PrimeGrid
Please visit donation page to help the project cover running costs for this month

Toggle Menu

Join PrimeGrid

Returning Participants

Community

Leader Boards

Results

Other

drummers-lowrise
1) Message boards : Number crunching : International Day of Education Challenge (Message 136574)
Posted 1 day ago by Profile AzmodesProject donor
T-Armstrong, looking at your computer, you're running 321 sv single-threaded, but also PPS sv on GPU. Maybe you're confusing the two here? To my knowledge, MT is exclusively for LLR tasks.
2) Message boards : Generalized Fermat Prime Search : GFN-14 Consecutive Primes Hunt Season is open! (Message 136448)
Posted 5 days ago by Profile AzmodesProject donor
I haven't tried the transfer thing, though I did restart BOINC and also reboot the computer. Neither worked.
3) Message boards : Generalized Fermat Prime Search : GFN-14 Consecutive Primes Hunt Season is open! (Message 136446)
Posted 5 days ago by Profile AzmodesProject donor
One of my hosts has over 4,000 tasks to report, but everytime I request an update it says "Scheduler request failed: HTTP internal server error". All my other machines work fine, by the way.

EDIT: Alright, resetting the project seems to have fixed it. Too bad about those 4,000+ tasks, though.
4) Message boards : Generalized Fermat Prime Search : GFN-14 Consecutive Primes Hunt Season is open! (Message 136380)
Posted 10 days ago by Profile AzmodesProject donor
Still some build-up of pending tasks resulting in the occasional dry spell for the faster hosts, but it's not as bad as when I first posted here.
5) Message boards : Generalized Fermat Prime Search : GFN-14 Consecutive Primes Hunt Season is open! (Message 136361)
Posted 11 days ago by Profile AzmodesProject donor
That makes sense, I hadn't even considered that. Thanks for the explanation, stream.

Provided a continuous supply of work, I estimate that I can do around 200,000+ GFN-13 tasks per day. I'll let it run for a couple days at the very least now, but would of course appreciate any DC help I can get. :) If anyone is ever running into the same problem, feel free to send me a PM and I'll try to help.
6) Message boards : Generalized Fermat Prime Search : GFN-14 Consecutive Primes Hunt Season is open! (Message 136337)
Posted 12 days ago by Profile AzmodesProject donor
I'm having trouble getting a consistent stream of work for my faster machines (ex. two 2070S running four tasks each, with each task taking about 10 seconds). I assume that the problem is that those hosts chew through so many tasks so quickly that the server refuses to send more for a certain period. I have of course PG as a back-up project for that case, but I'm wondering if there is anything I could do to ensure that I get a steady work flow? Setting a higher work cache doesn't really solve it, since the problem persists when I inevitably run dry again (plus the client doesn't particularly like it when there's thousands of tasks queued).

The event log isn't very helpful, by the way. It just says "got 0 new tasks" whenever a request for more is sent.
7) Message boards : Number crunching : Google Colab and its free GPUs (Message 135424)
Posted 47 days ago by Profile AzmodesProject donor
I successully ran a GFN21 task in about five hours on a P100. This is faster than a 2080 Ti; evidently those 4 TeraFLOPS of DP compute flexing their muscles. As far as I know, GFN21 and 22 (not DYFL) are the only projects that can still make use of DP on account of the OCL transform.

Anyway, could be useful for the upcoming challenge. If you can get a P100. :)
8) Message boards : Number crunching : Any information how much faster a RTX 20XX Super is compared to the "old" RTX 20XX GPU card's ?. (Message 134231)
Posted 91 days ago by Profile AzmodesProject donor
Here's some numbers I got from my new 2070 Super. Running the newest Ubuntu and one core kept free for the GPU. Times in seconds. The numbers in parentheses are the durations from my regular 2070 (though it is a Founder's Edition, so a tad faster).

AP27: 740 (849)
PPS sieve: 291 (161)* (1 task)
GFN15: 29 (25)**
GFN16: 56 (61)
GFN17Low: 188 (210)
GFN17Mega: 206 (231)
GFN18: 686 (818)
GFN19: 1675 (1917)
GFN20: 6432 (7603)

I also managed close to 700 P/day on manual GFN22 sieving (b13), cf. a bit over 600 for the regular 2070.

The Super is running driver version 435.21, with an i7-3770. The regular 2070 is running driver version 430.50 with an i9-9900k, on Ubuntu 18.04.

*Not sure why the regular version is so much faster here...?
**I'm assuming the numbers here aren't really all that indicative of performance, since GFN15 is very inefficient in GPU usage and maybe stuff like CPU and/or RAM comes into play?
9) Message boards : Number crunching : Google Colab and its free GPUs (Message 134043)
Posted 98 days ago by Profile AzmodesProject donor
AP27: little bit under an hour
GFN17Low: ~790 secs
GFN17Mega: ~850 secs
GFN18: ~3000 secs
PPS sieve: 600 secs (one task running)
10) Message boards : Number crunching : Google Colab and its free GPUs (Message 134003)
Posted 99 days ago by Profile AzmodesProject donor
GFN15: 75 secs
321 sieve: ~5,100 secs


Next 10 posts
[Return to PrimeGrid main page]
DNS Powered by DNSEXIT.COM
Copyright © 2005 - 2020 Rytis Slatkevičius (contact) and PrimeGrid community. Server load 3.45, 3.41, 3.60
Generated 25 Jan 2020 | 11:47:48 UTC