My New Computer. Impressed with the O/c.

[blackroseMD1]

On the subject of "how many FPS can we see" ... there is no definite answer simply because our eyes do not see "frames," but rather a continuous stream of light and motion. And yes, the threshold of FPS before we can't tell the difference is way way way WAY above 25.

That's very true...I have my avi movies running at 48 FPS, and it looks a lot more natural, especially during high action scenes.

 

[Lordbob75]

Dannyboy, you need to listen.

You WILL melt your CPU. Turn down the OC, and voltage.
And do some research about how to OC first, since you obviously don't know how to do it correctly.

~Lordbob

 

[dannyboy2005]

Dannyboy, you need to listen.

You WILL melt your CPU. Turn down the OC, and voltage.
And do some research about how to OC first, since you obviously don't know how to do it correctly.

~Lordbob

All the settings are set back to Default , nothing like 'wearing the chip in' lol

 

[BCXtreme]

Dannyboy, you need to listen.

You WILL melt your CPU. Turn down the OC, and voltage.
And do some research about how to OC first, since you obviously don't know how to do it correctly.

~Lordbob

All the settings are set back to Default , nothing like 'wearing the chip in' lol

More like "wearing the chip OUT". How long exactly did you have it running at the higher settings??

 

[dannyboy2005]

Dannyboy, you need to listen.

You WILL melt your CPU. Turn down the OC, and voltage.
And do some research about how to OC first, since you obviously don't know how to do it correctly.

~Lordbob

All the settings are set back to Default , nothing like 'wearing the chip in' lol

More like "wearing the chip OUT". How long exactly did you have it running at the higher settings??

lol. About 20 hours

 

[BCXtreme]

All the settings are set back to Default , nothing like 'wearing the chip in' lol

More like "wearing the chip OUT". How long exactly did you have it running at the higher settings??

lol. About 20 hours

Then hopefully you've avoided any permanent damage.

 

[pparks1]

if i drop the voltage, i sacrifice performance

I'm a little late to the party, but what good is that performance if it kills the chip in a month or so. Not much to brag about there.

Apparently every chip is unique in that way. No idea why.

Well, the processing of the chip and it's ability to run stable is what makes it the CPU it becomes. When Intel runs these off the line....chips that perform better are rated with higher speeds...while those that don't perform as stable get a lower clock speed. At least that is my understand. Thus, a Q9650 and a Q9550 might have come from same line..but one handled 3.0ghz better and the other is set to 2.83Ghz.

Your System Specs don't seem to indicate that you have any form of liquid cooling; in my experience, any quad-core will run hot without that. That's what burned out my last CPU (Core 2 Quad with fan only), and it wasn't even overclocked...

Not sure about that. I run a Core 2 Quad Q9550 at home with a minor overclock 2.83ghz to 3.2ghz (no voltage increase, just moved FSB to 400 and dropped from 8.5x to 8.0x). On the stock Intel cooler, I run low 30's at idle and mid-upper 50's under load. No liquid cooling needed.

 

[Lordbob75]

My CPU is definitely either messed up, or I have bad sensors. My idle temps sit at about 52C...

That is with a non-stock cooler too.

~Lordbob

 

[Fumz]

They don't call them "suicide runs" for nothing.

Hopefully with only 20 hours at 1.5 you did no lasting damage. Your first indicator that things weren't going as planned should have been your first benchmark. Yes, you were able to idle at 4.4GHz, but when you actually asked the cpu to do something (linpack), because it was so hot, it scaled itself back to 3.8GHz... thereby tossing out the window the theory that you would be "sacrificing performance" should you scale down from 4.4; you were never really there anyway.

There's a sticky at the top of this forum, you should read it, three or four times until you dream about it and can spit this stuff out without thinking. Then proceed to clock back up, at default voltage, to see how high you can go. You might be pleasantly surprised with the results.

... and yes, one does not need water cooling just because you want to clock a quad. There are many after market coolers which provide excellent cooling.

 

[chev65]

I just did a Stress test with IntelBurn Test. It breezed it no problem.

You can see that the stress test had used 3681Mhz (3.68Ghz) when stress testing, Why is this?

One scary thing is, the hottest the CPU got was 97*c. This was only for a second or 2 and probably only occurred twice in the whole process, but i herd that these chips have a very sensitive sensor's on them and a majority of the times, give a false reading? What would be a acceptable temperature when its under load (Which i guess it'll never be under 100% load ever again as i don't own any games that is that much FPS hungry.)

Testing with Prime Blend would not get nearly so hot although you will still need a better heat sink. That much Vcore won't work long term either. Four passes of IBT isn't even close to being stable, use Prime Blend test and run it 10 hours if you want it stable. Luckily your CPU throttled down the GHz to save it from being cooked although it still got way too hot. These are just the standard beginner mistakes, no harm "hopefully" no foul. Consider this a learning experience.

Overclocking has it's safe limits but you are well beyond them at this point. CPU Vcore isn't the only required adjustment for overclocking the i7. QPI and CPU vtt voltage play a major role here as well, getting those settings correct will allow you to lower your Vcore considerably. I would keep it under 70C although some guys go to 80C on the i7 while stress testing only.

 

[A Guy]

My CPU is definitely either messed up, or I have bad sensors. My idle temps sit at about 52C...

That is with a non-stock cooler too.

~Lordbob

Lordbob, are you using built in temperature monitoring? I don't trust those temps. Use something like Real Temp, or Core Temp. You will likely get a more accurate reading. Although it's not likely to be significantly lower than what you have. A Guy

 

[Lordbob75]

My CPU is definitely either messed up, or I have bad sensors. My idle temps sit at about 52C...

That is with a non-stock cooler too.

~Lordbob

Lordbob, are you using built in temperature monitoring? I don't trust those temps. Use something like Real Temp, or Core Temp. You will likely get a more accurate reading. Although it's not likely to be significantly lower than what you have. A Guy

I am using RealTemp

~Lordbob

 

[A Guy]

My CPU is definitely either messed up, or I have bad sensors. My idle temps sit at about 52C...

That is with a non-stock cooler too.

~Lordbob

Lordbob, are you using built in temperature monitoring? I don't trust those temps. Use something like Real Temp, or Core Temp. You will likely get a more accurate reading. Although it's not likely to be significantly lower than what you have. A Guy

I am using RealTemp

~Lordbob

Wow, and the V8? Do you have considerable variation between the cores? A Guy

 

[Lordbob75]

Lordbob, are you using built in temperature monitoring? I don't trust those temps. Use something like Real Temp, or Core Temp. You will likely get a more accurate reading. Although it's not likely to be significantly lower than what you have. A Guy

I am using RealTemp

~Lordbob

Wow, and the V8? Do you have considerable variation between the cores? A Guy

The V8? No, I have a different cooler (that one is in the sig because PooMan put it there, and it looks cool). It is not stock, but I would not say it is as good as the V8. Should be almost there though.

My cores are: 49, 51, 56, 56


~Lordbob

 

[smarteyeball]

I'm not sure if the Q8xx series had the same 'sticky/stuck/unreliable' sensors issue that the Q9xx did.

If they do, those idle core temps shown in realtemp may not be totally accurate. (they had a 'floor idle temp reading' that were reported at a 'stuck' temp when the actual temp was lower. *NB, as long as the load temps move and are within the limit, there's nothing to be fussed about.)

You could double check with another app that shows the CPU temp (tcase) - this temp should be roughly 10c lower than the core temps.

ie BIOS, Everest, HWmonitor, speedfan


Also, what are your current ambient temps like where you are?

 

[Lordbob75]

Seem to match:


I would image half of it is because of this:
1. I live in Arizona, and it is Summer. It is currently 00:30 right now, and still 91F (33C) outside.
2. My dad likes to keep the house warm to save money, so it is roughly 86F upstairs (where my room is)
3. My door is closed about half the time (I should check this...) and my computer heats my room very efficiently.

So I would image that is the biggest factors, but it still feels like it runs too hot.

~Lordbob

 

[Fumz]

Those idle temps do seem warm, but they're certainly within the "well and safe" zone. I'd be more interested to hear what your load temps are? The Q8 series TJ Max is 100C, and you show that you have Core Temp adjusted correctly, so unless load temps are above 70-75, then I think there may be little cause for concern?

 

[Lordbob75]

Those idle temps do seem warm, but they're certainly within the "well and safe" zone. I'd be more interested to hear what your load temps are? The Q8 series TJ Max is 100C, and you show that you have Core Temp adjusted correctly, so unless load temps are above 70-75, then I think there may be little cause for concern?

I actually don't think that my load temps get above about 65, but let me run a quick test.

Just started Prime 95 Blend test, and they immediately jumped up to 68 70 72 71 or so. Looking at that, I don't believe I have ever even hit 75C.

~Lordbob

 

[Fumz]

If you only have a 20C rise at load, when ~ 30 is the norm, then I'd say you're doing pretty good. I'd chalk up the high idle temps to ambient (summer in AZ).

 

[smarteyeball]

Yep, those load temps seem fine.

And with high ambient temps, a rise like that is pretty normal.

From memory, with 30c-ish ambient temps my Q9550 @ 3.8 1.3x? volts used to idle around the 40c+ -ish mark with a noctua NH-U12P in a push/pull config.