Page 5 of 6 FirstFirst ... 3456 LastLast
Results 121 to 150 of 178

Thread: TurboBench - Back to the future (incl. LzTurbo)

  1. #121
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    Anyway, is this hard to compile? has anyone tried?
    I'm also using mingw-w64 to test and benchmark every turbobench version on windows.
    If you can successfully compile and run a c "hello world" program on windows,
    then I'see no reason why turbobench shouldn't work.

    - Download and Install: Mingw-w64
    Test the installation by typing "gcc -v" and "make -v" ( or "mingw32-make -v")

    - clone turbobench (do not forget "recursive") and compile with make:
    git clone --recursive git://github.com/powturbo/TurboBench.git
    cd TurboBench
    make (or "make STATIC=1" )

    - test you compile by typing:
    turbobench --h


    - read "https://github.com/powturbo/TurboBench" for other compile options


    Some codecs do not compile, you must pass options to exclude them.
    Witch one? All default codecs should compile without errors on windows.


    The libraries libgcc_s_seh-1.dll, libstdc++-6.dll and libwinpthread-1.dll are in the same directory.
    None of the previous versions is working?
    use the static "turbobenchs.exe" (instead of turbobench.exe)
    Please read the "README" file included in your "turbobench.zip" or "turbobench.7z" package

  2. #122
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    TurboBench Compressor benchmark

    TurboBench Compressor benchmark github 2019 update:

    - NEW: ULZ codec added
    - Thanks to Turbobench submodule architecture, all codecs with latest version as usual
    - Benchmark updated with latest version of all compressors.
    c-blosc2
    fast-lzma2
    isa-l
    libdeflate
    lizard
    lz4
    lzma
    miniz
    snappy
    zlib-ng
    zstd


  3. Thanks (4):

    Conor (30th March 2019),encode (30th March 2019),JamesB (5th June 2019),Mike (29th March 2019)

  4. #123
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    zlib-ng is now integrated into TurboBench - Compression Benchmark.

    The current zlib-ng (SSE4.2) in also benchmarked in the Static/Dynamic web content compression benchmark.
    Better compression and decompression speed than zlib. Decompression is even faster than brotli.
    Last edited by dnd; 4th June 2019 at 22:55.

  5. Thanks:

    Jyrki Alakuijala (4th June 2019)

  6. #124
    Member SolidComp's Avatar
    Join Date
    Jun 2015
    Location
    USA
    Posts
    369
    Thanks
    133
    Thanked 57 Times in 40 Posts
    Are your Peak Memory columns really bytes, or are they kilobytes?

    Also, what OS do you use for your Sandy Bridge and Skylake benchmarks? And how much RAM?

  7. #125
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    Quote Originally Posted by SolidComp View Post
    Are your Peak Memory columns really bytes, or are they kilobytes?
    Bytes as mentioned in the header

    Also, what OS do you use for your Sandy Bridge and Skylake benchmarks? And how much RAM?
    Ubuntu.
    Memory usage is low for this benchmark. Max. peak memory is 33MB (zopfli).
    The average html page length is 84k. See the description in the Static/Dynamic web content compression benchmark
    ~200MB is used by turbobench as Input/Output buffer for the 100MB file.
    Last edited by dnd; 6th June 2019 at 14:40.

  8. #126
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    I've generated some plots about the transfer speedup : Static/Dynamic web content compression benchmark

    You can see in the "transfer+compression+decompression" plot how the different pareto codecs can speed up the transfer with dynamic compression.

    - brotli, 11 until ~13K/s
    - brotli,5 until ~1,5MB/s
    - libdeflate from ~1,5MB/s to 4MB/s
    - igzip for fast network speed > 4MB/s

    for a connection of 500KBytes/s corresponding to 4G mobile = 4Mbits/s we have a speed up:
    - brotil 5 = 510
    - libdeflate 6 or 7 = 475
    You get (510/475=1.07) ~7% improvement by switching from gzip to brotli.

    Last edited by dnd; 6th June 2019 at 14:41.

  9. #127
    Member
    Join Date
    Sep 2008
    Location
    France
    Posts
    892
    Thanks
    492
    Thanked 280 Times in 120 Posts
    By the way @dnd,
    something changed since you initiated the web content benchmark :
    zstd has become an official web-encoding format, since November 2018.

    While there is a classical chicken and egg problem for such deployments,
    it's already a bit more than theoretical :
    on the server side, there is a zstd compression module for nginx, caddy web server supports zstd compression, and hhvm generates zstd-compressed web pages on-demand.
    On the client side, wget2 supports zstd-compressed assets, and Facebook's iOS mobile apps (including Instagram) already request zstd-compressed resources.

    zstd on the web is more targeted at dynamic content.
    We expect brotli-11 to remain the king of the hill for static content, a use case for which it was optimized.
    In our tests, zstd is proving competitive with brotli-5, and is expected to overtake faster connection speeds.
    It also decompresses faster, with corresponding energy savings on the client side.
    Also, the current web-approved version doesn't use dictionary, which can save some memory, should memory be constrained on the target device.

    In the future, we hope that zstd can become an attractive alternative for fast connection speeds and low-energy clients.

    Notes on dictionary support :
    brotli benefits from its default dictionary, which can be more effective for certain file types, so depending on the exact content of your corpus, results can vary.
    In our tests, we try to vary file types, since a typical web page is more that just html : it actually ships more css, and even more js contents.
    As a special case, Facebook's dynamic content features a lot of [graphql](https://graphql.org/) resources, which is a kind of json format.

    zstd has been denied so far the advantage of a default dictionary, on the ground that a default dictionary would benefit some contents more than others.
    While the statement is true, it's unclear why it would prove a blocker for one solution only.
    A long term answer is that web should consider dynamic dictionary compression in the future, so that each website / toolkit can ship a dictionary optimized for it.
    In which case, compression efficiency can be dramatically improved.
    This discussion is ongoing, but will likely require a few years before acceptance.
    Last edited by Cyan; 8th June 2019 at 09:33.

  10. Thanks (4):

    dnd (7th June 2019),JamesB (7th June 2019),Jarek (7th June 2019),Mike (7th June 2019)

  11. #128
    Member
    Join Date
    Jun 2015
    Location
    Switzerland
    Posts
    976
    Thanks
    266
    Thanked 350 Times in 221 Posts
    Quote Originally Posted by Cyan View Post
    Facebook's iOS mobile apps (including Instagram) already request zstd-compressed resources.
    Do you have user experienced latency numbers in comparison with brotli?

    I am expecting zstd-based compression to give worse user experienced latency and I am quite surprised that you are deploying it for user facing applications. I'd like to learn what went differently in benchmarking your product teams have done and what our product teams have done.

    Quote Originally Posted by Cyan View Post
    zstd has been denied so far the advantage of a default dictionary
    Consider just plugging in the brotli dictionary with transforms, it is fully compatible with zstd and it is already in the browser, .NET, Android, Windows, iOS, etc. Replace it with a custom dictionary when necessary like brotli is doing.

  12. #129
    Member SolidComp's Avatar
    Join Date
    Jun 2015
    Location
    USA
    Posts
    369
    Thanks
    133
    Thanked 57 Times in 40 Posts
    Quote Originally Posted by dnd View Post
    Bytes as mentioned in the header


    Ubuntu.
    Memory usage is low for this benchmark. Max. peak memory is 33MB (zopfli).
    The average html page length is 84k. See the description in the Static/Dynamic web content compression benchmark
    ~200MB is used by turbobench as Input/Output buffer for the 100MB file.
    I'm confused as to why your reported memory usage for compressors like zstd and brotli is at least an order of magnitude less than what Matt Mahoney reports in his Large Text Compression Benchmark. Any idea why? Is it because of the size of the enwik9 dataset? It doesn't make sense. And zlib only uses 14 KB of memory to decompress?

  13. #130
    Member
    Join Date
    Sep 2008
    Location
    France
    Posts
    892
    Thanks
    492
    Thanked 280 Times in 120 Posts
    Do you have user experienced latency numbers in comparison with brotli?
    I guess you can understand we can't publish internal monitoring results just like that.
    For the time being, I'm limited to say that the gains where sufficient to warrant this switch.

    But that's not the end of the story.
    In the near future, I would like to open source the technology behind this.
    At which point, we'll be in good position to detail how it works, and why it provides better latency.

  14. #131
    Member
    Join Date
    Aug 2008
    Location
    Planet Earth
    Posts
    1,040
    Thanks
    104
    Thanked 420 Times in 293 Posts
    Code:
          C Size  ratio%     C MB/s     D MB/s   Name            
        25136364    49.9       0.52       0.51   zpaq 5          
        25220100    50.1       0.95     324.47   brotli 11       
        25228361    50.1       5.84     463.72   lzham 4         
        25245808    50.2       3.90      42.24   lzlib 9         
        25255739    50.2       6.76     470.77   lzham 3         
        25255991    50.2       4.54      42.23   lzlib 8         
        25266130    50.2       8.68    3906.52   zstd 22         
        25266548    50.2       8.18    3864.53   zstd 21         
        25267102    50.2       8.18      60.67   flzma2 10       
        25268266    50.2       8.56      59.09   flzma2 9        
        25270720    50.2       7.66    3912.90   zstd 20         
        25271008    50.2       8.37      46.46   csc 5           
        25276177    50.2       8.77      59.06   flzma2 8        
        25277117    50.2       6.93      55.05   lzma 9          
        25277117    50.2       6.92      54.36   lzma 8          
        25277147    50.2       6.99      55.04   lzma 7          
        25277494    50.2       8.66      57.72   flzma2 7        
        25296873    50.3       2.04     321.77   brotli 10       
        25307650    50.3       9.48      44.91   csc 4           
        25329822    50.3       7.95      54.26   lzma 6          
        25330612    50.3      12.32      45.75   csc 3           
        25337245    50.3       1.70       1.63   zpaq 4          
        25337625    50.3       4.80      42.12   lzlib 7         
        25366735    50.4       8.43     459.68   lzham 2         
        25377607    50.4       8.88     457.94   lzham 1         
        25389419    50.4      10.22      44.70   csc 2           
        25397666    50.5       7.49      54.74   lzma 5          
        25421396    50.5       9.56      59.15   flzma2 6        
        25421396    50.5       9.57      58.43   flzma2 11       
        25425185    50.5      14.20      44.75   csc 1           
        25484483    50.6       8.10     593.34   brotli 9        
        25527210    50.7       8.44    4042.05   zstd 19         
        25560120    50.8       5.50      41.73   lzlib 6         
        25575980    50.8      11.31    4311.80   zstd 18         
        25576106    50.8      12.03      13.98   bsc 0           
        25584532    50.8       6.60      41.79   lzlib 5         
        25597975    50.9       4.30       8.17   zpaq 3          
        25612508    50.9       7.30      11.14   bsc 6           
        25613550    50.9      12.05      61.07   flzma2 5        
        25621992    50.9      12.74      11.20   bsc 5           
        25639806    50.9      13.49    5158.52   zstd 17         
        25656526    51.0      20.93     732.39   brotli 8        
        25659567    51.0       8.04      41.20   lzlib 4         
        25662668    51.0      18.20      11.58   bsc 4           
        25666633    51.0      27.38     728.54   brotli 7        
        25688220    51.0      18.84    5522.45   zstd 16         
        25701943    51.1      46.10     731.04   brotli 6        
        25705827    51.1       6.00    7046.29   lizard 49       
        25753632    51.2      20.72    5785.91   zstd 15         
        25759326    51.2      24.63    5667.98   zstd 14         
        25770906    51.2      27.76    5655.88   zstd 13         
        25778079    51.2      44.03    5583.72   zstd 12         
        25784737    51.2      11.40      53.81   lzma 4          
        25787718    51.2      48.96    5563.97   zstd 11         
        25799422    51.3       6.29    8501.97   lizard 29       
        25799706    51.3      72.59    5506.74   zstd 10         
        25803215    51.3       9.68      41.63   lzlib 3         
        25864360    51.4     108.16    5586.82   zstd 9          
        25879867    51.4     116.06    5586.20   zstd 8          
        25905229    51.5     116.13    5467.27   zstd 7          
        25934894    51.5      19.07       9.96   bsc 3           
        25934971    51.5      68.90     747.05   brotli 5        
        25939211    51.5      15.91      61.49   flzma2 4        
        25956189    51.6     122.20    5029.64   zstd 6          
        25974929    51.6     163.91    5004.64   zstd 5          
        26004380    51.7     155.56     800.43   brotli 4        
        26019990    51.7      11.96      40.81   lzlib 2         
        26166014    52.0       9.43     121.97   zpaq 2          
        26231605    52.1     379.17    5079.90   zstd 4          
        26330354    52.3      22.51      52.99   lzma 3          
        26346691    52.3      24.22     120.23   zpaq 1          
        26357777    52.4      20.74      63.58   flzma2 3        
        26393993    52.4      28.81      76.66   flzma2 2        
        26451166    52.6      12.72      40.23   lzlib 1         
        26694963    53.0     416.13    5129.60   zstd 3          
        26862380    53.4      28.68    1264.52   yalz77 12       
        27167085    54.0      32.97      75.31   flzma2 1        
        27302382    54.2      44.86    1251.32   yalz77 6        
        27742743    55.1     742.96    5327.79   zstd 2          
        27872327    55.4      23.55      49.34   lzma 2          
        27970663    55.6      21.26      40.76   bzip2           
        28031978    55.7     233.91     500.43   brotli 3        
        28099376    55.8     319.64     500.69   brotli 2        
        28132769    55.9      28.35      49.48   lzma 1          
        28258413    56.1      35.33      49.03   lzma 0          
        28325849    56.3       0.85     531.17   zopfli          
        28345559    56.3      94.61     899.13   lzfse           
        28356382    56.3      27.50     929.47   libdeflate 12   
        28384269    56.4      52.76     928.13   libdeflate 9    
        28404266    56.4      60.91     918.36   libdeflate 8    
        28417226    56.5     140.32     920.66   libdeflate 7    
        28425914    56.5     938.48    5781.26   zstd 1          
        28431693    56.5      48.70     526.27   zlib 9          
        28433739    56.5     626.41     787.98   brotli 1        
        28434536    56.5      52.72     528.54   zlib 8          
        28439704    56.5     158.49     925.78   libdeflate 6    
        28449443    56.5      65.73     520.83   zlib 7          
        28461478    56.5      70.30     524.53   zlib 6          
        28476007    56.6      13.76    7596.08   lizard 39       
        28477594    56.6     164.61     927.93   libdeflate 5    
        28517633    56.7      77.58     521.00   zlib 5          
        28533846    56.7       5.63     687.92   brieflz 9       
        28571692    56.8     173.94     914.12   libdeflate 4    
        28585953    56.8     176.73     914.57   libdeflate 3    
        28587137    56.8       2.81    1147.63   lzg 9           
        28590104    56.8      40.30     676.78   brieflz 6       
        28596461    56.8      14.19    8500.53   lizard 19       
        28613551    56.9      80.35     518.74   zlib 4          
        28618374    56.9     181.00     914.44   libdeflate 2    
        28620676    56.9      96.60    1228.77   yalz77 1        
        28630279    56.9      50.19    8732.07   lz4 12          
        28630279    56.9      50.08    8698.87   lz4 16          
        28643303    56.9      79.86    8811.56   lz4 9           
        28679357    57.0     184.61     912.27   libdeflate 1    
        28745401    57.1      86.83     574.06   zlib 3          
        28797227    57.2      89.08     570.03   zlib 2          
        28849591    57.3      92.15     569.79   zlib 1          
        28992983    57.6      39.34    1125.84   lzg 8           
        29033788    57.7     752.79     439.66   brotli 0        
        29158463    57.9      46.74    1119.60   lzg 7           
        29346782    58.3    2383.35    8892.52   lz4 1           
        29375028    58.4      55.06    1115.93   lzg 6           
        29422415    58.5    3273.39    8600.76   lz4 0           
        29500714    58.6      57.08    1123.85   lzg 5           
        29517643    58.6     767.05    1081.24   slz 8           
        29517643    58.6     767.87    1080.84   slz 9           
        29517643    58.6     768.01    1080.68   slz 7           
        29519472    58.6     775.08    1084.73   slz 6           
        29522801    58.7     767.88    1091.96   slz 5           
        29527611    58.7     766.05    1079.71   slz 4           
        29546954    58.7     764.48    1082.24   slz 3           
        29566487    58.7      59.35    1115.56   lzg 4           
        29569228    58.7     761.29    1080.08   slz 2           
        29610559    58.8     754.39    1080.63   slz 1           
        29626047    58.9      61.52    1106.75   lzg 3           
        29688602    59.0      64.51    1155.69   lzg 2           
        29744311    59.1      65.76    1121.90   lzg 1           
        29823467    59.3     753.25    3962.50   fastlz 2        
        29967771    59.5     155.58     677.10   brieflz 3       
        30141019    59.9    2455.44    5001.16   snappy_c        
        30188133    60.0     340.80    4035.90   fastlz 1        
        30701990    61.0     234.67     680.80   brieflz 1       
        30792704    61.2     932.29     835.60   density 3       
        30813182    61.2    1555.70    8875.27   lizard 10       
        32171762    63.9     141.16    8394.20   yappy           
        32659008    64.9     483.55   10362.70   srle 0          
        32659009    64.9     484.60   10414.16   trle            
        32685693    64.9     551.85   10106.76   srle 8          
        32737290    65.0     335.27    2039.12   mrle            
        32883993    65.3    2781.83    7144.31   srle 16         
        32978518    65.5    5416.08   10133.21   srle 32         
        33101325    65.8    9024.86   10920.30   srle 64         
        34578204    68.7     432.37     462.47   zlibh           
        40799740    81.1    3816.47    3108.43   chameleon 1     
        40800434    81.1    2056.87    3410.70   chameleon 2     
        40901352    81.3    2740.33    3070.13   density 1       
        50331652   100.0   12222.35   12333.17   memcpy          
        50334989   100.0     162.08     183.52   zpaq 0          
        50347018   100.0    2729.33    3016.22   slz 0

  15. Thanks:

    Mike (8th June 2019)

  16. #132
    Member
    Join Date
    Aug 2008
    Location
    Planet Earth
    Posts
    1,040
    Thanks
    104
    Thanked 420 Times in 293 Posts
    Bible text in 107 different languages:

    Code:
          C Size  ratio%     C MB/s     D MB/s   Name            
        59079328    11.2       0.63       0.62   zpaq 5          
        63027116    11.9       2.10       2.06   zpaq 4          
        71435994    13.5      13.69      21.42   bsc 0           
        73816028    14.0      45.48      17.11   bsc 6           
        75527236    14.3      17.23      63.64   bzip2           
        77954896    14.7      43.79      18.88   bsc 5           
        79182878    15.0       2.22     143.74   lzlib 9         
        79220683    15.0       2.36     144.22   lzlib 8         
        79590021    15.1       2.56     144.54   lzlib 7         
        79809347    15.1       2.99     201.88   lzma 9          
        79809347    15.1       2.87     201.33   lzma 8          
        79818511    15.1       2.98     201.81   lzma 7          
        80675011    15.3       4.30     192.49   flzma2 10       
        80810536    15.3      50.72      22.56   bsc 4           
        80839997    15.3       2.68    1507.58   zstd 22         
        80923251    15.3       3.10    1531.52   zstd 21         
        80942763    15.3       3.24    1542.13   zstd 20         
        81178672    15.4       5.35     191.16   flzma2 9        
        81447630    15.4       5.78     191.03   flzma2 8        
        81487639    15.4       1.42     631.65   lzham 4         
        81590538    15.4       3.42    1575.12   zstd 19         
        81937231    15.5       0.87     977.20   brotli 11       
        82061352    15.5       6.12     190.92   flzma2 7        
        82145357    15.5       3.88    1577.68   zstd 18         
        82337563    15.6       3.06     166.86   csc 5           
        83170839    15.7       4.56    1600.26   zstd 17         
        83838468    15.9       2.12     847.19   brotli 10       
        84024109    15.9       3.34     139.79   lzlib 6         
        84026301    15.9       6.88     189.09   flzma2 6        
        84026301    15.9       6.85     188.90   flzma2 11       
        85036580    16.1       6.50    1599.71   zstd 16         
        85211105    16.1       8.58     185.81   flzma2 5        
        85444818    16.2       4.06     192.60   lzma 5          
        85460431    16.2       3.87     192.35   lzma 6          
        87642604    16.6      11.04     177.37   flzma2 4        
        88398236    16.7      48.49      27.15   bsc 3           
        89902107    17.0       6.93      17.19   zpaq 3          
        90921756    17.2       7.72    1626.92   zstd 15         
        90979193    17.2       4.87     133.53   lzlib 5         
        91905185    17.4      10.68    2110.90   zstd 14         
        91927181    17.4      14.28     983.25   brotli 9        
        92647580    17.5      11.97    1594.07   zstd 13         
        93396314    17.7      20.96     967.66   brotli 8        
        95071664    18.0       6.90     136.52   csc 4           
        95259455    18.0      28.35     947.23   brotli 7        
        95354431    18.0       2.49    2238.36   lizard 49       
        95809612    18.1      16.16     157.16   flzma2 3        
        97132307    18.4      20.63    1531.66   zstd 12         
        97611537    18.5       7.44     126.99   lzlib 4         
        97722899    18.5      42.77     939.96   brotli 6        
        99341893    18.8      36.89    1500.46   zstd 11         
        99499365    18.8      24.20     154.30   flzma2 2        
        99582683    18.8      40.89    1503.51   zstd 10         
       100052744    18.9      44.82    1503.92   zstd 9          
       100192970    18.9      21.53     145.03   csc 2           
       100287995    19.0      10.27     182.69   zpaq 2          
       100771002    19.1      58.58     907.38   brotli 5        
       102144396    19.3      10.07     120.53   csc 3           
       102680875    19.4      73.39    1462.20   zstd 8          
       104315865    19.7      96.96    1387.21   zstd 7          
       105025903    19.9      29.96     147.62   lzma 4          
       105255665    19.9       0.50     639.72   zopfli          
       105331042    19.9      32.71     146.13   lzma 3          
       105379344    19.9       6.31    1582.49   libdeflate 12   
       106125676    20.1      12.76     118.22   lzlib 3         
       106182510    20.1     111.54     926.92   brotli 4        
       106765611    20.2      34.56     143.31   lzma 2          
       108906144    20.6      13.76    1580.86   libdeflate 9    
       109041403    20.6     139.21    1306.91   zstd 6          
       109950844    20.8       6.39    2918.55   lizard 39       
       110721696    20.9      38.41     134.07   lzma 1          
       111392582    21.1       2.59    3201.08   lizard 29       
       111531030    21.1      19.24    1576.68   libdeflate 8    
       111547479    21.1      31.48     123.98   csc 1           
       111862810    21.2      12.89     601.33   zlib 9          
       111924503    21.2      13.65     600.37   zlib 8          
       112534688    21.3      33.61     127.39   flzma2 1        
       112669023    21.3      73.17    1580.61   libdeflate 7    
       112985313    21.4      24.43     596.56   zlib 7          
       113584535    21.5     193.06    1230.48   zstd 5          
       114074765    21.6     104.46    2103.22   libdeflate 6    
       114529021    21.7      33.73     589.63   zlib 6          
       116269820    22.0     141.08    1556.92   libdeflate 5    
       116819641    22.1     327.14    1273.42   zstd 4          
       116825616    22.1      19.89     108.78   lzlib 2         
       117491951    22.2     335.63    1288.50   zstd 3          
       117492700    22.2      42.83     119.36   lzma 0          
       118785758    22.5     159.95    1740.07   libdeflate 4    
       119970613    22.7      43.88     162.63   zpaq 1          
       119974980    22.7      58.40     563.53   zlib 5          
       120703339    22.8     156.65     808.55   brotli 3        
       122212971    23.1     199.50    1685.30   libdeflate 3    
       124010447    23.5      25.93     100.31   lzlib 1         
       124059589    23.5     186.02     759.00   brotli 2        
       126335273    23.9     411.29    1225.57   zstd 2          
       126351332    23.9     242.36    1625.05   libdeflate 2    
       127388041    24.1     479.40    1366.56   zstd 1          
       127443258    24.1      14.30    4815.11   lz4 16          
       127443258    24.1      14.41    4755.13   lz4 12          
       127546857    24.1      93.17     544.78   zlib 4          
       127636289    24.1       6.70    3990.54   lizard 19       
       130104030    24.6      26.72    4608.14   lz4 9           
       131889805    24.9      90.62     566.89   zlib 3          
       132132279    25.0     350.06     682.33   brotli 1        
       132715257    25.1       1.12    1187.71   lzg 9           
       134767213    25.5     279.28    1510.90   libdeflate 1    
       140511658    26.6     517.77     686.95   brotli 0        
       143018043    27.0     130.29     526.88   zlib 2          
       148205835    28.0       7.75    1063.91   lzg 8           
       152584611    28.9     146.07     504.95   zlib 1          
       153503040    29.0      54.47     558.16   yalz77 12       
       156969840    29.7      13.56     977.99   lzg 7           
       158265407    29.9      86.67     542.84   yalz77 6        
       166927863    31.6      27.23     894.12   lzg 6           
       174765844    33.1      41.71     836.81   lzg 5           
       179697570    34.0     547.25     497.77   density 3       
       182473112    34.5     190.15     497.67   yalz77 1        
       183716842    34.7      58.39     785.80   lzg 4           
       193356315    36.6     721.51    4298.82   lz4 1           
       193556663    36.6     652.27    4040.77   lizard 10       
       193794420    36.7     390.38     475.93   slz 9           
       193794420    36.7     390.35     471.23   slz 8           
       193794420    36.7     390.42     470.91   slz 7           
       193824136    36.7      75.47     745.57   lzg 3           
       193931630    36.7     394.04     470.75   slz 6           
       194213445    36.7     393.24     475.35   slz 5           
       194768913    36.8     394.74     469.74   slz 4           
       195899990    37.1     399.33     467.95   slz 3           
       196171590    37.1     715.47    1718.03   snappy_c        
       198156506    37.5     409.80     468.49   slz 2           
       202739672    38.3     415.08     463.58   slz 1           
       204681324    38.7      95.31     710.89   lzg 2           
       208021721    39.3     768.60    4316.93   lz4 0           
       214150733    40.5     120.93     693.86   lzg 1           
       289096720    54.7    4439.22    6502.16   chameleon 1     
       289096736    54.7    2369.65    2834.78   density 1       
       289098220    54.7    2635.04    5208.41   chameleon 2     
       508137687    96.1     281.32    1522.23   mrle            
       512033650    96.8     373.99    7895.80   srle 8          
       512033651    96.8     341.49    7898.04   srle 0          
       526717025    99.6    1976.26    5590.84   srle 16         
       527373853    99.7    4384.88    7966.47   srle 32         
       528677477   100.0    7434.20    8563.32   srle 64         
       528742404   100.0   12201.00   12102.41   memcpy          
       528776795   100.0     162.75     182.82   zpaq 0          
       528903770   100.0    2699.04    2864.31   slz 0

  17. Thanks (2):

    dnd (9th June 2019),Mike (8th June 2019)

  18. #133
    Member
    Join Date
    Jun 2015
    Location
    Switzerland
    Posts
    976
    Thanks
    266
    Thanked 350 Times in 221 Posts
    Quote Originally Posted by Sportman View Post
    Bible text in 107 different languages:

    Code:
          C Size  ratio%     C MB/s     D MB/s   Name            
        59079328    11.2       0.63       0.62   zpaq 5          
        63027116    11.9       2.10       2.06   zpaq 4          
        71435994    13.5      13.69      21.42   bsc 0           
        73816028    14.0      45.48      17.11   bsc 6           
        75527236    14.3      17.23      63.64   bzip2           
        77954896    14.7      43.79      18.88   bsc 5           
        79182878    15.0       2.22     143.74   lzlib 9         
        79220683    15.0       2.36     144.22   lzlib 8         
        79590021    15.1       2.56     144.54   lzlib 7         
        79809347    15.1       2.99     201.88   lzma 9          
        79809347    15.1       2.87     201.33   lzma 8          
        79818511    15.1       2.98     201.81   lzma 7          
        80675011    15.3       4.30     192.49   flzma2 10       
        80810536    15.3      50.72      22.56   bsc 4           
        80839997    15.3       2.68    1507.58   zstd 22         
        80923251    15.3       3.10    1531.52   zstd 21         
        80942763    15.3       3.24    1542.13   zstd 20         
        81178672    15.4       5.35     191.16   flzma2 9        
        81447630    15.4       5.78     191.03   flzma2 8        
        81487639    15.4       1.42     631.65   lzham 4         
        81590538    15.4       3.42    1575.12   zstd 19         
        81937231    15.5       0.87     977.20   brotli 11       
        82061352    15.5       6.12     190.92   flzma2 7        
        82145357    15.5       3.88    1577.68   zstd 18         
        82337563    15.6       3.06     166.86   csc 5           
        83170839    15.7       4.56    1600.26   zstd 17         
        83838468    15.9       2.12     847.19   brotli 10       
        84024109    15.9       3.34     139.79   lzlib 6         
        84026301    15.9       6.88     189.09   flzma2 6        
        84026301    15.9       6.85     188.90   flzma2 11       
        85036580    16.1       6.50    1599.71   zstd 16         
        85211105    16.1       8.58     185.81   flzma2 5        
        85444818    16.2       4.06     192.60   lzma 5          
        85460431    16.2       3.87     192.35   lzma 6          
        87642604    16.6      11.04     177.37   flzma2 4        
        88398236    16.7      48.49      27.15   bsc 3           
        89902107    17.0       6.93      17.19   zpaq 3          
        90921756    17.2       7.72    1626.92   zstd 15         
        90979193    17.2       4.87     133.53   lzlib 5         
        91905185    17.4      10.68    2110.90   zstd 14         
        91927181    17.4      14.28     983.25   brotli 9        
        92647580    17.5      11.97    1594.07   zstd 13         
        93396314    17.7      20.96     967.66   brotli 8        
        95071664    18.0       6.90     136.52   csc 4           
        95259455    18.0      28.35     947.23   brotli 7        
        95354431    18.0       2.49    2238.36   lizard 49       
        95809612    18.1      16.16     157.16   flzma2 3        
        97132307    18.4      20.63    1531.66   zstd 12         
        97611537    18.5       7.44     126.99   lzlib 4         
        97722899    18.5      42.77     939.96   brotli 6        
        99341893    18.8      36.89    1500.46   zstd 11         
        99499365    18.8      24.20     154.30   flzma2 2        
        99582683    18.8      40.89    1503.51   zstd 10         
       100052744    18.9      44.82    1503.92   zstd 9          
       100192970    18.9      21.53     145.03   csc 2           
       100287995    19.0      10.27     182.69   zpaq 2          
       100771002    19.1      58.58     907.38   brotli 5        
       102144396    19.3      10.07     120.53   csc 3           
       102680875    19.4      73.39    1462.20   zstd 8          
       104315865    19.7      96.96    1387.21   zstd 7          
       105025903    19.9      29.96     147.62   lzma 4          
       105255665    19.9       0.50     639.72   zopfli          
       105331042    19.9      32.71     146.13   lzma 3          
       105379344    19.9       6.31    1582.49   libdeflate 12   
       106125676    20.1      12.76     118.22   lzlib 3         
       106182510    20.1     111.54     926.92   brotli 4        
       106765611    20.2      34.56     143.31   lzma 2          
       108906144    20.6      13.76    1580.86   libdeflate 9    
       109041403    20.6     139.21    1306.91   zstd 6          
       109950844    20.8       6.39    2918.55   lizard 39       
       110721696    20.9      38.41     134.07   lzma 1          
       111392582    21.1       2.59    3201.08   lizard 29       
       111531030    21.1      19.24    1576.68   libdeflate 8    
       111547479    21.1      31.48     123.98   csc 1           
       111862810    21.2      12.89     601.33   zlib 9          
       111924503    21.2      13.65     600.37   zlib 8          
       112534688    21.3      33.61     127.39   flzma2 1        
       112669023    21.3      73.17    1580.61   libdeflate 7    
       112985313    21.4      24.43     596.56   zlib 7          
       113584535    21.5     193.06    1230.48   zstd 5          
       114074765    21.6     104.46    2103.22   libdeflate 6    
       114529021    21.7      33.73     589.63   zlib 6          
       116269820    22.0     141.08    1556.92   libdeflate 5    
       116819641    22.1     327.14    1273.42   zstd 4          
       116825616    22.1      19.89     108.78   lzlib 2         
       117491951    22.2     335.63    1288.50   zstd 3          
       117492700    22.2      42.83     119.36   lzma 0          
       118785758    22.5     159.95    1740.07   libdeflate 4    
       119970613    22.7      43.88     162.63   zpaq 1          
       119974980    22.7      58.40     563.53   zlib 5          
       120703339    22.8     156.65     808.55   brotli 3        
       122212971    23.1     199.50    1685.30   libdeflate 3    
       124010447    23.5      25.93     100.31   lzlib 1         
       124059589    23.5     186.02     759.00   brotli 2        
       126335273    23.9     411.29    1225.57   zstd 2          
       126351332    23.9     242.36    1625.05   libdeflate 2    
       127388041    24.1     479.40    1366.56   zstd 1          
       127443258    24.1      14.30    4815.11   lz4 16          
       127443258    24.1      14.41    4755.13   lz4 12          
       127546857    24.1      93.17     544.78   zlib 4          
       127636289    24.1       6.70    3990.54   lizard 19       
       130104030    24.6      26.72    4608.14   lz4 9           
       131889805    24.9      90.62     566.89   zlib 3          
       132132279    25.0     350.06     682.33   brotli 1        
       132715257    25.1       1.12    1187.71   lzg 9           
       134767213    25.5     279.28    1510.90   libdeflate 1    
       140511658    26.6     517.77     686.95   brotli 0        
       143018043    27.0     130.29     526.88   zlib 2          
       148205835    28.0       7.75    1063.91   lzg 8           
       152584611    28.9     146.07     504.95   zlib 1          
       153503040    29.0      54.47     558.16   yalz77 12       
       156969840    29.7      13.56     977.99   lzg 7           
       158265407    29.9      86.67     542.84   yalz77 6        
       166927863    31.6      27.23     894.12   lzg 6           
       174765844    33.1      41.71     836.81   lzg 5           
       179697570    34.0     547.25     497.77   density 3       
       182473112    34.5     190.15     497.67   yalz77 1        
       183716842    34.7      58.39     785.80   lzg 4           
       193356315    36.6     721.51    4298.82   lz4 1           
       193556663    36.6     652.27    4040.77   lizard 10       
       193794420    36.7     390.38     475.93   slz 9           
       193794420    36.7     390.35     471.23   slz 8           
       193794420    36.7     390.42     470.91   slz 7           
       193824136    36.7      75.47     745.57   lzg 3           
       193931630    36.7     394.04     470.75   slz 6           
       194213445    36.7     393.24     475.35   slz 5           
       194768913    36.8     394.74     469.74   slz 4           
       195899990    37.1     399.33     467.95   slz 3           
       196171590    37.1     715.47    1718.03   snappy_c        
       198156506    37.5     409.80     468.49   slz 2           
       202739672    38.3     415.08     463.58   slz 1           
       204681324    38.7      95.31     710.89   lzg 2           
       208021721    39.3     768.60    4316.93   lz4 0           
       214150733    40.5     120.93     693.86   lzg 1           
       289096720    54.7    4439.22    6502.16   chameleon 1     
       289096736    54.7    2369.65    2834.78   density 1       
       289098220    54.7    2635.04    5208.41   chameleon 2     
       508137687    96.1     281.32    1522.23   mrle            
       512033650    96.8     373.99    7895.80   srle 8          
       512033651    96.8     341.49    7898.04   srle 0          
       526717025    99.6    1976.26    5590.84   srle 16         
       527373853    99.7    4384.88    7966.47   srle 32         
       528677477   100.0    7434.20    8563.32   srle 64         
       528742404   100.0   12201.00   12102.41   memcpy          
       528776795   100.0     162.75     182.82   zpaq 0          
       528903770   100.0    2699.04    2864.31   slz 0
    What would be the results if the same lz77 window size or the same memory use was used across the algorithms?

    Brotli defaults to 4 mb because browser manufactures have expressed that wish. It is not helpful to benchmark a gigabyte window against 4 mb, you learn nothing and end up fooling a few bystanders.

  19. #134
    Member
    Join Date
    Aug 2008
    Location
    Planet Earth
    Posts
    1,040
    Thanks
    104
    Thanked 420 Times in 293 Posts
    Quote Originally Posted by Jyrki Alakuijala View Post
    Brotli defaults to 4 mb because browser manufactures have expressed that wish.
    Understandable but why not release a separate optimized BrotliXL with 1GB memory size default?
    Last edited by Sportman; 12th June 2019 at 12:35.

  20. #135
    Member
    Join Date
    Jun 2015
    Location
    Switzerland
    Posts
    976
    Thanks
    266
    Thanked 350 Times in 221 Posts
    Quote Originally Posted by Sportman View Post
    Understandable but why not release a separate optimized BrotliXL with 1GB memory size default?
    The flag parsing (for both -w or -large_window) is pretty fast and doesn't save time if there is an optimized binary for that. For ease of use you can try saving these two lines to a file called BrotliXL (didn't try if it actually works):

    #!/usr/bin/env bash
    brotli -large_window 30 $@

    With windows you would likely use in the brotlixl.bat file: 'brotli -large_window 30 %*' (Sorry if this is wrong. This is my first Windows script.)

    Note, that the problem in this benchmark doesn't go away if you fix it for brotli. You are still comparing for example zstd vs. lzma at different window sizes. It is meaningless to compare different LZ77 algorithms against each other with different window sizes. In practice you end up comparing the window sizes. It trumps the differences in entropy coding efficiency. Window size effect is typically around 20 % in density, while context modeling and entropy coding differences are in about 7 % in density. For example, brotli:11:30 often wins over LZMA at most aggressive settings and currently averages over large amount of compression corpora within 0.6 % of LZMA's density. Zstd is not much worse at large window, usually 5 % more bytes than LZMA.

    First, decide how much memory you can afford in your use case per on-going stream. Second, use that amount of memory in your benchmark settings for every codec. Otherwise density reading in the compression benchmark has 20 % noise in its density values for enwik9 and more for more repetitive data.

  21. #136
    Member
    Join Date
    Aug 2008
    Location
    Planet Earth
    Posts
    1,040
    Thanks
    104
    Thanked 420 Times in 293 Posts
    Quote Originally Posted by Jyrki Alakuijala View Post
    The flag parsing (for both -w or -large_window) is pretty fast and doesn't save time if there is an optimized binary for that.
    I installed Linux to compile TurboBench, here Bible text results for Brotli with -d30 parameter:

    Code:
       140331730    26.5     509.95     679.88   brotli 0 -d30      
       131946021    25.0     345.19     673.56   brotli 1 -d30      
       124009883    23.5     184.99     751.03   brotli 2 -d30 
       120601797    22.8     132.23     627.54   brotli 3 -d30   
       106007450    20.0      95.18     705.90   brotli 4 -d30     
       100507709    19.0      48.21     701.00   brotli 5 -d30     
        97478112    18.4      36.17     717.48   brotli 6 -d30    
        94736016    17.9      22.46     728.75   brotli 7 -d30   
        92791238    17.5      13.10     739.46   brotli 8 -d30  
        90972423    17.2       6.63     741.87   brotli 9 -d30 
        81589142    15.4       1.23     690.48   brotli 10 -d30 
        79318260    15.0       0.63     777.28   brotli 11 -d30
    Code:
          C Size  ratio%     C MB/s     D MB/s   Name
        80840053    15.3       2.69    1496.14   zstd 22 -d30
        80926183    15.3       3.17    1521.38   zstd 21 -d30
        80941864    15.3       3.30    1537.78   zstd 20 -d30
        81577834    15.4       3.41    1568.22   zstd 19 -d30
        82142955    15.5       3.94    1572.92   zstd 18 -d30
        83171584    15.7       5.48    1586.11   zstd 17 -d30
        85036875    16.1       6.53    1582.80   zstd 16 -d30
        90908641    17.2       7.78    1611.11   zstd 15 -d30
        91903344    17.4      10.72    1593.01   zstd 14 -d30
        92648437    17.5      12.16    1586.70   zstd 13 -d30
        97132308    18.4      20.36    1518.30   zstd 12 -d30
        99341685    18.8      35.57    1487.36   zstd 11 -d30
        99582758    18.8      39.29    1488.87   zstd 10 -d30
       100046544    18.9      45.14    1487.37   zstd 9 -d30
       102673812    19.4      72.71    1452.87   zstd 8 -d30
       104308232    19.7      96.15    1399.92   zstd 7 -d30
       109035732    20.6     140.22    1294.89   zstd 6 -d30
       113578796    21.5     195.60    1244.05   zstd 5 -d30
       116819637    22.1     331.41    1280.98   zstd 4 -d30
       117491947    22.2     330.13    1286.00   zstd 3 -d30
       126335269    23.9     406.37    1213.13   zstd 2 -d30
       127388037    24.1     483.83    1372.67   zstd 1 -d30
    Last edited by Sportman; 23rd June 2019 at 23:34.

  22. #137
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    Quote Originally Posted by SolidComp View Post
    I'm confused as to why your reported memory usage for compressors like zstd and brotli is at least an order of magnitude less than what Matt Mahoney reports...
    Here the memory usage for compression is depending on the lz77 dictionary/window size used by the compressor. For this benchmark, the largest compressed page size is very small ~200k (average is 84k).
    Hence the reported small peak memory.
    And zlib only uses 14 KB of memory to decompress?
    Yes,Input and Output buffer not accounted. So for a 200k page you need the compressed input buffer + 200k for output + 14k extra memory (huffman decoding tables,...)

  23. Thanks:

    Jyrki Alakuijala (26th June 2019)

  24. #138
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    New Turbobench - Compression Benchmark update:
    - Turbobench is now using large window brotli in the default mode.
    Now, you don't need to explicitly specify the dictionary/window size with the -d option.

    lzma use the dictionary as a sliding lz77 window and a large size is automatically set depending on the compression level and the input buffer size.

    You can always overwrite the default dictionary setting with the option "-d" in turbobench.

    Quote Originally Posted by Sportman View Post
    I installed Linux to compile TurboBench
    You can also use migw64 to compile turbobench on windows.

    for a nice output to "encod.ru", you can use "turbobench -p5 -o file.tbb"

  25. Thanks:

    Jyrki Alakuijala (26th June 2019)

  26. #139
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts
    TurboBench - Compression Benchmark updated.
    - lz4ultra + smallz4 integrated but must be manually activated (see plugins.cc)

    New ARM Cortext A73 Static/Dynamic web content compression benchmark.
    - very fast compression for slz and brotli,1
    - brotli decompression is ~2 times slower than
    zlib,libdeflate
    - zlib-ng decompression is significantly faster than zlib
    - igzip slower than on the intel benchmark


  27. Thanks:

    Jyrki Alakuijala (26th June 2019)

  28. #140
    Member
    Join Date
    Dec 2012
    Location
    japan
    Posts
    170
    Thanks
    31
    Thanked 71 Times in 44 Posts
    could you include lzsa?

  29. #141
    Member
    Join Date
    Apr 2017
    Location
    United Kingdom
    Posts
    82
    Thanks
    68
    Thanked 33 Times in 21 Posts
    Quote Originally Posted by xezz View Post
    could you include lzsa?
    I hope you appreciate that only 8-bit decompressors have been tuned for speed for LZSA? We did not make any special effort to speed-up the PC decompressor - it is mainly there to recover compressed data as well as for testing.

  30. #142
    Member
    Join Date
    Mar 2013
    Location
    Worldwide
    Posts
    565
    Thanks
    67
    Thanked 199 Times in 147 Posts

  31. Thanks (2):

    introspec (28th June 2019),xezz (26th June 2019)

  32. #143
    Member
    Join Date
    May 2008
    Location
    USA
    Posts
    47
    Thanks
    0
    Thanked 5 Times in 5 Posts
    Quote Originally Posted by introspec View Post
    We did not make any special effort to speed-up the PC decompressor - it is mainly there to recover compressed data as well as for testing.
    Would you like a speed-tuned PC decompressor? (I wrote the speed-tuned LZ4 and ZX7 8088 decompressors)

  33. #144
    Member
    Join Date
    Apr 2017
    Location
    United Kingdom
    Posts
    82
    Thanks
    68
    Thanked 33 Times in 21 Posts
    Quote Originally Posted by Trixter View Post
    Would you like a speed-tuned PC decompressor? (I wrote the speed-tuned LZ4 and ZX7 8088 decompressors)
    We would really appreciate all the help we can get. My Z80 skills should be pretty good, but my understanding of 8088 is insufficient to help Emmanuel to optimize it properly, neither for speed nor for size. Thus, your input would be invaluable.

  34. #145
    Member
    Join Date
    May 2008
    Location
    USA
    Posts
    47
    Thanks
    0
    Thanked 5 Times in 5 Posts
    Okay, I'll wait for a release 1.0.3 (a bug in the compressor was fixed?) that comes with a new windows .exe for the compressor, then I'll start optimizing. I've written up a test harness that lets me code and profile speed directly on 8088 hardware.

  35. Thanks:

    introspec (1st July 2019)

  36. #146
    Member
    Join Date
    May 2017
    Location
    Spain
    Posts
    27
    Thanks
    22
    Thanked 5 Times in 5 Posts
    Does anyone nows if ZX7 has an official repo somewhere?

  37. #147
    Administrator Shelwien's Avatar
    Join Date
    May 2008
    Location
    Kharkov, Ukraine
    Posts
    4,133
    Thanks
    320
    Thanked 1,396 Times in 801 Posts
    Maybe this? (seems to be newest): https://github.com/z88dk/z88dk/tree/...T/compress/zx7

  38. #148
    Member
    Join Date
    Apr 2017
    Location
    United Kingdom
    Posts
    82
    Thanks
    68
    Thanked 33 Times in 21 Posts
    There is no official repo for ZX7. The author has released the compressor via http://www.worldofspectrum.org/infos...cgi?id=0027996

  39. #149
    Administrator Shelwien's Avatar
    Join Date
    May 2008
    Location
    Kharkov, Ukraine
    Posts
    4,133
    Thanks
    320
    Thanked 1,396 Times in 801 Posts
    The link I posted has newer updates in ZX7 history (2016).

  40. #150
    Member
    Join Date
    May 2017
    Location
    Spain
    Posts
    27
    Thanks
    22
    Thanked 5 Times in 5 Posts
    Yes I found that, I got the files from WOS, but found new version (2016) in the z88dk git, also there is ZX7b by "antoniovillena", https://github.com/antoniovillena/zx7b

    and Saukav, "saukav is a generalization for zx7b algorithm" :
    Saukav is a generalization algorithm from zx7b. Basically is for advanced users,because the decompression routine is generated by the compressor.
    When you compress a file with saukav.exe, a file called d.asm is generated with the decompression code. For this reason is recommended to process multiple files in the same command line call. All these files will be compressed with the same variant.
    This generalization contains 2*9*2= 36 different compression codings, each one can be speed optimized with 5 variants, so a total of 180 decompression routines are available.
    It has a nice table to compare:

    Code:
             Shrinkler  deexov4   apcLib  BBuster  zx7mega   saukav   zx7bf2
    --------------------------------------------------------------------------
    lena1k      13757267   303436   176472   106746    95255    76547    81040
    lena16k    238317371  4407913  2964139  1908398  1727095  1646032  1462568
    lena32k    484967405  8443253  5846174  3651800  3300486  3231882  2803116
    alice1k     10060954   274111   132816    98914    89385    70869    73459
    alice16k   131592504  2973592  2132835  1812259  1614225  1338287  1328886
    alice32k   249719379  5378511  4152512  3614393  3230255  2550243  2654236
    128rom1k    13773150   249124   132763    82637    74110    60222    62000
    128rom16k  197319929  3571407  2295235  1550682  1407478  1392317  1180569
    128rom32k  394594060  7355277  4606385  3107867  2825773  1926027  2381847
    --------------------------------------------------------------------------
    routine size     245      201      197      168      244     ~200      191
    Anyway, thanks to TurboBench author, it's a very nice and unique tool

Page 5 of 6 FirstFirst ... 3456 LastLast

Similar Threads

  1. EBOLA AND THE FUTURE OF CIVILZATION
    By biject.bwts in forum The Off-Topic Lounge
    Replies: 20
    Last Post: 3rd November 2014, 11:34
  2. Future-LZ as last-step compression scheme
    By Piotr Tarsa in forum Data Compression
    Replies: 18
    Last Post: 3rd December 2011, 01:55
  3. BCM's future
    By encode in forum Data Compression
    Replies: 17
    Last Post: 9th August 2009, 02:00
  4. Future Bandwidth.
    By Tribune in forum The Off-Topic Lounge
    Replies: 9
    Last Post: 10th October 2008, 23:56
  5. LZPM's future
    By encode in forum Forum Archive
    Replies: 129
    Last Post: 3rd March 2008, 20:23

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •