how to get the best out of 7z compression

[ideas]

hey am new to 7z, but i would like to know the best method, settings and / or command or parameter to use and get the most compression on any file type.

Edited February 13, 2006 by ideas

[LLXX]

Use LZMA, maximum dictionary size and maximum compression option.

(It'll be very slow but have good results...)

[Jeremy]

I'd only recommend the time consuming process of using stuff like max dictionary size if you have a 6 year old 4 GB HD, and even then it will take centuries compared to using a 160GB SATA drive. If you have a huge HD, I wouldn't be so fussy about getting 700 MBs down to... 500...

[TravisO]

The larger the "Word Size" and the larger the "Dictionary Size" the better compression you'll get. And you'll want to use the LZMA "Compression Method".

But keep in mind the harder you compress it the longer it will take, mostly because of memory requirements. A 5.2meg Excel file I had, when set to the highest settings needed 1.3gigs of RAM to compress.

Obviously if you don't have enough ram you're system will swap and compression can take 5x longer, so choose a setting only as high as your memory allows. And don't forget we're talking about available memory, if you currently have 1gig of memory, you probably only have 512-768megs available at any given time so I wouldn't go beyond that (unless you don't mind the excessive wait).

When it comes to dict size vs word size keep this in mind, a larger word size will give you better compression with virtually no change to memory usage (consider this a freebie) but dict size will balloon your memory requirements very quickly.

Fortunately un-7Zip'ing is very memory friendly, it doesn't get higher than 130megs of ram needed to uncompress so the only person suffering here is the compressor, not the uncompressor, both ram and speed wise.

[ideas]

thanks alot guys

[eidenk]

You forgot something guys : solid archive option. That's where you gonna get the best results, especially if you compress loads of similar files in the same archive.