[QUOTE=Madpoo;410691]Besides, what if this guy is on the level and he's really on to something incredible? Like, some groundbreaking new way of doing LL sequences in a non-sequential way.

[snip]

Like I said, if he could do what it takes a dual-core 2.4 GHz CPU a month to do in only 1/100th of a second, wouldn't that be impressive?

In short, I don't think any of that is possible, and all of this was said by him 4 years ago and I'm sure by now he's realized the problems inherent in that line of thinking, but I'm still curious where he wound up.[/QUOTE]

Couldn't agree more, he's full of it.

1. The LL test is by definition [U]sequential[/U], the output of step 1 is the input of step 2
2. That is a speed up by a factor of 267,840,000 or only [B]8 orders of magnitude[/B] :w00t:

[QUOTE=Madpoo;410659]MPCG might as well not exist... it's the type of thing that, as the saying goes, generates more heat than light. It'd be like someone today announcing their new invention, the horse drawn carriage. It'll get you from here to there at an amazing 5 miles per hour![/QUOTE]

And since it can only find 64-bit factors, it also doesn't need to exist (since all candidates in the range of interest to the author have already been checked to 64 bits, or could be in hours with the right tools).

Seems to me if author had appeared on this forum 3.5 years earlier in his 4-yr effort, he might've had some bitcoin by now.

[QUOTE=Gordon;410694]Couldn't agree more, he's full of it.

1. The LL test is by definition [U]sequential[/U], the output of step 1 is the input of step 2
2. That is a speed up by a factor of 267,840,000 or only [B]8 orders of magnitude[/B] :w00t:[/QUOTE]

14=16-2

194 = 16^2-(2*(2*(16-2))+2^2+2) was somewhere around what I thought.

edit: though the proof of correctness can use a formula for each iteration and you can do binary and modular exponentiation on both separately and come to the answer.

[QUOTE=science_man_88;410696]14=16-2

194 = 16^2-(2*(2*(16-2))+2^2+2) was somewhere around what I thought.

edit: though the proof of correctness can use a formula for each iteration and you can do binary and modular exponentiation on both separately and come to the answer.[/QUOTE]

In regards to the authors comments (over on the older Nvidia posts) about what method he's using to do the multiplications (and since we can't see the source to his app, that's all we have to go on):

Any comments from the folks smarter than me on whether FFT and specificaly the GMP LIB is better/worse than the Karatsuba method he had proposed using at one point?

And consider the limitations of hardware (memory usage, # of cores, etc) in the analysis. That is to say, would he have been right in saying that the Karatsuba approach to multiplying large #'s is particularly better suited to GPU's than the FFT method?

And also, would his criticisms of the GMP library be valid (specifically his mention of lots of "TODO" comments)? Much has changed in the past 4 years.

[QUOTE=science_man_88;410696]14=16-2

194 = 16^2-(2*(2*(16-2))+2^2+2) was somewhere around what I thought.

edit: though the proof of correctness can use a formula for each iteration and you can do binary and modular exponentiation on both separately and come to the answer.[/QUOTE]

I just did (31 days * 24 hours * 3600 ) / 0.01

[QUOTE=Gordon;410698]I just did (31 days * 24 hours * 3600 ) / 0.01[/QUOTE]

technically you should be doing 31 days * 24 hours/day *3600 seconds/hours and get units of seconds not days hours.

[QUOTE=science_man_88;410699]technically you should be doing 31 days * 24 hours/day *3600 seconds/hours and get units of seconds not days hours.[/QUOTE]

31*24*3600 = 2,678,400 SECONDS :smile:

Claim is perform entire calculation in 1/100 second.

2,678,400 / 0.01 = 267,840,000

[QUOTE=Gordon;410700]31*24*3600 = 2,678,400 SECONDS :smile:

Claim is perform entire calculation in 1/100 second.

2,678,400 / 0.01 = 267,840,000[/QUOTE]

Oh I get the math ( wasn't quite sure why you originally commented like that but that's another thing), but saying a sequential operation can't be done as parallel sequential calculations isn't exactly giving another way of doing things leeway.

[QUOTE=CEMPLLA Author;410541]The MPCG program, which is installed by the CEMPLLA Install program if the machine does not have five or more GPU boards, is the "factoring" part of the CEMPLLA System. It's not nearly as advanced as the CEMPLLA program that's installed on machines with five or more GPU boards.

That said, it's operating principle is very simple. It attempts to find a modulus of zero for the selected prime candidate (2^P-1) by dividing it by a divisor (D) that is calculated from a "K Factor" (K), according to the formula: D = 2 * K * P + 1.

Each of these "K Factors" are tried, one by one, in numerical order by the MPCG program. It starts with a K Factor of one, and increments the "Current K Factor" by one for each "iteration". And, of course, it keeps track of the "Current K Factor" for each candidate.
[/QUOTE]

A [b]very[/b] inefficient way to do it.

Are you sure that you read about the math before you wrote code?

BTW, please describe the convolution algorithm you use in your LL code.

[QUOTE=pepi37;410689]Does that "good wiling people" have your own thoughts , or they need someone else to told him what they should/need to do?
What you think?[/QUOTE]

Apparently enough people think for themselves given the amount of flak he is receiving. But there are or might always be people around who prefer to have their own opinion formed after having been informed by more sceptical people about hefty risks based on techniques used that do not seem to serve any purpose other than to obfuscate the true objective of that computer program.

[QUOTE=science_man_88;410701]Oh I get the math ( wasn't quite sure why you originally commented like that but that's another thing), but saying a sequential operation can't be done as parallel sequential calculations isn't exactly giving another way of doing things leeway.[/QUOTE]

Well according to our projects [URL="http://www.mersenne.org/various/math.php#lucas-lehmer"]how it works[/URL] it is sequential, period.

You can't start iteration 2 until iteration 1 finishes

You can't start...you get the drift :smile: