Only to show I´m still working sometimes on my Disc-Cataloger (it´s not dead!) I decided to prove its incredible speed (I used it as base calculation in former benchmarks at the beginning of this thread) with an old tester. This should be a little bit faster than the version I used for the posted benches, but my newest routine is even faster.
There are some important things to do:
1.) Create c:\tmp
2.) Read some drives with "Read" - if the screen seems to freeze you only have to wait a while. This isn´t important for testing. Rereading of same drives is not possible after searches in this version.
3.) After reading some drives/volumes/CDs/DVDs/Networks you can search for a name. Included is only a case insensitive substring-search. Type in something in the File field for filenames and/or the directory field for directory filtering (you can also type several dirs like "\test\emulation" and click search.
4.) The first search reads all datas from the tmp-folder and in parallel caches the files to memory (you´ll be able to decide how to handle caching in the real release). Next searches will be done much faster.
5.) The strings per second says it all. Each string is a file-entry.
6.) "... seconds searchtime" is the time for the search itself - not taking care of file-activities. Only the first 10.000 lines are shown (release will shows all). This will be faster at the first release version and you´ll see the first results immediately on the screen. A live-search while typing is planed.
7.) Volumes and other infos are not shown in the results. The lines are counted as results - not the real amount of suiting entries! Hey, this is only a version I use for benchmarks and tests for several routines!
If you want to see how fast it would work a.e. with 1000 times your full HD-Drives, you can easily fake this with cloning the lines in the file Volume.vols as often as you like. The amount of searched entries (files, dirs and volumes) are shown at the top. Take care not to insert so many volumes that your memory starts swapping to harddisc!
My goal is something in the style of Lazycat, Locate32 or Everywhere but with different new features and an unbeatable search speed.