cvs commit: src/sys/i386/cpufreq est.c

[Nate Lawson]

Kevin Oberman wrote:
>> Date: Sat, 25 Feb 2006 15:42:06 -0800
>> From: Colin Percival <cperciva at freebsd.org>
>>
>> Kevin Oberman wrote:
>>>> cperciva    2006-02-25 04:55:39 UTC
>>>>   Log:
>>>>   Add frequency-voltage tables for Intel 778, 758, 773, 753, and 733J
>>>>   processors.
>>>>   
>>>>   Obtained from:  Intel Datasheet 302189-008
>>> Is this a new datasheet or did someone find it in some odd corner of
>>> Intel land? I'll admit that I'm baffled as to why Intel seemed to keep
>>> this information under wraps for so long.
>> Intel updates their datasheets from time to time -- the information for
>> these processors was added in a new version of an old datasheet.  Intel
>> still hasn't released the frequency/voltage tables for their 7x0 series
>> processors or the formula used to compute them.
> 
> Ack! I was not paying nearly enough attention! Still no joy for my 760
> CPU, then.
> 
> Thanks anyway. Hopefully some day it will be available. I still don't
> understand why Intel would withhold this sort of information on an
> out-of-date series of chips from those who would buy the chips! (Or, at
> least system containing them.)

I'm not defending this, but I think I know what happened.  The voltage 
to VID translation table was probably only present in the BIOS Writer's 
Guide.  This is because in the past, most of the code to use it was in 
the BIOS and the OS can control it via ACPI.  Intel doesn't publish the 
BIOS Writer's Guides.  To release the info, they probably had to extract 
just that table, get approval, etc.  The table itself doesn't seem that 
secret to me but I guess the bureaucracy could be the real problem.

-- 
Nate