1) Message boards : Number crunching : Why is SSE3 or PNI and SSE2 used for the same work units? (Message 24990)
Posted 3 Dec 2012 by Fred J. Verster
Looking at some results, also on BOINC (7.0.28), I see SSE2 and SSE3(PNI*)
Used for the same work?
*Presscot New Instruction Set.
Tasks are done by an I7-2600, with a very small OC and an C2Extreme X9650,
running at 3.6 GHz. (20% OC)

The 2 hosts.

I don't see any difference in completion time, so I assume there is another
Maybe 'older hosts' are doing better/faster with SSE2 and later hosts SSE3?
And are quite old if only SSE2 can be used.
Or AMD CPUs, older types doesn't have SSE3 or PNI

Or just testing which is better overall?
SSSE3x and SSE4.1/4.2 usually is faster then SSE3/PNI.
2) Message boards : Number crunching : Sorry to have to return those great wu's (Message 24959)
Posted 16 Nov 2012 by Fred J. Verster
Well I just increased recourses for LHG, SixTrack and can run on 10
CPU cores with 3390 to 3560MFLOPS (Whetstone), Drystone =>10GFLOPS/CPU(core).

Einstein is a big recourse hog, especially BRP4ATI/CUDA...;-)
But do a good job in finding a lot of Pulsars, maybe they can detect
a Quasar or Magnetar which emits a different "signal" or radiation or
(very)strong magnetic field (>100Teslas/m?).

3) Message boards : Number crunching : error -177 resource limit exceeded (Message 23794)
Posted 28 Dec 2011 by Fred J. Verster

My Computing preferences: for SixTrack and Test4Theory:

Disk and memory usage:

Use at most 542 GB disk space

Leave at least:

0.001 GB disk space free

(Values smaller than 0.001 are ignored)

Use at most 100% of total disk space

Tasks checkpoint to disk at most every 60 seconds

Use at most 100% of page file (swap space)

Use at most 100% of memory when computer is in use

Use at most 100% of memory when computer is not in use

my computer

Number of processors 8
Coprocessors NVIDIA Quadro 2000 (961MB) driver: 27536
Operating System Microsoft Windows 7
Ultimate x86 Edition, Service Pack 1, (06.01.7601.00)
BOINC client version 6.12.34
Memory 3325.58 MB
Cache 256 KB
Swap space 6649.45 MB
Total disk space 148.24 GB
Free Disk Space 113.99 GB

Hmmm ... what else could be wrong?

Hmmm, nothing IMO, except isn't giving all memory-resources
to BOINC, a little too much, Windows has some Services and Processes
, too
, which also need memory resources.

Imagine what happens when you start using your computer with all
memory resources assigned to BOINC, Windows can't use it's memory, swap-space
and the entire disk-space available, a BSOD is what I would expect!?

Trying to avoid this situation, I made a ~250GByte Disk-Partition for
BOINC (6.12.34; 64bit) only, on a third (1TByte) (system;SATA II) hard-disk. But any (system) hard-disk will do, even an USB =/> 2.0/3.0 disk
should work.
Also, 8 threads, but only 4 GByte system memory, leaves 0.5GByte DRAM
for each process or thread, using 8 GiG, on my I7(2600) host and running a
memory hungry simulation ends up using 7.6GByte RAM!

Answer is a bit late, but still applies ;-)

By the way: Wishing all a Happy, fortunate and healthy 2012.
4) Message boards : Number crunching : no more work? (Message 23785)
Posted 21 Dec 2011 by Fred J. Verster
Haven't seen work for the last couple of days, last work was send back a few days ago.
Maybe I should add another host account, have 2 switched off and 2 running.
5) Message boards : Number crunching : error -177 resource limit exceeded (Message 23746)
Posted 27 Nov 2011 by Fred J. Verster
Maximum elapsed time exceeded

Might be a configuration error on project side, each result has a value "rsc_fpops_bound" (or similar), where the project guys can configure when a result is aborted. This is thought to avoid endless loops / iterations that never reach their desired target value but it should not knock out a result that still is working properly.

Afaik. the benchmark results influence the value that is compared against this rsc_fpops_bound but your benchmark values do not look unusually high.

One possible cause on client side would be a power saving mode, where the host runs at a reduced clock speed.

Dynamic turbo mode on some CPU cores could have a similar effect, if the benchmark has been carried out on a higher clocked core. I have read somewhere that AMD has something like that on later Bulldozer chips but I'm not sure what this exactly means for BOINC.

Isn't a -177 error a timing error? I remember this happening at SETI@home, when using a GPU.
I also run Rosetta and was amazed by the >700MByte WU RAM use.
This host, now only does CPU jobs, a.t.m. (It has 2 EAH5870 GPU's)
Sandy Bridge CPU's, use a dynamic turbo, too, if CPU load is low, clock frquency
goes down.
6) Message boards : Number crunching : Faulty Computers or Modified BOINC ?? Huge Credits (Message 23745)
Posted 27 Nov 2011 by Fred J. Verster
Hi, I agree with former speakers, this should be impossible, displaying a
normal 'run-time' and an unusual large CPU-time!

If you were to drop the CPU speed to 1%, run-time would be affected, too.
I've seen weird situations running several projects, with (very) different
memory use, very high, >700MByte (Rosetta) per WU'.
Also the use of (a) GPU(s), if possible, is reducing CPU load,
time and (total)runtime.

And I agree that one is only guilty if proven so. And computers can display
weird or odd results, although, very uncommon.
When seriously OC'ed, heat, a gamma-ray :), etc.

Since I'm not familiar with the credit granting, anything with a 10 or 100
times higher runtime v.s. CPU time, or vice-versa, should raise a question and be looked at!

7) Questions and Answers : Windows : Can't reach site on Comcast network (Message 22562)
Posted 26 Sep 2010 by Fred J. Verster
Hi, looks like some \'weird\' limitation, from ComCast ISP,
since I\'ve got no problem reaching it.

I do have trouble sometimes in reaching, sites on the East Cost of the U.S.A., f.i. SETI, when people from California can connect, I still can\'t, maybe the longer distance is a reason.
It\'s hard to say, so many \'stations\' between my place (Zutphen in the Netherlands) and Berkeley.
And policies from ISP\'s aren\'t that clear, sometimes.
8) Message boards : LHC@home Science : Anyone getting work units? (Message 22406)
Posted 26 Jun 2010 by Fred J. Verster
0 units for my Linux box. But I have 5 other BOINC projects giving me work, plus a virtual machine running SETI on Solaris. So I am never out of work.

Yep, had some WU\'s on this VISTA 32 QUAD (Q6600), but just read a message :Server Status
Up, Out of work
1009 workunits in progress
9 concurrent connections.

Hope they soon do have some work. Do have some (9) back-up projects, too.
9) Message boards : LHC@home Science : Anyone getting work units? (Message 22391)
Posted 18 Jun 2010 by Fred J. Verster
Yeah I just add the LHC and I haven\\\'t got units yet... Thanks for letting me now.....

Since a few days, I got some work, not much, still better then nothing at all.
Have a few hosts, still connected to LHC@Home and after 6 month\'s, maybe more, I got work, again.
Seems ages, but it took them quite some time, to get the Collider working again!
10) Message boards : Number crunching : lhcathome.cern.ch blacklisting IP addresses? (Message 21531)
Posted 6 Oct 2009 by Fred J. Verster
Hi, long time since I visited the LHC fora.

My last, attempt to get some work, looks not so good:

6-10-2009 16:50:57 lhcathome update requested by user
6-10-2009 16:51:01 lhcathome Sending scheduler request: Requested by user.
6-10-2009 16:51:01 lhcathome Requesting new tasks
6-10-2009 16:51:06 lhcathome Scheduler request completed: got 0 new tasks.

NO contact problem, here in Holland.

[ADDED] Blocking IP adresses ???
11) Message boards : Number crunching : Can't get work (Message 21530)
Posted 6 Oct 2009 by Fred J. Verster
Hi, still hoping to get some work.

After having some trouble with BOINC versions 6.6.38 & 36.
Especially on WIN VISTA 32bit Home Edition. It\'s a HP Pavillion with a Q6600 and 2 GiG DDR2 RAM

Still missing SP2, have to be downloaded and installed, it took 34 times, to, finally install SP1, though.

Seeing the last post, in this forum, it looks like there isn\'t much of a chance, getting work.

Is the Large Hadron Collider, still out off order, or in some testing fase?
12) Message boards : Number crunching : Fortran error reported, when running concurrently with a 64 BIT app. (Message 20897)
Posted 31 Dec 2008 by Fred J. Verster
Hello, my first post here, but I noticed a rare fault, when running, 3 64BIT SSSE3 optimized SETI MB app. and 1 LHC Work Unit.
No BSOD, but a notice about a FORTRAN error, ending in faulty result.
A X9650 @ 3690MHz (410x9; (2T) memory timings.
Machine is stable and hasn\'t made an error, yet, that\'s why I, finally reported it.
Fred J. Verster.

©2024 CERN