M17X R3 + GTX 580M - Random Lag's - Help

ambient temperature : 25ºC

 

So ... More than anything I can do?

Sent from my GT-I9000

 

You can try to repaste if you feel comfortable with it. Hopefully someone else will come in here with a link to a repaste guide or further information on it. You can find videos on youtube by searching M17x R3 repaste. I would recommend arctic cooling paste, and either the MX-2 or MX-4 paste should do perfectly fine. They are non-conductive so no worries about shorting anything, just put a thick line of goop all the way across the GPU die (the part under the heatsink, be sure the watch a video and understand what you are doing) and when you clamp down the heatsink it will spread itself perfectly and any extra will just get squeezed out but wont hurt anything.

Hopefully the Dell reps I PMed will get back to us with a response, maybe they will be fixing this issue with a bios update. If they decide not to fix it you might want to consider calling Dell and swapping to a 6990m since they have no throttling at all. If you cannot keep your temperatures down there is no current method to get rid of the 77c throttle.

 

Here's a good guide to repasting both GPU and CPU in the R3, with pictures, prepared by our friend DaneGRClose. http://forum.notebookreview.com/alienware-m17x/561599-m17x-r3-cpu-gpu-re-pasting-guide-w-pics.html

 

My computer is coming tomorrow. I'm worried the throttling issue will hit me at 78 C as well. Any word on potential fixes for the secondary throttle yet?

 

Wow just come across this thread. I am waiting on sales calling me today about a 3d ready m17x with 580m which i thought was the cream of the crop card. Judging by this thread i should just ditch that and go with a 6990?

 

Nothing yet, I assume the Dell reps have read my message, hopefully they are getting it investigated and will give us some kind of response soon. You guys can also all feel free to PM them, Mr. Fox posted the usernames of the Dell reps a few pages back. I would think the more people who bring it up and make it an issue the more motivated they will be to fix it.

Hard to say, you lose 3D with the 6990m, and it is possible that Dell will do the right thing and fix the problems. However, personally I wish I had gone with the 6990m so that I could have a solid card and a solid machine without the problems of the 580m. If Dell will actually resolve these issues then I think the 580m is better, just for the more frequent and generally better driver updates and perhaps a tiny bit more performance. If these issues are not resolved then the 6990m is probably better, you can actually OC it and use it without getting throttling that kills your performance.

 

I spent over 2 hours on the phone with Dell today, was transferred no less than 5 times, including once early this morning (around 8:45pm last night CST) to what was supposed to be the final person I needed to talk to (which of course they put me on hold for 20 minutes and then told me they were closed and to call back during normal business hours. Normal business hours end at 9pm, imagine that).

So 13 hours later I call again and finally get the right person who tells me that no, I can't change the video card in my order because the computer that I ordered 4 days ago (Saturday) is already "In Production" and that I can cancel the order or hope it works fine.

Needless to say I canceled the order and it will be a cold day in Hades before I ever consider another Dell product.

Good luck with your computers folks...........I'll be having one built by Xotic PC or another company.

 

I can certainly sympathize, the whole getting transferred around thing is annoying, I have found it easier just to find out who I should be calling then hang up and call them, as half the time transfers go to the wrong place.

However, you want to really consider your decision, right now no one else really offers a comparable product to the M17x. One that is reasonably light, gets good battery life, has good cooling, and is available with a world class warranty. Not to mention with the discounts you can get Alienware is by far the best value gaming laptop.

If it was me I would call in and explain how you are angry about your experience and they will likely expedite your order and give you next business day shipping. They might even throw in an extra year of warranty or something. With an expedited order you would probably get your newly ordered computer with 6990m before the one you cancelled. They really only had 1 day to work on your computer (Monday) so re-ordering today you don't lose hardly any time, and with the upgrades you could get out of this you actually come out ahead.

 

wow, quite a storm I stirred up with my post on page one . People shouldn't worry this much. However it would be nice if some 580m owners would come in and defend their card a little

I'm not going to cancel my order because of some issues some peope are having. Not to sound like I'm ignoring people who do experience issues, but when you start counting the number of people complaining it's actually quite low. Even if you add other websites. Only when I experience real problems I will send it back for the hd6990 which should give 99-101% of the performance . I've read on a different forum from multiple people with the 580m they had no problems whatsoever.

 

Well when you get it, check how it works. If there is a throttle at quite an early stage, around 65Cish, do the fix in my sig. I love the 580m. Shame that my replacement is nowhere near as fast

 

I followed the tips on the topic throttling, I thought the problem was gone, but he continues, only changed the time to appear ... Putting VBIOS 580m92v.ROM, any different from normal? The only fix would be repaste?

 

92v would just make it worse, since it would run hotter. Basically, what you did was to eliminate the 63c throttle but you are still hitting the 77c throttle. As you said all it did was delay it.

Elevate your laptop and try a repaste, hopefully Dell will fix it, or you might want to consider a 6990m if you continue to have issues.

 

could also try HWinfo and run the fans at full speed. This seems to drop both my CPU and GPU temps a bit. Not sure how much it will help with the 580m though.

 

I will try but the noise is too high

 

Gave your advice a try, they said "sorry but that's life" basically. No big deal, I realize they're a big business and they are out to make a profit, they just won't make it off of my dime due to lousy customer service. Hopefully they fix the known issues with the GTX580m and the M17 and you guys that got them can use them as intended.

 

It worked when I increased the fan, the temp was at 72 ° ... But the noise is too high .. If I make the exchange of thermal paste, I will have success?

Sent from my GT-I9000

 

Proper application of high-end thermal paste (MX-4, PK-1, etc.) will improve temperatures, but fans are an integral part of the cooling system, regardless of the paste used. As loads increase, heat is generated and the fans increase (or should increase) speed in response to facilitate heat transfer and dispersion. Fans are your friend. They are crucial to the performance and longevity of high powered cards like the 580M and 6990M. A little noise is a necessary evil, although the fans in my R3 are not really objectionable even when pushed by HWINFO64 to full speed.

 

Hi all. We went through this with M18x, and the symptoms are pretty much the same. That issue was fixed. I will escalate this to the appropriate engineers for review, but I need to tick off some boxes first.

Please confirm if a re-paste changes the throttle points.
Has anyone had a heatsink replaced and still see throttling at the aforementioned temps?
Please confirm whether you see throttling in games. Synthetic benchmark results are immaterial for the most part to the conversation at this point, so please focus on performance in normal usage and gaming scenarios.

 

glad to see you guys are looking into it.

 

Thanks for taking an interest and responding! You give me hope that this issue will be fixed.

A re-paste does not change the 77°C throttle, in fact, nothing at all changes that throttle point. If at any point the GPU hits 77°C then it will throttle to the P8 state which is not usable for games. I have not tested the lower 63°C as much (since it can be bypassed with modification), but I am pretty sure nothing changes it either, once the GPU passes 63°C the card throttles down to P1.

I have a brand new machine with 580m, just a few weeks old, and mine is unusable for gaming because of the 77°C throttle. It has the newest heat sink and has been re-pasted. I had the tech out to replace the card and heat sink but he bungled it and broke everything, either way it was the same heat sink I already had that he was trying to put on. I will be getting a new exchange system probably next week, and can report if I get the same level of problems with it.

It is basically impossible not to hit the 63°C throttle. Any normal game will hit it, the fans don't even kick up very high by that point to try and keep it below 63°C. This means without modification the card is running something like 30% slower all the time because it always hit the 63°C throttle and uses the P1 state.

I can also hit the 77°C in any modern games just by not propping up the back of my laptop. Just using the laptop normally on a flat table it will hit the 77°C after maybe 10 minutes in any modern game I own (and my ambient temp at home is around 20°C). If I prop the laptop up in the back for better airflow then less intense games won't hit 77°C, but something like Battlefield 3 will still easily trigger it. Ambient temperature is a huge factor, other members can give their input about this, but I know several guys on this forum in hotter areas, and anything over like 26°C in Ambient temperature will result in hitting the 77°C throttle no matter what on every game even with re-paste and raising the laptop.

What is really annoying about the 77°C is it can be unpredictable. You might play fine for a while but your temps are hovering in the 70's°C, then something stresses the card, it hits 77°C and suddenly your game is running at 2 frames per second and totally unplayable for a long time while the card cools. Nothing is more frustrating than having it hit while flying a jet in BF3 or something similar, by the time the throttle is over you have been dead a long time. It is also annoying that I don't feel like I can just pull my laptop out, set it on a table, and game. If I do that I will throttle every time.

 

^ THIS. I haven't repasted mine yet, although I probably will just because I like having the best performance possible. However, the average Joe probably doesn't know or want to repaste, and shouldn't have to to use their product the way it's intended. I am very glad Dell has noticed this thread and decided to take a look into it. Looking forward to updates. Thanks!

 

So am I better off ordering mine with the 6990???

 

Thanks, Yoda. Exactly what I needed. I have sent the issue over. I think they will do intake some time next week, so we'll know more towards next weekend hopefully.

 

It's up to you. I still stand by NVIDIA. I've always used their products and I think they are better innovators, with both hardware and software. At least now we know that Dell is looking into it, and hopefully we'll be able to use our GTX580M beasts to their full potential.

 

It is your choice, before Bill responded I would have said yes for sure, but now that Dell is going to look into this, and hopefully provide the same fix as they did for the M18x my attitude is not nearly as clear cut. I really like Optimus in general (it is better than the whole switch graphics wait 15 seconds things), I like Nvidia drivers better, and technically with the throttle removed or raised to higher temps then the 580m is the faster card.

Since Dell is looking into this I would probably say get the 580m, if for whatever reason Dell chooses not to fix this issue then you should be able to get a 6990m through tech support and an in-home tech. That is what they have told me about my exchange system, which they insisted on building with a 580m, but said if I got it and still had problems they would send me a 6990m to swap out.

Good deal, thanks again for getting behind this issue. The dream of us M17x 580m owners is to get the same fix as the M18x, this card with the throttle removed or raised would be amazing.

 

@Yoda I have also pm'd Bill just to get at least more than one user involved. I've been working on the 580m GTX throttle for a few months now, longer than any other user here and have a lot of experience. I've given him a low down of what's wrong and what we want. I'm sure you have as well so lets hope this gets fixed

 

Hi all ,

I'm having issues for high temp even in idle 56-59° and 69° on game .
I can't say if i have throttling issues , but temp are very high and the performance are low .

The GTX580M can work at full speed with only 0,800v instead of 0,870 , with lower heat?

Thank you

 

Thanks Guys, I started reading through this thread and it started to worry me. I have the 580 in my PowerPro and I love it so I will stick with it and hope Dell does the right thing.

 

Great thank you, hope that both throttling points (66°c & 77°c) can be resolved! a bit sad to have such a powerful card limited at such low temps!

 

Good deal, I have been trying to encourage everyone with these problems to say something so that they understand it is a systemic issue, not just a few bad machines.

Full speed requires .87, what the 63c throttle is doing is dropping it to .85v and lowering clock speeds upwards of 30%. If you are always over 63c while gaming then your card is always running in slow mode. Try the vBios fix in Widezu's signature, it should let you at least get good performance, the problem will be if you start hitting the 77c throttle. Really though, your temps don't seem that high to me, these cards are fine up to 100c, the only issue is the Dell imposed throttles, which is what you are running into.

While gaming my card is almost always over 70c, and I have done quite a bit to try and cool it, as I said in my earlier post it will often just hit a stressful area, burst up to 77c and bam the game is unplayable. Hence why we are pushing to get these throttles removed, they are low enough to effect even normal gaming at stock speeds which is the problem.

 

Thankyou i'll try it.

 

Update: The issue is entering quality engineering intake as we speak. We will investigate and determine root cause of the concerns. If it does turn out to be a Dell issue, we will fix it for sure. If it's an Nvidia problem, we will work with them toward a fix and let you know as much as we can.

Federico may or may not reach out to a few of you to request capture units in exchange for new units.

A couple of things from you guys could help out in the meantime.
1. Please post pictures of the GPU heatsink installation on stock units. In other words, we would like detailed photos of factory GPU installations on systems in the field, if at all possible.
2. If you have re-pasted your heatsinks, post back and let us know if it had any affect on throttling at 77C. (I'm assuming many of you have re-pasted to no improvement, but let's make it official in posts I can link to, please.)

Not to worry, if there is a fix, we'll find it, and working with you guys here has proven to be the quickest way to get things done.

 

Thanks Bill, I'll definitely get to work on getting these things for you later today, is others haven't already taken care of it.

I can tell you though, re-pasting doesn't affect the throttle. It does lower temperatures, which helps keep us from hitting the throttle mark by having lower temps, but it still hits the throttle around 77/78 degrees either way. So re-pasting does keep temperatures cooler, but it doesn't remove the throttle.

 

Thanks for the update.

I will be getting a brand new exchange system with 580m, hopefully sometime this week, and will be happy to post a picture of the GPU installation on it right out of the box. That way you can see exactly what customers are getting straight from the factory.

As blink said, repasting can help to lower temperatures, but if at any point you hit 77c then it will throttle, all that seems to matter for the high end throttle is temperature, we have tried just about everything to get it to go away.

 

I'm sure they have/will think of this, but I feel it is a VBIOS/BIOS issue. Since we can remove the lower temp throttle through flashing a simple VBIOS update, I would assume that a new fixed up VBIOS and/or BIOS update would remove the throttle. It could be more complicated than that, but that's my thoughts on it. Just something to pass along.

 

pau1ow from the MSI forums installed a Dell 580M in his MSI machine. He reported no signs of throttling whatsoever, therefore it has to be a Dell Mobo/BIOS issue. Nothing to do with Nvidia

 

There is also the fact that the exact same 580m cards in the M18x no longer throttle since the A04 system bios update. The M17x has the exact same system bios GPU throttling that the M18x had.

 

The problem will then be, the new A09 bios downclocks the SATA ports from III to II. If Dell release bios A10, sure the throttle may be fixed but what about SATA speeds?

 

Now you're just being cantankerous.

 

I honestly had to look that word up

I know I'm being disagreeing but seeing as all the bioses before A09 had SATA III functionality why did A09 have only SATA II and will it be in all future bioses? It'll suck for many people who want to stop their 580m's from downclocking but then their SSDs are running half speed.

Doesn't make sense to me. Just my $0.02

 

haha. I was mostly kidding about posting off topic. Totally joking, though. They have a SATA II/III thread open on the issue now. We are currently unable to replicate the issue with Dell validated SSD's.

 

Oh I see so the A09 allows SATA III on certain SSDs I see. So I presume the Samsung 830 OEM version which is what ships with current laptops I think.

Still though, It looks like we are headed the right way for the M17x R3. Hopefully bios A10/A11 should have everything fixed. I'd love to see proper SATA III for all drives and full speed 580m GTX.

 

Another request- tell us your ambient temperatures when you hit 78C.

 

I don't hit 78C because I'm a student and I live in a stone cold flat. Temperatures less than 20C in this house. Less than 15C on a bad day I did hit it when I was in a friends house and the ambients were over 21C. Don't know precisely how hot but a warm comfortable house will cause the GPU to hit 78C without fail.

I'm also repasted with IC Diamond.

 

If you need a test case my current laptop has this issue (and I already have a new exchange unit coming, so you will be getting this unit back anyway). I have a Dell PM830 installed from the factory. On A08 it is running at 6Gb/s, if I flash to A09, not touching or removing the drive, it runs at 3Gb/s, flash back to A08 and 6Gb/s again.

Another issue that you might want to look into is that those of us buying the Dell solid state drive right now are getting the really nice PM830 6Gb/s drives in our new units, but our warranty is listing the PM810 3Gb/s drive as the replacement. I have not really tried to get this corrected, but another user was having a heck of a time because his drive failed and tech support was pretty much refusing to replace his drive with the same one. It is kind of a big deal for us to buy the $400 Dell PM830, but if it breaks have it replaced with a $200 PM810 that is half the speed. I only purchased the Dell SSD because I knew it was a PM830 and felt it was the worth the money.

I don't know if that is your department, but it is probably something that should be corrected. I am hoping my SSD doesn't fail, but if it ever did I would be extremely upset if I was forced to settle with a drive that is half the speed of my original.

EDIT: Also I can confirm they actually do send out PM810's for replacement. I sent this unit into the Repair Depot and they thought my SSD was bad, they sent it back with my SSD installed and a note that I would get a new one in a few days. Turns out they had set bios to AHCI which is why it wasn't working, so my SSD was still fine. They had already shipped out the replacement though, and it is indeed a Samsung PM810 3Gb/s versus my original drive which is a Samsung PM830 6Gb/s.

 

Is that covered in the SATA II/III thread somewhere? I've been looking for this kind of confirmation.

 

It is mostly being talked about here: http://forum.notebookreview.com/alienware-m17x/628584-alienware-m17x-r3-a09-bios-discussion.html

You will have to hunt for posts, but here is one I found pretty quick, there are a ton of people who have confirmed this behavior. There is also the whole issue of only the middle port running at full 6Gb/s, you will see some talk about that in the thread as well.

Another user that confirms the behavior: http://forum.notebookreview.com/ali...7x-r3-a09-bios-discussion-15.html#post8273423

Any thoughts on the whole warranty issue of replacing PM830 drives with PM810 drives? Is that something you can bring up to get looked at or is there another department or person I should contact. I figure while I actually have the ear of someone at Dell I should try to get it resolved, if I ever actually need an SSD replacement I won't be satisfied with getting a PM810 to replace my PM830.

 

Some M18x owners have experienced the same problem, but not all of us have.

 

I just got my R3 with a 580m in it on Monday. I too can confirm both of these issues.

I have not re-pasted, and I installed a Crucial M4 64GB HDD. The GPU throttles at 63 degrees Celsius and then again at 77 degrees Celsius. This is without an overclock or anything of that nature. I fixed the 63 degree C throttle by flashing the Vbios found here on the boards.

To be fair, I've only been able to ramp it up to 77 degrees C in FURMARK, but it is winter here and my house is cold(20 Celsius). I would hate to hit a wall this summer, that would be bad.

As for the SATA speeds, I was wigging out because my M4 was only getting SATA II speeds. Thanks to the lovely people here I saw it was an issue with A09. I downgraded to A08 from the dell website and magic, SATA III again. So there is an issue with the A09 bios. My SSD is in port 1.