[OpenPower-Firmware] P8 overtemp behavior

[Daniel M Crowell]

I'm speaking a little outside my expertise but my understanding is that 
the systems are built such that it shouldn't have a thermal issue when 
nothing is running (checkstopped).  Any checking beyond that would be 
driven by the BMC side so you'd need to check with the different vendors 
specifically.

--
Dan Crowell
Senior Software Engineer - Power Systems Enablement Firmware
IBM Rochester: t/l 553-2987
dcrowell at us.ibm.com



From:   Sergey Kachkin <s.kachkin at gmail.com>
To:     Daniel M Crowell <dcrowell at us.ibm.com>
Cc:     openpower-firmware at lists.ozlabs.org
Date:   05/31/2018 04:00 AM
Subject:        Re: [OpenPower-Firmware] P8 overtemp behavior



Hi Daniel, 


thanks a lot for your reply, but AFAIK we do not power off completely in 
case of checkstop i.e there is still risk of damage. Am I wrong? 

regards,
Sergey 

On Thu, May 31, 2018 at 1:36 AM, Daniel M Crowell <dcrowell at us.ibm.com> 
wrote:
>From one of our power/thermal management folks - 
HTMGT/OCC does not power off due to CPU overtemp it was always stated that 
the system would checkstop due to thermal runaway before causing damage.  
We would rather let it checkstop in that real case vs. shutting systems 
down due to a bad DTS reading.  The BMC can decide to do a shutdown based 
on the CPU temperatures that they get from the OCC, I know in P9 the 
OpenBMC did add this I don't recall if AMI BMC did this in P8.

--
Dan Crowell
Senior Software Engineer - Power Systems Enablement Firmware
IBM Rochester
dcrowell at us.ibm.com



From:        Sergey Kachkin <s.kachkin at gmail.com>
To:        openpower-firmware at lists.ozlabs.org
Date:        05/25/2018 07:54 AM
Subject:        [OpenPower-Firmware] P8 overtemp behavior
Sent by:        "OpenPower-Firmware" <openpower-firmware-bounces+dcrowell=
us.ibm.com at lists.ozlabs.org>




Hi Team, 

I've got a question regarding a P8 overtemp behavior.    After processors 
reach 85C (OPEN_POWER_PROC_DVFS_TEMP_DEG_C) OCC start frequency 
throttling. 
 I'm trying to figure out what would happen if temperature continues to 
grow? (broken fan/ condition).

POWER8 User manual says: "Automated hardware thermal overtemperature 
protection is not supported. The real-time OCC firmware accomplishes this 
function."

Is there any emergency power off mechanism due to overtemp on P8? 

thanks, 

regards,
Sergey _______________________________________________
OpenPower-Firmware mailing list
OpenPower-Firmware at lists.ozlabs.org
https://lists.ozlabs.org/listinfo/openpower-firmware







-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ozlabs.org/pipermail/openpower-firmware/attachments/20180531/e516a651/attachment.html>