mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   GPU Computing (https://www.mersenneforum.org/forumdisplay.php?f=92)
-   -   Prime95 performance suffers after installing GT 640 (https://www.mersenneforum.org/showthread.php?t=18000)

Rodrigo 2013-03-25 00:26

[QUOTE=kracker;334837]mm? don't you mean cm? Hmm...
[/QUOTE]
No, no -- the PSU is very tiny, you see... :smile:

Yes, I meant cm. :smile: Haste makes waste.

Rodrigo

P.S. If the picture link I gave doesn't work, then try the one Aramis Wyler provided (thanks for that). If that doesn't work either, let me know and I'll attach the file to a thread post (there's no way to embed an image file here, is there).

Rodrigo 2013-03-25 00:28

[QUOTE=tServo;334836]I just took a piece of thin cardboard or heavy paper and poked holes in
it when flush against a standard power supply where the screws go
and then put it up against the machine where the power supply screws
are and saw there were at least 2, maybe 3 that didn't match;
and then started cursing.[/QUOTE]
So it sounds like my PSU is in fact an oddball?

Rodrigo

chalsall 2013-03-25 00:32

[QUOTE=Rodrigo;334840]So it sounds like my PSU is in fact an oddball?[/QUOTE]

Actually, it looks (to me) like the mounting screw positions are pretty standard.

Rodrigo 2013-03-25 00:45

[QUOTE=chalsall;334842]Actually, it looks (to me) like the mounting screw positions are pretty standard.[/QUOTE]
Cool! Assuming that's correct, what else do I need to take into account when shopping for a new PSU?

BTW, I reported upthread that the "good" dx7500 has the exact same PSU model [B]and[/B] a GT 630 with the same 20A spec -- and yet it's working fine with no Prime95 performance hit. Which would it be more accurate to say: that this is just good luck, or that my experience with the "bad" dx7500 is just bad luck?

I think I know the answer, but want to make sure.

Rodrigo

chalsall 2013-03-25 00:54

[QUOTE=Rodrigo;334846]Assuming that's correct, what else do I need to take into account when shopping for a new PSU?[/QUOTE]

Try to calculate your total power draw, taking into account the maximum draw of each voltage. (Note that PSUs are normally measured in watts, which is simply the summation of each voltage times amps drawn.) Try to get a PSU which supplies at least 1.5 times your aggregated draw and the draw at each voltage.

[QUOTE=Rodrigo;334846]BTW, I reported upthread that the "good" dx7500 has the exact same PSU model [B]and[/B] a GT 630 with the same 20A spec -- and yet it's working fine with no Prime95 performance hit. Which would it be more accurate to say: that this is just good luck, or that my experience with the "bad" dx7500 is just bad luck?[/QUOTE]

PSUs are analog devices. A great deal of variance can be experienced between identical models.

Edit: But, either way, you were drawing more than you should have been.

Rodrigo 2013-03-25 02:01

[QUOTE=chalsall;334847]PSUs are analog devices. A great deal of variance can be experienced between identical models.[/QUOTE]
Ah, I see, thanks.

[QUOTE=chalsall;334847]Try to calculate your total power draw, taking into account the maximum draw of each voltage. (Note that PSUs are normally measured in watts, which is simply the summation of each voltage times amps drawn.) Try to get a PSU which supplies at least 1.5 times your aggregated draw and the draw at each voltage.[/QUOTE]
According to my Kill-A-Watt, with the 630 in place the wattage with nothing running is 52. It's about 87w with Prime95 running, and 124w with both Prime95 and mfaktc (self-test) running.

How do I determine the maximum draw at "each" voltage? (Learning a lot here, hopefully this will be useful to future readers coming across this discussion.)

Rodrigo

chalsall 2013-03-25 04:46

[QUOTE=Rodrigo;334851]According to my Kill-A-Watt, with the 630 in place the wattage with nothing running is 52. It's about 87w with Prime95 running, and 124w with both Prime95 and mfaktc (self-test) running.[/QUOTE]

Hmmm... That indicates that the card is consuming only 37W (or about 3A @ 12V) under load.

But, considering according to the nvidia site the maximum draw is 65W (not 20A), that's not terribly unreasonable.

[QUOTE=Rodrigo;334846]How do I determine the maximum draw at "each" voltage? [/QUOTE]

For DC circuits, that's a bit tricky -- to measure power the meter has to be "in series". Normally what is done instead is what's called a "power budget".

Rodrigo 2013-03-25 05:49

[QUOTE=chalsall;334863]Hmmm... That indicates that the card is consuming only 37W (or about 3A @ 12V) under load.

But, considering according to the nvidia site the maximum draw is 65W (not 20A), that's not terribly unreasonable.[/QUOTE]
This doesn't sound so bad. The existing 300-watt PSU theoretically provides 2.4 times the wattage that's being drawn (124).

So, Doctor, :smile: taking these factors into account along with all the other things we've discovered -- what do you recommend? PSU transplant? "Watchful waiting" (doing nothing and monitoring the situation)? Maybe adopting a lean diet (not using the GPU)? Something else?

Rodrigo

Antonio 2013-03-25 07:35

[QUOTE=chalsall;334863]Hmmm... That indicates that the card is consuming only 37W (or about 3A @ 12V) under load.

But, considering according to the nvidia site the maximum draw is 65W (not 20A), that's not terribly unreasonable.



For DC circuits, that's a bit tricky -- to measure power the meter has to be "in series". Normally what is done instead is what's called a "power budget".[/QUOTE]


I have been following this thread and would point out the following.
1. Both the 630 and the 640 are powered only from the PCIe slot so the maximum power draw is 75W, this can consist of (at maximum) 10W from the 3.3V supply and 65W from the 12V supply. So it looks like nVidia are just quoting the PCIe spec!

2. I have a Zotac GT640, and from measurements I have made this card draws 8W on idle and 48W when running mfaktc (0.20), and has no effect at all on the timings of 2*(P-1) and 1*LL tests running on Prime95. In fact I have the GT640 and a GTX650Ti in the same computer, both running mfaktc without affecting Prime95 timings.

3. Considering the sequence of events outlined in the thread I would imagine that the GT640 was in fact faulty from the beginning and has now effectively died. Possibly a bad (or reverse polarity) capacitor in the voltage convertor circuit. Continuing to try to use it could damage the motherboard and/or the power supply. Trying it with a more powerful supply could increase the possibility of damaging the motherboard as more amps would be available to burn out the PCIe power tracks.

LaurV 2013-03-25 08:30

[QUOTE=Rodrigo;334869]So, Doctor, :smile: [/QUOTE]
This related to power supplies...
An old man goes to the doctor, accusing a strong pain in his left leg. "It is because of aging, Dad.", the doctor said trying to calm him down, "yeah, but you see", answered the old man, "the right leg has the same age, and it does not hurt..."

chalsall 2013-03-25 13:59

[QUOTE=Rodrigo;334869]Maybe adopting a lean diet (not using the GPU)?[/QUOTE]

I think taking Antonio advise would be sound. Perhaps contact the manufacturer and see if you can RMA it.

(And, to put on the table, I made a bit of a mistake in my counsel above. When you said the box mentioned 20A, I assumed that's what the card drew. Drilling down further, however, it appears that's what the minimum PSU should be -- the card draws much less.)


All times are UTC. The time now is 13:04.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.