ZFP(1)

💖 ZFP(1) 💖

okay bestie, welcome to the zfp manual page — this is the User Commands girlie and she's here to compress or expand files using the zfp algorithm. no cap.

---

✨ NAME ✨

zfp — compress or expand files using the zfp algorithm 🗜️

---

✨ SYNOPSIS ✨

zfp <options>

---

🌸 DESCRIPTION 🌸

zfp version 1.0.0 (August 1, 2022) — yes she has a birthday and it's literally August 1, 2022. 🎂

---

🫶 General options:

• -h — read/write array and compression parameters from/to compressed header
• -q — quiet mode; suppress output 🤫
• -s — print error statistics

---

🎀 Input and output:

• -i <path> — uncompressed binary input file ("-" for stdin)
• -o <path> — decompressed binary output file ("-" for stdout)
• -z <path> — compressed input (w/o -i) or output file ("-" for stdin/stdout)

---

🦋 Array type and dimensions (needed with -i):

• -f — single precision (float type)
• -d — double precision (double type)
• -t <i32|i64|f32|f64> — integer or floating scalar type
• -1 <nx> — dimensions for 1D array a[nx]
• -2 <nx> <ny> — dimensions for 2D array a[ny][nx]
• -3 <nx> <ny> <nz> — dimensions for 3D array a[nz][ny][nx]
• -4 <nx> <ny> <nz> <nw> — dimensions for 4D array a[nw][nz][ny][nx]

---

💅 Compression parameters (needed with -i):

• -R — reversible (lossless) compression
• -r <rate> — fixed rate (# compressed bits per floating-point value)
• -p <precision> — fixed precision (# uncompressed bits per value)
• -a <tolerance> — fixed accuracy (absolute error tolerance)
• -c <minbits> <maxbits> <maxprec> <minexp> — advanced usage
  • minbits : min # bits per 4^d values in d dimensions
  • maxbits : max # bits per 4^d values in d dimensions (0 for unlimited)
  • maxprec : max # bits of precision per value (0 for full)
  • minexp : min bit plane # coded (-1074 for all bit planes)

bestie -c is the "advanced usage" one — she's giving final boss energy. you don't have to understand her, but she's iconic. 👑

---

⭐ Execution parameters:

• -x serial — serial compression (default)
• -x omp[=threads[,chunk_size]] — OpenMP parallel compression
• -x cuda — CUDA fixed rate parallel compression/decompression

---

🌈 EXAMPLES 🌈

like, here's the tea on how you'd actually call her:

-i file : read uncompressed file and compress to memory

-z file : read compressed file and decompress to memory

-i ifile -z zfile : read uncompressed ifile, write compressed zfile

-z zfile -o ofile : read compressed zfile, write decompressed ofile

-i ifile -o ofile : read ifile, compress, decompress, write ofile

-i file -s : read uncompressed file, compress to memory, print stats

-i - -o - -s : read stdin, compress, decompress, write stdout, print stats

-f -3 100 100 100 -r 16 : 2x fixed-rate compression of 100x100x100 floats

-d -1 1000000 -r 32 : 2x fixed-rate compression of 1M doubles

-d -2 1000 1000 -p 32 : 32-bit precision compression of 1000x1000 doubles

-d -1 1000000 -a 1e-9 : compression of 1M doubles with < 1e-9 max error

-d -1 1000000 -c 64 64 0 -1074 : 4x fixed-rate compression of 1M doubles

-x omp=16,256 : parallel compression with 16 threads, 256-block chunks

and then — not me about to explain this — she ALSO spits out this whole thing again, like the usage + every single option a second time, which is honestly so extra of her 💀 so here it is, in full, exactly as she says it:

Usage: zfp <options> General options:

-h : read/write array and compression parameters from/to compressed
       header

-q : quiet mode; suppress output

-s : print error statistics

Input and output:

• -i <path> — uncompressed binary input file ("-" for stdin)
• -o <path> — decompressed binary output file ("-" for stdout)
• -z <path> — compressed input (w/o -i) or output file ("-" for stdin/stdout)

Array type and dimensions (needed with -i):

• -f — single precision (float type)
• -d — double precision (double type)
• -t <i32|i64|f32|f64> — integer or floating scalar type
• -1 <nx> — dimensions for 1D array a[nx]
• -2 <nx> <ny> — dimensions for 2D array a[ny][nx]
• -3 <nx> <ny> <nz> — dimensions for 3D array a[nz][ny][nx]
• -4 <nx> <ny> <nz> <nw> — dimensions for 4D array a[nw][nz][ny][nx]

Compression parameters (needed with -i):

• -R — reversible (lossless) compression
• -r <rate> — fixed rate (# compressed bits per floating-point value)
• -p <precision> — fixed precision (# uncompressed bits per value)
• -a <tolerance> — fixed accuracy (absolute error tolerance)
• -c <minbits> <maxbits> <maxprec> <minexp> — advanced usage
  • minbits : min # bits per 4^d values in d dimensions
  • maxbits : max # bits per 4^d values in d dimensions (0 for unlimited)
  • maxprec : max # bits of precision per value (0 for full)
  • minexp : min bit plane # coded (-1074 for all bit planes)

Execution parameters:

• -x serial — serial compression (default)
• -x omp[=threads[,chunk_size]] — OpenMP parallel compression
• -x cuda — CUDA fixed rate parallel compression/decompression

and the examples, round two, same energy, same words:

-i file : read uncompressed file and compress to memory

-z file : read compressed file and decompress to memory

-i ifile -z zfile : read uncompressed ifile, write compressed zfile

-z zfile -o ofile : read compressed zfile, write decompressed ofile

-i ifile -o ofile : read ifile, compress, decompress, write ofile

-i file -s : read uncompressed file, compress to memory, print stats

-i - -o - -s : read stdin, compress, decompress, write stdout, print stats

-f -3 100 100 100 -r 16 : 2x fixed-rate compression of 100x100x100 floats

-d -1 1000000 -r 32 : 2x fixed-rate compression of 1M doubles

-d -2 1000 1000 -p 32 : 32-bit precision compression of 1000x1000 doubles

-d -1 1000000 -a 1e-9 : compression of 1M doubles with < 1e-9 max error

-d -1 1000000 -c 64 64 0 -1074 : 4x fixed-rate compression of 1M doubles

-x omp=16,256 : parallel compression with 16 threads, 256-block chunks

---

and that's the whole page, bestie 🫶 you ate. go compress something. periodt. 💅✨

zfp version 1.0.0 — August 1, 2022, so basically a Leo ♌ — stamped August 2022, ZFP(1). mwah, bye 💖🎀