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

[largon]

LiquidReactor,
Underclock your card and convince yourself the problem is plugged into a PCIe x16 slot.

So, this program is not a power virus since it is a benchmark/stresstest?
Does it do everything a power virus does?

Looks like by the definition of the word it's not a power virus since the app is not malicious and is user controlled. But sure, it acts like one.

TBH, I did not know before this thread the term "power virus" refers to a malicious app. I've encountered the term to be used for unusually higher power draw inducing apps in the past...

[Tetedeiench]

I won't enter the debate "power virus or not". Again, it is not constructive. My goal is not to fry anything. Nor to damage anyone's reputation.

For anyone to know, my goal is objectivity. Even if, as you all guys do, i do have my likings in terms of brands, i am NOT a fanboy. I had a 9700Pro for years and was really happy with it. Currently, i have a GTX285 and i'm happy with it.

My previous CPU was a A64 3800+ and i was really happy with it. Now i do own a Core i7 and i love it. So on. Again, it this was Nvidia at stake, this thread would be all the same.

That being said, about the performance of the shader code : yes, indeed, the fact i'm getting 52fps instead of 80fps on the Nvidia cards is weird. I actually never noticed that. I am going to investigate this, and try to see if i can make the test run better on both cards. That's indeed the next step. and as soon as i achieve that, i'm going to release it.

Now, i'm too far in the release process to not release it. The test is heating Nvidia cards all the same. The FPS-es are less, yes, the the temps are still going up the roof, and beating Furmark's temps on my computer. So this is still an improvement, and i don't see that as an issue preventing the release, which i planned today.

However, rest assured that i'm going to investigate as to WHY this is happening. and quickly.

I'm going to respond to liquidreactor now :

Please, read all the tests that have been done. They prove that the card hardware are at stake here, from all we could gather. Truly. I can't convince you, and don't want to, and if you don't want to believe what i'm saying, you'll never believe it. Your card is affected by what i'm reporting, and i'm sorry it is. We must have reached the 20 cards affected by now. maybe more. not 2.

And Yes, i do not have a 4870 to test with. Why is that ? I don't have fundings. I don't have contacts with AMD/ATI (they basically ignore my emails asking for testing hardware). And i don't see myself investing into testing hardware, where OCCT's donations and advertisement covers hosting expenses, and... hosting expenses. I love OCCT, i work on it for free, but i don't see myself investing into hardware on my personal funds for it to grow, just because i need testing hardware. That's the limit i've always put. Some companies have always been supporting me when i needed it, some don't. ATI/AMD is in the "don't" category. So i'm relying on the good will of my users. That's why you see threads such as this one, asking for everyone's kindness.

Would you believe me if i told you i spent about 3 to 4 days sending exe files to people, so that they could launch them on their HD4870 and tell me what's happening ? That's the only way i've got to debug the thing.

Anyway, let's try to isolate the issue. i'm trying to get more accurate readings done, and to get a website with testing means doing research on the thing, because i'm convinced we touched something there.

[Shadov]

Is it just me or is someone looking for attention?

This is yet another program that puts unrealistic load on the components and is clearly misleading some of the audience.

Move along... Nothing to see here.

[Nightstar]

Thanks. Mine is 3 phase.

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

Good question Sparky. If I might venture a guess it would be that your 512MB
Reference card doesn't draw as much current as the 1024MB cards. Furthermore If I'm not mistaken you're excellent cooling(water?) reduces the current draw of your card.

[Shintai]

Is it just me or is someone looking for attention?

This is yet another program that puts unrealistic load on the components and is clearly misleading some of the audience.

Move along... Nothing to see here.

Why is it unrealistic? Because nVidia/AMD havent made a driver with enough idle states in a profile for it?

[Shintai]

Tetedeiench:

I dont know if you can use this for your bughunting. In EvE Online after an expansion this year with changed shadercode. Alot of AMD/nVidia owners had cards that went through the roof with temps and classified as overheating. Screens turning black or blinking black (Like every 15-60secs. And all was related to some changed shadercode in that expansion.

Enabling of Intermidiate one aka v-sync fixed it for most people until the shadercode was fixed again.

So its certainly not just your app or furmark.

[kromosto]

Is it just me or is someone looking for attention?

This is yet another program that puts unrealistic load on the components and is clearly misleading some of the audience.

Move along... Nothing to see here.

are you kidding.

if you buy a card it should run at any load at stock speeds. this load maybe realistic or unrealistic it is not important. company has to guarantee to to this.

and who will define unrealistic load. what is unrealistic load? will chip company define it, game makers, benchmark companies, users, who?

Tetedeiench is a programmer and making a tool and the load he is making with his code is realistic for him but he cant use this load with some cards. now how can you say it is unrealistic because no games uses that much load? this is your way of looking this maybe unrealistic for you but for others it may be realistic.

again if you buy a card it should run at any load at stock speeds. otherwise intel, amd, nvidia, etc. etc. can say you cant use our cpu - gpu with this application because the load it is making is unrealistic

[largon]

I'm convinced this is a vDDC OCP that causes the blackout.
I just measured on my HD4890 that vDDC drops to 1.14v from 1.31v at the moment of the crash. That's one of the typical ways OCP works.

Why is it unrealistic? Because nVidia/AMD havent made a driver with enough idle states in a profile for it?

How is it not unrealistic, as saying otherwise equals to saying HD4870 is ~60% more powerful that GTX285, as HD4870 manages 60% higher FPS in this test.

are you kidding.

if you buy a card it should run at any load at stock speeds.
(...)

Don't worry. Future catalyst update will fix this. By inserting those idle cycles.

[xixou]

are you kidding.

if you buy a card it should run at any load at stock speeds. this load maybe realistic or unrealistic it is not important. company has to guarantee to to this.

and who will define unrealistic load. what is unrealistic load? will chip company define it, game makers, benchmark companies, users, who?

Tetedeiench is a programmer and making a tool and the load he is making with his code is realistic for him but he cant use this load with some cards. now how can you say it is unrealistic because no games uses that much load? this is your way of looking this maybe unrealistic for you but for others it may be realistic.

again if you buy a card it should run at any load at stock speeds. otherwise intel, amd, nvidia, etc. etc. can say you cant use our cpu - gpu with this application because the load it is making is unrealistic

I totally agree, it seems that the margin taken on the ATI cards was too low,
it should exceed at least 10% of the stock frequency !

[Olivon]

Thx for your soft tetedeiench

You've made some greats improvements and it's a must have now !

Respect !

[largon]

http://ht4u.net/reviews/2009/leistun...mes/index4.php

OCCT GPU test goes ~50W higher that Furmark.
Is OCCT GPU test realistic?

[Nightstar]

I won't enter the debate "power virus or not".

But I will.

So, this program is not a power virus since it is a benchmark/stresstest?
Does it do everything a power virus does?

A virus must self replicate, even in the context of computer malware. Does OCCT self replicate? No? Ok then we can agree, it's not a power virus.

While we're at it it's not a trojan either, those are either inhabitants of Troy, condoms or malware which provides clandestine access ala the Trojan horse of legend. I think we can all agree OCCT doesn't protect your willy.

Did I miss some form of malware with which you will label this stress testing tool next?

and in other news

Pray tell how are my posts irrational?

The answer was in the post, you bidder of post re-reading. You'r posts are irrational due to your fallacious arguments.

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.

I'm sorry for going OT I know I'm not supposed to feed the trolls. I rcognize that I have a impulse control problem and I'll go sit in the corner now.

[Shintai]

I'm convinced this is a vDDC OCP that causes the blackout.
I just measured on my HD4890 that vDDC drops to 1.14v from 1.31v at the moment of the crash. That's one of the typical ways OCP works.
How is it not unrealistic, as saying otherwise equals to saying HD4870 is ~60% more powerful that GTX285, as HD4870 manages 60% higher FPS in this test.
Don't worry. Future catalyst update will fix this. By inserting those idle cycles.

Yet it happend in an MMO with over 300000 subscribers

[Solus Corvus]

Look at it this way, nVidia cards just can't process these shaders fast enough to push them to the limit.

*runs away from angry mob*

I'm convinced this is a vDDC OCP that causes the blackout.
I just measured on my HD4890 that vDDC drops to 1.14v from 1.31v at the moment of the crash. That's one of the typical ways OCP works.

So how do we defeat the OCP on a 4890?

[largon]

Yet it happend in an MMO with over 300000 subscribers

I'm sure something happened in EVE online. Though, there's no reason at all to take for granted these incidents are related and their mechanism would be the same.

Look at it this way, nVidia cards just can't process these shaders fast enough to push them to the limit.

So how do we defeat the OCP on a 4890?

It's possible.
But I personally will not do it on my card as the mod will only gain stability in this test, it will not gain anything else. Well, maybe Furmark too...
But it will cause even higher VRM temps and a VRM failure is pretty much guaranteed as current draw is around double that of nominal, even with the OCP.

[Shintai]

I'm sure something happened in EVE online. Though, there's no reason at all to take for granted these incidents are related and their mechanism would be the same.

The was symptoms surely the same

[Bo_Fox]

Is it just me or is someone looking for attention?

This is yet another program that puts unrealistic load on the components and is clearly misleading some of the audience.

Move along... Nothing to see here.

I added that thread-crapper to my ignore list.


We Xtreme Enthusiasts should always support ingenious programmers of freeware tools that are valuable pieces of software allowing us to further test or tweak our hardware that allows us to be Xtreme enthusiasts in the first place. If it allows us to find video cards that are of better QUALITY than other cards of a different brand or design, then we should be grateful.

Like my Dad said, if you cannot say somehting nice, then do not say anything at all.

In the end, it's nice to be constructive, for the greatest number, for the greatest good, for the truthful essence of itself. You might think that you are doing the constructive thing by criticizing it, believing it to be a useless piece of software, but in reality, it is actually more constructive to allow it to be useful to Xtreme enthusiasts rather than try to destructively bury it into oblivion. Therefore, criticizing it with an overall result of constructiveness is, in all of its honesty, false. We are endowed with a gift of knowing right and wrong, for we have already eaten from the tree of knowledge of good and evil.

Just kidding with all of the philosophical religious stuff!!!

[Bo_Fox]

But I will.



A virus must self replicate, even in the context of computer malware. Does OCCT self replicate? No? Ok then we can agree, it's not a power virus.

While we're at it it's not a trojan either, those are either inhabitants of Troy, condoms or malware which provides clandestine access ala the Trojan horse of legend. I think we can all agree OCCT doesn't protect your willy.

Did I miss some form of malware with which you will label this stress testing tool next?

and in other news



The answer was in the post, you bidder of post re-reading. You'r posts are irrational due to your fallacious arguments.




I'm sorry for going OT I know I'm not supposed to feed the trolls. I rcognize that I have a impulse control problem and I'll go sit in the corner now.

LOL!!! Trojan condoms!!!

[Tetedeiench]

Tetedeiench:

I dont know if you can use this for your bughunting. In EvE Online after an expansion this year with changed shadercode. Alot of AMD/nVidia owners had cards that went through the roof with temps and classified as overheating. Screens turning black or blinking black (Like every 15-60secs. And all was related to some changed shadercode in that expansion.

Enabling of Intermidiate one aka v-sync fixed it for most people until the shadercode was fixed again.

So its certainly not just your app or furmark.

V-sync, on ATI cards, doesn't reduce the load. Every frame is still calculated, just not sent to the screen. I was surprised at first, but you give it some thought, it's normal

[Bo_Fox]

V-sync, on ATI cards, doesn't reduce the load. Every frame is still calculated, just not sent to the screen. I was surprised at first, but you give it some thought, it's normal

Hmm, that's new for me.. I thought that enabling Vsync always lowered the temps for me.. when i disabled vsync, the temps seemed to rise more. Are you 100% positive on this?

[Solus Corvus]

It's possible.
But I personally will not do it on my card as the mod will only gain stability in this test, it will not gain anything else. Well, maybe Furmark too...
But it will cause even higher VRM temps and a VRM failure is pretty much guaranteed as current draw is around double that of nominal, even with the OCP.

For one thing, I wouldn't do it with stock cooling.

That seems to be something that a lot of people in this thread are forgetting. They say that this kind of behavior would be unacceptable from a CPU or RAM. But I personally have experience with several cases where high stress programs can crash certain CPUs at stock speeds with the stock coolers. Same thing with RAM - load up all four slots and you can find some memory that will overheat without additional cooling. It has also been the case where some hardware, including certain cpus, can fail even when given 100% valid instructions because of design errors. Those of you that don't think it has happened repeatedly in computer history haven't either don't have enough experience with computers or are willfully deluding yourselves.

Some other people seem to just be using this thread to bash AMD or to bash Tetedeiench, depending on their angle. You guys should just ignore this thread. You aren't being constructive and are just cluttering up the thread for the rest of us that just want to figure out what's going on.

And lastly, it's not as bad as some of you are making it out to be. For the 99% of users that don't use OCCT the 3 phase design is fine. For the rest of us there are 4 phase cards, OCP mods, luck, and/or...nVidia cards.

[largon]

The was symptoms surely the same

Considering the Eve online problem did not cause a hardlock, caused the picture blink every 15-60 seconds and manifested on both nV and ATi HW, I don't see a connection nor any similarity in symptoms. What OCCT causes is a state that's most likely a hardwired protection mode. It causes a hardlock, it does not cause blinking screen and it does not affect both vendors.

btw,
Got a link to more info on this Eve issue?

[Shintai]

V-sync, on ATI cards, doesn't reduce the load. Every frame is still calculated, just not sent to the screen. I was surprised at first, but you give it some thought, it's normal

When set in the game engine I am sure it only makes the GFX card do x frames with interval one. Other 2 settings are interval default and interval immidiate.

My first post also needs correction since the fix was interval one in the game engine that is equal to v-sync type of behaviour.

The fun part here was it was equal for nVidia and AMD cards. Both "overheating" and having black screen blinks.

Without interval one you could sit with 350-1300FPS in stations where the issue was most severe.

[Tetedeiench]

Hmm, that's new for me.. I thought that enabling Vsync always lowered the temps for me.. when i disabled vsync, the temps seemed to rise more. Are you 100% positive on this?

I'm using DXUT, where VSync is enabled by default.

i'm disabling it by using D3DPRESENT_INTERVAL_IMMEDIATE.

I tried that on ATI cards, crippled by the bug. Temps did not lower at all. It was a crossfire, mind you, but still. Same behaviour.

So yes, all the frames are still calculated, just not sent to the screen. I asked somebody from Hardware.fr, who is writing their reviews, and he confirmed this.

So yes, i'm positive about this. This is true for crossfire configurations, i don't know about Single-GPU ATI cards though.

[LordEC911]

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

How so?
Where does it say that OCCT doesn't meet the criteria of a power virus?
Simply, in this case, instead of excessive heat causing the problem it is excessive power draw, supposedly.

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

Again, how so?
I guess people are interpreting the definition of a power virus differently...
If defending my POV and opinion is harming my rep, so be it.

Looks like by the definition of the word it's not a power virus since the app is not malicious and is user controlled. But sure, it acts like one.

TBH, I did not know before this thread the term "power virus" refers to a malicious app. I've encountered the term to be used for unusually higher power draw inducing apps in the past...

I think people are getting too hung up on the word malicious.
I also agree with the last part, for that is what I am referring to.

So you use the app, that is the same as a power virus, except that you simply have the ability to abort the app, if needed. It still is a power virus at the root.

A virus must self replicate, even in the context of computer malware. Does OCCT self replicate? No? Ok then we can agree, it's not a power virus.

While we're at it it's not a trojan either, those are either inhabitants of Troy, condoms or malware which provides clandestine access ala the Trojan horse of legend. I think we can all agree OCCT doesn't protect your willy.

Did I miss some form of malware with which you will label this stress testing tool next? u bidder of post re-reading. You'r posts are irrational due to your fallacious arguments.

I'm sorry for going OT I know I'm not supposed to feed the trolls. I rcognize that I have a impulse control problem and I'll go sit in the corner now.

I'm sorry you think I am a trolll...
I also didn't realize that trolls take the time to gather results to post and try to discuss things in a civil manner. I will do my best to stop these actions I previously mentioned...
Again people are focusing too much on the individual words and not the whole. Please see above.