Team drive #9 k=1005-2000 n=50K-350K

[IronBits]

Just added 6 more cores w/20 cache set

Too bad the I7 is stuck on sieving till 11 Jan, or I would have tossed it on there to.

[em99010pepe]

I would like to reserve 4k's to be run on my own llrnet server. Please send the file via PM or email. Thank you.

Carlos

[henryzz]

Originally Posted by IronBits

Just added 6 more cores w/20 cache set

Too bad the I7 is stuck on sieving till 11 Jan, or I would have tossed it on there to.

if you put your I7 on this at some point i would suggest using hyperthreading as more cores should overall be faster for so low numbers
i would guess slow computers wont be that much slower on this drive

[kar_bon]

some 2 ct's from me:

- no manual reservations for this drive:
the range is small but to handle it (split files, send/receive results, verify completeness, display primes/done ranges in pages) is to heavy to handle

- i suggest only to put some 'slower' cpu's on this drive because we need the 'runners' for the higher n-ranges (Drives #5-#7)!

- keep the order: ascending k-values and within a k the ascending n-values:
it's quite easier for me to update the PrimeDatabase!

[gd_barnes]

OMG. We have blown away over half the file that I sent already! I think it's already handing out k=1029 or 1031. I only sent up to k=1049.

David, I'm going to sent you a file over 3 times as big. The last file was k=1005-1049. The next one will be k=1051-1199. After that depending on the continued response, I'll send them in k=200 ranges, which will be 4 times as big as first file that I sent you. I'll try to make them last a week.

Karsten is right, this drive is better for slower machines but personally, I don't care as long as David's server can handle it. If people want to slay this range quickly, let's do it!

Yes, we will have a lot of results to match up but that's no big deal. We'll get it taken care of in good time.

Due to the huge response to these low n-ranges, I will keep my cores off of it unless it starts to wane.

Max, let's plan on processing the results on 100 k's at a time for this drive; perhaps even 200 k's. Otherwise we'll be doing them about every 1-2 days. lol


Gary

[gd_barnes]

One thing here: It'd almost be better if people move a core or two back to the 5th-6th-7th drives. Not that we mind but it'd be nice to continue finding top-5000 primes too.


Gary

[gd_barnes]

reserving k=1051-1199 for port 9000

[gd_barnes]

Originally Posted by em99010pepe

I would like to reserve 4k's to be run on my own llrnet server. Please send the file via PM or email. Thank you.

Carlos

Carlos,

With the way we are smoking through the ranges, are you sure you want to mess with loading 4 k's into your server? I think it'd be easier if you just stayed connected to port 9000 here. But if you really want them, I'll send you k=1201-1207 within a day.


Gary

[IronBits]

Loaded up what you sent me, without a single hiccup when I restarted the server.

Port 9000 has 178657 remaining knpairs!
knpairs.txt - Size: 2222009
Last Updated: 1/7/2009 7:55:39 AM
first k/n pair 1007 189180
last k/n pair 1049 199976
jobMaxTime = 1 * 24 * 3600 -- 1 days
prunePeriod = 1 * 1 * 3600 -- hourly

[IronBits]

Backing off 6 cores that I put on before I went to bed to test the server a little more.
If you guys want to toss on a couple more cores, let er rip.

edit: ok, 6 cores moved back to port 5000

[Flatlander]

Originally Posted by IronBits

Loaded up what you sent me, without a single hiccup when I restarted the server.

Port 9000 has 178657 remaining knpairs!
knpairs.txt - Size: 2222009
Last Updated: 1/7/2009 7:55:39 AM
first k/n pair 1007 189180
last k/n pair 1049 199976
jobMaxTime = 1 * 24 * 3600 -- 1 days
prunePeriod = 1 * 1 * 3600 -- hourly

I think the daily totals reset when you restarted.
http://nplb.ironbits.net/progress_9000.html