[Conference Call] Nvidia GeForce GTX 580M 78C° Throttling [SEE POST #191 and later]

If your having thermal throttling use my guide, it fixes the issue and will work until Dell provides a fix.
http://forum.notebookreview.com/ali...x-r3-580m-thermal-throttling.html#post8345887

 

Yeah my vbios fix plus Yoda's temporary software solution eliminates all throttling. Unfortunately, it is not a perfect world. Yoda's fix isn't the most elegant and neither is mine. We still need a real fix from Dell.

 

Forgot to add, my model's fan profile seems to be fine. i see max fans around 65 depending on load but they have always been humming away at max for some time before it hit the wall.

 

Yeah but Yoda's fan profile overrides the EC allowing your temps to go over 78 without throttle.

 

I've done both and can't thank you both enough for workarounds that makes my machine usable. I still do get the top throttle but i have to have the card clocked around 720 and not prop it up so not the show stopper that it was.

Now I'm just fascinated by what the problem is. my hope is that a BIOS fix will be produced by Dell and all the delays and other strangeness is simply waiting for their programmers.

 

I doubt those are max fans, the max EC profile fans and actual max fan are a big leap. Max fan sounds like a fighter jet is taking off using your desk as a runway. I have never been able to get mine to gain heat at all running at true max fan, and even 3500 and 3900 RPM (which i set in my guide to be in the 60's) are faster than the built-in profiles fastest speed.

If you can't manually set your fan higher than it normally goes then it might be a defective fan. Try manually setting it to 5000, it should max out between 4300 and 4600 RPM which should be noticeably faster than it normally ever goes.

 

I need to go back and check i did it all as i still see the 78 one

Sounds right, I do remember the MX18 I had for a short period turning heads when the fans really kicked in. I'm checking out your link (that I missed above!) now. thanks again.

 

Hello,

I’m from Spain and yesterday I received my new laptop R3 with this graphic card.
I have the throttling problem…but I have one friend who received his laptop yesterday with the same configuration and he haven´t any problem with it!! And his max temps are 69°, how it´s possible?? Where’s the problem in my laptop??...the thermal compound??...i don´t understand…

(sorry for my english!)

 

Could be the paste job on his system was done better then yours. If you can, try redoing the paste and see if that helps. Also, make sure you guys are running it in the same environment (not sure if you are or not). If his house is cooler than your house, his might not reach the same temps. Also make sure you're running the same benchmarks or games.

 

I guess we will be battling two problems here. Some people will have a poor performing 580m that will reach 78C easily while just playing most games. These people will have a bad GPU, poor thermal compound, or problems with their thermal pads, etc.

The second group is those who have a properly performing 580m that will tend to pass this thermal threshold after playing a particularly demanding game for a couple of hours. I hope that Dell does not focus in on those with actual physical problems and assume that anyone breaking the 78C barrier has some sort of problem that can be solved with replacement parts and a technician visit.

Ultimately I am hoping for a more aggressive cooling profile and to be allowed to run the card up to 85C or so. I think over a couple of years of ownership the GPU requirements for average games will creep up where my card is being pushed pretty hard all the time. I don't want it arbitrarily throttling itself when it doesn't need to.

 

^ THIS. Some people will inevitably have some hardware problems, but PLEASE Dell, do not focus on this. Not everyone has these physical hardware issues, but we are still getting throttled. Replacing our hardware will not keep it from throttling at 78C. Now for the people that are having hardware issues, I do hope that you get your fixed and working properly, but even once you get a good paste/GPU, you will still throttle at 78C, no matter what.

 

Yes some people, but the rate of failure is really bad and they have to be looking at a link. Maybe the throttle most of us are looking to fix (78C) is protecting some other component that is failing? I, like many others am on my second R3 in as many months. The first throttled so bad far cry 2 almost halted after only 15 mins or so play. I never investigated it since when I called dell they just said 'we will collect it and refund you'. I gave it a second try and this one is the more 'normal' throttling that is fixed somewhat (I still haven't got HWINFO fully stable, but I'm sure I'll get it!) with the fixes listed here.

 

Update:

We just finished up a meeting on this after a few weeks of back and forth in IM and email. I now know more about this issue than I ever wanted to, but at least I have a good understanding of 78C.

We are still talking about how to word the message, but it's looking more like a conference call would be a better idea for reasons I'm not going to get into. (Sorry, but I can't.) So, if you guys want to get started, please delegate someone to pull a group of you guys together, come up with a time and date, and an agenda. The agenda should be easy: "explain 78C to us." Once you have someone named to string it together, I'll get with that person in email and work out the details.

 

I nominate widezu69 and YodaGoneMad, as these guys have been amazing with coming up with ways to deal with the issue so far. They really know their stuff when it comes to this. I could get together if we come up with a time to do a call, if even just to be one more voice confirming what we've talked about. I don't know as much as these masters of the throttle, but I do know computers, and I have experienced the 78 degree Celsius throttle.

Sorry to put you all on the spot We love you widezu and Yoda

 

I would be willing to. I am from the UK though so the time differences would be a bit of an issue. I can be bios guy and Yoda can be EC guy.

To all the community, the issue of the 78C throttle is being dealt with at the moment which is great. However some people face the power throttle around the 60's. Because not everyone has this issue, we don't know how easy it will be to diagnose and fix. I dunno what can be done about it. The fix is in my sig and it is a permanent fix with little to no risk. I wonder what Dell would do about this. Perhaps once the EC is rewritten, both types of throttle will disappear.

Edit: @Bill, I can help with a write up on what exactly is going on. I've been doing this for a while now and I can confidently say that I can present the issues as simply as possible.

 

Please get the 78C throttle temp removed guys! It is a big pain, especially for guys from hot climates.

widezu69 and Dell-Bill_B -- try your best guys! Get this throttle removed.

 

@widezu69 Do you have a way of reading the GPU card temp- not the core, but the card? Also, do you have a way to read wattage draw of the MXM slot? Just trying to bounce some ideas around internally, and it'd be nice to know that info.

 

No sorry not some expert or anything so I don't have access to necessary tools to measure internal temperatures. Sorry. The only reason we came to the conclusion that the 60ish degree throttle was power related is due to the fact that it occurred to a few on the default vbios and to every single person using an overvolted one unless they performed a fix for it. I for example never had the 60ish throttle when on stock, until I moved to an overvolted vbios, then the early throttle happened. Hence the reason I came up with a work around for both default and overvolted vbios users. The 78C throttle does happen to all owners though so that should be priority.

 

No worries. I was going after 78C there.

 

Great so I'm always around if you need me to help. You can PM me if you want further contact details if you wish for me to speak more directly to your colleges.

 

Searching around I'm finding a LOT of old threads on other forums about this but this 'fix' is a new one on me. Anyone tried it? Disable GeForce GTX 580 Power Throttling using GPU-Z | techPowerUp
There is a power consumption graph...
A guy with a M17x said it fixed it for him although I think this might be aimed at the desktop gpu

 

Dude, different type of GPU completely. Plus different architecture as well. Different throttling origin too. Won't work. Good research though keep at it you might stumble over something mindblowing

 

That's actually a GTX 580, the desktop card (not the 580 M)

 

Yes but the posts are interesting since the same problem is being discussed. Check out the thread below the download, saying that Nvidia introduced this technology with the 580 due to it's high power draw. Also the page I linked from was in a forum with a guy saying it fixed his M17x so i wondered if anyone had tried it?

 

Seems interesting though. But anyways we don't want anymore software fixes. We want permanent implementations from Dell.

 

Do we know for sure the problem was quickly fixed in the M18x? That has a much beefier power supply...

Totally agree, anything I find is aimed at background for a fix. . I'm not too optimistic one will be forthcoming but it sure won't be for lack of trying on the part of you guys here...

 

Forgive me if I am wrong but I have been following this throttle issue with the GTX580M graphics card for a while now. As far as I remember, a group of members here did once have a conference call with Dell before and the conclusion was that the throttle issue is due to the hardware limitations of the graphics card it self. Am I remembering things incorrectly or do I have the felling that this conference call will have the same results as the one before? I'm sorry if this is a negative post but why are we having an other conference call when the previous did not solve anything?

 

If there was a previous call and it didn't solve anything, shouldn't we try again and get something done? Why give up? And no it is not due to hardware limitations because the same GPU in any other machine - M18x, MSI or Clevo all are able to go above 80C without throttle.

 

I was just about to say the same I do think it's a power issue though and we are struggling with a cludge fix that dell originally put in to cover for the cards high power draw. We know that the fixes here cure it but I'm thinking that just removing the throttle will increase the overall failure rate, something dell will want to avoid.

 

Yeah but there can still be a throttle. If Dell sets it at like 87C then that would be fine for almost everyone.

 

It would and there is no obvious reason not to just do that except if by that time the power draw was overloading something else - like the power supply? If they start tripping out, or a risk of failure then dell might be liable to provide upgraded power supplies as the existing one being 'not fit for purpose'.Or some other M17x specific component. I'm looking for a max power draw comparison against the AMD ppl are running happily at 80C-90C.

 

You are absolutely right, we should try again and get something done about it. As I said before, I don't remember the exact details of the last conference call. All I can remember is the fact that Dell came back with some technical details in regards of the throttle issue and the guys that did have the conference call seemed very understanding with it and everything got forgotten a little while after. The details of that conference call should still be somewhere within this forum but I don't know where. All I am saying is that Dell know the exact reason to why the 78 degrees throttle is happening and once again want to have an other conference call to explain it to the members of this forum. What good is that going to be in resolving this issue? This all is like a big dejavu for me. I am once more sorry if this is not what you guys want to hear. I very much hope that I am wrong and that this issue gets resolved as soon as possible. I am just trying to shed some light in to this issue.

Regards.

 

I have to agree, Dell know exactly where the problem is and now they seem to be hunting for some reason to tell everyone without exposing them to any cost - although all this bad pub must hurt sales. Also hope I'm wrong but it does look that way.

 

I don't know if it's fair to postulate possible causes of the throttle and then criticize Dell for it. We don't yet know why it throttles at 78C. Dell is having meetings, discussing the issue, conferencing with us, and formulating a plan. One of the reasons I went with Alienware was because I saw and read that some of their technical folks hang out here and see that we are taken care of. So far they seem to be doing that.

 

Their sidestepping of the situation is costing them money. Assuming they put this limitation in place to cover their backsides because they don't want to keep replacing overheated cards, then why do they insist that replacing perfectly working cards will solve the throttling? Maybe that is just the Indians on the phone that don't really know much of anything.

Refusing to acknowledge and fix the problem will lose them customers as many forum members have ditched the M17x R3 for other machines because there is throttling in the 580m.

If we present all of these facts to Dell, it should be something for them to think about. They need to know that their strategy to avoid losing money is in fact losing them money.

 

Hi again everyone. I done some searching and found the thread where a conference call took place with Dell. It was actually on the M18x part of the Alienware forum, however it is intact for the GTX580M throttle issue. It might be a long read but here you go have a look at it your self http://forum.notebookreview.com/alienware-m18x/616150-dell-580m-throttle-contact-group-18.html

 

Ah but the Dell 580m throttle contact group was successful. Their 78C throttle was successfully removed.

 

At the time I did not own an Alienware. I was following every thread for the M18x as I wanted to buy one. After the conference call and the conclusion of it I stopped reading the M18x threads. It was some time back and all I could remember was that a conference call had previously take place regarding the throttle issue. Now after all that time reading about the same issue and seeing that dell is pretending of not knowing the cause of the problem and wants to have an other conference call to discuss things made me a little angry. Anyway, I hope that the conference call will be constructive and we all put this enoying issue behind us. Thanks.

 

Exactly! My only thought on their reasoning is it is still cheaper to replace cards on the people that complain (and hope they go away) rather than admit a design fault with a subsequent recall.
I also think that the 'good' dell guys here are stuck in the middle of a potentially disastrous PR. Onwards and upwards and good luck with a conference call (tape it!!!).

 

Honestly, there is no reason to throttle the 580m, I have no idea why or who thought it was necessary, but it isn't, at all. There has been some speculation the last few pages and I just want to clear it up a bit.

There are only a couple possible reasons to throttle. First, the higher temps might hurt the other components. This isn't the reason since they don't throttle the 6990m, which often runs in the 80's C and will go all the way to 99C thermal cut-off without problems.

Second, they are worried about some combination of temps and power draw. This isn't it either because the 6990m has the SAME 100 watt TDP and is allowed to run full speed as hot as it wants with no throttle.

Third, the higher temps could damage the 580m. This isn't it, because 580m cards are present in many other laptops and are allowed to run as hot as they want (and average a max temp of 85C, almost 10C higher than our throttle, in other gaming laptops). Obviously we don't have internal data, but just going on customer forums it looks like far more M17x owners kill their cards trying to get around the throttles or have their cards replaced to get rid of the throttles than any actual card failures on all the other MSI, Sager, etc gaming laptops that run the 580m with no throttle. I think we can pretty safely say that the 580m does not fail at temps of 85C, as there have been no reported problems with other manufactures. In fact, we have a pretty strong case that the throttle is costing Dell a lot of money, since it is causing people to try risky fixes or get a replacement card when they don't need one.

Put it all together and it is obvious, the 580m should not be throttled. Period. The card won't die at higher temps as proven by other manufactures, the M17x won't die at higher temps as proven by the 6990m, and power draw is not the problem as proven by the 6990m.

Depending on when this needs to happen I could do it. I feel like I have a pretty firm understanding of the entire issue so I could talk you guys through it.

 

^ THIS THIS AND THIS. End of discussion. Thank you Yoda, this is the best summary I've ever seen. Lets not get the thread off track suggesting what might be wrong. We know what's wrong. Simple as that. I hope the conference call goes well and they will fix us the way they did the M18x, because has Yoda summarized, there is no reason for us to throttle at this low of a temperature.

 

This is Dell pretending of not knowing the cause:

Simply put: it throttles for a reason. You are missing a lot of variables to the equation. It's not "x=78, then throttle, where x is GPU temp." It's more like "if x=78, then y=Z and A>58, then throttle" (not literally). Solve for y, Z and A. Just saying, you're missing 3 variables. It has zero to do with warranty costs. We shall discuss.

Read back through my posts here over the years, and you'll see there are few times when I don't spell things out in plain English. It's complicated beyond the hardware and the specs.

 

Thanks Bill. That is what I am seeing, it depends on GPU load as well as temp plus some other unknown variable. I also see that the problem is happening to 580 desktop GPU's, that thread (linked above) states that Nvidia introduced this 'feature' with the 580 series. maybe with Dell's power you could get Nvidia involved? They refuse to deal with us 'm' owners as 'a manufacture issue'.
Basically though all everyone seems to be seeking is the 'hard' 78C limit that always happens (although to get to that temp the GPU is always at 90% +). I'm thinking the other manufactures have implemented it differently to the reference design?

 

@Micky, it has nothing to do with the GPU. If you take the exact same GPU from a M17x and put it in a M18x or MSI or Clevo machine, you will not get throttling. The problem isn't in the video card as was the case with the Desktop 580, the problem is in the EC on the motherboard.

 

What I meant is that it is a poorly implemented GPU feature. Nvidia designed in this and it does not work properly on the M17x. Knowing the triggers will help Dell identify what they did wrong with the EC? Same result on some desktops though 78C drop to P8...

just for information the GPUZ switch does work on the 'm' in my M17x but I wouldn't advise it since it still does not address the fan mapping.

 

Good to know thanks.

You make a very good point about the triggers. My knowledge is much too limited to be able to comment on anything like that. I can only do what I can.

 

100% not Nvidia induced. Dell induced all the way. It's not what we would characterize as an "EC problem." The EC is functioning exactly the way we want it to. Once we get the facts on the table on the call, we can all move onto exploring what, if anything considering our design specs, can be changed. We shall discuss.

Who's my official point person? Should I check my PM box?

 

Hopefully at some point you can divulge some details, because I would be really interested in the actual decision process taking place. As I see it there aren't a lot of things those variables can be (fan speed, MXM slot draw, GPU load, GPU temp, that is my guess at the 4 variables), and whatever they are they must not matter very much since the 78C throttle is incredibly consistent. In other words, if the card hits 78C all the other variables are likely max, so basically the only active variable, or the one that matters, seems to be temperature.

It also still doesn't explain why the 580m gets a complex 4 variable throttling profile and the 6990m has nothing at all.

 

Yoda is correct. Whatever the variable, the only one that matters is the temp because regardless of the power draw or anything else, when the GPU hits 78C, it throttles, period.

PS I'm happy to be your official guy @Bill.

 

I sent him a PM and basically said it was you or me or perhaps both of us depending on when they want to do the call.