New_boy

Open doc... w00t? Yes open documentation, do we even need it?

I was working an escalation on an amd opteron workstation today, and while doing the basic elimination of known issues i just realised how lucky i am that i could easily access amd erratum pages.

Just go to amd.com choose product > processor > technical documentation and then revision guide
for example

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/25759.pdf

A nice simple pdf with all the known/public issues that amd has acknoledged,

in my case the customer was hitting erratum 154 with an incorrect diode ofset and was loging temperature events, that had already been documented in out internal bugster and the fix was already available.

That got me thinking does the big player intel have the same policy?
Tried google and the intel site and all i got was some old pentium sparse erratas, there was a collection of them on drdobs wesite as well but nothing official.

After a quick search on google i found no5 hit

http://www.sparc.com/standards/v9-errata.html

which lists more than 90 erratas for sparc v9 arch and in combination with sunsolve/opensparc.net
i found more than plenty bugs etc.

I was compelled to actually see if this is usefull information for anyone, and i realised after asking a few friends it is not. For the pure reason if you get hardware misbehaving you replace it.... I was quite amazed
how could people look for bugs/mistakes only in software when hardware/firmware could total your system.
This gave me an idea for the future to blog about techniques to investigate hardware failures.

Now the topic, since it all about openess i couch for open documentation for erratas, i have read endless stories about chipsets and bios "features" that annonyed many people.

And yes i leave myself wide open and i can see comments coming back saying you are not open etc etc.
Would be glad to hear as its good to know the other side.

Posted at 09:00πμ Δεκ 21, 2006 by Ioannis Psyllas in Sun  |  Comments[0]