![]() |
![]() |
#1 |
"University student"
May 2021
Beijing, China
32×29 Posts |
![]()
Statistics showed that yesterday only 291 first prime tests and 481 verified tests have been completed yesterday. But back in May and June, on average about 310 first prime tests and 600 verified tests were completed every day.
Meanwhile, I've also seen a significant reduction in GPU TF activity. Only months ago, there were days when thousands of numbers were factored. Yesterday only 129 numbers were factored. Why could this happen? |
![]() |
![]() |
![]() |
#2 | |
Jun 2003
19·283 Posts |
![]() Quote:
Last TF level reached the end of the GIMPS range and might have wrapped around to the next bit level/lowest exponents. That would reduce the thruput a lot. But also, some of the SRBase firepower is being redirected to low exponents in support of the "2k" project. |
|
![]() |
![]() |
![]() |
#3 |
Banned
"Luigi"
Aug 2002
Team Italia
3·5·17·19 Posts |
![]()
Also consider that the temperatures are quite high here in Europe, and most GPU owners stop their computers due to the overheating.
And that the 2 central weeks of August are often chosen for vacation. ![]() Last fiddled with by ET_ on 2021-08-15 at 10:17 |
![]() |
![]() |
![]() |
#4 |
"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest
3·5·439 Posts |
![]()
There are several factors. Here's another. Possibly more university computers are off for the summer in first half of August than in May. Summer session enrollment tends to be lower than in full length semesters September to May. Some computer labs can be turned off to save power / money. That's the American schedule, which may apply elsewhere too, and may affect curtisc's throughput (University of Central Missouri).
Lately it's been hot enough I'm selectively idling some gear. |
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
GPU Computing Cheat Sheet (a.k.a. GPU Computing Guide) | Brain | GPU Computing | 20 | 2015-10-25 18:39 |
Efficient modular reduction with SSE | fivemack | Computer Science & Computational Number Theory | 15 | 2009-02-19 06:32 |
CPU stats reduction | Prime95 | PrimeNet | 3 | 2008-11-17 19:57 |
Computing n-th power residue symbols | geoff | Sierpinski/Riesel Base 5 | 2 | 2006-10-24 00:09 |
Lattice Reduction | R.D. Silverman | Factoring | 2 | 2005-08-03 13:55 |