[Macchiato] MacchiatoBin GPU progress [#freetimeproject]

[Matt Spencer]

Understandable Leif!


Out of interest, I thought I would check to see if the response could be changed by modifying the memory clock speed.  I changed it to 2400 (0010 on the dip switches).  All was looking good as I hit 19 straight boots without failure, then the wheels fell of the cart and the next 20 boots started looking a lot like the earlier 1600 results.


Out of interest, is there any dynamic frequency scaling on the memory interfaces that could be causing this problem?


/Matt

________________________________
From: Leif Lindholm <leif.lindholm at linaro.org>
Sent: 09 June 2017 15:32:19
To: Matt Spencer
Cc: macchiato at lists.einval.com; Neta Zur Hershkovits
Subject: Re: [Macchiato] MacchiatoBin GPU progress [#freetimeproject]

I've ordered the same kind of riser (and think it's waiting for me in
the mail room), so I'll probbaly be testing it myself this weekend.

If it works, it's never coming out.
If not, I could bring it in once you're back :)

/
    Leif

On Fri, Jun 09, 2017 at 02:24:05PM +0000, Matt Spencer wrote:
> Sure, it Leif is Ok lending it to me?
>
>
> I won't be back in the office until the week of the 19th June, but Leif - can I borrow your card and see what results I get?
>
>
> /Matt
>
> ________________________________
> From: Ard Biesheuvel <ard.biesheuvel at linaro.org>
> Sent: 09 June 2017 15:22:30
> To: Matt Spencer
> Cc: marcin wojtas; Steve McIntyre; macchiato at lists.einval.com; Steve Capper; Leif Lindholm; Wookey; Igal Liberman; Neta Zur Hershkovits
> Subject: Re: [Macchiato] MacchiatoBin GPU progress [#freetimeproject]
>
> On 9 June 2017 at 13:54, Ard Biesheuvel <ard.biesheuvel at linaro.org> wrote:
> > On 9 June 2017 at 13:52, Matt Spencer <Matt.Spencer at arm.com> wrote:
> >> Ok, ran a reboot test as a background task whilst doing other things...
> >> Getting bored of it now - so here are the results (1=success 0=fail)
> >>
> >>
> >> Before/After atf/u-boot upgrade - (success/total) success % - cadence
> >>
> >>
> >> Before - (7/11) 54% - 1010011110001
> >>
> >> After - (50/60) 83% -
> >> 111011111111111101010111011011111111011001111111101111111111
> >>
> >>
> >> I wish the initial sample had more data in it, but at a superficial level,
> >> it looks like things have improved?
> >>
> >
> > Oh definitely! Only not to the level that would justify putting 'PCIe
> > support' on the box.
>
> Any chance you could try Leif's x1 graphics card on your box, with and
> without the riser?
> IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

> _______________________________________________
> Macchiato mailing list
> Macchiato at lists.einval.com
> https://lists.einval.com/cgi-bin/mailman/listinfo/macchiato

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.einval.com/pipermail/macchiato/attachments/20170609/fc70b5bc/attachment.html>