LZMA recompression

[twisted89]

Hi,

Could someone explain to me why this file doesn't match the original compressed file when decompressed then recompressed with LZMA on default settings? The properties are set to default as far as I can see so I've no idea why they're different.

Thanks

Compressed file: removed

[Bulat Ziganshin]

lzma is indeterministic algorithm, it doesn't gurantee that compressed data will be same for the same input data

[twisted89]

Ah right, thanks for the answer.

[Shelwien]

to be specific, there're lots of parameters which affect compression:

 -a{N}:  set compression mode - [0, 1], default: 1 (max)
 -d{N}:  set dictionary size - [12, 30], default: 23 (8MB)
 -fb{N}: set number of fast bytes - [5, 273], default: 128
 -mc{N}: set number of cycles for match finder
 -lc{N}: set number of literal context bits - [0, 8], default: 3
 -lp{N}: set number of literal pos bits - [0, 4], default: 0
 -pb{N}: set number of pos bits - [0, 4], default: 2
 -mf{MF_ID}: set Match Finder: [bt2, bt3, bt4, hc4], default: bt4
 -mt{N}: set number of CPU threads

fortunately there's mostly one encoder implementation, but its still necessary to know all the parameter values used for encoding.
lc/lp/pb can be detected with lzmarec, but there're still many others.

Also note that .7z has some additional filters which can be applied before lzma compression, which adds even more parameters.

[Bulat Ziganshin]

m/t lzma results may vary because one thread calculates probabilities used in another thread so it may update probs earlier or later. or may not vary. definitely m/t and s/t versions produce different results