Questions on coding my FFT&LLR implementation - Page 3

sylvester · 2008-09-13, 18:55

Quote:

Originally Posted by nuggetprime

What's wrong?

Your compiler. Please get a compiler that really supports long double, like icc/icpc.

nuggetprime · 2008-09-14, 08:07

Quote:

Originally Posted by sylvester

Your compiler. Please get a compiler that really supports long double, like icc/icpc.

Doesn't the gcc support long double if it has a m128-bit-long-double (it doesn't change anything). I'm quite sure the gcc can do that.

--nuggetprime

retina · 2008-09-14, 09:50

Quote:

Originally Posted by nuggetprime

Doesn't the gcc support long double if it has a m128-bit-long-double (it doesn't change anything). I'm quite sure the gcc can do that.

nuggetprime · 2008-09-14, 11:37

Quote:

Originally Posted by retina

Sorry, I've somewhat looked through the gcc manual and found nothing helping.
Can you please tell me what it is?

thanks,
nuggetprime

retina · 2008-09-14, 12:02

Quote:

Originally Posted by nuggetprime

Sorry, I've somewhat looked through the gcc manual and found nothing helping.
Can you please tell me what it is?

The reason I didn't directly answer is because I don't know. I'm just a simple ASM programmer that occasionally dabbles in C when I have to.

Perhaps you can try gnu.gcc.help, perhaps you can try google/yahoo/MSN. You will probably find your answer faster that way anyway. I just used google now and I think I have an idea about your question, but since I am not expert I will let you read about it yourself to fully understand the reason.

rogue · 2008-09-14, 13:18

Quote:

Originally Posted by nuggetprime

Doesn't the gcc support long double if it has a m128-bit-long-double (it doesn't change anything). I'm quite sure the gcc can do that.

--nuggetprime

I think it depend upon the OS and the version of gcc. I know for a fact that some switches in gcc work on some systems, but not others. This could be the case of support added for a specific system.

nuggetprime · 2008-09-14, 16:48

I'm understanding my error.

It should have been 1.0L/7.0 instead of 1.0/7.0.

--nuggetprime

nuggetprime · 2008-09-14, 17:17

I'm currently not really understanding IBDWT. The formula attatched would always give a complex number number,or? Can you, if possible,provide a small example?

Thanks,
nuggetprime

nuggetprime · 2008-09-15, 15:02

..up
can anyone you please provide a small IBDWT example,like in "FFT explanation for non-math guys" would probably let me grasp it,for example Mod(1739^2,2^11-1=2047).

Also(if possibe) how can this be generalized to k*b^n+/-c as gwnum can do?

Many thanks,
nuggetprime

ewmayer · 2008-09-15, 20:24

The best place to find a nice explanation of [including a worked example] the IBDWT is the original Crandall/Fagin paper.

Teaching yourself the rudiments of FFTs and working through the above paper are pretty much the price of admission to the IBDWT/FFT club.

nuggetprime · 2008-09-16, 14:31

Today I grasped that. But in the paper they're only talking about Fermat and mersenne numbers. I can generalize from 2^n-1 to b^n-1, but I can't generalize b^n-1 to k*b^n-1.
LLR uses IBDWT for LLR tests, so for k*2^n-1 IBDWT must be somewhat possible.
Are proth tests IBDWT capable?

I know I could study gwnum, but I think it's rather complex,so I'll probably get the answer here faster.

Thanks,
nuggetprime

2008-09-14, 17:17	#30
nuggetprime Mar 2007 Austria 12E₁₆ Posts	IBDWT help I'm currently not really understanding IBDWT. The formula attatched would always give a complex number number,or? Can you, if possible,provide a small example? Thanks, nuggetprime Attached Thumbnails

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Anyone know enough about coding to do this?	jrafanelli	Software	2	2018-01-11 15:16
Python Coding Help?	kelzo	Programming	3	2016-11-27 05:16
Zhang's OPQBT coding help?	flouran	Programming	0	2009-07-25 02:43
coding midlet for TF	starrynte	Programming	1	2008-12-30 22:31
Coding Challenges	R.D. Silverman	Programming	18	2005-08-09 13:14

2008-09-14, 16:48	#29
nuggetprime Mar 2007 Austria 302₁₀ Posts	I'm understanding my error. It should have been 1.0L/7.0 instead of 1.0/7.0. --nuggetprime

2008-09-15, 15:02	#31
nuggetprime Mar 2007 Austria 2·151 Posts	..up can anyone you please provide a small IBDWT example,like in "FFT explanation for non-math guys" would probably let me grasp it,for example Mod(1739^2,2^11-1=2047). Also(if possibe) how can this be generalized to k*b^n+/-c as gwnum can do? Many thanks, nuggetprime

2008-09-15, 20:24	#32
ewmayer ∂²ω=0 Sep 2002 República de California 103×113 Posts	The best place to find a nice explanation of [including a worked example] the IBDWT is the original Crandall/Fagin paper. Teaching yourself the rudiments of FFTs and working through the above paper are pretty much the price of admission to the IBDWT/FFT club.

2008-09-16, 14:31	#33
nuggetprime Mar 2007 Austria 2·151 Posts	Today I grasped that. But in the paper they're only talking about Fermat and mersenne numbers. I can generalize from 2^n-1 to b^n-1, but I can't generalize b^n-1 to kb^n-1. LLR uses IBDWT for LLR tests, so for k2^n-1 IBDWT must be somewhat possible. Are proth tests IBDWT capable? I know I could study gwnum, but I think it's rather complex,so I'll probably get the answer here faster. Thanks, nuggetprime