There's a nice comparison of compression algorithms (including zlib, zstd, brotli, snappy, etc) here: https://quixdb.github.io/squash-benchmark/ It's nice because it uses many datasets and machine platforms.

Unfortunately the graphs leave a bit to be desired, especially when you're trying to compare two algorithms. They provide the raw data, but it needs a little munging to make it workable.

For my day job I made the following graph to compare zlib level1 and zstd. This was using the "peltast" platform which is a Xeon based system, because that was most relevant for us.

http://imgur.com/a/0a3kK

I'll make one with brotli now.

Imgur album now compares brotli and zstd. Looks like zstd is always significantly faster, but Brotli compresses slightly better for every dataset except the XML one.

This is with brotli level 1, by the way. My understanding is that brotli is pretty quick through the first few levels, but the levels that ask for the highest compression are insanely slow (which is a valuable thing to have as an option, for things like game assets or something, which are compressed once and delivered many times!)

For "compress once, decompress many times" workloads like game assets, you might want to check out Oodle, a proprietary library that kills zstd in decompression speed while achieving better compression ratios.

http://cbloomrants.blogspot.com/2016/07/introducing-oodle-me...

I thought that brotli was tuned for typical web workloads, that it contained a dictionary tuned for web workloads. Our internal testing shows that it performs very poorly for binary 3D vector data.

So a test between zstd and brotli would show brotli in a poor light if it used a mixed corpus, but a test between zstd and brotli on a web corpus would give an advantage to brotli...

Not just tuned for web workloads in general, but for specific web workloads. The Brotli dictionary is mostly composed of English words and phrases, and fragments of HTML, CSS, and Javascript. It would perform poorly on non-English text.

I have a feeling that the dictionary was designed with the specific goal of performing well on a specific corpus similar to the Large Text Compression Benchmark[1]. It has quite a few words and phrases that I'd associate with Wikipedia's "house style".

[1]: https://cs.fit.edu/~mmahoney/compression/textdata.html

The dictionary also contains lots of Chinese, Russian and Arabic

By the numbers:

- 9216 phrases total

- 5857 (63.5%) pure ASCII phrases, mostly English with a few Spanish words thrown in

- 1372 (14.8%) code fragments -- mostly HTML, CSS, and Javascript

- 1027 (11.1%) CJK (Chinese, Japanese, and Korean, and mostly the first two) phrases -- it's very hard to tell Chinese and Japanese apart in this context; I didn't try

- 158 (1.7%) phrases containing extended Latin-1 characters (nearly all Spanish words)

- 303 (3.3%) Cyrillic script (probably Russian) phrases

- 322 (3.5%) Arabic phrases

- 172 (1.9%) Devanagari script (Hindi) phrases

Plus a few miscellaneous other scripts and generally unclassifiable content.

'tuned' for the brotli dictionary is a bit generous [1], but yes, it contains a smattering of strings [2] you'd find in plain text and web documents.

Because of its seemingly haphazard dictionary, I wouldn't rule out zstd outperforming brotli, if trained on a good dataset.

[1] https://news.ycombinator.com/item?id=12010313

[2] https://gist.github.com/klauspost/2900d5ba6f9b65d69c8e

>Our internal testing shows that it performs very poorly for binary 3D vector data.

Uh ? What format of data was this ?

I did a pretty large test (2gb+) on OBJ/STL 3d data and brotli compressed within ~5% margin of lzma, and this holds true on other binary data I've compared.

It also compressed better than zstd (as in compression ratio) on the same data at their highest respective compression settings:

bro -quality 11 -window 24

zstd --ultra -22

So I find it baffling that you find brotli to be poor for any binary data, could you share the data in question ?

What is the compression time for brotli to achieve compression ratios comparable to LZMA on these types of data sets: https://github.com/google/brotli/issues/165

I understand that brotli is incredibly slow as compared to LZMA when using these high ratio settings (q11, w24), so slow as to be impractical in production even in a write-once, read-many scenario if you have any non-trivial amount of data being regularly produced. I do not want to have a farm of machines just to handle the brotli compression load of our data sets just because it is 10x slower than LZMA.

>What is the compression time for brotli to achieve compression ratios comparable to LZMA...

Compression time is indeed the achilles heel of Brotli, which is why it's something I would only use for compress once (and preferably decompress very often) scenarios.

Compared to lzma at it's best compression setting for this particular data (-mx9 -m0=LZMA:d512m:fb273:lc8), brotli took 6 minutes and 4 seconds to compress, while the same data took 1 minute and 47 seconds for lzma.

On the other hand, brotli decompressed the same data in 0.6 seconds, while it took lzma 2.2 seconds.

Brotli is available in this graph - https://github.com/facebook/zstd/blob/master/images/DCspeed5...

This seems to be slightly old doc - from April.

The README at https://github.com/facebook/zstd mentions brotli

Slide 4 here has one: https://dl.dropboxusercontent.com/u/12405967/ANSsem.pdf

Would be especially interesting to have zstd work with brotli's dictionary for the comparison.