FX-8350 temps look fine... and then spike up to 255° C for no apparent reason

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

cytg111

Lifer
Mar 17, 2008
24,184
13,686
136
And then there was this lady that got 3 trizillion dollars suing mcdonalds cause the coffee was too hot.
 

Idontcare

Elite Member
Oct 10, 1999
21,110
59
91
And then there was this lady that got 3 trizillion dollars suing mcdonalds cause the coffee was too hot.

I would not take "3 trizillion dollars" in exchange for having my genitals scalded with hot coffee, not ever. Sure she got compensated in ways we can only dream of but at what personal cost? More to life than money and the money isn't going to take back the tissue damage.
 

cytg111

Lifer
Mar 17, 2008
24,184
13,686
136
and thats where ppl differ.
Id take a couple of burned nuts for that kinda money anyday (of course, i allready reproduced my quota so..)

But the point is really from the opposite view, why does this woman get this huge huge compensation from mcd? It has got less to do about her and her injuries and more about the penalty to the company at fault.
 

UaVaj

Golden Member
Nov 16, 2012
1,546
0
76
hot coffee right off the brewer. 3nd degree burn on your manhood, and no guarantee of actually getting that money.

sex will never be the same. unless you not getting any then it does not matter. how much is your manhood worth to you?

note: don't take the sarcastic comment personally. just a figure of speed to get the point across.
 

cytg111

Lifer
Mar 17, 2008
24,184
13,686
136
sex will never be the same..

Arh .. from a 3deg burn to the nuts? You'll recover, and sex will be just fine..(ill never have to shave again ...) But for the record and point in case, no I would not surrender my nuts to any sum of money!
 

Dankk

Diamond Member
Jul 7, 2008
5,558
25
91
MSI sent me this email today.

Hi Dan,

We can do advanced shipment for you and a credit card is required to secure the return of the board ($80 will be charged, and refunded to you within 7 working days after receipt of item), but our RMA policies by paying for the shipping sending the board back since it is same as if you were going to the store to exchange something by driving to the store at your expense.

But if this is the attitude you are going to have, then it may hinder the good will that we provide to you.

Please note that it's not false advertisement because the processor works, except it's not meeting your unreasonable demand because you wanted more than what the board can handle through unreasonable amount of practice/usage. Any V8 can perform only when the chassis allows it, but not exceeding its spec. The thermal protection is doing what it's supposed to do to protect the system from overheating therefore it's not a flaw.

You should ask AMD why they underrated their FX-8350 processor that mislead you into buying something that you thought was 125W but instead was performing at 140W.
Our engineers did what they were supposed to do to keep the board running at is designed capacity of 125W.

This is why we are offering you to get to the better board by performing an upgrade process for you since we normally don't provide this service.
Anyway, we will cover 2-way shipping for you this time as you've requested, but as stated in the previous email you do need to pay for the upgrade to either one of the boards.

If you really need 140W boards such as 990FXA-GD65/GD80 then we can offer you a courtesy step-up option by paying for the differences.Please note that your current board will need to come in through RMA service, you are responsible for 1-way shipping sending the board to us, and the replacement board will be factory-rectified board but not brand new. The only accessory that will come with the replacement board is the I/O shield.

990FXA-GD65 - $40
990FXA-GD80 - $70

Let me know if you wish to proceed with the upgrade and advanced shipment, thank you.

Sincerely,
MSI Service & Support Division

According to MSI, expecting your CPU to function properly under 100% load is an "unreasonable demand".

Also, "The thermal protection is doing what it's supposed to do to protect the system from overheating, therefore it's not a flaw." In other words, they're saying "Hey, we implemented a half-assed workaround that prevents the board from going up in flames under heavy use, so the board is obviously fine." :hmm:

Nonsense aside, at least they're now offering advanced shipping, and paying both ways. That's good enough for me. I'll probably take them up on their offer and upgrade to the 990FXA-GD65.
 

inf64

Diamond Member
Mar 11, 2011
3,883
4,666
136
Yep I would do that if I were you. Still you had to threaten them with legal action before they came through and offered that deal... Pretty lame from their side. This should be stickied somewhere so other conned buyers of this board can reference it and ask for the same deal.
 

itsmydamnation

Platinum Member
Feb 6, 2011
2,971
3,640
136
Arh .. from a 3deg burn to the nuts? You'll recover, and sex will be just fine..(ill never have to shave again ...) But for the record and point in case, no I would not surrender my nuts to any sum of money!


maybe someone needs to have the birds and bees talk with you, because i think your missing one of the major physiological difference between a man and a woman.
 

Idontcare

Elite Member
Oct 10, 1999
21,110
59
91
MSI sent me this email today.



According to MSI, expecting your CPU to function properly under 100% load is an "unreasonable demand".

Also, "The thermal protection is doing what it's supposed to do to protect the system from overheating, therefore it's not a flaw." In other words, they're saying "Hey, we implemented a half-assed workaround that prevents the board from going up in flames under heavy use, so the board is obviously fine." :hmm:

Nonsense aside, at least they're now offering advanced shipping, and paying both ways. That's good enough for me. I'll probably take them up on their offer and upgrade to the 990FXA-GD65.

I continue to be impressed with MSI if only for the degree of candor and frankness for which they are willing to engage in discussing the crux of the issue here, both technical/engineering as well as the business/sales support.

You don't come by that much these days. Most customer support teams are coached to hide behind boilerplate passages of text that have been stripped of any and all emotional expressionism as well as avoiding stating anything that could ever possibly be construed as "finger pointing".

And yet there it is. Clearly MSI is grappling with the issue as much as you are.

Their job is a difficult one, just because AMD slaps an FX-8350 label on the side of a box and silkscreens it onto the CPU's IHS doesn't mean the IC inside the CPU package actually adheres to the electrical specs for which the mobo was designed. MSI is stuck with the financial obligation of burned out mobo's in that case, and they rightly sought to limit their liability.

The blame game that can go on with enthusiasts pointing fingers at MSI can go on but really there isn't much that will come of it because that discussion is the Red Herring.

The real issue goes back to AMD, they need to spec their products based on what they intend to ship to the customer and stop this shady business of shipping out-of-spec parts.
 

jvroig

Platinum Member
Nov 4, 2009
2,394
1
81
@OP: It seems to me that MSI is dealing with you as best they can without completely assuming 100% responsibility of AMD's screw up (and they certainly should not assume the liability for AMD's screw up).

At the end of the day, it is AMD that screwed up, marking a product as "125W TDP" when it clearly goes over rather easily when loaded at 100% even on non-power-virus-like applications, or that the quality of the product varies by a lot that it may or may not end up surpassing the target TDP.

I know you feel that you've kinda been screwed. But they (MSI) clearly were screwed as well, and how to deal with it (from their POV) is a complex decision that has to take into account their product mix and their competitors' product mix as well. They can't just overspec all their boards to allow AMD's out-of-spec product to function as one normally expects at full load, because that would mean their own product mix might be less competitive in the market than their competitors (more expensive and/or less margins), but they also cannot strike out the 8350 from their supported CPU lists for 125W boards, otherwise their product mix becomes weaker than their competitors who wouldn't do such a thing.

I'm not trying to defend MSI. I am trying to show that their decision on how to handle an out-of-spec 8350 is not based on malice or on trying to screw customers.

AMD is the one doing the screwing here. It's unfortunate that you ended up on the receiving end of bad luck here (a mix of the "right" board plus the "right" product (out-of-spec 8350 chip) to make it happen). You should be angry, but take that anger into the real root cause, which is AMD's shady practice of selling out-of-spec chips as good.
 

deltree86

Member
Jun 2, 2011
34
0
0
I have seen a socket 754 amd spit a bit of smoke at 90c degrees plus with a accidental over volting to 1.7v and it was so sudden that I knew immediately I had to go back into the bios half way through startup! So at 255 ud have a fire even if it was for a second:p
 

VirtualLarry

No Lifer
Aug 25, 2001
56,566
10,181
126
@OP: It seems to me that MSI is dealing with you as best they can without completely assuming 100% responsibility of AMD's screw up (and they certainly should not assume the liability for AMD's screw up).

At the end of the day, it is AMD that screwed up, marking a product as "125W TDP" when it clearly goes over rather easily when loaded at 100% even on non-power-virus-like applications, or that the quality of the product varies by a lot that it may or may not end up surpassing the target TDP.

I know you feel that you've kinda been screwed. But they (MSI) clearly were screwed as well, and how to deal with it (from their POV) is a complex decision that has to take into account their product mix and their competitors' product mix as well. They can't just overspec all their boards to allow AMD's out-of-spec product to function as one normally expects at full load, because that would mean their own product mix might be less competitive in the market than their competitors (more expensive and/or less margins), but they also cannot strike out the 8350 from their supported CPU lists for 125W boards, otherwise their product mix becomes weaker than their competitors who wouldn't do such a thing.

I'm not trying to defend MSI. I am trying to show that their decision on how to handle an out-of-spec 8350 is not based on malice or on trying to screw customers.

AMD is the one doing the screwing here. It's unfortunate that you ended up on the receiving end of bad luck here (a mix of the "right" board plus the "right" product (out-of-spec 8350 chip) to make it happen). You should be angry, but take that anger into the real root cause, which is AMD's shady practice of selling out-of-spec chips as good.

I wouldn't put all the blame on AMD here. Everyone knows that the FX chips use a lot of juice. Even my ASrock 990FX Extreme4 board, has had reports of the FX chips throttling when overclocked a tad and run at full load. That's a 140W+ 8+1(2?) phase board too.
 

Idontcare

Elite Member
Oct 10, 1999
21,110
59
91
I wouldn't put all the blame on AMD here. Everyone knows that the FX chips use a lot of juice. Even my ASrock 990FX Extreme4 board, has had reports of the FX chips throttling when overclocked a tad and run at full load. That's a 140W+ 8+1(2?) phase board too.

This has nothing to do with overclocked situations. This is straight-up stock clocks, stock settings, and the chip throttles because it consumes more power than it is spec'ed to consume. That is a distinctly different situation.

It is one thing to expect the chip to use lots of power when overclocked, but it is entirely different to be expected to know the chip breaks its own spec even when running at stock. Who expects that to happen? I didn't.
 

crashtech

Lifer
Jan 4, 2013
10,589
2,153
146
AMD's fault for releasing a product with a false TDP, and MSI's fault for helping cover it up.
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
AMD's fault for releasing a product with a false TDP, and MSI's fault for helping cover it up.

I don't think it's MSI fault here. AMD exceeded their own specifications. The hour MSI goes on public and say that their 125W is unable to handle the 8350 AMD will cut them off.
 
Last edited:

sourn

Senior member
Dec 26, 2012
577
1
0
I don't think it's MSI fault here. AMD exceeded their own specifications. The hour MSI goes on public and say that their 125W is unable to handle the 8350 AMD will cut them off.

I disagree.. Both are at fault here. It's pretty clear why AMD is also at fault so no need to go into that.

But when this issue became known MSI could've did the right and issued at the very least a warning/caution.
 

coffeejunkee

Golden Member
Jul 31, 2010
1,153
0
0
I don't think it's MSI fault here. AMD exceeded their own specifications. The hour MSI goes on public and say that their 125W is unable to handle the 8350 AMD will cut them off.

I understand what you're saying, if they pull 8350 support it will put them in a bad position compared to other brand's budget boards. But the question now is whether these other manufacturer's boards can in fact handle 8350 properly without throttling.

These are Asus and Gigabyte boards with what looks like comparable vrm-circuitry:

this Asus board supports 8350: http://www.asus.com/Motherboard/M5A78LM_LX_V2/

this one doesn't: http://www.asus.com/Motherboard/M5A78LM_LE/

this Gigabyte board supports 8350: http://www.gigabyte.com/products/product-page.aspx?pid=4395&dl=#ov

this one doesn't: http://www.gigabyte.com/products/product-page.aspx?pid=4385#

What we now need to know is if these boards that claim to support 8350 in fact do so. If true it shows that MSI skimped out on the vrm-circuitry too much compared to Asus and Gigabyte. If not, it's a problem that's present across all brands.
 

Coward

Junior Member
Feb 10, 2013
1
0
0
Does this problem apply to FX-8350 processors with 125W TDP or does it apply to all 125W TDP processors specifically the Phenom II X 4 965 125W processor? In other words will the MSI 970A-G46 motherboard support the Phemon II X 4 965 125W processor without the problem that the FX-8350 has?
 
Last edited:

SPBHM

Diamond Member
Sep 12, 2012
5,059
413
126
Does this problem apply to FX-8350 processors with 125W TDP or does it apply to all 125W TDP processors specifically the Phenom II X 4 965 125W processor? In other words will the MSI 970A-G46 motherboard support the Phemon II X 4 965 125W processor without the problem that the FX-8350 has?

they claim the trouble is related to the 125w TDP for the 8350 not being a realistic number, as long as the 965 is not going over 125w it should work.

looking here it seems like the 8350 uses significantly more power than the 980
x264-power-peak.gif


965 should use even less, so I would expect the TDP to be significantly lower
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
I understand what you're saying, if they pull 8350 support it will put them in a bad position compared to other brand's budget boards. But the question now is whether these other manufacturer's boards can in fact handle 8350 properly without throttling.

It seems that there are others manufacturers doing exactly what should be done regarding the FX8350. MSI indeed shouldn't be put a 8350 stick on its boards and my first opinion on this matter was completely wrong.
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
The real issue goes back to AMD, they need to spec their products based on what they intend to ship to the customer and stop this shady business of shipping out-of-spec parts.

The clock differences between server and desktops parts are too high IMHO to be just a marketing issue. Assuming that AMD is factory overclocking their processors, e.g. exceeding their own specification, can this have any adverse effect over the life spam of 8350 processors?
 

Idontcare

Elite Member
Oct 10, 1999
21,110
59
91
The clock differences between server and desktops parts are too high IMHO to be just a marketing issue. Assuming that AMD is factory overclocking their processors, e.g. exceeding their own specification, can this have any adverse effect over the life spam of 8350 processors?

Definitely, you don't get anything for free there. It might explain why you can't find specs on the 8350 as well - no max temp or max Vcc specs exist, for example.

And given AMD's precarious financials, do they really care if 8350's die in 3yrs instead of 10yrs when they are scrambling to still be here in 2yrs time?

Crisis mode management puts a different focus on everything.

I find it telling that GloFo is basically making no provisions whatsoever for AMD in its roadmaps beyond 28nm. 20nm and beyond is all low power mobile, no plans for 20, 14XM or 10XM to support the existing HPP segment.

So that either means AMD has told them they are done with high performance MPU at 28nm or GloFo has told AMD they can go to TSMC for all their future HPP needs.

If you were a betting man, which do you think is more likely?
 

mrmt

Diamond Member
Aug 18, 2012
3,974
0
76
So that either means AMD has told them they are done with high performance MPU at 28nm or GloFo has told AMD they can go to TSMC for all their future HPP needs.

If you were a betting man, which do you think is more likely?

Surely with the former. GLF was *really* stringent in binding AMD in every way they can, I doubt they would allow AMD to manufacture even a vacuum cleaner at TSMC, much less chips there. HPP beyond 20nm is probably written off from AMD future roadmaps.

This move may also speed the demise of AMD GPU business, as they will be handicapped against Nvidia. Pie in the sky stuff is going to happen in AMD GPU division when GLF announces they will be late with their 20nm node, and worse, when GLF announces it will perform worse than TSMC's.
 

GeoN2K6

Junior Member
Feb 12, 2013
2
0
0
This post was very helpful in understanding what just happen to me few days ago, so I want to thank you and to share my story.

I own a MSI 970A-G45 and a PII 960T. I had a small overclock from 3.0 to 3.6 with a lower voltage 1.3 compared to stock 1.35.

Because I was getting crappy performance in a hotseat civ5 game I wanted to see if I can squeeze some more performance. So before going to bios I fired the intel burn test and guess what? ... after 2 iterations ... spike to 255 degrees and throttle to 800 Mhz.

I though something got damages so I started reducing frequency and voltage and only at 3Ghz and 1.28 volts it worked fine.

Because of this post and because the temps were like 5 degrees higher compared to the moment I did the overclock.(57 before the spike compared to 52 before) I started looking into the heating of the main-board and the thing is I had 2 fans on the back bringing air but one of them started to make some noise so I stopped it.

Long story short, after I cleaned all the dust from the radiators, bring back that extra fan I started the pc and temps were once again 5C lower, I could overclock to 3.6 and the throttling is not happening anymore.

Now because I was planning to get a FX I have a question. Will this crappy MSI be able to handle an 125W cpu, considering it cannot handle a 95W part? Sure I overclocked, but I also reduced the voltage.

According to http://www.extreme.outervision.com/tools.jsp#cpuoc the cpu should consume 106 W, which is significantly lower than the stock 125W of FX8350. And I can get 106W cpu to work with 2 extra fans and more than this the samurai zz fan blows also over the voltage circuitry.