Team drive #1: k=400-1001 n=333.2K-600K

[gd_barnes]

Originally Posted by MyDogBuster

So everyone has been challenged to finish this BEFORE the fat guy gets stuck in my chimmney again. I read Gary's analysis post and got a headache halfway thru, but all seems plausible barring any unforseen disasters.

He he. My estimate posts create headaches for many folks. Math is fun!

My estimates of times to complete ranges at CRUS get even worse. I won't go there. The problem is that people will sometimes reserve 50-100 CPU years of work and won't even realize it thinking that they can complete it in a few months on their single quad! lol

[henryzz]

Originally Posted by gd_barnes

He he. My estimate posts create headaches for many folks. Math is fun!

My estimates of times to complete ranges at CRUS get even worse. I won't go there. The problem is that people will sometimes reserve 50-100 CPU years of work and won't even realize it thinking that they can complete it in a few months on their single quad! lol

no matter how hard i try i cannot work how you get some of your results from some of your CRUS estimates

[gd_barnes]

Originally Posted by henryzz

no matter how hard i try i cannot work how you get some of your results from some of your CRUS estimates

It isn't easy. You have to remember that the testing times vary by the square of the exponent. n=200K takes 4 times as long to test as n=100K. Also, the k's remaining reduce at a reverse logarithmic rate (or a reverse compound-interest rate; not sure what the correct term really is there). That is at n=10K, there might be 48 k's remaining and at n=20K, there might be 36 k's remaining; which means that there is a 25% reduction in k's remaining for each doubling of the n-value. Therefore, the estimate is that you should have 36*.75=27 k's remaining at n=40K.

Since I put some of those estimates up, I realized that there is a slight error in my calculations: The k's remaining do NOT reduce at a reverse logarithmic rate because the average weight of the final few k's is much lower than previous k's. I do not know the proper way to make such an adjustment. Using the above situation with 36 k's remaining at n=20K, it could be that the estimate should be that there should be 29 k's remaining at n=40K but that is only a guess. Still, using the above estimate gives the least amount of time that it SHOULD take. It gives a general idea of the amount of work reserved.


Gary

[kar_bon]

my 2 cent :-)
last days pairs submitted to all server:
2008-11-25: 13427
2008-11-26: 13809
2008-11-27: 14163
2008-11-28: 15093
2008-11-29: 15169
2008-11-30: 14953

there're about 320,000 pairs remain (in server-pool or not reserved yet).
so with about 15,000 pairs per day we have 21 days left!

[MyDogBuster]

The problem is that people will sometimes reserve 50-100 CPU years of work and won't even realize it thinking that they can complete it in a few months on their single quad! lol

That's why when I start something new, like I just did this week over at CRUS, I ALWAYS start small and build up. That way I know I can finish before the house falls down around me. (Probably caused by the fat guy being stuck in the chimmney).

[em99010pepe]

575.2K-576K is 50 % completed.
I think today I'll receive a quad-core for my work....hehehe

[Lennart]

578.0-578.1
578.1-578.2
578.2-578.3
578.3-578.4
578.4-578.5

They are done.

/Lennsrt

[henryzz]

what pc has Wabbit98 got on port 400
he is only managing 1 or 2 pairs an hour
i would guess he is puting an old amd to use

[em99010pepe]

Is it possible to put more work available?

[mdettweiler]

Originally Posted by em99010pepe

Is it possible to put more work available?

Okay, I can add more shortly. How much are you specifically interested in? If you're looking for a larger range, I can make a big file of that size for you.

[mdettweiler]

Originally Posted by mdettweiler

Okay, I can add more shortly. How much are you specifically interested in? If you're looking for a larger range, I can make a big file of that size for you.

Okay, I've added another n=1K worth of work (10 files).