48 hr rally

[arminius]

next stage: FFT length 32K

[axn]

Originally Posted by arminius

next stage: FFT length 32K

Only the SSE2 machines have switched over to 32K. Non-SSE2 machines are still doing 28K tests.

[tnerual]

only 25 minutes before the 21 utc stats update and the end of the riesel prp 48H rally

[tnerual]

results by test done in 48 hours :

1423 grafzahl
1394 masser
680 tnerual
675 axn1
643 xentar
521 geoff
321 jasong
320 free-dc mercenaries
315 arminius
269 yoyo
238 birdman2584
185 bananeweizen
58 robert44444uk
35 ladysilver
3 vfrey

results by points:

382,697941 grafzahl
379,532434 masser
204,973722 geoff
183,049091 tnerual
182,124938 axn1
172,506086 xentar
86,995996 free-dc mercenaries
86,547315 jasong
83,989093 arminius
65,696204 yoyo
64,092505 birdman2584
47,819515 bananeweizen
15,672750 robert44444uk
8,979794 ladysilver
0,789864 vfrey

the difference is here: geoff gain 2 place (probably offline results on his k) and small gain for carlos (free-dc mercenaries) probably because he started later than jasong and so he get higher point value test.

congrats everybody for participating in this push !

[em99010pepe]

I didn't move my machines to the rally because I forgot. It was a friend who helped Mercenaries with his dual-core machines.

Carlos

[jasong]

Originally Posted by em99010pepe

I didn't move my machines to the rally because I forgot. It was a friend who helped Mercenaries with his dual-core machines.

Carlos

lol, the guy who starts it ends up forgetting to join. lmao

[masser]

382,697941 grafzahl
379,532434 masser

Congrats GrafZahl! I almost caught up with you, but fell just short...


Next time, we should work on the Sierpinski queue.

[tnerual]

yes, it's a good idea, now that riesel is a little in advance.

thinking loud (or aloud ? don't know the exact word in english):
2 rally a month, one for sierpinski, one for riesel is a good idea. first week end for riesel, third week end for sierpinski. sieving or testing between ... all k at n=200000 for the end of the year ... yeah but it's 2500 test a day for riesel only and we have to add sierpinski ... but there will be prime ... but testing time increase (3000sec a test at n=190000 on amdX2 @ 2400) ...rhaaa will it be possible ...

[jasong]

Originally Posted by tnerual

yes, it's a good idea, now that riesel is a little in advance.

thinking loud (or aloud ? don't know the exact word in english):
2 rally a month, one for sierpinski, one for riesel is a good idea. first week end for riesel, third week end for sierpinski. sieving or testing between ... all k at n=200000 for the end of the year ... yeah but it's 2500 test a day for riesel only and we have to add sierpinski ... but there will be prime ... but testing time increase (3000sec a test at n=190000 on amdX2 @ 2400) ...rhaaa will it be possible ...

I agree about the second and fourth weekend(but what if the month starts on a Sunday, does that count as a weekend).

But instead of alternating Sierpenski and Riesel, let's have someone announce which one has the lowest tests on the Thursday before and have everyone concentrate on that side. That will maximize the chance of someone finding a prime and will, therefore, be more fun.