• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** Official AMD Phenom II X6 1055T & 1090T Overclocking Thread ***

1273423728.png


I'll push further soon
 
alexp999 said:
i know amd said its right one, Im just saying to take temp readings with a pinch of salt.

Coretemp gives you the "core" temp, yet goes on about only being one sensor

Amd reply to you saying its not core temp you need to watch.

On Intels, CPU temps is the IHS temp, core temp is exactly that. Seems to be dispute and confusion over phenom II sensors.

That was my meaning behind saying we dont know which is which, cus we cant say for sure exactly where each is being read, can just speculate and hope for best.

Id say as long as both of the two temps are under 62*C stress testing we should be ok.
Click to expand...

I will attempt to answer alexp999 about which is the right CPU temperature reading, leave your system idle for a few minutes, until the temperatures drops, and then run Fritz Chess Benchmark. Whichever reading jumps up the fastest and cools the quickest refers to the Core Reading.

The Core Reading is the internal temperature from their die inside the chip i.e. the silicon. Unlike Intel Chips AMD have no such thermal sensors on the die therefore any so called core readings on AMD Chips are theoretical. AMD Chip thermal sensor monitors the Chip itself as a whole which will be the heat from all six cores. This reading lags behind the Core Reading but it is the one you need to watch in regard to the 62 degrees max.

With 6 cores instead of 4 the lag between the Core and AMD Chip readings will be greater. So the reading which lags the most is the most genuine reading from the AMD Chip itself.

I understand where alexp999 confusion comes from. I use Speed Fan which shows the AMD Chip Temp as the Core Temp. Just stick with reading that shows the greatest lag in heating up and cooling then you cannot go wrong. This advice applies to AMD only.

-----------------------------------------------------------------
Gigabyte GA-MA790FX-UD5P, 955 BE(C2)@3.77GHz 1.45V, Geil Value 4GB DDR3-1600,
ATI 5850 1GB, CM90 II Advanced, 1TB Western Digital Caviar Black, 4 HDDs, Jeantech Storm 700w,
Asus Xonar Essence STX, Pinnacle PCTV 3010iX Dual Analogue + DVB-T.
 
Ok, just run Fritz like you said, the "core" temp, as reported by everest and the program CoreTemp, shot up massively as it went under load, and dropped massively as the test finished.

So I guess that means "core" temp is in fact the true sensor reading on the phenom?

If thats the case, why would AMD say to monitor cpu temp, and why would the mobo be set not to use the cpu's sensor?

Is there still a socket sensor like back in the old days? So "cpu" temp is in fact a mobo sensor?

Im still a bit :confused: lol.
 
mayebaza said:
I will attempt to answer alexp999 about which is the right CPU temperature reading, leave your system idle for a few minutes, until the temperatures drops, and then run Fritz Chess Benchmark. Whichever reading jumps up the fastest and cools the quickest refers to the Core Reading.

The Core Reading is the internal temperature from their die inside the chip i.e. the silicon. Unlike Intel Chips AMD have no such thermal sensors on the die therefore any so called core readings on AMD Chips are theoretical. AMD Chip thermal sensor monitors the Chip itself as a whole which will be the heat from all six cores. This reading lags behind the Core Reading but it is the one you need to watch in regard to the 62 degrees max.

With 6 cores instead of 4 the lag between the Core and AMD Chip readings will be greater. So the reading which lags the most is the most genuine reading from the AMD Chip itself.

I understand where alexp999 confusion comes from. I use Speed Fan which shows the AMD Chip Temp as the Core Temp. Just stick with reading that shows the greatest lag in heating up and cooling then you cannot go wrong. This advice applies to AMD only.
.
Click to expand...
yeah thats right

amd says basicly core temp reading is nothing.. the one to watch is cpu temp reading..
 
Last edited:
Oh how muddy is this water !!
My take on it is this.

The true temp sensor is the one reading die temp, sometimes referred to as tcase.
It is this we should monitor.

Coretemp the program, AOD, Everest etc measure this but in the case of the latter 2, display it somewhat misleadingly as individual 'coretemps'

So most of us running at 4GHZ with 1.4 - 1.45 vcore are seeing high 40's low 50's so a fair bit of headroom there.

Now we have the ASUS BIOS reading higher than this, run up AI suite or any program that reads TMPIN2 and you see much higher readings.
BTW the version of Everest i am running displays this as AUX temp.

Why would the MB use the wrong temp sensor ? and make thermal throttling decisions based on it.

Well i have a theory.

What happens when you use core unlocking ?
The temp sensor measuring die temp stops working.

I tried this on my Crosshair and although i had no more cores to unlock just enabling it disabled my die temp sensor.
Coretemp, AOD etc read my temps as 0.

Now if the MB was using the die temp sensor to monitor CPU thermals there would be a BIG problem if someone used the core unlocker.
They could go looney tunes with vcore and the MB would have no way to know when to apply the brakes or shut the PC down.

To counter this scenario MB manafacturers appear to be using an alternative temp sensor that still functions when core unlocking is used.
It reads higher but at least it always reads.
I believe this is what is referred to as TMPIN2.

It would appear that when pushing our CPU's that the MB is applying the brakes based on readings from TMPIN2.
We still have headroom accoring to die temp readings but the TMPIN2" readings are high and the MB goes into 'nanny' mode.
 
I've decided on buying the Crosshair iv now and then the 1055T later which has the advantage of seeing whether AMD releases anything new in the meantime:D So much love for AMD ;)
 
PCZ said:
Oh how muddy is this water !!
My take on it is this.

The true temp sensor is the one reading die temp, sometimes referred to as tcase.
It is this we should monitor.

Coretemp the program, AOD, Everest etc measure this but in the case of the latter 2, display it somewhat misleadingly as individual 'coretemps'

So most of us running at 4GHZ with 1.4 - 1.45 vcore are seeing high 40's low 50's so a fair bit of headroom there.

Now we have the ASUS BIOS reading higher than this, run up AI suite or any program that reads TMPIN2 and you see much higher readings.
BTW the version of Everest i am running displays this as AUX temp.

Why would the MB use the wrong temp sensor ? and make thermal throttling decisions based on it.

Well i have a theory.

What happens when you use core unlocking ?
The temp sensor measuring die temp stops working.

I tried this on my Crosshair and although i had no more cores to unlock just enabling it disabled my die temp sensor.
Coretemp, AOD etc read my temps as 0.

Now if the MB was using the die temp sensor to monitor CPU thermals there would be a BIG problem if someone used the core unlocker.
They could go looney tunes with vcore and the MB would have no way to know when to apply the brakes or shut the PC down.

To counter this scenario MB manafacturers appear to be using an alternative temp sensor that still functions when core unlocking is used.
It reads higher but at least it always reads.
I believe this is what is referred to as TMPIN2.

It would appear that when pushing our CPU's that the MB is applying the brakes based on readings from TMPIN2.
We still have headroom accoring to die temp readings but the TMPIN2" readings are high and the MB goes into 'nanny' mode.
Click to expand...

Now that makes sense :)

Altho the actual temp sensor seems to vary from board to board, i.e. Whether its TMPIN 0/1/2/...etc
 
alexp999 said:
Im getting two conflicting things here, one person says the reading which moves most with load is one to watch, someone else says core temp is nothing. :confused:
Click to expand...
even mayebaza as basicly said theres no core sensor, the sensor inside the chip monitors the whole chip..

tmpin0 or tmpin1 (depending on motherboard) is the real cpu temp...

idle


underload



room temp 17c..
 
Last edited:
gareth170 said:
even mayebaza as basicly said theres no core sensor, the sensor inside the chip monitors the whole chip..

tmpin0 or tmpin1 (depending on motherboard) is the real cpu temp...

idle


underload



room temp 17c..
Click to expand...

Its called core by many progs, regardless of where it is, thats the one mayebaza identified by me running fritz.

I'm not saying it is in the core, but the "core" temp is the sensor inside the cpu chip.

I'm agreeing with PCZ, makes the most sense. In which case its prob as I thought, a sensor in socket or somewhere round the cpu to monitor temps.

And the "core" reading is AMDs sensor inside the chip itself. So you could argue that the one to be kept under 62*C is the one reported by most as "core" temp.

Altho I think its best to keep both temps below that.
 
well im keeping to what amd told me, to monitor the cpu temp reading, not core temp reading.

i've had no issues by doing that.

when i had a 965 i let the core temp go over 62c and it didn't throttle down. but if i let the cpu temp go over 62c it would throttle
 
gareth170 said:
well im keeping to what amd told me, to monitor the cpu temp reading, not core temp reading.

i've had no issues by doing that.

when i had a 965 i let the core temp go over 62c and it didn't throttle down. but if i let the cpu temp go over 62c it would throttle
Click to expand...

Ill be keeping to whatever is hotter, which seems to be the "cpu" temp. The one the BIOS reads.

The one the BIOS reads will be the one in control of throttling anyway.
 
Just got myself a 1090t and I cant overclock it :S I know what I am doing, it just wont overclock. Very strange.

I have updated to the latest BIOS and the CPU is supported. I have a CrossHairII (nvidia chipset, not AMD so i cant use AMD overdrive).

When I go into the BIOS and change the multiplier to anything above x16 it will just stay at 3.2ghz... any ideas?
 
core temps is just giving you the reading of the sensor.

They dont have to make sense, they are there to protect ur cpu.

To different sensors in different locations (imo) calibrated differently, Bound to give different readings. Not meant to be used like we use em for :)
 
jrwagh333 said:
Just got myself a 1090t and I cant overclock it :S I know what I am doing, it just wont overclock. Very strange.

I have updated to the latest BIOS and the CPU is supported. I have a CrossHairII (nvidia chipset, not AMD so i cant use AMD overdrive).

When I go into the BIOS and change the multiplier to anything above x16 it will just stay at 3.2ghz... any ideas?
Click to expand...

I had exactly the same problem with an ASUS M4N72-e nvidia 750 chipset.
Couldn't take the multi over 16.
If you tried it would boot on defaults.
ASUS did release a newer bios for the M4N72-E which does allow multi and vcore changes but it is still extremely buggy.
 
You must log in or register to reply here.
Back
Top Bottom