Where is x3 speed acceleration? :))
mrantonion opened this issue · 2 comments
mrantonion commented
There are so little arguments to use. Thus i can't see the acceleration as written earlier. :)
With keyhunt i used 15 threads and got a speed of 1x10^15 per 1 (one!!!) thread. (1000 trillion per thread).
Running this code i see just 500 trillions per 15 threads... Very slow.
May be i need to create a bPfile with much more elements in it?
I created it with 50M points.
iceland2k14 commented
- The elements in bPfile will affects the speed as it will add more items in bloom to check in each loop, so more speed. With 6 billion items in bPfile added to bloom you should get x120 times faster compared to 50M items.
- Performance improvement was related to earlier python only code which i moved to folder. Not related to keyhunt. Keyhunt is faster.
- This code is still 1 thread only. Multicpu will be added later after more changes.
- So little arguments, i don't know about it??
mrantonion commented
- The elements in bPfile will affects the speed as it will add more items in bloom to check in each loop, so more speed. With 6 billion items in bPfile added to bloom you should get x120 times faster compared to 50M items.
- Performance improvement was related to earlier python only code which i moved to folder. Not related to keyhunt. Keyhunt is faster.
- This code is still 1 thread only. Multicpu will be added later after more changes.
- So little arguments, i don't know about it??
- Got it.
- I run it on python 3.8.
- Now i understand. Thus i saw a 1-thread speed.
- I meant smth like quantity of cpu threads, or predefined range, or continuously/random search switch.
Anyway thank you very much for software. Waiting for multicpu version! :)