Temp rises 10 degrees after updating BIOS
- Thread starter Valhalla2000
- Start date
If you happen to be talking about the Asus A7V, then yes, everyone has experienced this when they updated the BIOS. Not sure WHY this is happening, but it seems to be a normal 'side effect'.
The Reason Kt7(and possibly other VIA boards) temps rose 10C or so, is because Asus and others(abit included) recognized that the under CPU Thermistor is faaaar too inaccurate to be relied on. It is far too variable, and can't be trusted on to show CPU temp. They tried to compensate with a 10C "compensation", but who knows if this is enough...
Mike
Mike
A probe under the socket, even if it is touching the cpu, produces inaccurate results, often much lower than what it actually is. For example, i was running between a soyo 7VCA, which has a thermistor underneath the socket, and a CUSL2, which reads a p3's internal diode. Under identical situations, the CUSL2's internal diode readings could be as high as 10-15C higher than an underneath socket thermistor.
Calling Asus stupid merely shows your ignorance. A lot of people seem to believe that this thermistor is accurate. Most people ignore the fact that with FC cpus, the hottest part contacts the heatsink. Even where the thermistor is touching the back of the cpu, you are only touching substrate. Still a vastly inaccurate reading.
You can't trust thermistor based readings from thermistors underneath the socket. They aren't accurate. This Compensation helps, but it still does not show fast changes in cpu temp(such as going from idle to 100% load in one second, which results in a quick cpu temp change).
Mike
Calling Asus stupid merely shows your ignorance. A lot of people seem to believe that this thermistor is accurate. Most people ignore the fact that with FC cpus, the hottest part contacts the heatsink. Even where the thermistor is touching the back of the cpu, you are only touching substrate. Still a vastly inaccurate reading.
You can't trust thermistor based readings from thermistors underneath the socket. They aren't accurate. This Compensation helps, but it still does not show fast changes in cpu temp(such as going from idle to 100% load in one second, which results in a quick cpu temp change).
Mike
Actually it is a good thing. It was rediculous to try and rate the CPU temp... by having a probe resting against it. When I changed my thermistor (made it sandwiched between the CPU and heatsink)... I got a 10degree higher reading.
A 10 degree adjustment should actually be just about right!
A 10 degree adjustment should actually be just about right!
Cyberdolphin,
10C is probably about right, but it highly depends on heatsink/other factors. There is even a potential problem with a thermistor jammed up against the side of the cpu. You could potentially be resting on the l2 cache(which runs at a much lower temp), and on top of that, the thermistor is also reading ambient temp(unless you seal it with some non-heat transferring material, with the only open spot touching the cpu..
).
But anywhere from 10-15C would probably still be a correct adjustment.
Mike
10C is probably about right, but it highly depends on heatsink/other factors. There is even a potential problem with a thermistor jammed up against the side of the cpu. You could potentially be resting on the l2 cache(which runs at a much lower temp), and on top of that, the thermistor is also reading ambient temp(unless you seal it with some non-heat transferring material, with the only open spot touching the cpu..
).But anywhere from 10-15C would probably still be a correct adjustment.
Mike
i dont understand why every company that includes sensors under the chip dosent just read it directly off the cpu, i mean.. its gonna be ALOT more accurate that way, and probably cheaper because they dont need to put a sensor on the board.
am i right or am i right?
am i right or am i right?
My P3B-F seems to be showing some high numbers regarding my cpu-temp. If I run a software from within Windoze it measures the temp 10-15 degrees below the Temp in the BIOS. And by the way, the temp in the BIOS seems too inaccurate.
It looks like they haven't found a solution to that problem
AoD
It looks like they haven't found a solution to that problem
AoD
<< i dont understand why every company that includes sensors under the chip dosent just read it directly off the cpu, i mean.. its gonna be ALOT more accurate that way, and probably cheaper because they dont need to put a sensor on the board.
am i right or am i right? >>
Oh yeah, you're right... too bad the AMD chips don't have an on-chip thermistor!
MikeW,
Check around the forums, the BIOS didn't just report a higher temp, the chip actually gets hotter. Some guy confirmed this with an indepentant probe and digital thermometer. The increases are normally under low cpu load, full load temps don't change much.
Check around the forums, the BIOS didn't just report a higher temp, the chip actually gets hotter. Some guy confirmed this with an indepentant probe and digital thermometer. The increases are normally under low cpu load, full load temps don't change much.
Yep, that guy was juggernaut i think.
He reports that the 1.003 bios when PCI Master read caching is enabled actually lowers the CPU temperature when idle, unlike the 1.004 bios. Something to do with HLT commands being issued. I wonder whether this problem would occur in Windows 2K?
He reports that the 1.003 bios when PCI Master read caching is enabled actually lowers the CPU temperature when idle, unlike the 1.004 bios. Something to do with HLT commands being issued. I wonder whether this problem would occur in Windows 2K?
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 21K
-
-
-
Facebook
Twitter