SweClockers: Geforce GTX 590 burns @ 772MHz & 1.025V

[LiuKangBakinPie]

repost from "official" thread.

OK guys, I want you to pay attention to this video.

The reviewer in this video has his gtx590 at 809 core (30% overclock) memory at 2128 (20% overclock) , @ 1.05 volts (max afterburner voltage) running at 86c. Pulling over 500 watts, just like the 6990 did in the Anandtech review, so no surprises there.
Driver version 267.71.

http://www.linustechtips.com/ltt-vid...inus-tech-tips

It is VERY OBVIOUS that using the older driver was the problem. It not bad vrm's , overheating, or any other of these BS fanboy excuses.

In fact the same reviewer that blew up his gtx590 THEN used the right driver and the card ran fine.

QUOTE!
"The Swiss team tried again using the 267.71 drivers, which has the overcurrent protection built in and the GTX590 didn't fry itself."

The mistake that Nvidia made was packing the older drivers in with the gtx590 to some reviewers.

Bottom line?
Always use the latest drivers when reviewing a brand new card and for Nvidia, don't send the card to reviewer with the wrong drivers!

I now look forward to some real overclocking reviews like the one in the video I linked above.

edit: I found a good one..
Gtx590 overclocked vs 6990 overclocked using a 2600k @ 5.0 ghz.

http://www.microsofttranslator.com/B...the-titans/8

http://www.microsofttranslator.com/B...-frecventa-cpu

edit 2 : now back to my vacation. see ya around.

hi don't wanna be rude but you don't know how the driver works or how its implemented.
The voltage table is inprinted in the bios of the card where the driver reads the voltage of a table. Same way it reads the performance level of the table. A junk driver alone can't make it to go poof unless the voltage tables in the bios was wrong. Doesn't matter if its 1966 drivers. It knows nothing about the hardware. It reads it off the table which is in the bios.
So who added the table in the bios?
Or
Who modded the bios?
If the card was a Asus one then blame them. They are so stupid by zapping the pointer to the performance table by setting it to 0.

 

[bryanW1995]

Intel and Nvidia both have serious issues in the same year. AMD must have Secret Agents or something. :hmm:

 

[bryanW1995]

Did you ask me if I were for real and bullshitting afterwards?

Sincerely why?

You 're taking an extreme stance that a lucky few manage to overclock their cards. And that it is a non issue that many cards blow if you try to overclock them.

All other cards in history where you rolled down the over clock when you hit the your limit instead of throwing it in the garbage can? They haven't existed?

Nvidia s not agree with you they told sweclockers that the card should be able to take the overclock ann explicitly asked them to proceed so with the second card.

You'r right in that it is a non issue for the seller since warranty is no more. But that's a very narrow view of it.

You picked the wrong dude to pick a fight with. Here's a list of BFG10k's recent articles at ABT:

http://alienbabeltech.com/main/amd-radeon-6000-series-image-quality-analysis
http://alienbabeltech.com/main/amd-6850-bottleneck-investigation
http://alienbabeltech.com/main/amd-radeon-6850-performance-test
http://alienbabeltech.com/main/call-of-duty-black-ops-review
http://alienbabeltech.com/main/nvidia-260-99-driver-test
http://alienbabeltech.com/main/nvidia-400-series-image-quality-analysis

Why don't you link a few of your recent gpu review articles on a reputable hardware site?

 

[bryanW1995]

This is entirely true. Only English-speaking people are allowed to pronounce upon exploding video cards. Every knows that speaking foreign languages seriously detracts from your ability to test video cards.

Sorry, it's because nvidia and, now ati/AMD are American companies. Translations to swedish PLUS the metric system conversions are simply too complicated to prevent card explosions willy-nilly all over the place. Hopefully in the future both companies will realize this and limit testing to reviewers in english speaking countries (and preferably North America only).

 

[themodernlife]

5970 was great, they were made to OC and most managed to reach and exceed 5870 speeds. no problems with overheating or dying cards.
Again, 6990 are the same, most are able to be manually OC beyond 6970 speeds and dont die from overheating or component failures. Perhaps there is a reason why these cards and their PCBs are so long and robust.

If these cards keep on dying even with the new fixed drivers, it will be a huge problem.

Yeah my 5970 overclocks to 1000mhz/1300mhz on stock cooling (i believe the sapphire has a 1ghz bios lock on core iirc). Never hits 100c ever even after 4 hrs crysis. I've gamed on it everyday for the last year at high resolution with aa enabled pushing it as hard as i can with a 5ghz intel chip.

I think it's a cooling issue. How do they go from 580gtx cooler celebration to a disaster with the 590? I bet there is alot of yelling going on behind closed doors atm. . .

 

[bryanW1995]

something tells me w1zzard just got taken off jhh's xmas list... lol... EPIC FAIL!

 

[LiuKangBakinPie]

Nvidia ->The way it meant to explode

 

[TerabyteX]

How did I miss that, Ive seen that page before and never noticed!

lmao it spells out Epic Fail.

Scary cryptic message, subliminal!!

 

[tincart]

Sorry, it's because nvidia and, now ati/AMD are American companies. Translations to swedish PLUS the metric system conversions are simply too complicated to prevent card explosions willy-nilly all over the place. Hopefully in the future both companies will realize this and limit testing to reviewers in english speaking countries (and preferably North America only).

"What the hell is wrong with this graphics card? It only renders 3D graphics in english!"

 

[cusideabelincoln]

^ 0_0 at both... why can't AMD and nVidia focus on power efficiency and not raw power? you'd think that they would due to the fact that PC games are going to be going nowhere for the next few years graphics-wise. (talking about console ports, of course.)

Well, they are, but they are stuck at 40nm. As a result of being stuck at 40nm and having pressure put on them by the gaming community, something had to give. We weren't magically going to get substantially better performance without increasing power draw on the same process.

But what we got wasn't too bad. Nvidia increased their efficiency top to bottom with the 500 series over the 400 series. AMD did the same with most of their lineup. The 6800 series provides better performance/watt than the 5800 series. The 6950 is also pretty efficient, and has a better performance/watt than the 5800 series. These latest dual cards, though, are just for bragging rights.

 

[Tempered81]

nvidia ->the way it meant to explode

rofl !!!!

 

[Grooveriding]

Nvidia ->The way it meant to explode

 

[SlowSpyder]

Testing? I doubt they had much chance for it. Clearly they did not expect 6990 to scale so well, so they thought that 2x downclocked 580s on 1 pcb will be good enough to beat 6990... Remember the 2 day delay due to drivers? I bet that's because they never designed the 590 to go above 500MHz and so, they had to do a 6 months worth of catching up in 2 weeks, to push it to semi competitive range of 6xxMHz... The problem with 590 is exactly the same problem as is with 580, it is too power hungry.

590 can be a great card if:
1. You clock it at 580 speeds
2. You give it 580 voltages
3. You install much much better power circuitry (This is what is failing now)
4. You install a cooler capable of 600+ watts of thermal capacity
5. You give it 700W of electrical power.

I have to think Nvidia built the card and waited for AMD to release the 6990 to see it's clocks and performance before pushing out the GTX590. They had to clock the 590 at the very top of the spectrum to have it compete. In the end it used too much power for the board and ended not being tested thoroughly enough.

This is all just speculation on my part.

 

[TerabyteX]

Well, your speculation might be closer to be accurate than you think. The GTX 590 can't be overclocked properly, which means that the card is pretty much maxed, and only matches the HD 6990 at stock, and the HD 6990 overclocks reasonably well. Not spectacularly but 960MHz/1390MHz is nothing to sneeze at.

 

[96Firebird]

Since when is overclocking guaranteed? And what is "overclocked properly"?

 

[Barfo]

Lol @ fermi based cards in general. ALWAYS something goes wrong with them. First compatibility issues and now this.

wat? the 460 has been one of the greatest value cards in recent times, to give just one example of successful fermi cards

 

[AnandThenMan]

Since when is overclocking guaranteed? And what is "overclocked properly"?

Being able to raise voltages at least 10% without having the hardware self destruct would be a starting point. And no overclocking is never guaranteed, but unless you are doing some extreme stuff, it's rare to have outright hardware failures.

Just to be clear, anyone with the GTX590, Nvidia is telling you to only use stock voltage:

We recommend anyone using the GTX 590 board with the reference aircooler stick with the default voltage while overclocking

Pretty weak stuff from what is supposed to be an ultra enthusiast card.

 

[bryanW1995]

I have to think Nvidia built the card and waited for AMD to release the 6990 to see it's clocks and performance before pushing out the GTX590. They had to clock the 590 at the very top of the spectrum to have it compete. In the end it used too much power for the board and ended not being tested thoroughly enough.

This is all just speculation on my part.

i don't think so, the whole "drivers cooking the card" fiasco reeks of a rush job to me. NV knew that amd had a head start on the sandwich card and they just pushed as hard as they could to get the card out as quickly as possible. I'm sure that any later production runs of the gtx 590 will have the bugs fixed, but if I was in the market for this type of card I'd definitely avoid the first batch even if I was an NV-only buyer.

wat? the 460 has been one of the greatest value cards in recent times, to give just one example of successful fermi cards

yeah, it's probably the best card from either camp since 8800gt. It gave NV a huge win, and was possibly even more important for them than any card since 8800gtx b/c it halted amd's momentum and changed the conversation about fermi.

 

[tincart]

yeah, it's probably the best card from either camp since 8800gt. It gave NV a huge win, and was possibly even more important for them than any card since 8800gtx b/c it halted amd's momentum and changed the conversation about fermi.

More aggressive pricing on the 6870's would have probably cut that victory short. I don't think AMD pushed the 68xx nearly enough.

 

[LiuKangBakinPie]

Lets get back to reality here. Ok so it exploded but it still traded blows with the 6990 afterwards.

 

[bryanW1995]

they had supplier issues or they probably might have. however, at that time nvidia was under more pressure to grab immediate market share than amd was, so there's no guarantee that amd would have dropped their prices regardless.

 

[Tempered81]

Lets get back to reality here. Ok so it exploded but it still traded blows with the 6990 afterwards.

hahahah

 

[cusideabelincoln]

More aggressive pricing on the 6870's would have probably cut that victory short. I don't think AMD pushed the 68xx nearly enough.

They could have pushed the 5850 a bit harder before the 6800 came out. Was it supply issues? I don't know, but it seems logical enough AMD would have dropped the price on the 5850 if it felt any pressure from the 460. Maybe in retrospect they would change their tactic, if they had a choice at the time.

 

[Tempered81]

From Nvidia:

Note on GeForce GTX 590 Overcurrent Protection and Overclocking

Question

Note on GeForce GTX 590 Overcurrent Protection and Overclocking


Answer

In the web release driver of GeForce GTX 590, we have added some important enhancements to our overcurrent protection for overclocking. We recommend anyone doing overclocking or running stress apps to always use the latest web driver to get the fullest protection for your hardware. Please note that overcurrent protection does not eliminate the risks of overclocking, and hardware damage is possible, particularly when overvoltaging. We recommend anyone using the GTX 590 board with the reference aircooler stick with the default voltage while overclocking, and avoid working around overcurrent protection mechanisms for stress applications. This will help maintain GTX 590's great combination of acoustics, performance, and reliability. NVIDIA has worked with several watercooling companies to develop waterblocks for GTX 590, and these solutions will help provide additional margin for overclocking, but even in this case we recommend enthusiasts stay within 12.5-25mV of the default voltage in order to minimize risk.

These are guidelines only - any overclocking/overvoltaging can void your manufacturer's product warranty.

http://nvidia.custhelp.com/cgi-bin/nvidia.cfg/php/enduser/std_adp.php?p_faqid=2947

 

[ArchAngel777]

Are you for real? Whether it’s 100% or 0% that fail overclocked, it’s still not the card’s fault.

The card is rated to operate at the volts and clocks it ships from the factory. Anyone that goes above those is taking a risk. If the card blows up as a result it’s not “faulty”, it’s the individual at fault for operating it incorrectly outside of its rated parameters.

The only way the card would be faulty is if it failed at factory voltages/clocks, but that doesn’t seem to be happening here.

Don’t confuse the overclocking hysteria on the internet with something that is mandatory and guaranteed from any piece of hardware, because it isn’t. Just because little Timmy can overclock something it doesn’t mean you can, should expect it to, or is “faulty” because it can’t.

I can't say I agree with you BFG. If these things are burning up with a slight clock speed increase, then chances are, even at stock these things are close to frying (which appears to be true). I understand your view point and I agree with it only to a point, that one should not execpt a successful overclock. But, having a card blow up and fail, that is taking it to the extreme. I understand that you don't overclock anything (CPU or GPU) and I respect that, but a LOT of enthusiests do and the MFGs know this. If the 6990 does not blow up, why should the 590 blow up when overclocking?

Hey, I already know you won't agree, but I at least want to state my view point.