Saw this in a result of mine:

de_separation_10_3s_fix20_2_1352770_1308235885_0

<stderr_txt>
Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '<name>' expected near '4'
Error reading astronomy parameters from file 'astronomy_parameters.txt'
Trying old parameters file
Using SSE2 path
Integral 0 time = 32865.269092 s
Integral 1 time = 7446.439993 s
Likelihood time = 5.785704 s
<background_integral> 0.000305972286305 </background_integral>
<stream_integral> 92.469354042897564 1736.230670756624500 205.096107473016390 </stream_integral>
<background_likelihood> -3.063092013626213 </background_likelihood>
<stream_only_likelihood> -5.857693463583072 -4.498928880820719 -3.983584257926393 </stream_only_likelihood>
<search_likelihood> -2.826062392501479 </search_likelihood>
<search_application> milkywayathome_client separation 0.84 Windows x86 double </search_application>
21:55:56 (1600): called boinc_finish

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

Saw this in a result of mine:

That's normal and OK

Interesting: Same gpu, same clocks, same mem size but 10% slower on the calculations.
You are using cat 11.2 right? Mine runs with cat 11.3.
Wonder if it's the different cat version or something else slowing your gpu down.

I think I know why. Despite what the log file reports, according to GPU-Z the figures for the defaults and current settings are :-

Clock 725Mhz - 775Mhz
Memory 1000MHz - 500MHz

As you can see I have upped the clock by an extra 50MHz, and lowered the memory by 500Mhz to keep the heat down. Also my stock memory speed was 1000Mhz not 1150Mhz. It depends upon the make of card what the stock speeds are, some are higher than others, and run 10% faster out of the box.

A fully committed CPU can bring it down a bit eg two CPU WUs running on a Dual CPU, and a GPU application running at the same time (2WUs + (say) 2x0.05 GPU assist = overcomitted CPU). Not going to be a lot for sure, but it will definitely slow the GPU app(s) a little.

As to how much, variations are nearly infinite, core useage, GPU efficiency yaddie yadda. Nontheless the effect would still be there to a greater or lesser degree - even comparing similar PCs if different CPU apps were run on both or non-BOINC useage was different.

Regards
Zy

I run 1 x Milkyway GPU @ 0.05 CPU, plus 2 x Seti CPU on a Pent D3.2. GPU-Z says I'm using 92-94% GPU utilisation, so I expect that explains it as well. I could run 2 x MW at a time for 100% utilisation but the card is quite happy at 73C and I'm inclined to leave it as it is.

If I had 1/2 dozen of these cards an extra 10% RAC across the board would be useful, as it is I'm nowhere near the high crunching figures so I'm happy with what I get, coupled with a longer card life.

If I had 1/2 dozen of these cards an extra 10% RAC across the board would be useful, as it is I'm nowhere near the high crunching figures so I'm happy with what I get, coupled with a longer card life.[/quote]


I've been overclocking the bejeezus out of many cards for just over a year now without a single card failure(except for a few fans replacements)
You have to TRY and do something crazy to kill a good graphics card!
I've even woke up to find my four 5870's below my cpu sitting in pooled water from the cpu watercooling blowout with no lost anything!!
(yeah, gotta say I got lucky there,LOL)
and sorry, didn't mean to spam this thread!!!

...... coupled with a longer card life.

Thats the bit that gets my Vote Big Time ..... often gets forgotten. Cards these days are resilient beasts at component level, and dont instantly go *poof*, they slowly burn out (hence the claims of some card makers over "military standard" components, they are trying to covince us their cards cantake the hassles).

Unfortunately this increase resilience, and lack of instant burnout, gives rise to a false sense of security in many. The card is degrading if its operating over recommended safe limits, it just takes longer as components are more resilient - but it is slowly going bang. Eventual Burnout happens 1-3 months later.

That delay will give rise to a false sense of security in many running extreme overclock, it may not go bang immediately .... but be under no illusion, its going downhill and the card will melt - period.

All that is not a statement against overclocking, I'm in there with the best of 'em when I get the urge, but it is a very cautionary statement to those just starting overclocking, about ignoring safe limits. Just because there is not instant fireworks, it doesnt mean there are no problems - there are, terminal ones.

Hence:

... coupled with a longer card life.

.. gets my vote every time.

Regards
Zy

I'm running an nvidia GTS450 and tried updating the drivers to the latest WHQL release, 275.33, hoping 0.82 had taken CUDA V4 into account. SETI liked the new drivers, MW not so much. My crunch times doubled yet MSI Afterburner didn't show downclocking, just erratic GPU utilization averaging about 50%.

Are there any work arounds or do we nvidia folk have to stick to the older 266.58 drivers?

Hi,

My computer can't finish this work : MilkyWay@Home v0.82 (ati14)
It says : Computing error

:(

Somebody know why?



Thanks!

Hi,

My computer can't finish this work : MilkyWay@Home v0.82 (ati14)
It says : Computing error

:(

Somebody know why?

Yes, see this thread, in particular the descriptions of different attempts to work around the problem.

---

Hi,

My computer can't finish this work : MilkyWay@Home v0.82 (ati14)
It says : Computing error

:(

Somebody know why?



Thanks!

Yes. Please have a look here:

http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=2468&nowrap=true#49994

for further help pM me

regards Franz

System is Linux 64 - Catalyst 11.6 - HD5830

Last night was getting "maximum time limit elapsed" errors on all GPU WUs.

Upgraded to 6.12.18
now I'm getting this -

<core_client_version>6.12.18</core_client_version>
<![CDATA[
<message>
Maximum elapsed time exceeded
</message>
<stderr_txt>
Error loading Lua script 'astronomy_parameters.txt': [string "number_parameters: 4..."]:1: '<name>' expected near '4'
Error reading astronomy parameters from file 'astronomy_parameters.txt'
Trying old parameters file
Using SSE3 path
Found 1 CAL devices
Chose device 0

Device target: CAL_TARGET_CYPRESS
Revision: 2
CAL Version: 1.4.1417
Engine clock: 825 Mhz
Memory clock: 1050 Mhz
GPU RAM: 1024
Wavefront size: 64
Double precision: CAL_TRUE
Compute shader: CAL_TRUE
Number SIMD: 14
Number shader engines: 2
Pitch alignment: 256
Surface alignment: 256
Max size 2D: { 16384, 16384 }

Estimated iteration time 152.765152 ms
Target frequency 30.000000 Hz, polling mode 1
Dividing into 4 chunks, initially sleeping for 0 ms
Integration range: { nu_steps = 640, mu_steps = 1600, r_steps = 1400 }
Using 4 chunk(s) with sizes: 400 400 400 400
Integration time = 127.293641 s, average per iteration = 198.896314 ms
Integral 0 time = 130.222443 s
Likelihood time = 5.456142 s
<background_integral> 0.000125639127066 </background_integral>
<stream_integral> 409.950098617738433 13.735900627600021 </stream_integral>
<background_likelihood> -3.063023849357760 </background_likelihood>
<stream_only_likelihood> -4.212400995965620 -173.914236733988474 </stream_only_likelihood>
<search_likelihood> -3.026010903920815 </search_likelihood>
<search_application> milkywayathome_client separation 0.82 Linux x86_64 double CAL++ </search_application>
01:31:06 (9368): called boinc_finish

</stderr_txt>
]]>