <dir> | zstd -T0 > <a>.tar.zst
Pipe tar into zstd
zstd -dc <a>.tar.zst | tar -xf -
Pipe zstd into tar
zstdmt <file>
Alias for multithreaded zstd
zstd -c <file> | ssh <host> 'zstd -d > <out>'
Compress over ssh

Dictionaries and patches

zstd --train <files> -o <dict>
Build a dictionary from samples
Needs many small, similar sample files to work well.
zstd --train --maxdict=<bytes> <files>
Limit the dictionary size
zstd -D <dict> <file>
Compress with a dictionary
zstd -D <dict> -d <file>.zst
Decompress with the same dictionary
zstd --patch-from=<old> <new> -o <patch>
Make a binary diff against old
zstd -d --patch-from=<old> <patch> -o <new>
Apply that patch to old
zstd --filelist <list>
Read file names from a list
zstd --output-dir-flat <dir> <files>
Put outputs in one directory
zstd --output-dir-mirror <dir> -r <src>
Mirror the source tree structure

Recipes

Fast backup archive
tar -I 'zstd -T0 -10' -cf backup.tar.zst <dir>zstd --test backup.tar.zstzstd -l backup.tar.zst
Maximum compression for a release
zstd --ultra -22 -T0 --long=27 <file> -o <file>.zst# decompress with the same windowzstd -d --long=27 <file>.zst
Dump a database through zstd
pg_dump <db> | zstd -T0 > <db>.sql.zstzstd -dc <db>.sql.zst | psql <db>
Dictionary for many small files
zstd --train logs/*.json -o logs.dictzstd -D logs.dict logs/*.jsonzstd -D logs.dict -d logs/*.zst