[GTALUG] Extending EOL by modding intel power attributes

[Russell]

On March 6, 2018 1:33:54 PM EST, "D. Hugh Redelmeier via talk" <talk at gtalug.org> wrote:
>| From: o1bigtenor via talk <talk at gtalug.org>
>| 
>| On Tue, Mar 6, 2018 at 7:44 AM, Russell via talk <talk at gtalug.org>
>wrote:
>| 
>| > Apparently Coffee Lake CPU's power attributes can be reset by
>overriding
>| > the management engine on Skylake MB's.
>| >
>| > http://www.overclock.net/forum/8-intel-general/1665657-
>| >
>coffeelake-completely-working-100-200-series-chipsets.html#/topics/1665657
>
>
>This is definitely interesting.  Peering into a box intended to be
>black.

That's what piqued my interest. From my recent reading. I can't remember the exact number of undocumented intel features but I believe the count was 60-70.

>
>Why would Intel intentionally obsolete a bunch of boards?  It would
>seem to reduce the number of CPU sales.  Perhaps there is something
>inferior about a system where the two pins are ignored (or whatever
>happens to the two pins after the hack).
>
>Intel might sell more support chips this way.  But they have to split
>the 
>proceeds of a motherboard sale with the motherboard manufacturer while 
>they get all the proceeds from a processor sale.
>
>I assume that the motherboards in question were meant for Kaby Lake
>CPUs (Kaby Lake preceded Coffee Lake).  After the firmware hack, could
>you still use a Kaby Lake processor?

Good question. It seems to be primarily a  voltage hack, perhaps the overclockable K series could withstand the burn?

>
>| Your quoted info seems specifically for a MSI product. As they are
>only one
>| of more than a few mobo suppliers I'm wondering as to the possibility
>of
>| using this on other manufacturer's products.

I haven't read the entire thread yet but I believe others have worked.

>
>My guess is that this blob within the firmware is supplied by Intel
>and is the same for all motherboards.  Except that Intel may make
>changes over time but the copy in a particular firmware may not be the
>latest.  But I would not want to act on a guess.  If I were trying
>this on another board, I'd first verify that that board's initial blob
>exactly matched MSI's initial blob.
>
>| Ideas, thoughts please?
>
>Why would you actually do this hack?

Personally I like tinkering and extending the life of the tech I use. At this point its more about the possibility of frankensteining something when prices bottom out and new old stock makes the rounds.
>
>- this is from reverse engineering, not specification.  There is no
>  guarantee that it is 100% functional.

There are warnings about the possibility of burning the cpu. So not for prime time or the faint of heart.

>
>- hacking firmware is the last thing you should try UNLESS you find it
>  a fun challenge.  It isn't usually an optimal route if there are
>  simpler choices.
>
>For example, buying a new motherboard and selling your old motherboard 
>  and old processor is simpler and may not be expensive.
>
>- if a new firmware is released, you have to ignore it or hack it all
>  over again, with a new set of risks.

This hack wouldn't be a production choice for any real enterprise. I just like the mix and match possibilities when the stuff hits salvage pricing. Definitely it's an end of life hail mary pass.

>---
>Talk Mailing List
>talk at gtalug.org
>https://gtalug.org/mailman/listinfo/talk

-- 
Russell