Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Possible overheat vunerability in AMD X2 3800+ CPUs
37 views
Skip to first unread message
Skybuck Flying
May 14, 2013, 10:52:29 PM5/14/13
to
Hello,
While playing Compay of Heroes Tales of Valor I noticed the computer
shutdown a few times.
I am now starting to believe that 100% cpu utilization + 100% gpu
utilization could have caused a strange/rare case over CPU overheat.
(It could also simply be a cause of too many amperes being drawn).
However overheat cannot be ruled out.
Windows 7 is now a risk to overheat. Some of it's services will cause 50% to
100% cpu utilization, when running a game in the 50% case this will also
become 100% or even higher, when 100% is shown is it truely 100% ?, perhaps
different branches taken in CPU and true 100% cpu utilization is something
else than windows is showing but ok. Or how about: usually some threads my
relinquish control back or might start running a bit slower only when idle,
but because of defects in windows 7 they do not run while idle, but always
run.
Anyway now the question is: Could it be possible that AMD X2 3800+ dual core
CPUs have a overheat vunerability ?
This hypothesis is as follows:
The AMD CPUs were only stress tested with MAXIMUM CPU load.
The AMD CPUs were never stress tested in combination with MAXIMUM GPU load.
This could mean that the AMD CPUs have an OVERHEAT vunerability in there CPU
<-> GPU communication transistors section/design/chip area.
So the assumption is now:
To make an AMD CPU overheat it requires:
MAXIMUM CPU + MAXIMUM GPU utilization.
Now the question for CHIP/CPU/GPU designers is:
Could there be some thruth in this hypothesis ?
Is there indeed some section inside the AMD CPU which is responsible for
CPU<->GPU communication ? And could it be possible that this section
overheats the CPU ?
I do know that AMD CPUs have an integrated memory controller, these used to
be seperately and I have seen some motherboard designs that included special
heatsinks even fans to cool these MEMORY CONTROLLERS.
The memory controller inside AMD was called: "hyper transport" or something
like that... probably responsible for communication with main system ram,
this could also be the culprit.
Perhaps the AMD cpu also has a special transport towards the GPU ?
Bye,
Skybuck.
While playing Compay of Heroes Tales of Valor I noticed the computer
shutdown a few times.
I am now starting to believe that 100% cpu utilization + 100% gpu
utilization could have caused a strange/rare case over CPU overheat.
(It could also simply be a cause of too many amperes being drawn).
However overheat cannot be ruled out.
Windows 7 is now a risk to overheat. Some of it's services will cause 50% to
100% cpu utilization, when running a game in the 50% case this will also
become 100% or even higher, when 100% is shown is it truely 100% ?, perhaps
different branches taken in CPU and true 100% cpu utilization is something
else than windows is showing but ok. Or how about: usually some threads my
relinquish control back or might start running a bit slower only when idle,
but because of defects in windows 7 they do not run while idle, but always
run.
Anyway now the question is: Could it be possible that AMD X2 3800+ dual core
CPUs have a overheat vunerability ?
This hypothesis is as follows:
The AMD CPUs were only stress tested with MAXIMUM CPU load.
The AMD CPUs were never stress tested in combination with MAXIMUM GPU load.
This could mean that the AMD CPUs have an OVERHEAT vunerability in there CPU
<-> GPU communication transistors section/design/chip area.
So the assumption is now:
To make an AMD CPU overheat it requires:
MAXIMUM CPU + MAXIMUM GPU utilization.
Now the question for CHIP/CPU/GPU designers is:
Could there be some thruth in this hypothesis ?
Is there indeed some section inside the AMD CPU which is responsible for
CPU<->GPU communication ? And could it be possible that this section
overheats the CPU ?
I do know that AMD CPUs have an integrated memory controller, these used to
be seperately and I have seen some motherboard designs that included special
heatsinks even fans to cool these MEMORY CONTROLLERS.
The memory controller inside AMD was called: "hyper transport" or something
like that... probably responsible for communication with main system ram,
this could also be the culprit.
Perhaps the AMD cpu also has a special transport towards the GPU ?
Bye,
Skybuck.
MitchAlsup
May 14, 2013, 11:10:20 PM5/14/13
to
On Tuesday, May 14, 2013 9:52:29 PM UTC-5, Skybuck Flying wrote:
>
Perhaps a bigger heat sink and more airflow from cooler room conditions would be in order.
>
Perhaps a bigger heat sink and more airflow from cooler room conditions would be in order.
Skybuck Flying
May 15, 2013, 12:56:41 PM5/15/13
to
"MitchAlsup" wrote in message
news:f83a75ba-2699-4917...@googlegroups.com...
On Tuesday, May 14, 2013 9:52:29 PM UTC-5, Skybuck Flying wrote:
>
"
Perhaps a bigger heat sink and more airflow from cooler room conditions
would be in order.
"
What if there is a hotspot that's so hot that it cannot be cooled in that
manner ? ;)
Bye,
Skybuck :)
news:f83a75ba-2699-4917...@googlegroups.com...
On Tuesday, May 14, 2013 9:52:29 PM UTC-5, Skybuck Flying wrote:
>
"
Perhaps a bigger heat sink and more airflow from cooler room conditions
would be in order.
"
manner ? ;)
Bye,
Skybuck :)
MitchAlsup
May 16, 2013, 12:37:46 PM5/16/13
to
On Wednesday, May 15, 2013 11:56:41 AM UTC-5, Skybuck Flying wrote:
>
Then you are running at too high a Vdd.
>
Then you are running at too high a Vdd.
0 new messages