No problem !
I'll release a new version soon
joerg (15th May 2014)
@nania:
by the way your program zcm version 091a and version 091b works like a charm for me
- compress and uncompress my datafile correctly
- i have used only the standard syntax without any extra switches:
"zcm a archivefile datafile" and "zcm x archivefile"
and it compress my datafile (oracle-dmp-file) better then 7zip (max compression) - in one word - it is wonderful
but, for using "zcm" in a regular backup-process
- it seems neccessary for me
to save the original filedate and original filetime of the datafile within the archivefile
@nania: is there any chance to solve this?
best regards
Dear joerg thank you for the compliments.
With regard to the fact that it is not an archiver with all the features that you said (filetime, filedate, empty folder) will remedy shortly. I intend to build a structure that can use it for all my future plans. I had some problems that I couldn't easily resolve as if you consider well, each individual file must have all the data you've pointed out.
For now you'll want to compress a TAR file
ZCM 0.92 released!
I had to remove from the search buffer 128 MB
News:
- New CM core!
- More stable !
- Corrected some bugs
Only from:
http://heartofcomp.altervista.org/
@Matt
Please if possible test ZCM in LTCB and 10GB Benchmark's
Last edited by Nania Francesco; 17th May 2014 at 01:02.
PiPPoNe92 (18th July 2014)
LTCB results. http://mattmahoney.net/dc/text.html#1629
Testing 10GB now.
Nania Francesco (17th May 2014)
Thanks for all test Matt!
I did the interventions for multi-processing correcting some bugs, I think should succeed the test 10GB as from me with 4 core works perfectly
Last edited by Nania Francesco; 17th May 2014 at 07:06.
Nice improvement using 1 thread. http://mattmahoney.net/dc/10gb.html
2969410664 4545 3875 4 zcm 0.92 -m7 -t1 -r -s
3234943741 4312 3804 4 zcm 0.90 -m7 -t1 -r -s
Nania Francesco (17th May 2014)
This time my choice to change the structure of the data analysis gave its fruit. The program is a little slower because I used a unique hash for search latest buffer 2 UNSIGNED LONG while before they used two different types. It will be difficult to improve but it's not called the last word. The program has achieved the great giants of compression but ZPAQ is always very powerful.
I also tested on 10GB with 2 threads.
2969410664 4545 3875 4 zcm 0.92 -m7 -t1 -r -s
3278149230 3694 3042 4 zcm 0.92 -m7 -t2 -r -s
Nania Francesco (18th May 2014)
Thanks Matt !
@nania: can you compile a linux-version (for example redhat 6.x) of your wonderful compression program in the near future?
if it is too difficult to implement all features, so may be you can implement only a few features
"-t1 = only one task" and "-m4 = only method 4"
best regards
Joerg
Intensity (3rd August 2015)
I am testing in Ubuntu using Wine 1.6. Works OK.
I realize that creating an interesting archiver requires me to release a full version of all the features of archiver. Definitely do not want to make the mistake that I made with other archiver's which I have released which do not maintain backward compatibility and does not allow to store folder empty or full attributes (date and time).
Bulat Ziganshin (23rd May 2014)
@nania:
thanks for your answers.
---
archiver requires me to release a full version of all the features of archiver
... maintain backward compatibility .. store .. filetime, filedate, empty folder ..
---
Of course it would be wonderful to have such a full zcm-archiver
further you wrote: "For now .. compress a TAR file"
I have tested it with a tar-file and with a zip-file (without internal compression)
the result is a archivefile.tar.zcm and a archivefile.zip.zcm
and the result is for me okay
@matt: thank you very much for your hint "using wine"
best regards
Joerg
for the moment I am leaving to work at ZCM because I'm dedicating a new compressor LZ77 decompression with great speed (type LZ4, ZHUFF etc) and with a simple Bytecoder seems to get good results.
Last edited by Nania Francesco; 28th May 2014 at 13:28.
Version 0.92 release x64 compatible with the 32 bit version. Adding only option-m8 that requires maximum 2.9 GB!
@Matt
Please test this version in LTCB and 10GB benchmarks!
Biozynotiker (4th July 2014),Stephan Busch (4th July 2014)
enwik8 19,700,970 bytes, 27.788 sec. - 28.336 sec., zcm_x64 -m8 -t1
enwik9 160,848,538 bytes, 243.064 sec. - 248.084 sec. , zcm_x64 -m8 -t1
enwik8 21,582,135 bytes, 6.447 sec. - 6.317 sec., zcm_x64 -m8 -t8
enwik9 170,176,223 bytes, 53.018 sec. - 53.137 sec., zcm_x64 -m8 -t8
Nania Francesco (4th July 2014)
So far, ZCM_x64 is 400 sec faster than the 32-bit variant on the squeezeChart testset.
But I think timer.exe is not taking the correct timings.. Which timers do you use?
Nania Francesco (4th July 2014)
I used AcuTimer v2.0 for that test.Quote Originally Posted by Stephan Busch View PostWhich timers do you use?
I redid tests (two times each) with zcm 32-bit and 64-bit versions in WinZpaq benchmark mode (.NET stopwatch timings):
enwik8 19,700,970 bytes, 27.786 sec. - 28.329 sec., ZCM - 8 - 1 - zcm 0.92 64-bit
enwik8 19,700,970 bytes, 27.788 sec. - 28.329 sec., ZCM - 8 - 1 - zcm 0.92 64-bit
enwik8 19,803,545 bytes, 27.213 sec. - 27.125 sec., ZCM - 7 - 1 - zcm 0.92 32-bit
enwik8 19,803,545 bytes, 27.015 sec. - 27.089 sec., ZCM - 7 - 1 - zcm 0.92 32-bit
enwik9 160,848,538 bytes, 243.582 sec. - 248.075 sec., ZCM - 8 - 1 - zcm 0.92 64-bit
enwik9 160,848,538 bytes, 243.925 sec. - 248.403 sec., ZCM - 8 - 1 - zcm 0.92 64-bit
enwik9 163,246,657 bytes, 240.857 sec. - 240.637 sec., ZCM - 7 - 1 - zcm 0.92 32-bit
enwik9 163,246,657 bytes, 240.858 sec. - 240.777 sec., ZCM - 7 - 1 - zcm 0.92 32-bit
enwik8 21,582,135 bytes, 6.287 sec. - 6.076 sec., ZCM - 8 - 8 - zcm 0.92 64-bit
enwik8 21,582,135 bytes, 6.130 sec. - 6.239 sec., ZCM - 8 - 8 - zcm 0.92 64-bit
enwik8 21,582,135 bytes, 6.324 sec. - 6.083 sec., ZCM - 7 - 8 - zcm 0.92 32-bit
enwik8 21,582,135 bytes, 6.309 sec. - 6.346 sec., ZCM - 7 - 8 - zcm 0.92 32-bit
enwik9 170,176,223 bytes, 53.112 sec. - 52.361 sec., ZCM - 8 - 8 - zcm 0.92 64-bit
enwik9 170,176,223 bytes, 53.733 sec. - 53.415 sec., ZCM - 8 - 8 - zcm 0.92 64-bit
enwik9 170,176,223 bytes, 53.363 sec. - 53.298 sec., ZCM - 7 - 8 - zcm 0.92 32-bit
enwik9 170,176,223 bytes, 53.300 sec. - 53.801 sec., ZCM - 7 - 8 - zcm 0.92 32-bit
Nania Francesco (4th July 2014)
Dear Sportman.. could you please provide a download link for AcuTimer 2.0? I didn't find it.
Quote Originally Posted by Stephan Busch View Postcould you please provide a download link for AcuTimer 2.0?
Stephan Busch (5th July 2014)
ZCM 0.92x64
@Matt
Please test this version in LTCB and 10GB benchmarks!
The 64 bit version compresses the same but is a little faster. http://mattmahoney.net/dc/text.html#1629
Edit: likewise on 10GB: http://mattmahoney.net/dc/10gb.html
Last edited by Matt Mahoney; 8th July 2014 at 07:23.
Nania Francesco (8th July 2014)
Thanks Matt for all test, but option -m8 not work?
The help message says -m0 ... -m7 so I didn't try -m8.
maybe he meant -t8
Excuse me Matt x64 version is not official !
x64 -m8 has better compression, close to Pareto frontier on 10GB system 4. Since it is not official do you want me to remove it?
http://mattmahoney.net/dc/10gb.html
Nania Francesco (9th July 2014)
You can also leave. I meant 'not official release' because not inserted into my site. This version was a kind of experiment, considering that I have no great experience with 64-BIT versions.
Nice improvement on LTCB. It now beats nanozipltcb so it makes the Pareto frontier. :)
http://mattmahoney.net/dc/text.html#1610
Nania Francesco (10th July 2014)