475+ Mhash/s 5870 - Voltage mod and overclock

[rethaw]

I've been having a hard time keeping my 5870 below 80c with the fan at 100% atlhough that could be because as an eyfinity edition i have a rather small vent on the card (1/2 of one slot) i've been thinking about simply replacing the case side with a box fan though! 

I also have an eyefinity 5870. What I like about it is that all of the ports are recessed (minidisplay port). Because of this you can place a fan right up against the back which helps immensely with the exhaust. This is what I did:

[1713887662]

1713887662

[1713887662]

1713887662

[1713887662]

1713887662

[1713887662]

1713887662

[1713887662]

1713887662

[1713887662]

1713887662

[deslok]

i'll look into that however mine still has some DVI ports on it(not a flat surface) but that shows an answer to my question

[teukon]

Broke 500MH/s.

Congrats!  A significant milestone!  You also beat my highest stable core clock: 1110 MHz (I say stable but in reality I only ran it for 3 hours so who knows).  I see that you took your screenshot shortly after starting the miner.  How long did you run it for at that speed?  Perhaps you are still running it at that speed.

I did get my 5850s core clock up to 1140 MHz (473.4 MH/s) in testing but it only lasted for 30 seconds.  I'd love to try and compete but with a 5850 and air cooling I'm a little out of my depth.

[Mousepotato]

I've been having a hard time keeping my 5870 below 80c with the fan at 100% atlhough that could be because as an eyfinity edition i have a rather small vent on the card (1/2 of one slot) i've been thinking about simply replacing the case side with a box fan though! 

Do you have other cards right next to it in the case?  If so, get one of those "vortex" type fans ($15 or so at Target) and force some air across your boards.  It'll make a world of difference.

[deslok]

Actually there are no other cards in the case. it's an ITX size motherboard in a full ATX case with the one video card

[scifimike12]

Congrats!  A significant milestone!  You also beat my highest stable core clock: 1110 MHz (I say stable but in reality I only ran it for 3 hours so who knows).  I see that you took your screenshot shortly after starting the miner.  How long did you run it for at that speed?  Perhaps you are still running it at that speed.

I did get my 5850s core clock up to 1140 MHz (473.4 MH/s) in testing but it only lasted for 30 seconds.  I'd love to try and compete but with a 5850 and air cooling I'm a little out of my depth.

I ran it about an hour after taking the screenshot.  After checking my Kill-A-Watt, I'm saving ~40w by lowering the clocks to 1100/300 @ 1.300v.

It's hard messing with overclocking while mining because every second counts.   

[rethaw]

It's hard messing with overclocking while mining because every second counts.   

Haha so true. That's why I'm trying to get as much data before I start flashing again.

[Zagitta]

I've been having a hard time keeping my 5870 below 80c with the fan at 100% atlhough that could be because as an eyfinity edition i have a rather small vent on the card (1/2 of one slot) i've been thinking about simply replacing the case side with a box fan though! 

I also have an eyefinity 5870. What I like about it is that all of the ports are recessed (minidisplay port). Because of this you can place a fan right up against the back which helps immensely with the exhaust. This is what I did:

1) You're sucking air away from the card right?
2) Idk what that cover on the backside of the PCB is made of but you might be able to lower your temps even further if you can remove it (i say if because it might be an integral part of what keeps the fan attatched)
3) what RPM/CFM fan is that?

[rethaw]

1) You're sucking air away from the card right?
2) Idk what that cover on the backside of the PCB is made of but you might be able to lower your temps even further if you can remove it (i say if because it might be an integral part of what keeps the fan attatched)
3) what RPM/CFM fan is that?

1) Yes removing air.

2) Interesting thought, though temps are not so much of a problem that I would do that. Perhaps it even conducts some heat away from the card?

3) It is a .1 A Antec silent fan of some kind. It is running at 100% and is still silent so I'm sure you could easily outperform it with any other kind of fan. Having trouble finding the specific model it came from an Antec Sonata case.

[Zagitta]

I really think you should take a look at your VRM temps though... Because my asus 5850 had no issue overclocking up to 1050 mhz on the core when i increased the volt to 1.2 however my VRM's were incredibly hot, even with 100% fan, 125-129 degrees wich is just under the 130 degrees max supported operating temp...

But then again there are only 3 of them running in phase on a 5850, there might be more on a 5870 since it's able to go to as much as 1.35v...

[rethaw]

Did you measure the VRM temperatures manually?

[neotrino]

I was wondering if you would be willing to play around with some electromigration numbers. The Mean Time to Failure (MTTF) due to electromigration has been estimated empirically using:



Where A is the cross-sectional area, J is the current density, n is a scaling factor that is usually set to 2, k is the Boltzmann constant, Ea is the activation energy of the material, and T is the temperature. (link)

Lets take a VERY conservative MTTF based on your estimation of two years. You are welcome to provide your own adjustments, but based on my experience the temperature has not changed much raising the voltage and overclock. This may just be due to favorable cooling circumstances but I will assume the temperature sensor readings are accurate.

So Ea, T, k, and A are all constant or nearly so. This leaves us with the current density being the most important change. For non-ohmic materials like semi-conductors J can be expressed as:



Where sigma is the capacitance, E is the electric field, D is the diffusion constant, q is elementary charge, and n is the electron density. Therefore in calculating the current density D, q, and n are constant for the same component. Then we can look at the electric field inside our component. For the sake of simplicity we assume that we are dealing with a small wire where E = V / d. Again d remains the same and the electric field grows linearly with voltage.

The purpose of the analysis above is to show that current density grows more or less linearly with increasing voltage in our case. This means for for n=2, MTTF will reduce by the inverse square of the change in voltage. Doubling the voltage reduces the MTTF of the junction by 4, quadrupling by 16, etc.

Therefore by increasing the voltage by .1V we reduce the MTTF by roughly 15%. For the given two years the MTTF for any given junction will reduce by 3.5 months.

(change in MTTF) ~ (modded V / stock V)^-2

Given 2 years is a fairly conservative estimate of the life of your card you may be reducing the total lifetime by a much higher amount, but the percent reduction in mean time to failure will be the same. For the sake of completeness I will estimate the difference in hashes over the life of a card in the two circumstances. I will give a very bullish reduction in life based on the calculations above of 2 weeks (even though in fact it was about 100 hours).

Overclock at stock voltages of a 5870, 425 Mhash/s for 2 years (104 weeks) ~= 27 x 10^18 hashes
Overclock at 1.25 V, 470 Mhash/s for 90 weeks ~= 26 x 10^18 hashes
Untested! 1.35 V, 490 Mhash/s for 77 weeks ~= 23 x 10^18 hashes

Therefore, you are correct you will reduce the overall hashes your card will be able to perform based on the MTTF. Also, there are numerous other points of failure and I believe electromigration in the junctions is just one to consider.

I would like to further complicate this picture by suggesting that having an extra 20 MHash/s now will be worth much more than the entire hashing power of the card at the end of two years given difficulty increases. I have not calculated this and difficulty changes are not predictable that far ahead. But you can see that increasing difficulty is an exponential and decreasing MTTF is quadratic.

I look forward to other considerations in ramping up the voltage before I do so.

Wow.. this post its without doubt the best argument reasoning that I have read about the topic. Kudos to you 

[rethaw]

Little update. This is 1055Mhz at 1.25V. Not all cards are created equally. The other 5870 will not stay at over 1020 MHz for any significant length of time.

Code:

___________ [482.306 MH/s (~487 MH/s)] [Rej: 6/1432 (0%)] 

I also wanted to make a little attempt at arguing that 20 Mhash/s now is worth more than two months of mining two years from now.

Here is one an excellent graph of network computational power by sipa available here.

As you can see that over the course of two years it is extremely steep, especially in the last six months. Whether this will be maintained is anyone's guess.

You can very quickly see the importance of speed now. 20 MHash/s may seem small, but it is not difficult to imagine 200 Mhash/s being small in a matter of months. Current estimates project the next jump to be modest, however, just in the last two increases we've seen the same hardware giving half the payout. If this maintains it means that in say, 4 months, a 400 Mhash/s card will be paying out as much as 25 Mhash/s does today. Even if the difficulty increases slowed dramatically, evidence would suggest that in less than 18 months that extra 20 Mhash/s would be significant.

Obviously this does not take into account resale value of the card, and I wouldn't attempt to guess the exchange rates down the road. Neither is it a given an overvolted card will fail, but it does increase the likelihood.

[cyberlync]

Perhaps I am doing something wrong, better ask.

I have a couple of 5850's and a 5870 in a system, and whenever I try to OC the cards to 900 core, I seem to get more stales etc. The weird thing is, both the 5850's and the 5870 are hitting the same "wall" 900/300 they seem to get more stales, they restart a tad more often I think, 898/300 they seem fine.

I haven't done extensive testing, but somehow I feel it's not really the cards, but more my way of OC'ing. Any tips and tricks? How do I see if I pushed the card to the limit, only 1 of them is connected to a monitor, so I am guessing it's the only one that could make the system "hang" or am I wrong? Im using Trixx to OC, Afterburner to monitor temps, gpu load. The 5850's are Sapphire Xtreme's and the 5870 is a XFX.

Windows 7 32bit, AMD 11.7, AOCLBF 1.75, Phoenix 1.5, PhatK 07-17-11 (BFI_INT, Vectors, Aggression 13, Worksize 256)
5850a   898/300  ~369MH/s
5850b   898/300  ~369MH/s
5870     898/300  ~408MH/s

Thanks for any help/tips in advance.

[Mousepotato]

WTF, I was getting 495 MH/s @ 985/300 for about an hour.  I'm not sure what it is I did, but was mining Namecoins in GUIminer, then switched back to mining Bitcoins with Phoenix.  I've done it 2 or 3 times now and it will mine around 495 MH/s for about an hour before it locks up.  I wish I could repro this stably.

[mike678]

WTF, I was getting 495 MH/s @ 985/300 for about an hour.  I'm not sure what it is I did, but was mining Namecoins in GUIminer, then switched back to mining Bitcoins with Phoenix.  I've done it 2 or 3 times now and it will mine around 495 MH/s for about an hour before it locks up.  I wish I could repro this stably.

You can get the $190 5870's from newegg up to 495? 0.o Even if its only for a hour at that speed that's a huge bargain. Wish I had the money when they had that sale.

[wallet.dat]

WTF, I was getting 495 MH/s @ 985/300 for about an hour.  I'm not sure what it is I did, but was mining Namecoins in GUIminer, then switched back to mining Bitcoins with Phoenix.  I've done it 2 or 3 times now and it will mine around 495 MH/s for about an hour before it locks up.  I wish I could repro this stably.

You can get the $190 5870's from newegg up to 495? 0.o Even if its only for a hour at that speed that's a huge bargain. Wish I had the money when they had that sale.

This was with a 5870 reference card.  The $190 5870s run normally at around 460 MH/s at 1010/350.