PDA

View Full Version : turns out my g75 was defective :(



kiba
07-07-2012, 11:13 AM
so i posted this a while ago...

http://rog.asus.com/forum/showthread.php?18047-G75VW-really-strange-behaviour

and ive been having similar problems ever since. yesterday, i almost hulked out and ripped it in half when it started freezing up while booting both normal and safe mode, it had been freezing up in windows for a while now.

so after i heard a loud pop come from under the keyboard, i took it back to micro center where i got it, and listed just anout every problem that i know of w/ the machine (and there are many; gpu throttling problem, audio driver issue, etc. among my probs i was having), although i think this is being caused by the mobo, but it turns out i used to work at micro center and the manager remembered me and aggreed w/ me about the mobo and offered me an exchange even though ive had it for 2.5 months.

anyways, i had the choice between another one w/ the 670m or one w/ the 660m. i chose the 660m (even though it has 1g less ram and less bittage) because i consider it a newer, better card, and i like the kepler architecture. also i got some money back :) . the point is though, that as i was leaving, they told me that they had a lot of exchanges w/ the 670m models w/ defective mobo's, so just a heads up, im not sure if this is true or not, but i did try a memtest, 2x fresh installs of windows, 2x F9 restores, and different HD's, so that leaves the processor or the mobo, food for thought i guess.

im pretty happy w/ the 660m model, i stuck in my SSD and extra ram and it purrs like a kitten, way more stable and seems a lot faster than the 670m model, but it WAS a turd, i must admit.

dstrakele
07-07-2012, 01:06 PM
Umm - we've all seen the picture of your original G75 that you posted in http://rog.asus.com/forum/showthread.php?18561-Remove-bottom-cover-G75VW-Quick-question-video-link-to-making-a-recovery-pa&p=126233&viewfull=1#post126233. I'm not pointing fingers here, but to me it looks like there may have been some CID (Customer-Induced Damage)...

But seriously, this quote:


. . the point is though, that as i was leaving, they told me that they had a lot of exchanges w/ the 670m models w/ defective mobo's, so just a heads up, im not sure if this is true or not, but i did try a memtest, 2x fresh installs of windows, 2x F9 restores, and different HD's, so that leaves the processor or the mobo, food for thought i guess.

is indicative that ASUS has SERIOUS quality control problems that cannot be explained away by saying only those with problems post to this forum. I want ASUS to be the best laptop manufacturer. That will require they seriously address this problem. It can no longer be swept under the rug.

Holy
07-07-2012, 04:35 PM
Jsl1ce
welcome to the 660m side lol.

Liquidus
07-07-2012, 06:09 PM
TURNS OUT YOU'RE DEFECTIVE! hah!

Brian_ROG
07-07-2012, 06:11 PM
Jsl1ce
welcome to the 660m side lol.

I lol'd, glad you were able to get a new machine. Now i'm a tad nervous with my 670m, but i havn't had any issues so far and it's been almost 2 months now.

kiba
07-07-2012, 10:42 PM
Umm - we've all seen the picture of your original G75 that you posted in http://rog.asus.com/forum/showthread.php?18561-Remove-bottom-cover-G75VW-Quick-question-video-link-to-making-a-recovery-pa&p=126233&viewfull=1#post126233. I'm not pointing fingers here, but to me it looks like there may have been some CID (Customer-Induced Damage)...
lolz, what damage? looks fine to me? (that looks like an old IBM thinkpad)


Jsl1ce
welcome to the 660m side lol.

Thanks, Im really happy with the 660m over the 670m, imho it seems better in all aspects.


I lol'd, glad you were able to get a new machine. Now i'm a tad nervous with my 670m, but i havn't had any issues so far and it's been almost 2 months now.

Thats about how long I had mine when I started having problems with it, either the mobo or processor, I wish you luck that you dont have the same thing happen bro. honestly i almost ripped the screen off lol.

kiba
07-07-2012, 10:48 PM
But seriously, this quote:



is indicative that ASUS has SERIOUS quality control problems that cannot be explained away by saying only those with problems post to this forum. I want ASUS to be the best laptop manufacturer. That will require they seriously address this problem. It can no longer be swept under the rug.

I completely agree, I love ASUS (obviously since I have 3 of their laptops), but as far as the problems that Ive had with my laptops, I think that you are right. Who knows what quality standards they have, when small, probably underfed and sick children are probably building these things in sweat shops in china. Its hit or miss, sometimes you get a good one, sometimes you get a bad one. imo all of that really needs to change.

Especially since some people, on this forum even order these things from halfway around the world, it takes weeks to get to them, and when they open the box, it turns out to be damaged or defective in some way and they have to send it back, wait many more weeks for it to be exchanged (if theyre lucky) or RMA'd. This is extremely frustrating, if the quality standards were better, many people would be a lot happier.

Shawnnepc
07-07-2012, 10:56 PM
Neither us or our distro channels have seen any mass failures with the G series

Out of all of our G75/55 series sold only 2 had enough issues to warrant exchange.

kiba
07-07-2012, 11:05 PM
Neither us or our distro channels have seen any mass failures with the G series

Out of all of our G75/55 series sold only 2 had enough issues to warrant exchange.
well, like i said i dont know if what they told me is true or not, but i heard a loud *POP* come from under the keyboard, on the CPU side, followed by a quick freeze-up, then a shutdown. that combined w/ the other problems i had, its obvious that at the very least I got a bad one.

Shawnnepc
07-07-2012, 11:09 PM
well, like i said i dont know if what they told me is true or not, but i heard a loud *POP* come from under the keyboard, on the CPU side, followed by a quick freeze-up, then a shutdown. that combined w/ the other problems i had, its obvious that at the very least I got a bad one.

No doubt there.


If I remember correctly ASUS has a defective rate of about 6-11% (well below industry)

ASUS also has some of the best QA that we've dealt with

kiba
07-07-2012, 11:14 PM
No doubt there.


If I remember correctly ASUS has a defective rate of about 6-11% (well below industry)

ASUS also has some of the best QA that we've dealt with

yep, dont get me wrong, i love asus, theyre one of my favorite manufacturers for notebooks, mobos and gpu's.

also sone of the metal near the top of the keybd was pushed up like it had been assembled wrong or at the very least, dropped during transit.

Shawnnepc
07-07-2012, 11:24 PM
yep, dont get me wrong, i love asus, theyre one of my favorite manufacturers for notebooks, mobos and gpu's.

also sone of the metal near the top of the keybd was pushed up like it had been assembled wrong or at the very least, dropped during transit.

Yeah see that's why we're what is considered a VAR (Value Added Re-seller)

We don't stock any computers in out store, everything is ordered in from distro.

After we receive the unit from distro we inspect the unit and burn it in for 48 hours. During this time we also inspect the thermal paste on the GPU/CPU, check the voltages and timings of the ram, and we test the output of the AC.

When we ship or sell a unit we verify to the best of our ability that EVERYTHING is perfect.

Online retailers don't do that so it's also a gamble when your purchase stuff from them

kiba
07-08-2012, 04:28 AM
yeah, i do that on any computer i buy/build as well as benchmark it w/ furmark, i dont check the actual thermal paste on notebooks as its unneccessary (ive repasted a lot of mine and friends laptops and whenever there is something wrong w/ the paste, there is ALWAYS something off w/ the temp), but i think its really cool that you guys do that.

onesuperrob
08-13-2012, 07:35 AM
Well put me down as a g75 swap over.

Straight out the box the stupid GPU would **** it self as soon as it hit 63 degrees.

Place i got it didn't even check it, they just straight swapped it for a new one and this one is good so far.

Techinstein
08-13-2012, 04:22 PM
Yes the 660m is cheaper and good graphics. But I don't understand why you think its newer than the 670m. I don't even really know the differences. Which G75 did you have, or get? I got the G75VW-BBK5, and now I am looking into SSD's so I would have spent less money and have the beast.

c_man
08-13-2012, 05:45 PM
670M is Fermi (40nm), 660M is Kepler (28nm).

Kepler is the successor of Fermi. Kepler is 2012, Fermi is 2010.

670M has its performance roots more or less in 470M. The fact that they could optimize this so much from 470M to 570M and finally 670M is just amazing. These GPUs are very solid.

Thing is no one wants Fermi tech. Kepler's shaders should be up to twice as power efficient, but the real life once again tells a different story.

Kepler is more power efficient, but it's not exactly by half. Also in theory Kepler should be much cooler, but again it's not exactly the case.

I had Kepler 650M. This should be very cool, right? Well, I got so hot, my leg was burning. Just buying a Kepler does not mean that the laptop will be ice cold or that it will OC like crazy - I could not OC the 650M at all since it's all dependent on cooling.

Asus G75, the one with 660M will reach 71C under Furmark compared to 75C for 670M. 670M needs more power but not like double, it's more like +15-20% (give or take).

You can OC 660M very high, but only on laptops designed for top GPUs (they can handle the heat and extra power) and only if you hack the vBIOS. If something goes wrong and you will not be able to access the card, bye bye warranty.

670M is easier to OC, it has no limits other than plain hardware limitations. I have seen that some can go stable up to 800/1750 and that means a score in 3dmark11 of ~4k points. Everyone needs to test the limits for the card in the laptop, but a 700-720-750 should be OK as long as temps are very close to stock and no artifacts are present.

kiba
08-13-2012, 09:01 PM
^i have a work laptop m17x w/ a 570m, which is a very similar card to the 670m, its kinda like me having two of the same laptops, i swapped it out for a 660m because of that and because its newer tech. in my experience the fermi cards arent very solid at all, my 570m has been replaced twice and they always run hot, the second 670m i had, i took back because it ran hot, and both are very unstable at these high temps, to me this doesnt speak well for OCing. everyone has their own opinions i guess though.

c_man
08-13-2012, 09:31 PM
If it runs hot it had a defective cooling system (thermal paste or whatever). My 670M is about the same with what others have with 660M and they all consider this to be very good temps (let's say it's 5C more, even if mine has less and I'm not the only one here).

But I've seen on this forum 670M running +90C. I had 650M doing the same and that is a Kepler running at very low clocks. While the 670M had a defective cooling, the 650M did not, Dell thinks it's adequate.

Now the problem with GPU is not the temp itself (even if after a while high temps have their own problems), but the difference between low and high. If this difference is big, when the GPU cools down there will be some microfractures. After a while the GPU will die. There are methods of resurrection, but with a solded GPU, it's a job for a pro with special equipment. If it's very well done, it should work for 1-2 years. If not, maybe a few months.

So what I say is this. See exactly how much power your game needs. I play Inversion with low clocks and I get a very low temp (from 50 idle to 60 in game). There is no need to run it at full speed or even OC. Then again if I want Max Payne 3 with all max, I need to OC the card to keep those settings. Another thing. Do I really need to run MP3 at maximum? Is there a big difference? Well, not exactly, so I'll try lower settings until there is little difference, but I can run my card with lower clocks (this kinda sucks with MP3 as it restarts every time if you just take small steps). This is not Fermi vs. Kepler, but good practice using GPU. I never got the chance to write the topic about setting and using 670M.

I'm not a Fermi guy, I had 670M, but also 650M and 660M. Just saying Kepler on the card means nothing.

kiba
08-14-2012, 04:02 AM
Yes the 660m is cheaper and good graphics. But I don't understand why you think its newer than the 670m. I don't even really know the differences. Which G75 did you have, or get? I got the G75VW-BBK5, and now I am looking into SSD's so I would have spent less money and have the beast.

your BBK5 has a kepler card (660m) same as mine.


Just saying Kepler on the card means nothing.

from what i have seen in furmark, my 570m and 670m, as well as a freinds 570m alienware and my coworkers 670m g75, run a good 12C hotter than the 660m, which increases w/ time, and they are unstable at those high temps especially for extended periods, unlike the 660m which runs low temps for extended periods (8hrs +), and runs very stable OC'd at higher temps (which OCing on the 660m doesnt increase the temps by too much), from what i have seen Its more versatile and IMO a better card, otherwise i would have kept my 670m g75, thats all i am saying.

c_man
08-14-2012, 08:22 AM
This is my 4th laptop. I have not seen any "unstable" behaviour with 670M OC @ 755/1650 playing all day long. And temps are just a few degrees up. 660M OC +135Mhz (similar amount) is the same. The power brick is just as hot on both. They are both warm. There is a difference in temp, battery time, power usage, but not as much as people think.

If 670M runs 12C hotter, it's not the GPU's fault. Mine has about a 3rd of that..

kiba
08-14-2012, 05:56 PM
This is my 4th laptop. I have not seen any "unstable" behaviour with 670M OC @ 755/1650 playing all day long. And temps are just a few degrees up. 660M OC +135Mhz (similar amount) is the same. The power brick is just as hot on both. They are both warm. There is a difference in temp, battery time, power usage, but not as much as people think.

If 670M runs 12C hotter, it's not the GPU's fault. Mine has about a 3rd of that..

the only way to know that for sure is to run furmark on the two notebooks, side by side, same room same ambient , i did this right after switching to the 660m, first time 10c difference, second time 12c difference, once at home w/ a freinds, and once at work w/ my coworkers. first one was xoticpc, second one was micro center like the first 2 g75's i had. i know this is not a thermal paste issue because i got similar results on my first 2 g75's(670m). not to mention that this is all done w/ the inadequate older driver which doesnt allow it to throttle/spool up all the way, i have heard they are working on a new driver, but im guessing when they get the driver right it will make a difference.

one thing that i remembered & thought about before switching to the 660m, they have been using the fermi across 3 generations (400, 500, 600) if the new nvidia cards are going to be kepler, thats what i'd rather have, not to mention all the perks that it has over the fermi - smaller, faster, cooler & more stable. like i said i consider it a better card and im thankful to micro center for letting me switch.

c_man
08-14-2012, 07:24 PM
This will be a long story. My wife kinda hated me for all these laptops I got home and spent hours on them testing.

I could only run them one after another since I can't take home more than one at a time. So the room and everything else was the same. Same OS, same programs (except brand related), routine, games and so on.

Now, let's say Kepler in general are better. So here I am thinking the same thing as you, few months back. And I wanted a good cool all-around laptop that does not sound like a jet (I have a Clevo for that). A 650M Kepler, low clocks, it should run extremly cool, right? So I got the laptop. Now, it's funny you saying that Fermi can't keep up performance for long. So here I am some time ago, "happy" with my Kepler laptop, buying the "smaller, faster, cooler and more stable" card (not to mention Ivy, 3610QM, 16Gb RAM 1600Mhz, SSDs, in a word, a fiesta). Now, at what temps do you think this Kepler runs under heavy gaming load, not Furmark? Well, after a while it will hit ~90C. OK, I say, what do I really know about Kepler? Not much, only what they say. But 90 sounds bad. I knew this much. Can it hold performance? No. Clock drops to around 600Mhz. What, where's the 835 I say? OK, one might do that, let's test another one. Still 90. OK, how about others. The same.

OK, never mind, I don't have to keep it. Now I want a cooler laptop. Kepler 640M. What do you know, this also runs hot as hell, bad performance, throttling. I even forgot about this one. OK, one might be bad. 2nd, 3rd, the same. OK, not good.

So I've started searching and asking around. What's going on, are all laptops with problems? This testing has some limits, I can't get every laptop, I need to stop soon. I found the thread here about Lenovo Y580 versus G55 saying how cool the Lenovo is. Well, now I know it's not exactly like that.

Next I went for G series, G55 was my first choice.

660M cool at last. I was in love. First G55 tops 71C. Now we're talking. Other user said the exactly the same value on G75. Searching and asking around some more, seems like this is it then. Still some throttling message from Furmark, but who cares, clocks are OK, everything is fine.

670M. Fermi, no one wants Fermi anymore. It was cheap so I wanted to see how well it does in G75. As things turned out, I could test a few G75s with 670M. First, it was just a bit hotter than 660M. In all of them. BUT it's a Fermi, it should BURN! Or not. That is stange as some people here had +90C, I mean my experience with some Keplers (but I've tested more than one and I had other owners' confirmation that all of them were hot). This means that some laptops can be hot, Fermi or Kepler, it does not matter.

Then I've started searching for power needs. Kepler laptops should need a lot less power, right? As it turns out, not quite. OK, then battery life should be much better. Again, it's really not like twice as much or close. So, Kepler might look great on paper, but in real life ... . It might just be too new with bad drivers. It might get better in time. But what about those laptops where Kepler works great (see G55)? I guess maybe they are what they are.

Right now I've found little to sustain that Kepler blows the Fermi on every level. Except the 680M with the overvolting thing, but then again, it's from another world. The card itself = my G75.

In the end I don't care about Kepler or Fermi. In G55 is a Kepler, works great. In G75 is a Fermi, works great. Strange fact, I've paid less for G75 with Fermi because it's Fermi, than what I've paid for G55 with Kepler. With Fermi I don't have to hack the vBIOS. My Fermi is stable at 800Mhz, but I won't go that high. Nothing more to say. And stock it's about 4 hotter that 660M stock, G75 as . And it's not the only one on the forum. And this means your G75 tops a 59C after 15 min. Furmark benchmark.

UltimaRage
08-14-2012, 07:30 PM
My 660M never went higher than 68 C when running Furmark for 20 minutes.

c_man
08-14-2012, 07:45 PM
I'll post the screen in 15 min. now. It should be about the same with yours if you did the 1920x1080 @ 8xMSAA.

LE: So, like I've said, it's about the same.

~30 min. I think, room temp. ~32C, it hit max. 74C, but for most part was 73 and you can see some drops to 70. NvidiaInspector says clocks were constant at 620/1495Mhz.

http://postimage.org/image/k5bhb0xyz/full/

Still far from a constant 12 more and I would say that my room temp is above average.

Not bad at all for this old Fermi.

c_man
08-14-2012, 08:16 PM
delete please

kiba
08-16-2012, 12:14 PM
as far as what you are saying i agree w/ you, no argument here, and also untill ive run my own benchmarks im not ready to admit that one is better than the other or decide on one or the other, but i have done extensive benchmarks side by side w/ other G75's/m17x's and so far this card not only outperforms them for longer but it gets numbers in furmark that arent far off from my wc'd gaming rig w/ twin 560ti's. IMO the 660m is the most well made laptop gpu ive had (and my company buys me a new one every year, and also i have an alienware m17x at work w/ a 675m).

as far as the other laptops go, as you know the biggest concern w/ a gaming laptop being viable is the cooling system, and very few of them are adequate, even the msi gt780r has the cpu thermal pipe running across the gpu, and they both share a corner vent, but a lot of these so-called gaming laptops (samsung 7-gamer, hp envy17, and even an older origin that i opened up for my buddy to clean out the pipes & vents) are like that, w/ stuff just haphazardly packed in there. IMO the best cooling systems are 1) alienware m18x w/ watercooling (blows everything else away) 2)tie between Falcon NW DRX and origin eon 17 3)tie between alienware m17x and g75 (very similar). IMO all the rest (sager/clevo, msi, samsungs 7-gamer, and of course at the bottom of the barrell - hp envy and fujitsu lifebook gamer) are poorly designed. i mainly help run a server farm but another part of my job is also to do upgrades and repairs for my coworkers (when something breaks, personal laptops included, i or one of my coworkers sees it first), so i have seen a lot of so-called high end laptops, and a lot are just plain poorly designed, even clevo/sager. a lot have crossed thermal pipes, or pipes that run over top of the cpu/gpu or thermal vents that arent set far enough apart or even seperated at all. some of them have an ok cooling design, only to cover it up with a peice of plastic w/ only tiny inadequate holes to let the air escape and no intakes to speak of, its like they just threw all the components in there in a hurry or w/out thinking about it.

So IMO unless youre benchmarking on an alienware, falcon nw, origin or asus, most likely it will be a poor representation of the card because theres so many out there w/ a bad cooling design.

c_man
08-16-2012, 01:05 PM
I had them both on Asus in the same room, same spot, same install (I do it myself), same room temp and so on.

How can I say that 660M is much cooler and stable when 670M was about the same? And this is for several laptops. I just can't. The laptops are all warm, power bricks about the same temperature. All have the same power brick. I can't say that.

I agree with you and it's what I'm saying, the name Kepler does not mean a cool GPU. It takes more than that. And "old" Fermi are not exactly the very first ones years ago.