20131215, 23:50  #1 
Dec 2013
2_{16} Posts 
Help with FFT length
I have searched high and low for a way to control the FFT length of the exponent I have been assigned. Based on my benchmark test, my computer runs fastest (by far) on a FFT of 1024K. I calculated the rate using the following: [avg. time]/[FFT] and sorted the results to find the lowest sec/K.
I have read all the .txt files including undoc.txt and cannot find a way to force it to 1024K. Can anyone help? Thanks. 
20131216, 00:25  #2  
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2
41·229 Posts 
Quote:
The program will use the latter, automatically. You don't have to do anything. See this thread, for example. (There were other similar threads, periodically.) 

20131216, 01:11  #3  
Dec 2013
2 Posts 
Thank you. I will step away from Mordor.
Quote:


20131216, 01:53  #4 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2
41×229 Posts 
Welcome to the forum!
The good thing is that you haven't even entered Mordor. It is not there  where you were headed. That way, there were only the tar pits of lost time and frustration. ;) 
Thread Tools  
Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Factorization length test?  siegert81  Math  17  20111220 15:33 
FFT Length  Samoflan  Information & Answers  8  20100216 22:05 
FFt length  mack  Information & Answers  1  20090906 03:24 
llr: FFTlength not monotone in n?  hhh  Software  4  20081220 09:49 
Q: FATAL: length 840 K not available. Halting... ?  MartinHvidberg  Software  14  20030723 16:14 