Posts by Ananas
log in
1) Message boards : Number crunching : Updates on BoincStats (Message 2279)
Posted 17 Jan 2016 by Ananas
FYI :

I just found this in the BoincStats board :
No, there was a real problem here.

The stats should not be three days old. So if that's the case please show examples.

You are however correct about SRBase. They have a dynamic IP and the DNS server of the datacenter was lagging and still pointing to an old IP. I changed DNS to Google en OpenDNS and the IP is now correct, stats will be imported on the next run.

So everything should be fine by this evening :-)
2) Message boards : Number crunching : How do i know if i found a prime? (Message 2246)
Posted 4 Jan 2016 by Ananas
Riesels seem to run much longer if they suspect a prime in the 1st pass.

So it might be interesting to have a look at the task details if you catch one of those by chance.

p.s.: I guess that's not a reportable one :
1411147994*3^243348-1 may be prime, trying to compute gcd's U((N+1)/3) is coprime to N! 1411147994*3^243348-1 is prime! (116116 decimal digits, P = 9) Time : 5408.691 sec. 08:44:21 (2916): llr.exe exited; CPU time 5401.109375
3) Message boards : Number crunching : Too long calculation for the Sierpinski/Riesel Bases - long (Message 2243)
Posted 2 Jan 2016 by Ananas
...
Its better to disable HT. --> runtime ~392210s

But only 8 concurrent tasks then instead of 16, overall that would be a loss.

I guess I need an AVX machine some day :-/
4) Message boards : Number crunching : Too long calculation for the Sierpinski/Riesel Bases - long (Message 2241)
Posted 2 Jan 2016 by Ananas
...
    8*800^951038-1, bit: 8250000 / 9171677 [89.95%]. Time per bit: 34.432 ms.

...


It is getting harder to keep the deadline ...

    8*800^973840-1, bit: 50000 / 9391576 [0.53%]. Time per bit: 41.762 ms.
    8*800^973840-1, bit: 100000 / 9391576 [1.06%]. Time per bit: 40.380 ms.
    ...
    8*800^973840-1, bit: 3250000 / 9391576 [34.60%]. Time per bit: 40.744 ms.
    8*800^973840-1, bit: 3300000 / 9391576 [35.13%]. Time per bit: 147.026 ms. <= oops!
    8*800^973840-1, bit: 3350000 / 9391576 [35.67%]. Time per bit: 40.288 ms.
    ...
    8*800^973840-1, bit: 7100000 / 9391576 [75.59%]. Time per bit: 41.622 ms.
    8*800^973840-1, bit: 7150000 / 9391576 [76.13%]. Time per bit: 41.693 ms.

same computer, 20% longer processing time/bit

5) Message boards : Number crunching : Too long calculation for the Sierpinski/Riesel Bases - long (Message 2222)
Posted 29 Dec 2015 by Ananas
...
You can easily check the line in stderr.txt in slot directory, find this line:

8*800^914450-1, bit: 50000 / 8818828 [0.56%]. Time per bit: 34.238 ms.

...

The last stderr line seems to be even more interesting (taken from an active run) :

    8*800^951038-1, bit: 8250000 / 9171677 [89.95%]. Time per bit: 34.432 ms.


BoincView shows only somewhat more than 60.5% (in yellow, my color setting for "will miss deadline")

p.s.: If the wrapper could read that value, it could report that one to the core client. That's how the RNA wrapper does it - unfortunately the RNA applications don't have such a linear loop characteristic so the values are still not too exact there. Even though I do know the expected numbers of loops exactly, the progress line is not straight.

6) Message boards : Number crunching : Too long calculation for the Sierpinski/Riesel Bases - long (Message 2191)
Posted 16 Dec 2015 by Ananas
Thanks, I will use that formula next time to fight my (not the core client's) deadline panic :-)
7) Message boards : Number crunching : Too long calculation for the Sierpinski/Riesel Bases - long (Message 2187)
Posted 16 Dec 2015 by Ananas
...
Did they run to 100% or have they been finished earlier?
...

I can answer it myself now, they end somewhere between 95.5% and 95.8% - I finished all 6 before a second one had been sent out and it gave my RAC temporarily a nice boost :-)
8) Message boards : Number crunching : Too long calculation for the Sierpinski/Riesel Bases - long (Message 2186)
Posted 15 Dec 2015 by Ananas
I'have stopped the Sierpinski / Riesel Base - long v0.01 WUs because with my Phenom II X6 1100T under Ubuntu 14.10 64 bits the time of calculation was far too much long especially the last 10% where progress moved only 0.001 % forward every 2 - 3 seconds.
...

Did they run to 100% or have they been finished earlier?

I have lix long ones on a lowpower Xeon (much slower than your box, especially the int speed - and no AVX on chip) and I'm afraid that the last 10% will run much longer than the first 90%

The result with the best progress is 93.75% / 79 hours, this morning (~10 hours ago) it was at about 91% so it isn't too much happening per hour :-/
9) Message boards : Number crunching : short deadlines cause panic mode (Message 2177)
Posted 14 Dec 2015 by Ananas
Something that would help - but I'm not sure if it can be done so easily :

Delay the resend of the long ones after a timeout by 2 days without changing the deadline itself.

This would give the slower hosts a better chance to return them before the server side scheduler sends them out a second time, even if they are somewhat behind the deadline.

According to the server status page, they currently have a worst runtime of 52+ hours, hard to do within the deadline if they are not on top of the stack on client side.

p.s.: if it is hard to do by application ... this behaviour wouldn't hurt for any kind of results, I even think it should be default for BOINC servers. The client still would do its best to stay within the deadline but the server would be more tolerant. I made that proposal a few years ago but the Berkeley guys didn't pick up the idea :-/
10) Message boards : Number crunching : server overloaded ??? (Message 2130)
Posted 1 Dec 2015 by Ananas
...
Good to hear, nothing changed on server side.

Lots of tiny results seem to have a major impact on the server's connection to the internet, when a lot of the 2-3 minutes Riesel results are available, the server becomes somewhat sluggish sometimes.

I guess it is not necessarily the throughput but rather the number of concurrent connections that causes this.




Main page · Your account · Message boards


Copyright © 2014-2017 BOINC Confederation / rebirther