OCCT 3.1.0 shows HD4870/4890 design flaw - they can't handle the new GPU test !

[Zucker2k]

This is to eliminate the pci-e providing inadequate power. PCI-E specifications require that cards pull 80w max from the pcie iirc. So the argument the card tries to pull more power from pci-e resulting in an ocp and crash is not valid. Any card that resorts to pci-e for more power than specified is operating out of pci-e spec and deserves what it gets = shut down (that is IF the pci-e does trigger the shutdown at all).

This also means that those who overvolt their pci-e to pass tests are running their cards out of specification.

This test should only be run with stock everything to eliminate any other variables being the culprit.

Also those who continue to beat about the bush, please get in line: The argument here is this: Certain cards are limited to drawing no more than 82.x watts. These are the cards the OP is addressing. He suggests this may be a serious design flaw. In an age where profiles are lock and step with game performance, who knows what else these profiles are doing with the power delivery in gaming sitiuations and subsequent impact on overall fps output.

[Andrew LB]

Man... what a tiresome read. If some of my fellow XS members had actually been polite enough to simply help the OP out with his theory... or just not chime in at all... this thread would be about 1/4 its size and FAR more productive.

I actually am glad that I read through this thread since i'm currently in the market for a new high end video card and truly don't know which one I'm going to end up buying. So I'd like to thank the OP for spending so much time and effort writing this program as well as the time he has spent trying to discover whether there is a defect in some of these cards designs or if it was an issue with his own coding. The OP definitely approached the problem he experienced with objectivity and instead of just making accusations against ATi, asked for those of us on XS with the card to either confirm or deny the validity of his findings.

As for the thread crappers. You guys did absolutely nothing to aid in finding ths solutions requested by the OP. Next time you guys get the urge to put on your fanboy hats and act like children, please refrain from adding your thoughts in future discussions. Thx.

OK.....I see. So the vgpu amperage, at like 82A at the 1.26V is only like 103W of power. DUH! My stupidity.

Don't feel so bad, I did the exact same thing while reading through this thread.

Would be nice if op could provide more evidence

If you read the OP, you'd see that he doesn't have unlimited funds and hardware.... and is ASKING NICELY of all of us on XS to help him by testing this software on OUR video cards to either confirm or deny his theory.

Do you have a card you can test for him? The more people contribute to the raw data, the better.

Thus far we've got about a dozen cards tested and a pattern is starting to show itself with the power regulation. I was definitely skeptical when I began this read, but the OP's theory is looking like it just may be valid.

[LordEC911]

Get Out!

http://en.wikipedia.org/wiki/Power_virus "A power virus is a malicious computer program that executes specific machine code in order to reach the maximum CPU power dissipation (thermal energy output for the central processing unit). Computer cooling apparatus are designed to dissipate power up to the Thermal Design Power, rather than maximum power, and a power virus could cause the system to overheat; if it does not have logic to stop the processor, this may cause permanent damage."

OCCT is not malicious code.
The wiki entry even makes a distinction between a power virus and a stability program in that a stability program is directly under user control. And OCCT is.
It has not been proven that this causes the GPU to overheat, rather what people suspect is that OCP is being reached, or degradation of an inductor has occurred shutting down the graphics card. Thereby heat is not the main issue.
^^ this is the hypothesis under question though more broadly it is why this has ocurred.

You have not contributed a single piece of data, nor have you augmented the hypothesis and analyzed the results. Your only conclusion is that this is caused by temperature(because you cited power virus), which is wrong.

You're useless, and blatantly lying so you're absolutely worthless. Destroying you is fun.

You might want to actually re-read the thread and also rethink what this app is...
I also never said anything about temperature in any of my posts...
People if you are going to argue, at least argue with what I am saying, not about stuff I didn't say.

[Praxis1452]

You might want to actually re-read the thread and also rethink what this app is...
I also never said anything about temperature in any of my posts...
People if you are going to argue, at least argue with what I am saying, not about stuff I didn't say.

I gave you the definition of the word you used; it includes temperature.

Fail at dodging.

[LordEC911]

I gave you the definition of the word you used; it includes temperature.

Fail at dodging.

Unfortunately you don't seem to keen to read my previous post throughly...
I mentioned that supposedly, due to the app being a power virus, that it overloads the card which was the OPs conclusion. The side effect of temp didn't seem to be the problem, though with the findings on this board, it cannot be overlooked.

[cowie]

power virus? is that more like swine flu or the y2k bug.

[Bo_Fox]

Bravo Tetedeiench I've grown quite fond of your OCCT. Don't let the fanboys and shills shout you down.



I've been told that GDDR5 stuff is power hungry, could be your 512MB card draws less than a 1024MB reference design card and doesn't trigger OCP.



Or a manufacturing problem. Nonetheless /agree wholeheartedly



Nah, here is the sort of thing that would deter me from buying Radeons:

"Your limited testing is flawed. If the VDDC's overheat they will shut down the card to protect it. If you dont want to have a crash when testing artificial means such as this stability benchmark then increase the fan speed to cool the VDDC's. Obviously underclocking acomplishes a similar goal as they dont get as hot. No games are affected just artificial stability tests."

I was considering a 4890 however I won't abide this kind of BS and censorship. I certainly won't encourage such with my hard earned $.

This has been a very revealing thread. I've learned a lot about individuals that compose the XS community

Yes, let's just all start blocking LordEC911's posts. Those people just dont change, so once you see an immature post on a long thread, just block the poster and it should shorten the thread considerably!! It could just be a temporary block, and then unblocked afterwards..

I'm so grateful for such a honorable programmer who has created OCCT. Exploiting such a controversial issue (which is officially made "controversial" by the deleting and banning that is done on AMD forums) is always a good thing. Heck, it could actually explain some of the crashes in older "simple" games, like EVE, as experienced by Shintai, and others..

I'm surprised that Nvidia cards can handle it just fine. Seems that the shaders arent as demanding on NV cards. If it were a problem with NV cards instead of ATI cards, then Charlie over at the INQ would be praising about it for the next 2 months with 10-page drunkely-slurred articles!

[zerazax]

Maybe when the VRM's hit the amperage limit the card tries to increase power draw from the PCIe resulting in an over-current condition and shuts down the port, that still puts the alleged fault with the card not the board. Overvolting the PCIe bus might get around this on some systems.

Try using the three data points makes a line in two dimensional space approach to formulate a test:

3 3-phase cards from 3 vendors
3 4-phase cards from 3 vendors
3 different X58 boards
3 different Dragon boards
3 different P45 DDR3 boards
3 different sets of DDR3 RAM
3 different OSes Win-7, Vista, XP
3 different DX9 builds

Test each card on each platform with each set of RAM with a fresh install of each OS and each DX9 build. If the 3-Phase cards all fail and all the 4-Phase cards all pass in all configurations. Then you can be fairly sure that the 3-Phase cards are incompatible with your test.

Then the question becomes 'is it your test software, DirectX or the cards that is at fault?'

If you rewrite the test from scratch in OpenGL or DX10(.1) will the fault still occur?

Does data from GPU-z corroborate what you have found with RivaTuner?

Does data from multi-meters corroborate with software monitoring apps?

I'm not saying that your findings are wrong, I doubt that they are but convincing ATi and their board partners that 3-Phase power is not enough for will probably take the level of scientific rigour I've proposed above.

Even then is it relevant if there are no commercial (let alone AAA) games that are capable of stressing the hardware to this level while retaining playable framerates?

BRAVO! This is how science is done!

The rest of the fanboy fest in this thread is so full of utter fail

I can't even begin to fathom how any logical conclusion can be made without a scientific approach that eliminates the possible variables as initialised wrote.

[Nullack]

Unfortunately you don't seem to keen to read my previous post throughly...
I mentioned that supposedly, due to the app being a power virus, that it overloads the card which was the OPs conclusion. The side effect of temp didn't seem to be the problem, though with the findings on this board, it cannot be overlooked.

No offense, but honestly your arguments are irrational.

The software in question does not meet the definition of a power virus. It is a stress test, which is a totally different class of software.

Your attempts at providing red herrings into the debate about performance is again illogical.

You also cling to straw man arguments where you misrepresent other's arguments in an attempt by you to "defeat" them.

If you care so much about this and arent just another fanboy who's in denial, I honestly suggest you read up on how to construct a logically rational argument.

[Warboy]

Maybe when the VRM's hit the amperage limit the card tries to increase power draw from the PCIe resulting in an over-current condition and shuts down the port, that still puts the alleged fault with the card not the board. Overvolting the PCIe bus might get around this on some systems.

This is where your post fails. It would draw more power from the rails. Not the slot.

[Sparky]

You can tell by look at the backside of the card. If the marked caps are missing it's 3 phase, if they are there it's 4 phase.

Thanks. Mine is 3 phase.

So this begs the question, why does my reference 4870 have no problem, while others do

[initialised]

Would be nice if op could provide more evidence

He so far tested 2(?) HD 4870 cards and one of them was defective (?)

All this without any serious pics or any screenshots of the tests etc...besides the rivatuner shot of HD 48xx series.

See my previous post!

Multiple HD 4870 owners in this thread tried his test and their cards passed.

I'm sure I have a few 4870s, 4850's and 4890s I can try the test on. All our 4890s are reference and there are odd compatibility issues with some 4850s and 790FX DDR2 and P43 boards. I have a feeling that this is a similar problem given the number of 4870s that pass. Or it could just be bad code!

[Warboy]

It's "Vitec" not "Vtec" and that's just an inductor.
A passive component. It's a coil of copper wire. It cannot fail in any to contribute to this issue.

It was a joke, Google it.

[initialised]

This is to eliminate the pci-e providing inadequate power. PCI-E specifications require that cards pull 80w max from the pcie iirc. [snip]

[/snip]
Also those who continue to beat about the bush, please get in line: The argument here is this: Certain cards are limited to drawing no more than 82.x watts.

Ergo the cards that are failing are trying to pull more than 82.xW from the PCIe bus!

[Bo_Fox]

Any hardware can fail with special software...

How obnoxious of a statement..

It's like saying that any system memory (DDR sticks) can fail with special software like Memtest86 or Linpack or Prime95 or OCCT or SuperPi, or anything you could think of, which is a poor statement after all...

And then after you underclock your memory, voila it works! You upgrade your motherboard with one that has an 8-phase PWM circuitry, and you find that you do not have to underclock your memory for it to work. So, you scratch your head, pondering about this issue for the rest of your life until you become bald, because you refuse to allow yourself to understand the issue, like the rest of us can.

Many of the guys here can get OCCT to work by just lowering the memory clock, and not the GPU clock. I mean, the memory is not supposed to take that much power, but GDDR5 is known to be power-hungry, and not supplying adequate VRM's (or at least allowing them to handle the current) is probably going to cause this issue in a game or two. As Shintai brillantly clarified, it could have been the very issue causing crashes with EVE Online. Some of the older games use very simple shader calculations, and could eat insane power even when the menu screen is running.

Recently, I was playing NOLF games at 4x4 supersampling AA, and it heated my card hotter than ever. It actually made the screen artifact just a tiny little bit after an hour or so of running @ max overclocked speeds. I know that Nvidia does not officially support 4x4 mode, but it ran along without crashing at even hotter temperatures than Furmark. (No, I'm not a NV fanboi at all.. before any of you guys assume me to be one.)

Bottom line: I'm grateful that there's such an application testing tool that can help us to know which cards are of better quality than others. Thank you, Tetedeiench!!!

[Praxis1452]

Unfortunately you don't seem to keen to read my previous post throughly...
I mentioned that supposedly, due to the app being a power virus, that it overloads the card which was the OPs conclusion. The side effect of temp didn't seem to be the problem, though with the findings on this board, it cannot be overlooked.

Alright I'll pretend you didn't write what you wrote.

[STaRGaZeR]

BRAVO! This is how science is done!

The rest of the fanboy fest in this thread is so full of utter fail

I can't even begin to fathom how any logical conclusion can be made without a scientific approach that eliminates the possible variables as initialised wrote.

See below, you'll see a pure demonstration of that "science" you're so proud of:

Ergo the cards that are failing are trying to pull more than 82.xW from the PCIe bus!

82.xW what? PCIe what? Seriously WTF are you talking about? The cards crap out when the core wants to suck more than ~82 Amperes (current), and the PWM can't provide that current. The best possible explanation is OCP until someone confirms it with measurements.

[LordEC911]

No offense, but honestly your arguments are irrational.

The software in question does not meet the definition of a power virus. It is a stress test, which is a totally different class of software.

Your attempts at providing red herrings into the debate about performance is again illogical.

You also cling to straw man arguments where you misrepresent other's arguments in an attempt by you to "defeat" them.

If you care so much about this and arent just another fanboy who's in denial, I honestly suggest you read up on how to construct a logically rational argument.

Pray tell how are my posts irrational?
So, this program is not a power virus since it is a benchmark/stresstest?
Does it do everything a power virus does?

All I care about is the lack of detective work done by certain individuals.
I guess people don't/can't understand a different POV.

Until there is substantial proof of an actual problem, I don't know why people are agreeing with a hypothesis...

[Praxis1452]

Pray tell how are my posts irrational?
So, this program is not a power virus since it is a benchmark/stresstest?
Does it do everything a power virus does?

All I care about is the lack of detective work done by certain individuals.
I guess people don't/can't understand a different POV.

Until there is substantial proof of an actual problem, I don't know why people are agreeing with a hypothesis...

Get Out!

http://en.wikipedia.org/wiki/Power_virus "A power virus is a malicious computer program that executes specific machine code in order to reach the maximum CPU power dissipation (thermal energy output for the central processing unit). Computer cooling apparatus are designed to dissipate power up to the Thermal Design Power, rather than maximum power, and a power virus could cause the system to overheat; if it does not have logic to stop the processor, this may cause permanent damage."

OCCT is not malicious code.
The wiki entry even makes a distinction between a power virus and a stability program in that a stability program is directly under user control. And OCCT is.
It has not been proven that this causes the GPU to overheat, rather what people suspect is that OCP is being reached, or degradation of an inductor has occurred shutting down the graphics card. Thereby heat is not the main issue.
^^ this is the hypothesis under question though more broadly it is why this has ocurred.

You have not contributed a single piece of data, nor have you augmented the hypothesis and analyzed the results. Your only conclusion is that this is caused by temperature(because you cited power virus), which is wrong.

You're useless, and blatantly lying so you're absolutely worthless. Destroying you is fun.

You are now officially a bigot. And you complain about others not reading your writing.

I not only defined power virus and cited the source, I proved that OCCT does not meet the criteria, and yet you keep pushing the point.

EPIC FAIL

[AMDDeathstar]

Shouldn't the GTX 285 setup be kicking my butt in frames per sec.?
my average was 77
his was 53

How long do I have to run it for it crash?
I only let run for an hour

For the one who requested it :

GTX285, stock frequencies, Shader Complexity 3, 1680x1050 FullScreen Mode :


Sorry for the french header, forgot to remove it.

had about 53FPS average.

[kl0012]

Just for statistic: Sapphire 4850 reference design (1920x1200) - no problems there. 44 fps average.

[Bo_Fox]

Pray tell how are my posts irrational?
So, this program is not a power virus since it is a benchmark/stresstest?
Does it do everything a power virus does?

All I care about is the lack of detective work done by certain individuals.
I guess people don't/can't understand a different POV.

Until there is substantial proof of an actual problem, I don't know why people are agreeing with a hypothesis...

LordEC, just drop this power virus thing. It's only harming your reputation here on XS forums.

[LiquidReactor]

So what have we found so far? Just a few bad cards that cant seem to handle 80A+ meaning they're probably drawing 50% more watts then they're rated for?!?!

Very impressive.

[Bo_Fox]

So what have we found so far? Just a few bad cards that cant seem to handle 80A+ meaning they're probably drawing 50% more watts then they're rated for?!?!

Very impressive.

Not "few", but better put as "several"...

[LiquidReactor]

Not "few", but better put as "several"...

Well you can add my Visiontek HD 4870x2 to that list.

My card switched off/lost feed to the monitor as soon as I launched the app. Used the settings that op suggested; highest native resolution, error checker off and shader optimal 3. Considering the fact that this new occt version didnt even run for a sec and switched my card off right away I'd say its most likely the code that is bad.

Seriosly op maybe you should recheck all the algorithms and your coding once more. Seems like something in your coding causes the ati drivers to go berserk and shut down the card.