All of lore.kernel.org
 help / color / mirror / Atom feed
From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
To: buildroot@busybox.net
Subject: [Buildroot] Raspberry Pi OpenGL/Qt5 10x slower than Raspbian
Date: Thu, 4 Aug 2016 18:28:25 +0200	[thread overview]
Message-ID: <20160804182825.4c9ff4ff@free-electrons.com> (raw)
In-Reply-To: <CA+-urNQqnUdBvsMERgNxnAKWSGpiUNekvuSZADvyw7_HpYO41A@mail.gmail.com>

Hello,

On Thu, 4 Aug 2016 08:35:24 -0400, Frank Hunleth wrote:

> > Change -Os by -O2 in the Buildroot configuration, by using
> > BR2_OPTIMIZE_2. I even believe we should stop using BR2_OPTIMIZE_S by
> > default, and switch to BR2_OPTIMIZE_2 instead.  
> 
> Done, but no luck for this particular issue. I have another piece of
> code that differs in performance from Debian by 5% and is more CPU
> bound. I'll try it on that too.

Gaah.

> > Also, are you sure it's really using OpenGL with the Buildroot build?  
> 
> On the Buildroot build, I've tried running the Qt application with
> "-platform eglfs", but that seemed to be the default already. I'm not
> sure how to check deeper than that.

I guess if eglfs support has been built and you're able to use, it
means that it's using the GPU.

> > Under Raspbian, are you also running the eglfs build of Qt, or is it
> > with X.org? I'm not sure Debian has an eglfs capable build of Qt5.  
> 
> You're right. On Raspbian, I'm running with X.org. I am naively
> assuming that Qt5 w/ eglfs isn't 10x slower than Qt5 w/ X.org. I
> really hope that's not the case.

I also hope, but using X.org on one side and doing eglfs on the other
side is really a very different configuration. I would also assume that
there should not be a 10x difference between Qt5/X.org vs Qt5/eglfs,
but I have never tested, so I don't know.

Which toolchain are you using ? Are you building everything optimized
for ARMv6 (assuming you're using Rpi1) ?

But a 10x difference really feels like acceleration is not used at all.
You can also check the CPU consumption while running your Qt app to see
what it looks like.

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

  reply	other threads:[~2016-08-04 16:28 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-08-04  2:39 [Buildroot] Raspberry Pi OpenGL/Qt5 10x slower than Raspbian Frank Hunleth
2016-08-04  7:17 ` Thomas Petazzoni
2016-08-04 12:35   ` Frank Hunleth
2016-08-04 16:28     ` Thomas Petazzoni [this message]
2016-08-04 17:09       ` Frank Hunleth
2016-08-05  2:09         ` Khem Raj
2016-08-06 21:49         ` Frank Hunleth

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=20160804182825.4c9ff4ff@free-electrons.com \
    --to=thomas.petazzoni@free-electrons.com \
    --cc=buildroot@busybox.net \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.