I just released v0.3.2 of my multi-threaded platform-agnostic CPU-based tripcode searcher!
On older architectures, it should run at least 10% faster than the previous version!
https://github.com/microsounds/tripforce
It also makes a nice benchmark!
Here's some random xeon system running on 32 threads at once!TRIP: '!aKaRIjHnQw' -> PASS: 'Mtw,"xl[' @ 4.10 mTrip/s
TRIP: '!AKarIMLuzE' -> PASS: '1|P?HKg%' @ 4.14 mTrip/s
TRIP: '!IcaKaRiY3A' -> PASS: 'U$)$Ka5e' @ 4.14 mTrip/s
TRIP: '!YakAriRyrM' -> PASS: '{`V5Lhcf' @ 4.14 mTrip/s
TRIP: '!kboMakaRiA' -> PASS: '!x5m`,QO' @ 4.17 mTrip/s
TRIP: '!9YiakArIdE' -> PASS: 'Et2vFceU' @ 4.17 mTrip/s
TRIP: '!waghAkarIM' -> PASS: 'oF4e*:-[' @ 4.17 mTrip/s
>>54485006
dumb tripfag cancer
Cpu-based? Why do I want THAT?
>>54485037
Maybe some people have a fuckton of CPU power without the need for a beefy GPU?
testing
But I can do easily over 40 million on CPU with this and 300 on my mobile GPU
Yours doesn't even seem as fast as the old school trippers like tripcrunch or 4brute
>>54485138
yours is almost 40 times slower
>>54485138
Those programs use platform specific extensions like AVX and SSE2 tho.
>>54485006
Build one that uses the am3 extension in AMD processors. Wet need to be superior somehow!
Try generating tripcodes in large batches and then searching through them and printing out matches. This should be faster than doing them one by one because of less locking and IO
>>54485392
Searching isn't really a bottleneck tho.
The real bottleneck is the crypt function, which eats up about 75% of the running time.