FX its tougher than you think.

During the daytime hours it can get pretty damn uncomfortable up here.

Click to expand...

I can imagine, last year when doing the Chimp Challenge I had 4 rigs going in my office and the windows were open with a fan blowing air into the room and it was still 85-90f in the room.

Well I finally decided it was time to clean out under the desk, dust off my radiators, and install my UPS, so the system got shut down today. However, I did make this SS just before shutting down. 24+ days at full load without a crash or a lockup. Pretty stable IMO.

ssjwizard said:

Well I finally decided it was time to clean out under the desk, dust off my radiators, and install my UPS, so the system got shut down today. However, I did make this SS just before shutting down. 24+ days at full load without a crash or a lockup. Pretty stable IMO.

View attachment 136646

Click to expand...

You sure thats stable SSJ? You may want to run it a few more hours just to make sure.

Heh, once I get done toying around with my drivers and making some tweeks to the system Its likely going to stay up until the end of january unless I have a major power outage. So time will tell. I might try to up my clocks a bit today also. I have been tuning my buddies rig all month so Im kind of in the mood for beating up my system a bit.

I'd like to help aid this thread with some actual data in concern to what would be safe temperatures for processors to run.

I shall quote TheStilt and give 100% testimony to his postings to enlighten me and many other people with his vast knowledge of Computer hardware and his teachings which I personally very much enjoy reading and learning from.

And I directly quote :

"Tctl is a processor temperature control value used for processor thermal management. Tctl is accessible through D18F3xA4[CurTmp]. Tctl is a temperature on its own scale aligned to the processors cooling requirements. Therefore Tctl does not represent a temperature which could be measured on the die or the case of the processor. Instead, it specifies the processor temperature relative to the maximum operating temperature, Tctl,max. Tctl,max is specified in the xxx thermal datasheet."

For example:

FX-8370E
tCTLMax = 70
tCaseMax = 70.5°C
THERMTRIP on tCTL scale = 85

FX-8370
tCTLMax = 70
tCaseMax = 61.1°C
THERMTRIP on tCTL scale = 85

FX-9590
tCTLMax = 70
tCaseMax = 57.0°C
THERMTRIP on tCTL scale = 85

The tCTL output algorithm (scalars) vary between different models

Click to expand...

That the CPU can go above and beyond 62c and run just fine is a given, everyone that's spent a few days overclocking and tweaking their rig can attest to that. I had my running at 5ghz peaking at 78c overnight Prime95 blend and not a single problem. I thought we told people to not go over that temp was simply because it causes degradation (chip lasts less time) which is very sensible to the majority of users which for the most part have no idea what their doing.

But that also raises a question, how long do you really want it to last ? i still have a very decrepit but still very much working Compaq AMD-200mhz that i use to teach my kids how to mess around with windows that has to be what, over 10 years old ?

Most of us nowadays upgrade roughly every 2y so even if you abuse it to hell and back (say for example running it 75c 24/7 like OP did) it SHOULD last at least that long or more no ? YMMV on how much abuse you give it of course.

Yes we RECOMEND 62*C & 72*C temps for 24 / 7 use and will always do so. Pushing above that in some / most cases does induce some instability and the higher VCore as well as temps and degradation of the chip just plain does not make sense. Most people do not need to go here and a big part of this argument as well as our recommendations are geared towards those people that come here looking for advice on how to OC their gear.

Do I push my gear harder, Johan, Manny, RGone .... etc and what about caddie daddie? I know I do at times but the problem with giving advice is people will follow it. If I push my stuff real hard and I burn out a motherboard or CPU I can go down to the store today and grab a new one no problems. Can you? I would be willing to bet most people that come in here can't afford that so we keep things safe. For that reason I will not post my voltages and temps of when I push that hard and the same can be said for CD as well as others nor recommend to do so.

the basics of 62c/72c is i am going to feel REALLY, REALLY bad if I fry another guys gear.
if I want to risk mine that is up to me, I just can't risk the others gear.

caddi daddi said:

the basics of 62c/72c is i am going to feel REALLY, REALLY bad if I fry another guys gear.
if I want to risk mine that is up to me, I just can't risk the others gear.

Click to expand...

Exactly. I would bet that most of the newbies go above and beyond as well be it by accident or design.

So my question stands in line with this thread, how far can you push it beyond the "safety margin" KNOWING FULL WELL that you're shaving lifespan off a chip (and can possibly blow up/melt your setup) and still call it a 24/7 "safe-ish" overclock. Use a 63**/83** for comparison testing and make a chart/list of the results ?

Maybe put this in a no-newbies thread or something or put a disclaimer like the OP

The problem, guys, correct me if I am wrong, is not so much processor life and preventing magic smoke, but the instability that tends to happen for a lot of overclockers above that value...

We tell people in the intel world about 10c less than its throttling point because of the same reason...you can see instability creep in above it not because it will kill anything.

A no newbies thread?? Like anyone would listen to that!

What usually happens more often, magic smoke or CPU dying ? Instability that seems to be corrected with more vcore which adds to temp, so vicious circle, yes. Soooooooooooooo question remix : how far can you push it beyond the "safety margin" KNOWING FULL WELL that you can possibly blow up/melt your setup and still call it a 24/7 "safe-ish" overclock.

I would say from personal experience with my 8370 ~75c, but would ofc need to test it further, hence the new thread for anyone wanting and willing to participate.

If a chip has throttling in it, I would say up to that temp. There is a reason that temperature throttling is there so it doesn't go up in smoke.

The reason why most say 62c is because of instability above that point, not because the cpu will die.

I'm not sure how you are going to get the data you are looking frlord though (it's relevence). But I'm likely missing something.

EarthDog said:

If a chip has throttling in it, I would say up to that temp. There is a reason that temperature throttling is there so it doesn't go up in smoke.

The reason why most say 62c is because of instability above that point, not because the cpu will die.

I'm not sure how you are going to get the data you are looking frlord though (it's relevence). But I'm likely missing something.

Click to expand...

You're not missing anything IMO.

Ok so it's up to silicon lottery again ? Some chips are perfectly fine running 80c 24/7, others spaz out the second they hit 63c + (or sometimes even less) ?

All new chips have additional overheating protection. FX will shut down at 100*C+ but before that will throttle. The same Intel chips. There is also additional protection from motherboard and additional software. It's really hard to burn CPU because of overheating. It simply won't let you.
You can kill CPU because of overvoltage but even that is harder with each new generation.

How high temps CPU will tolerate without generating errors depends from chip itself. There is no clear rule how high it will run. The same as graphics chips sometimes are scalling better or worse at higher/lower temps or voltage. However if you have some base of results then you can tell that most chips can work at about +15-20*C above temps declared by manufacturer ( just an example ).

That was what I was trying to do, acquire a baseline before throttle not CPU death.

If you guys would take notice to TCase temps.... The higher binned chip requires a lower temp.

FX-8370E
tCTLMax = 70
tCaseMax = 70.5°C
THERMTRIP on tCTL scale = 85

FX-8370
tCTLMax = 70
tCaseMax = 61.1°C
THERMTRIP on tCTL scale = 85

FX-9590
tCTLMax = 70
tCaseMax = 57.0°C
THERMTRIP on tCTL scale = 85

Observe the TCase on that FX-9590 clocked 4.7ghz with 5.0ghz turbo!! And people wonder why stability issues occur so quickly once reaching these type of clock speeds without adequate cooling.

In example, OP has running at most 4.4ghz with about 1.5000v. In this case with low clocks (ignoring voltage used) he may get away with TCase temps well into the 70c mark.

However as you overclock your processors, the goal is always to run a cooler temp with a higher clock speed.

Thermtrip on FX is still 85c. Core Temps displays TJMax of 90c. PC will shut down between 85-90c core temps and there is no work around.

Case in point..... Don't try running 4.8ghz at 80c. It's not even advised to run TCase anywhere near that number. This type of information should be at the Original post for factual purposes and disregard the statement.... "Look how FX is tougher than you think" while running such a low overclock with way too much voltage.

I'm not sure how you are going to get the data you are looking for...

Click to expand...

Well start here perhaps?

http://amd-dev.wpengine.netdna-cdn.com/wordpress/media/2012/10/42301_15h_Mod_00h-0Fh_BKDG1.pdf

http://support.amd.com/TechDocs/43374.pdf

• VID_VDD. The VID_VDD voltage is the VID-requested VDD supply level. Refer to the BIOS and
Kernel Developer’s Guide (BKDG) for AMD Family 10h Processors, order# 31116, for VID to
voltage translation specifications.

plenty of information out there.

EDIT:

This one is pin layout for AM3-
http://wfcache.advantech.com/www/certified-peripherals/documents/96mpaa-2.8g-2mam3t_Datasheet.pdf

A good read on AMD AM3 thermal Design Power
https://en.wikipedia.org/wiki/Thermal_design_power

Has there ever been a CPU that liked cold as much as the FX

Kenrou said:

Has there ever been a CPU that liked cold as much as the FX

Click to expand...

Yes actually Phenom II responded well to cold. Not as well as FX, but clocks exceeding 7ghz was possible.
The picture at HWbot is my 965BE de-lidded. I did not make 7ghz with it though lol....
http://hwbot.org/hardware/processor/phenom_ii_x4_965_be/

Also, I believe that Intel responded quite well in the 775 days as the 8ghz barrier was broken on a celeron I believe.

Kenrou said:

That was what I was trying to do, acquire a baseline before throttle not CPU death.

Click to expand...

thats.. well, tough at best. It would take forever to gather the data.. to the point where, imo, we use the same rules of thumb from our already gathered experiences. I don't even think you mentioned what temp we are talking about and what application to read it from. You talking hwmonitor and the cores or package? I don't know amd well, but do know that you need to specify what you want and it needs to be relevant (not sure what that is specifically..Intel guy here)

@ shrimp - good refresher on the info. But that isn't the data we need/that kenrou was trying to gather. In the end though, did you notice their thermal shutdown was all the same @ 85c?