overclocking e4300 after Orthos failure

T

the_transporter

Limp Gawd
Joined
Aug 19, 2004
Messages
223
I oc'ed my cpu to 3GHz~ (333x9) at stock vcore 1.325 and Orthos failed after 4 hours and 30minutes. I used the blend test (stress cpu and ram). This is the message it gave me:

1:FATAL ERROR: Final result was A866A498, expected: 32472A99.
1:Hardware failure detected, consult stress.txt file.
1:Torture Test ran 4 hours, 32 minutes 30 seconds - 1 errors, 0 warnings.
1:Execution halted.

I don't quite understand what that means and stress.txt is nowhere to be found. It looks like something went wrong with the memory? I'm running the timings at 4-4-4-12. Do you guys have any suggestions on what I should do next to ensure a clean Orthos run of 8+ hours? Should I down the FSB, increase vcore to 1.350 (which I don't want to do, but probably have to). My CPU load went to 70c 70c using CoreTemp during the course of running Orthos.

TIA

Here's my system specs

Intel® Core™2 Duo e4300 @ 3.0GHz
Gigabyte 965P-DS3 rev.3.3
Corsair XMS2 2x1GB PC5400 DDR667
Western Digital Caviar SE16 250GB SATAII
eVGA 8800GTS 320MB
Enhance ENP-5150GH 500W
Windows Vista Business and Windows XP Pro dual boot
 
I have an E4300 which I run at 9x335.

I had to bump my vcore up one increment from the default 1.325 (can't remember what it is offhand). And that solved the problem.

It might also just be because 333 is an FSB strap change which I'm told you should avoid. 266, 333, 400, and 500 are the ones. But I dunno much about FSB straps. However they say to just bump the FSB a bit higher and try again.

Or you might need to do both.
 
Set the divider to 1:1 and then the ram should run at rated speed at 333 FSB. I imagine you did this already, so I doubt the RAM is the issue. Give the RAM a tiny bit more voltage (like 0.1 volts more, or less if the board will allow) and bump the vcore up by 1 to 3 increments. I bet you'll be stable then.
 
the_transporter: Most likely the vcore isn't enough. I have to use 1.35v to get 3GHz stable (e4300), so bump that up. So small an increasment isn't going to affect your temps or anything alot, so you don't have to worry about that.
 
Before adding any cpu voltage I would make sure your memory is at a minimum of 2.1 volts. All DDR2 likes that number and good DDR2 like the D9 chips don't mind a little bit more. :D
 
I will try that thanks. And also, is there any downsides to overvolting a processor? I know that there's an increase in temps, but with all that aside, does it degrade the quality of the processor over time? Since it's not meant to be overvolted if you know what I mean.
 
the_transporter said:
I will try that thanks. And also, is there any downsides to overvolting a processor? I know that there's an increase in temps, but with all that aside, does it degrade the quality of the processor over time? Since it's not meant to be overvolted if you know what I mean.
Click to expand...

In short: yes. Running the lowest possible voltage is good for longevity. That said, you aren't going to hurt it below 1.5vcore or more (this is a very conservative number, and I'm sure some will disagree, but it's certainly 'safe'), so a couple of little bumps is really nothing.

In most cases CPUs aren't getting their actual rated voltage at stock settings anyway, due to the tendency of motherboards to provide less voltage to the chip than the BIOS settings indicate, a phenomenon known as 'vdroop'. So you could have vcore set to 1.4v in BIOS and only get 1.35 true vcore. It's very difficult to measure well, however.

I don't agree with unclewebb about memory voltages, however. Run your chips at what they're rated for at stock speeds -- like 1.8v. If they're still unstable, 1.9v is fine, but increasing it further is unlikely to improve stability at stock speeds and will only create more heat, shortening the RAM's lifespan.
 
As long as you aren't going super crazy (i.e. 1.6v+ on air) with it there shouldn't be a prob. I have an E4300 and I believe that it took 1 bump in vcore to get 333*9 stable for me (1.350v).
 
You might want to invest in an aftermarket heatsink if you decide to bump the vcore and you're already hitting 70C (you didn't list one in your sepcs, so I assume it's the stock one).

Even an inexpensive Arctic Cooling Freezer 7 Pro, Cooler Master Hyper Tx or Scythe Mine (~$20 - ~$30 range + shipping for any of those - see this link for store/price links and further discussion) should get your temps under controll and reduce total system noise in the process.
 
I just clicked on those links and... a few are now out of stock and some prices may have changed by a buck or so, but those were the best prices I found at the time. I also saw on Newegg comments (not the most reliable) that the Scythe Mine might not fit the caps and such on the DS3, so if you're interested in that one, ask around first.
 
I also saw on Newegg comments (not the most reliable) that the Scythe Mine might not fit the caps and such on the DS3
Click to expand...

I use a Scythe Mine with a 120mm and have a HR-05 on the DS3 ;)
 
blade52x said:
I use a Scythe Mine with a 120mm and have a HR-05 on the DS3 ;)
Click to expand...

Thanks :). I figured the problem was between chair and keyboard (or motherboard, in this case).

BTW, are you pretty happy with the Mine's cooling on your E6600 overclocks?
 
Alright, everything seems fine with 1.350v and 333*9. I ran Orthos for 8 hours and 30 minutes. Temps were 70c 70c based on CoreTemp and I was wondering if there was any harm caused to the CPU being 70c 70c for 8 hours straight. I'm definitely getting a new cpu cooler this week, but was just wondering if any harm was caused just want some comments to ease my mind (that I didn't degrade my CPU or anything) if you know what I mean. And also, my load temps never goes over 55c 55c when I'm playing a game say GRAW2 demo. It seems that Orthos stresses my comp a lot, so the temps reach so high 70c 70c. I may hold off getting a cooler for now
 
the_transporter: How much memory voltage did you end up using?

My old Pentium III - 450 has been running for the last 8 years overvolted and overclocked and it doesn't seem to have hurt it's life expectancy too much.

Your temperatures and voltages are nowhere near serious.
When you deliberately force the processor into throttling by turning down the cpu fan while continuing to run Orthos, well now that is serious! :eek:

bakinghotqr4.png


It's also serious when you knock the heatsink off your processor and the damn thing gets so hot that temperature monitoring programs go tilt and start using negative numbers, well now, that is definitely getting serious!

test3gs6.png


You've got a ways to go yet. :D
 
unclewebb...WOW, you're 2C away from the Tjunction there! whew...now that's hot shit!

to OP: glad to hear the voltage increase helped get it stable. I would expect quite the temperature drop (across the board, load and idle temps) from almost any aftermarket cooler that you added. that little block of aluminum Intel provides is only meant to keep the processor cool at it's stock voltages and speeds. Overclocking on that thing, I'm surprised that you got it to 3GHz, honestly (although I know many others also report the same scenario, it would scare me too much). What the other poster has said about "safe" voltages, it is true for these that 1.5v on air is generally the maximum "safe" operating frequency on aftermarket cooling. I personally have pegged 1.55v as my ceiling on my E4300, as at this voltage, I am running stable (Orthos Small-FFT for 12+ hours) at just over 3.2GHz. To each his own, but certainly replace the cooler, it will help a ton. Pay heed to unclewebbs advice on the memory voltage. When OC'ing, and running 1:1, 2.1v is a good reference point that most memory is more than happy to run at. My Firestix require this to run stable at their stock rated speed, DDR2-8000. I have to peg 2.2 to get my current speed of 1198 @5-5-5-15 @2T. It's trial and error for the most part, but if you get an error, and are within your voltage threshhold, try stepping it up, one notch at a time. The key to good overclocking is doing one setting at a time. if you're chaning multiple things, and it starts working like you wanted, you won't know which setting it was that caused it (or combination thereof), and you may end up running something with more votlage than it needs, thus increasing the heat output...potentially effecting the life of the system (for those that are concerned with that), or at the very least, the stability of the OC. Good luck!
 
sabregen said:
unclewebb...WOW, you're 2C away from the Tjunction there! whew...now that's hot shit!
Click to expand...
If you look at the graph things actually got hotter than that.

The DTS is a 7 bit number. After it gets down to zero and the processor continues to heat up it will continue counting. 000 0000 binary wraps around and becomes 111 1111 binary which is equal to 127.

TAT and SpeedFan both continue to use this formula:
Reported Temp = TjMax - DTS

The TjMax for my processor is assumed to be 85C so when the DTS gets over loaded, this calculation becomes.
Reported Temp = 85 - 127 = -42

which is exactly what the graph is showing. I let my C2D continue to heat up and I think core0 might have hit 97C before I turned the fan back on and cooled things down. It never rebooted or complained once. This seems to have loosened it up a bit :D because now it is running better than ever.

http://www.hardforum.com/showpost.php?p=1031178478&postcount=670

Tough to hurt the Core 2 Duo. I've tried [H]ard-er than most.
 
unclewebb said:
~
Click to expand...
sabregen said:
~
Click to expand...

Thanks a lot for the info. I'll be definitely looking to get a better CPU cooler this weekend.

I just left the RAM voltage at stock 1.8v. The only voltage I increased was the CPU one to 1.350 from 1.325. I don't think I'm going to o/c anymore higher. My main goal was to run at 1:1 stably, which I seemed to have achieved. Do you still recommend me upping that RAM voltage? Even though it's running at its rated PC5400 667MHz speed? I know you guys are recommending it, but I don't see any need for it, since it's pretty much stable. Don't want to add anymore power to something that's already running pretty well if you know what I mean, but I will definitely up it up otherwise.
 
the_transporter said:
Thanks a lot for the info. I'll be definitely looking to get a better CPU cooler this weekend.

I just left the RAM voltage at stock 1.8v. The only voltage I increased was the CPU one to 1.350 from 1.325. I don't think I'm going to o/c anymore higher. My main goal was to run at 1:1 stably, which I seemed to have achieved. Do you still recommend me upping that RAM voltage? Even though it's running at its rated PC5400 667MHz speed? I know you guys are recommending it, but I don't see any need for it, since it's pretty much stable. Don't want to add anymore power to something that's already running pretty well if you know what I mean, but I will definitely up it up otherwise.
Click to expand...

The only reason to add more voltage to the ram would be if you want to run it faster and get it stable. If you're happy with that speed and it's stable, then leave it at 1.8v.
 
the_transporter said:
Thanks a lot for the info. I'll be definitely looking to get a better CPU cooler this weekend.

I just left the RAM voltage at stock 1.8v. The only voltage I increased was the CPU one to 1.350 from 1.325. I don't think I'm going to o/c anymore higher. My main goal was to run at 1:1 stably, which I seemed to have achieved. Do you still recommend me upping that RAM voltage? Even though it's running at its rated PC5400 667MHz speed? I know you guys are recommending it, but I don't see any need for it, since it's pretty much stable. Don't want to add anymore power to something that's already running pretty well if you know what I mean, but I will definitely up it up otherwise.
Click to expand...

Did you try intel TAT? it gives you a better idea of a cpu running HOT! orthos is nothing compared to it.
 
You must log in or register to reply here.
Back
Top