Beta user stats for Phase 2 - Test of 210885
 

I have started working on user stats for 210885 based on the returned lresults.txt files. It's at [url]http://www.15k.org/homephase2stats.htm[/url]
Timings based on averages from a Piii 1GHz processor. I'm working on getting better averages / more data points for the timings.

Let me know what you think about the stats!
Also, does anyone know where the FFT crossover points for LLR are?

I have updated the stats formula. It's based on how long each iteration takes, and how many need to be done for each n. I am using the following crossover points / iteration times:
[CODE]
n=220,000 to 255,000 - 0.003697 seconds per iteration
n=255,000 to 300,000 - 0.004745 seconds per iteration
n=300,000 to 340,000 - 0.005360 seconds per iteration
n=340,000 to 42?,??? - 0.007828 seconds per iteration (approx)
n=42?,??? to 50?,??? - 0.009937 seconds per iteration (approx)
n=50?,??? to ???,??? - 0.012443 seconds per iteration (approx)
[/CODE]
I am working on testing n=400,000+ range. I'll have better information on that range tomorrow.
All crossover points are rounded down to the lower n.

How do I formulate the stats:
Each test takes n*(0.0xxxxx seconds per iteration) I calculate the total seconds for each test, add the seconds from all of your tests, and convert to days / years!