[SunHELP] SPARC Server 20...

[Michael Vang]

My "new" SPARCserver 20 died this morning...

***********************************************************************
          SMCC SPARCstation 10/20 UP/MP POST version VRV3.45 (09/11/95)


CPU_#0       TI, STP1021APGA(2.x)       1Mb External cache
CPU_#2       TI, STP1021APGA(2.x)       1Mb External cache

CPU_#1       ******* NOT installed *******
CPU_#3       ******* NOT installed *******


           >>>>> Power On Self Test (POST) is running .... <<<<<


CPU_00000000 >> error: dw stream data reg
    asi 2, addr 01c00000, exp 80000000 80000004, obs 80000000 00000004

***********************************************************************

My guess is it overheated when I ran my "validation suite"...

I know SS20s with 2xSM71s are at the limit thermal-wise, so I made sure
that the exhaust was not blocked and I made sure the cover was on when I
ran it... There were no other cards installed inside and only 1 stick of
memory... I also keep my home at 50% humidity and 72F and I have an APC
surge protector on everything... (No power conditioning yet...)

I usually test my boxes for 24-48 hours when I first get them... I run a
copy of Setiathome bound to each processor overnight... After that I run
a large mathematical calculation with a known answer using mlucas (Prime
number tester)... I figure if a computer can perform a 24-hour long
calculation on a number with literally millions of digits and get the
right answer then it is stable...

Setiathome is cool because it puts an incredible amount of stress on the
floating point unit and on the L2 cache...

I did check the temps last night real quick... The HDs were luke warm
and the CPUs were around 55C (I have calibrated my fingers +- 5C)... Too
bad there isn't a temperature monitor for the CPU dies...

Oh well... I've contacted the guy I bought it from to return it... I
don't anticipate any trouble, but just to be safe, do you think I
followed all the necessary precautions to prevent something like this
from happening? I feel that I should be able to run *any* program
without worrying about overloading the system to the point it locks up
or catches on fire...

Before I turned it on I read near every post Google has about the SS20,
so I was aware about the heat issue...

Finally, in reference to the SuperSPARC versus HyperSPARC issue, I found
a great reference here at SunHELP...

http://mbus.sunhelp.org/

I'm embarassed that I didn't see it before I posted that last message...

Thanks!