I will have to read up on how to save my bios so I can set it back, changed way too many things that I know nothing about...(ducking from Ned)

Ohh.. then better not to do it..

Normally (in my BIOSs) in the (1st) overview should be a point like SAVE & EXIT .
[EDIT: Maybe you have the possibility for to save one or two user settings.]


BTW.
You have the possibility of reading CPU and Core temp simultaneously?
So we could compare the CPU temp and TjMax. of the Cores (and my recommended > 10 °C CPU / >= ~ 20 °C Core)

I can but am about to leave for airport I will do temps when I get back with Grandaughter, and I do believe I can save bios profiles just haven't ever tried it. I downloaded real temp, are there any others you want me to try?

---

Official Abuser of Boinc Buttons...
And no good credit hound!

And You might like to read this thread here, Unclewebb knows quite a bit about Intel cpus, As He's the author of RealTemp. :D

While I appreciate your pointing out that someone has their own ideas on this subject, I'm not particularly interested in someone else's opinion.

The manufacturer (Intel) has done the work in the lab, and they publish specifications. If you exceed those specifications, you are inherently in a state of sin.

You may do what you wish.

Hiamps asked "what is the maximum temperature" and Intel publishes their recommended maximums. If he's running hotter, and he knows that, he can choose to take the risk, or he can add cooling.

One can't make an informed decision without information.

That's the thing, The temps measureability isn't quite as reliable as You'd think, As the top might be 100C or 105C with some cpus and different with others.

unclewebb wrote:

Intel also admits that there is a significant amount of error in how they set TJMax so the actual TJMax of your CPU is anyone's guess. There is no number written in stone and no documentation that is 100% accurate. Based on your results, my best guess is 100C or 105C. Many 45nm Core 2 desktop CPUs use 100C and the Core 2 mobile CPUs are documented as 105C. Flip a coin.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

I will have to read up on how to save my bios so I can set it back, changed way too many things that I know nothing about...(ducking from Ned)

Ohh.. then better not to do it..

Normally (in my BIOSs) in the (1st) overview should be a point like SAVE & EXIT .
[EDIT: Maybe you have the possibility for to save one or two user settings.]


BTW.
You have the possibility of reading CPU and Core temp simultaneously?
So we could compare the CPU temp and TjMax. of the Cores (and my recommended > 10 °C CPU / >= ~ 20 °C Core)

I can but am about to leave for airport I will do temps when I get back with Grandaughter, and I do believe I can save bios profiles just haven't ever tried it. I downloaded real temp, are there any others you want me to try?

If You must there's CoreTemp which is used for both Intel and AMD cpus, The RealTemp author has no AMD cpus or even Atom cpus for that matter.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

I can but am about to leave for airport I will do temps when I get back with Grandaughter, and I do believe I can save bios profiles just haven't ever tried it. I downloaded real temp, are there any others you want me to try?

Then come safe back.

If RealTemp or CoreTemp, both are well for to measure the Core temps.
You don't have a prog for to measure the CPU temp?

For to measure the CPU temp I used SpeedFan or the mobo manufacturer software.

---

Also..
There are progs out there which can control/measure all your equipment simultaneously.
CPU/Core/mobo/HDD.. and so on. And can say the whole settings.
Some can make also benchmarks.

But I can't remember well which I used in past..

I guess it was SiSoft Sandra.
Don't know if this is now a free- or maybe 'shareware prog'_(after some time few things disabled, you must pay little bit for to unlock).

---

Ok my Asus PC Probe says I am at 64C and Real temp says 64,62,62,58 They fluctuate but that is the average, so Asus's tool is pretty accurate. That is cool to see.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

Are you sure that the ASUS prog say the CPU-Core temp?
I guess this is the CPU temp.

The max. CPU temp is 67.9 °C for the 920.

[Message 958655] here in this thread.

This Core temp stuff is nothing officially from Intel.

So I wouldn't go higher than ~ 58 °C CPU temp if 24/7 crunching.

---

And You might like to read this thread here, Unclewebb knows quite a bit about Intel cpus, As He's the author of RealTemp. :D

While I appreciate your pointing out that someone has their own ideas on this subject, I'm not particularly interested in someone else's opinion.

The manufacturer (Intel) has done the work in the lab, and they publish specifications. If you exceed those specifications, you are inherently in a state of sin.

You may do what you wish.

Hiamps asked "what is the maximum temperature" and Intel publishes their recommended maximums. If he's running hotter, and he knows that, he can choose to take the risk, or he can add cooling.

One can't make an informed decision without information.

That's the thing, The temps measureability isn't quite as reliable as You'd think, As the top might be 100C or 105C with some cpus and different with others.

unclewebb wrote:

Intel also admits that there is a significant amount of error in how they set TJMax so the actual TJMax of your CPU is anyone's guess. There is no number written in stone and no documentation that is 100% accurate. Based on your results, my best guess is 100C or 105C. Many 45nm Core 2 desktop CPUs use 100C and the Core 2 mobile CPUs are documented as 105C. Flip a coin.

I don't disagree. The problem is there is only one way to determine the maximum junction temperature for any single processor: destructive testing.

You get it too hot, and when it fails, you know what the maximum temperature was for that particular chip.

Intel takes a bunch of chips, and runs them at various temperatures. At 100c (for example) some of them live, and some fail, and 80c, more live and fewer fail, and so on. They decide how many chips they're willing to replace under warranty (it's a small number, so their reputation is untarnished) and they publish the temperature based on that test.

It's the same with overclocking. When they stamp 2 GHz on a chip, it generally means it tested clean at 2.1 GHz, and so they have some "margin" -- they're sure that signals will be stable and decisions are made on a solid "1" or "0" and not "0.3" or "0.7" as the signal slews from one level to another.

You can overclock, and you can let the temperatures climb, and probably most of the chips will be okay.

But if you're wrong, you have a $280 paperweight.

---

And You might like to read this thread here, Unclewebb knows quite a bit about Intel cpus, As He's the author of RealTemp. :D

While I appreciate your pointing out that someone has their own ideas on this subject, I'm not particularly interested in someone else's opinion.

The manufacturer (Intel) has done the work in the lab, and they publish specifications. If you exceed those specifications, you are inherently in a state of sin.

You may do what you wish.

Hiamps asked "what is the maximum temperature" and Intel publishes their recommended maximums. If he's running hotter, and he knows that, he can choose to take the risk, or he can add cooling.

One can't make an informed decision without information.

That's the thing, The temps measureability isn't quite as reliable as You'd think, As the top might be 100C or 105C with some cpus and different with others.

unclewebb wrote:

Intel also admits that there is a significant amount of error in how they set TJMax so the actual TJMax of your CPU is anyone's guess. There is no number written in stone and no documentation that is 100% accurate. Based on your results, my best guess is 100C or 105C. Many 45nm Core 2 desktop CPUs use 100C and the Core 2 mobile CPUs are documented as 105C. Flip a coin.

I don't disagree. The problem is there is only one way to determine the maximum junction temperature for any single processor: destructive testing.

You get it too hot, and when it fails, you know what the maximum temperature was for that particular chip.

Intel takes a bunch of chips, and runs them at various temperatures. At 100c (for example) some of them live, and some fail, and 80c, more live and fewer fail, and so on. They decide how many chips they're willing to replace under warranty (it's a small number, so their reputation is untarnished) and they publish the temperature based on that test.

It's the same with overclocking. When they stamp 2 GHz on a chip, it generally means it tested clean at 2.1 GHz, and so they have some "margin" -- they're sure that signals will be stable and decisions are made on a solid "1" or "0" and not "0.3" or "0.7" as the signal slews from one level to another.

You can overclock, and you can let the temperatures climb, and probably most of the chips will be okay.

But if you're wrong, you have a $280 paperweight.

Or one that originally cost a few times more than that of the $280 paperweight, Mine hits 81-82C when the heat is on or in summer, But I need to filter a case and get the Corsair H50 for It, It's a dusty area.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

I don't disagree. The problem is there is only one way to determine the maximum junction temperature for any single processor: destructive testing.

You get it too hot, and when it fails, you know what the maximum temperature was for that particular chip.

Intel takes a bunch of chips, and runs them at various temperatures. At 100c (for example) some of them live, and some fail, and 80c, more live and fewer fail, and so on. They decide how many chips they're willing to replace under warranty (it's a small number, so their reputation is untarnished) and they publish the temperature based on that test.

It's the same with overclocking. When they stamp 2 GHz on a chip, it generally means it tested clean at 2.1 GHz, and so they have some "margin" -- they're sure that signals will be stable and decisions are made on a solid "1" or "0" and not "0.3" or "0.7" as the signal slews from one level to another.

You can overclock, and you can let the temperatures climb, and probably most of the chips will be okay.

But if you're wrong, you have a $280 paperweight.

The reason I'm not real worried is because I belong to a few overclocking forums and they have been out awhile. If there was a problem with them frying it would be all over them. Did a google search and nothing there either. I promise you I am no where near the voltage or temps that some of these guys do. But like you said, it could end up a paperweight. LOL I have overclocked every chip I have ever owned in my play machines. Been a boatload. Still have my Peltiers in the closet.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

Heres my coretemp screenshot.

HT Off Now Asus says 53C

---

Official Abuser of Boinc Buttons...
And no good credit hound!

I don't disagree. The problem is there is only one way to determine the maximum junction temperature for any single processor: destructive testing.

You get it too hot, and when it fails, you know what the maximum temperature was for that particular chip.

Intel takes a bunch of chips, and runs them at various temperatures. At 100c (for example) some of them live, and some fail, and 80c, more live and fewer fail, and so on. They decide how many chips they're willing to replace under warranty (it's a small number, so their reputation is untarnished) and they publish the temperature based on that test.

It's the same with overclocking. When they stamp 2 GHz on a chip, it generally means it tested clean at 2.1 GHz, and so they have some "margin" -- they're sure that signals will be stable and decisions are made on a solid "1" or "0" and not "0.3" or "0.7" as the signal slews from one level to another.

You can overclock, and you can let the temperatures climb, and probably most of the chips will be okay.

But if you're wrong, you have a $280 paperweight.

The reason I'm not real worried is because I belong to a few overclocking forums and they have been out awhile. If there was a problem with them frying it would be all over them. Did a google search and nothing there either. I promise you I am no where near the voltage or temps that some of these guys do. But like you said, it could end up a paperweight. LOL I have overclocked every chip I have ever owned in my play machines. Been a boatload. Still have my Peltiers in the closet.

The cpu that runs this cpu is a QX6700 B1(ES) or a prototype non production cpu w/a stock speed of 2.66GHz(10 multiplier), I run It at 3.52GHz(8 multiplier), I've heard of cpus of this type running as fast as 4GHz on Air, But I have no idea what the temp was only the fact they used a big heatsink at that speed and the multipliers on the QX cpus are unlocked.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

Intel states that the DTS value, the way most programs measure the core temperature, should not be used for temperature readings.
It's meant for power control regulations.

That is, I think it's a good indicator. I use it in my own TThrottle program as well.
But the best indicator is the case temperature.
My board has a nice blinking light indication a too high CPU temperature.
And a shutdown at 100C core temperature. I once had the pleasure of using that feature when a cooler was not properly seated.

I've seen a lot of data from other computers 80-90 for a I7 should be ok.

---

TThrottle Control your temperatures. BoincTasks The best way to view BOINC. Anza Borrego Desert hiking.

[...]
I promise you I am no where near the voltage or temps that some of these guys do.
[...]

Just curious..
Which is the stock Vcore and at which Vcore you are now?
(in BIOS and real in Windows)


I OC only to the max. without changing the Vcore.
More wouldn't be well for the performance/wattage ratio.

So I didn't changed the Vcore of my QX6700 (2.67 GHz) for to reach 3.14 GHz.
BIOS - 1.35 V
In Windows - ~ 1.26 - 1.27 V
All stock the CPU got ~ 0.01 V (~ 1.27 - 1.28) more.

So with OC my QX6700 will live longer, because of less real voltage..

The wattage of the whole system at the wall plug increased ~ 25 W.
The CPU temp increased ~ 3 °C.

---

The cpu that runs this cpu is a QX6700 B1(ES) or a prototype non production cpu w/a stock speed of 2.66GHz(10 multiplier), I run It at 3.52GHz(8 multiplier), I've heard of cpus of this type running as fast as 4GHz on Air, But I have no idea what the temp was only the fact they used a big heatsink at that speed and the multipliers on the QX cpus are unlocked.

You speak about my QX6700 ?

It's a normal bought CPU in a shop.
It's one of the first which was sold in Germany.
I was 'crazy' for some years and must have a QUAD-CPU so early as possible.. and this only because of SETI@home..
(To this time AMD had only a two socket DUO-CPU system.. And in my eyes it wasn't ready for to release/sell..)

And payed 1,050.- € (now ~ 1,500.- US$) for this toy/joy.. ~ 2,000.- € (now ~ 2,900.- US$) for the whole system.


BTW.
Maybe I got a prerelease CPU. Because if I change the multi in BIOS, Windows see the speed up (different GHz) but in benchmarks (also SETI WU crunching) it's the same or less than stock.
Maybe there is a BUG in the CPU or the Intel D975XBX2 mobo.
So I'm at 314 x 10 CPU (stock 266 x 10) & DDR2 944/5-5-5-15 (stock DDR2 800).

---

The cpu that runs this cpu is a QX6700 B1(ES) or a prototype non production cpu w/a stock speed of 2.66GHz(10 multiplier), I run It at 3.52GHz(8 multiplier), I've heard of cpus of this type running as fast as 4GHz on Air, But I have no idea what the temp was only the fact they used a big heatsink at that speed and the multipliers on the QX cpus are unlocked.

You speak about my QX6700 ?

It's a normal bought CPU in a shop.
It's one of the first which was sold in Germany.
I was 'crazy' for some years and must have a QUAD-CPU so early as possible.. and this only because of SETI@home..
(To this time AMD had only a two socket DUO-CPU system.. And in my eyes it wasn't ready for to release/sell..)

And payed 1,050.- € (now ~ 1,500.- US$) for this toy/joy.. ~ 2,000.- € (now ~ 2,900.- US$) for the whole system.


BTW.
Maybe I got a prerelease CPU. Because if I change the multi in BIOS, Windows see the speed up (different GHz) but in benchmarks (also SETI WU crunching) it's the same or less than stock.
Maybe there is a BUG in the CPU or the Intel D975XBX2 mobo.
So I'm at 314 x 10 CPU (stock 266 x 10) & DDR2 944/5-5-5-15 (stock DDR2 800).

No I don't, Mine. When I got Mine production models were not out yet.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

[...]
I promise you I am no where near the voltage or temps that some of these guys do.
[...]

Just curious..
Which is the stock Vcore and at which Vcore you are now?
(in BIOS and real in Windows)


I OC only to the max. without changing the Vcore.
More wouldn't be well for the performance/wattage ratio.

So I didn't changed the Vcore of my QX6700 (2.67 GHz) for to reach 3.14 GHz.
BIOS - 1.35 V
In Windows - ~ 1.26 - 1.27 V
All stock the CPU got ~ 0.01 V (~ 1.27 - 1.28) more.

So with OC my QX6700 will live longer, because of less real voltage..

The wattage of the whole system at the wall plug increased ~ 25 W.
The CPU temp increased ~ 3 °C.

Going over my srttings the CPU voltage is set at 1.30 but in my Asus tool it shows as 1.25 which I believe is stock. CPUID says I am at 1.248 so I am not worried here also. I paid $269.00 it is worth it to me to play.

---

Official Abuser of Boinc Buttons...
And no good credit hound!

I don't disagree. The problem is there is only one way to determine the maximum junction temperature for any single processor: destructive testing.

You get it too hot, and when it fails, you know what the maximum temperature was for that particular chip.

Intel takes a bunch of chips, and runs them at various temperatures. At 100c (for example) some of them live, and some fail, and 80c, more live and fewer fail, and so on. They decide how many chips they're willing to replace under warranty (it's a small number, so their reputation is untarnished) and they publish the temperature based on that test.

It's the same with overclocking. When they stamp 2 GHz on a chip, it generally means it tested clean at 2.1 GHz, and so they have some "margin" -- they're sure that signals will be stable and decisions are made on a solid "1" or "0" and not "0.3" or "0.7" as the signal slews from one level to another.

You can overclock, and you can let the temperatures climb, and probably most of the chips will be okay.

But if you're wrong, you have a $280 paperweight.

The reason I'm not real worried is because I belong to a few overclocking forums and they have been out awhile. If there was a problem with them frying it would be all over them. Did a google search and nothing there either. I promise you I am no where near the voltage or temps that some of these guys do. But like you said, it could end up a paperweight. LOL I have overclocked every chip I have ever owned in my play machines. Been a boatload. Still have my Peltiers in the closet.

It's all about informed decisions.

For most of the people here, if they fry a CPU, they have to buy a new one.

For me, if I fry a CPU, my income stops. For some of the machines here, if the CPU fries, it has to be replaced in the next few minutes, or customers leave and my income drops a lot.

I overclock during burn-in. I want to see the system run reliably at 10% over so when I go back to "stock" I know the whole system has a nice 10% margin.

I pick processors by performance an TDP. I like the lower power bills and longer UPS run times, but I also like systems that run cold. The new servers I'm building draw sixteen watts total.

I'm mostly concerned though by well-meaning "advice" from people who have relatively little to lose if they burn out a CPU. If it's a raw cruncher, not used for any critical purpose, and all you have to lose are some work units, then you can afford the risks.

---

Going over my srttings the CPU voltage is set at 1.30 but in my Asus tool it shows as 1.25 which I believe is stock. CPUID says I am at 1.248 so I am not worried here also. I paid $269.00 it is worth it to me to play.

Ahh.. in BIOS you set 1.30 Vcore.

Then in Windows you see the real voltage at the CPU, this is normally less than set in BIOS.

I have 1.35 in BIOS -> 1.265 in Windows -> ~ 0,085 difference.
It depend which system you have, some have less and some have more mobo Vcore loss.

You have 1.30 -> 1.248 -> ~ 0,052


You didn't changed the Vcore in your BIOS for to reach ~ 3.8 GHz (from stock 2.67)?

---

But if you're wrong, you have a $280 paperweight.

The reason I'm not real worried is because I belong to a few overclocking forums and they have been out awhile. If there was a problem with them frying it would be all over them. Did a google search and nothing there either. I promise you I am no where near the voltage or temps that some of these guys do. But like you said, it could end up a paperweight. LOL I have overclocked every chip I have ever owned in my play machines. Been a boatload. Still have my Peltiers in the closet.
[/quote]
It's all about informed decisions.

For most of the people here, if they fry a CPU, they have to buy a new one.

For me, if I fry a CPU, my income stops. For some of the machines here, if the CPU fries, it has to be replaced in the next few minutes, or customers leave and my income drops a lot.

I overclock during burn-in. I want to see the system run reliably at 10% over so when I go back to "stock" I know the whole system has a nice 10% margin.

I pick processors by performance an TDP. I like the lower power bills and longer UPS run times, but I also like systems that run cold. The new servers I'm building draw sixteen watts total.

I'm mostly concerned though by well-meaning "advice" from people who have relatively little to lose if they burn out a CPU. If it's a raw cruncher, not used for any critical purpose, and all you have to lose are some work units, then you can afford the risks.[/quote]

I agree, not one of my work computers is overclocked. I would only do that to my play computer. If this guy dies I will be very bummed and out a few bucks but business will go on as usual. I'm networked and go do lots of my work right here but everything is goes thru my server and is backuped even if this machine dies mid keystroke. You would be crazy to play with a mission critical machine like I do. I go nuts on my play machine so I can fix anything the girls do to our work machines, at times they keep me quite busy and with the only Microsoft RMS qualified people in the area based out of portland it would be cheaper to fry a couple machines than pay them travel time each way and the simple things it usually ended up being. Haven't had to call tech support )other than here) in quite a few years. Most of the time I am tied here so this gives me a lot of stress relief. I am fully aware that if I were to kill my machine it would be my fault and mean I would have to work for my older brother for a couple days to replace it and take his crap as he thinks I am NUTS to overclock. Believe me that is a fate I don't want, he loves to tell me "told you so" a few thousand times.

---

Official Abuser of Boinc Buttons...
And no good credit hound!