mailarchive of the ptxdist mailing list
 help / color / mirror / Atom feed
From: Clemens Gruber <clemens.gruber@pqgruber.com>
To: ptxdist@pengutronix.de
Subject: Re: [ptxdist] ptxdist with iMX6
Date: Thu, 2 Mar 2017 17:17:41 +0100	[thread overview]
Message-ID: <20170302161741.GA16522@archie.localdomain> (raw)
In-Reply-To: <CABDcavaFGrwGxFuozM0e+V5owQ3=gUkqgSYU7bM-VzgodZauDA@mail.gmail.com>

Hi Guillermo,

On Thu, Mar 02, 2017 at 03:50:19PM +0100, Guillermo Rodriguez Garcia wrote:
> Hi Alexander,
> 
> Thank you for your answer.
> 
> 2017-03-01 22:03 GMT+01:00 Alexander Dahl <post@lespocky.de>:
> > Hei hei,
> >
> > On Wed, Mar 01, 2017 at 06:36:38PM +0100, Guillermo Rodriguez Garcia wrote:
> >> Is someone using ptxdist with iMX6 based targets? I know phytec used
> >> ptxdist in the past but now they seem to be using Yocto instead.

We use PTXdist for an i.MX6Q based target. The toolchain we use is based
on OSELAS.Toolchain, but I updated it to use the most recent GCC 6.3.0,
glibc 2.25, binutils 2.27, gdb 7.12.1, kernel-headers, etc.
Also changed some configure options to enable a few additional features
and improve the reproducability of builds. Besides that, I also added a
few patches from the buildroot project.
Not sure if it is upstreamable because I can't test if the other
ptxconfigs still work and I only ever used it for our i.MX6Q board.

But if you are interested I could send a few patches or create a GitHub
repo to work on it.

> That is a great start. Some things that draw my attention, though:
> 
> - VFPv3-D16 selected whereas iMX6 can do VFPv3-D32 (or just "VFPv3")

For the i.MX6 you can use the full VFPv3-D32 but I'd recommend using
-mfpu=neon instead. Performance is a little bit better than with
-mfpu=vfpv3.

> - Thumb mode used by default (not sure why -- wouldn't it be more
> appropriate to use arm mode by default unless there is an explicit
> intent to optimize size over speed ?)

Yes, the performance is better than with Thumb 2. (Maybe related to the
conditional execution feature in the original ARM instruction set..
still being faster than the use of the IT instruction in Thumb-2?)

Space and memory are not the biggest concerns on our board (4GB eMMC,
1GB RAM), so using ARM mode made sense and works great.

Regards,
Clemens

_______________________________________________
ptxdist mailing list
ptxdist@pengutronix.de

  reply	other threads:[~2017-03-02 16:17 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-03-01 17:36 Guillermo Rodriguez Garcia
2017-03-01 21:03 ` Alexander Dahl
2017-03-02 14:50   ` Guillermo Rodriguez Garcia
2017-03-02 16:17     ` Clemens Gruber [this message]
2017-03-02 17:35       ` [ptxdist] GitHub and free software Uwe Kleine-König
2017-03-02 18:17         ` Guillermo Rodriguez Garcia
2017-03-02 18:19           ` Guillermo Rodriguez Garcia
2017-03-02 18:28         ` Clemens Gruber
2017-03-16 15:53         ` Clemens Gruber
2017-03-17 20:50           ` Roland Hieber
2017-03-18  6:01             ` Robert Schwebel
2017-03-02 16:31     ` [ptxdist] ptxdist with iMX6 Alexander Dahl
2017-03-08 10:34     ` Michael Olbrich

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20170302161741.GA16522@archie.localdomain \
    --to=clemens.gruber@pqgruber.com \
    --cc=ptxdist@pengutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox