CPU @ 100%
 

Just installed GIMP, and as soon as I started the program, my cpu maxed out. I restarted, and it maxed out again. Opened my process manager, and GIMP was using between 49-96%. I understood that it would not affect my performance very much, however with the CPU maxed, I can't do much. I closed the program, and my CPU dropped down to about 10%.
So what do I need to do?

[QUOTE=acerdeville;264078]Just installed GIMP, and as soon as I started the program, my cpu maxed out. I restarted, and it maxed out again. Opened my process manager, and GIMP was using between 49-96%. I understood that it would not affect my performance very much, however with the CPU maxed, I can't do much. I closed the program, and my CPU dropped down to about 10%.
So what do I need to do?[/QUOTE]You don't need to do anything and the behaviour you observe is normal. The program is designed to use all available CPU cycles at very low priority, which means it grabs the CPU only when nothing else wants it and gets out of the way the instant anything else wants to do something.

My machines spend virtually their entire life running at 100% CPU and yet hardly ever does it impinge on interactive work.


Paul

Yes, this is normal and it is fine.:smile::tu:

The Prime95 program only fills in the gaps when nothing else is running. As you saw, when it is not running typically between 90-95% of the CPU time is wasted, Prime95 only takes that time.

As you noted, you saw 49%-96%. The first was likely while Prime95 had started a thread on 1 core and was waiting to start it on another.

And since Paul neglected to say it, "Welcome to the world of Prime Crunching".

Thanks guys, I just misread the info. At this rate I'll find a prime in no time.

[QUOTE=acerdeville;264078]however with the CPU maxed, I can't do much.[/QUOTE]I think that if you'd actually tried doing something (word processing, browsing, whatever...), instead of only looking at the CPU-busy meter, you would have found that you could actually do whatever you wanted to do.

[QUOTE=acerdeville;264092]Thanks guys, I just misread the info. At this rate I'll find a prime in no time.[/QUOTE]

More likely you will find a non-prime in no time...the collective prime-crunchers have been looking for Mersenne Prime #48 (which may or may not be the 48th Mersenne Prime) for a couple years now...but nonetheless, good luck!

[QUOTE=Christenson;264199]More likely you will find a non-prime in no time...the collective prime-crunchers have been looking for Mersenne Prime #48 (which may or may not be the 48th Mersenne Prime) for a couple years now...but nonetheless, good luck![/QUOTE]

His chances of finding a prime are quite good: most found factors are prime.

[QUOTE=Mr. P-1;264207]His chances of finding a prime are quite good: most found factors are prime.[/QUOTE]

LOL!

Very funny, and equally true.

I think the OP meant "Mersenne Prime" though.

If you look at the total GHz-days reported on the Mersenne.org page, and divide that by 13 (the number of them found by GIMPS) that's the average GHz-days per prime.

Put your own GHz-days in the numerator, and that number in the denominator, and that is roughly your odds of finding a Mersenne.

[QUOTE=LiquidNitrogen;265251]If you look at the total GHz-days reported on the Mersenne.org page, and divide that by 13 (the number of them found by GIMPS) that's the average GHz-days per prime.

Put your own GHz-days in the numerator, and that number in the denominator, and that is roughly your odds of finding a Mersenne.[/QUOTE]
Hmm, that will be very very roughly.:smile:
A sizeable proportion of that total GHz-days will be factoring work-types which would not in themselves discover a Mersenne Prime, or double checking which would be extremely unlikely to (never has so far). So obviously your work-type is very important: it needs to be LL and highly preferably first-time testing.
And the LL-work which is being done now is on much larger numbers than in the early days of GIMPS, and these numbers require far more GHz-days to test, plus it's also reasonable to assume that the Mersenne Primes will be more sparsely distributed in this current working zone: so the LL-work now will discover far fewer primes relative to GHz-days than the early work did.

[QUOTE=Brian-E;265256]Hmm, that will be very very roughly.:smile:
A sizeable proportion of that total GHz-days will be factoring work-types which would not in themselves discover a Mersenne Prime, or double checking which would be extremely unlikely to (never has so far). So obviously your work-type is very important: it needs to be LL and highly preferably first-time testing.
And the LL-work which is being done now is on much larger numbers than in the early days of GIMPS, and these numbers require far more GHz-days to test, plus it's also reasonable to assume that the Mersenne Primes will be more sparsely distributed in this current working zone: so the LL-work now will discover far fewer primes relative to GHz-days than the early work did.[/QUOTE]

So I have 3 cores ddoing LL and 1 core doing D, so that means only the 3 cores are searching for Mersenne primes?

Yep, core #4 is, however, doing something important: Making sure that the first LL test indeed correctly reported that the exponent was a composite. There's a slight (probably worse than 1 in a million) chance that an exponent was incorrectly reported as composite -- it both has to be prime, and the first LL test has to have incorrectly reported it was composite.
P-1 is also good work for a CPU; I'm finding factors on average in significantly fewer GHz days than it would take to do two LL tests. You just have to give it a lot (a Gigabyte or so) of memory.

Chance of finding a prime via DC v first LL
 

People seem to be over pessimistic on this question.

The ball park error rate of LLests is 2%.

But DCs are performed on exponents half the size, which
means they take 4 times less GHz-Days, and are twice as
likely to be prime.

So by DCing, your chance of finding a MP per year is 16%
of your chance doing first time LLtests: not as dramatic a
difference as is often portrayed.

David

PS and as Christenson has mentioned elsewhere,
6 x miniscule = miniscule ;-)

[QUOTE=davieddy;265317]People seem to be over pessimistic on this question.

< snip >

But DCs are <snip > are twice as
likely to be prime.

So by DCing, your chance of finding a MP per year is 16%
of your chance doing first time LLtests:[/quote]

The most common error in probability calculations is failing to include the effects and consequences of knowledge.

Your calculation seems to assume that the probability of a DC assigned exponent's being that of a Mersenne prime is just as though it were the corresponding probability for first-time tests -- without using any knowledge gained from the first-time tests. That assumption's not valid!

In actuality, GIMPS [i]does not assign DCs for exponents which are already known to be Mersenne primes![/i]

Thus, a correct calculation would have to take into account that GIMPS is deliberately and knowingly _not_ assigning DCs for exponents of known Mersenne primes.

[quote]not as dramatic a difference as is often portrayed.[/QUOTE]... only because you neglect to account for the effects of prior knowledge!

I deliberately refrained from including the necessary pedantry.

Given that a first time LLtest is erroneous, (2% say), the probability
of the a DC being prime is doubled while the test time is quartered.

What earthly difference does "already found MPs" make to this
sound and simple conditional probability argument?

David

[QUOTE=davieddy;265351]I deliberately refrained from including the necessary pedantry.[/QUOTE]... which allowed misinterpretation, because your actual statement was not your intended statement. :-)

[quote]What earthly difference does "already found MPs" make to this
sound and simple conditional probability argument?[/quote]None, but "this" argument is not the same as the "So by DCing, your chance of finding a MP per year" statement to which I previously replied.

[QUOTE=cheesehead;265354]... which allowed misinterpretation, because your actual statement was not your intended statement. :-)

None, but "this" argument is not the same as the "So by DCing, your chance of finding a MP per year" statement to which I previously replied.[/QUOTE]

The statement stands.

[QUOTE=cheesehead;265349]... only because you neglect to account for the effects of prior knowledge![/QUOTE]

I'm skimming, so I may have missed something important.

But didn't David start with the prior knowledge that only 2% of the first tests are faulty, so if all things were equal, you would be 2% as likely to find a prime.

But these tests, if done correctly, are twice as likely to be prime, so in the same number of tests you are 4% as likely to find a prime.

But the tests are four times faster, so in the same time you are 16% as likely to find a prime.

I don't see what additional prior knowledge would be relevant to the calculation. In particular, knowledge of known Mersenne Primes is not knowledge about faulty tests - those tests are known to be correct, so not relevant to this analysis.

William

[QUOTE=cheesehead;265349]Thus, a correct calculation would have to take into account that GIMPS is deliberately and knowingly _not_ assigning DCs for exponents of known Mersenne primes.[/QUOTE]
My own understanding, and it is likely to be faulty since I have studied no probability theory in more than 25 years, is that the above statement is quite correct in that it identifies a biasing factor. However surely the bias in this case is so tiny (there are so few identified Mersenne Primes) that it will have no effect on the very rough 16% estimate which David arrives at.

[QUOTE=Brian-E;265369]the very rough 16% estimate which David arrives at.[/QUOTE]

THX Brian (and William).

Not all that rough though. OK someone can work out a more accurate
current "error rate" (E%) by the DCs whose residues don't match the
first LL, but I made it clear that "2%" was a ball park guess at E%.

Doubling the exponent increases the GHz-Days by slightly more than 4 times.
The probability of being prime is complicated by "how far TFed" etc, but it involves 1/exponent.

If I said "8E%" instead of "16%" I would not be too far out.

David

Blah blah blah blah blah, you guys should run for office.

Estimating the "error rate"
 

[QUOTE=cheesehead;265349]The most common error in probability calculations is failing to include the effects and consequences of knowledge.

Your calculation seems to assume that the probability of a DC assigned exponent's being that of a Mersenne prime is just as though it were the corresponding probability for first-time tests -- without using any knowledge gained from the first-time tests. That assumption's not valid!

In actuality, GIMPS [I]does not assign DCs for exponents which are already known to be Mersenne primes![/I]

Thus, a correct calculation would have to take into account that GIMPS is deliberately and knowingly _not_ assigning DCs for exponents of known Mersenne primes.

... only because you neglect to account for the effects of prior knowledge![/QUOTE]

I think all your qualms boil down to obtaining an unbiassed estimate
of the probability of the non-zero residue of an unverified LL being wrong.

No doubt a determined quibbler will find this too simple,
but I would take a suitable* large sample of "verified composite via
two matching residues", and find the % where the first test proved
to be wrong.

This totally bypasses your "known prime" "difficulty".

David

* I am "covering my back" this time by allowing
you to decide what "suitable" means!

[QUOTE=LiquidNitrogen;265406]Blah blah blah blah blah, you guys should run for office.[/QUOTE]

More than you needed to know ATM?

"Us guys" (comparative veterans) still enjoy "getting it straight".
:smile:

David

[QUOTE=davieddy;265491]I think all your qualms boil down to obtaining an unbiassed estimate of the probability of the non-zero residue of an unverified LL being wrong.[/QUOTE]Okay.

[QUOTE=Christenson;265277]P-1 is also good work for a CPU; I'm finding factors on average in significantly fewer GHz days than it would take to do two LL tests. You just have to give it a lot (a Gigabyte or so) of memory.[/QUOTE]

A Gigabtye is good. Mutliple Gigabytes are even better.

But half a Gigabyte per core, or even less, should be quite adequate for someone who wishes to specialize in this type of work.