On bookworm with the upcoming lzma 24.08:
$ time gzip --best -c 18x18ko.pcf > 18x18ko.pcf.gz
real 0m1.515s
user 0m1.507s
sys 0m0.008s
$ time ../build-output/lzma_24.08+dfsg-1_amd64/usr/bin/lzmp --best -c
18x18ko.pcf > 18x18ko.pcf.lzma
real 0m1.334s
user 0m1.912s
sys 0m0.078s
$ l 18x18ko.pcf*
-rw-r--r-- 1 nabijaczleweli users 2.8M 10-24 10:14 18x18ko.pcf
-rw-r--r-- 1 nabijaczleweli users 709.3k 10-24 10:16 18x18ko.pcf.gz
-rw-r--r-- 1 nabijaczleweli users 365.3k 10-24 10:16 18x18ko.pcf.lzma
$ hyperfine '/bin/gzip -dc 18x18ko.pcf.gz'
'../build-output/lzma_24.08+dfsg-1_amd64/usr/bin/lzmp -dc 18x18ko.pcf.lzma'
Benchmark 1: /bin/gzip -dc 18x18ko.pcf.gz
Time (mean ± σ): 27.8 ms ± 2.9 ms [User: 26.1 ms, System: 1.7 ms]
Range (min … max): 23.0 ms … 31.9 ms 96 runs
Benchmark 2: ../build-output/lzma_24.08+dfsg-1_amd64/usr/bin/lzmp -dc
18x18ko.pcf.lzma
Time (mean ± σ): 50.0 ms ± 4.5 ms [User: 47.0 ms, System: 3.0 ms]
Range (min … max): 43.9 ms … 55.8 ms 59 runs
Summary
'/bin/gzip -dc 18x18ko.pcf.gz' ran
1.80 ± 0.25 times faster than
'../build-output/lzma_24.08+dfsg-1_amd64/usr/bin/lzmp -dc 18x18ko.pcf.lzma'
compression gz=365k lzma=709k => 1.94x better and decompression speed is 1.8x
slower, so this is a win on both fronts now
(OP has same compression numbers but lzma is 4.42x slower, so this is
clearly the effect of speeding up the decompressor)
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/446431
Title:
unlzma is 4 times slower than gunzip
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/lzma/+bug/446431/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
