Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 17 Mar 2008 19:03:16 +0100
From:      Bernd Walter <ticso@cicely12.cicely.de>
To:        =?iso-8859-1?Q?Bj=F6rn_K=F6nig?= <bkoenig@alpha-tierchen.de>
Cc:        arm@freebsd.org, ticso@cicely.de
Subject:   Re: defining the main clock frequency of AT91 boards
Message-ID:  <20080317180315.GA72551@cicely12.cicely.de>
In-Reply-To: <51329.192.168.1.2.1205776085.squirrel@webmail.alpha-tierchen.de>
References:  <50161.192.168.1.2.1205540152.squirrel@webmail.alpha-tierchen.de> <20080316.154215.1387160441.imp@bsdimp.com> <20080317014143.GQ67602@cicely12.cicely.de> <51329.192.168.1.2.1205776085.squirrel@webmail.alpha-tierchen.de>

next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Mar 17, 2008 at 06:48:05PM +0100, Björn König wrote:
> > On Sun, Mar 16, 2008 at 03:42:15PM -0600, M. Warner Losh wrote:
> >> In message:
> >> <50161.192.168.1.2.1205540152.squirrel@webmail.alpha-tierchen.de>
> >>             Björn_König <bkoenig@alpha-tierchen.de> writes:
> The patch added the option to the BWCT kernel configuration. I just
> mentioned TSC because there is no default configuration in sys/arm/conf/.
> 
> >> I'd go one step further.  I'd require everybody to define this value
> >> since there's no 'standard' frequency and the value that's there is
> >> just the value of the boards we used for the port.
> >
> > Me too - that way noone can forget defining it and run with the wrong
> > frequency.
> 
> I agree. It sounds reasonable to make such an option mandatory. I tried to
> be conservative.

I'm more worried about users (inkluding myself ;-) taking a old config
without adding the new option, than about beeing conservative.
But since we all agree about it :)

> I think users will notice very quickly if they run their board with the
> wrong frequency, because they won't get output on the serial console -
> maybe already too late. ;-)

Console speed is about MCK.
Inside the kernel the xtal Clock is only used to setup PLLB for USB,
so it is just USB, which is not working.

> > But since we are already about RM9200 clocks.
> > Is it possible today to setup different clocks?
> > I remember that MCK was hardcoded some time ago, but now I saw that
> > it can be setup in the kernel conf.
> 
> It's just overriding the #define in at91rm9200.h.

I just remember it was hardcoded in somewhere some time ago.

> > E.g. the MCK can be 80MHz and the only reason it is 60MHz right now is
> > because PCK is 180MHz and MCK has to be divided from that.
> > I always wondered myself if it is faster to run the CPU at 160MHz and
> > have 80MHz MCK.
> 
> I think it depends on what you are doing with the board, but in general I
> would say that the CPU is the bottle neck. I used my board with an 80 MHz
> master clock and didn't notice better performance (Yes, I reconfigured the
> memory timings accordingly).

I was thinking about network performance.
Core can only be faster with higher PCK if running from cache, otherwise
it is slowed down by MCK based things.
I was hoping to get more speed for routing, but maybe the network code
is working too much inside caches.

What kind of application did you try with the higher MCK?
What kind of RAM settings are you talking about?
Almost all SDRAM on such boards is 133MHz (worst case I would expect
100MHz), so no need to slow memory settings down for just 80MHz.
And I don't see any reason to even add further waitstates, since 133MHz
SDRAM should be good for at least 100MHz without additional waitsates.
The only parameter you can tune is refresh, since the refresh frequency
doesn't need to increase as well, but I doubt that this would make more
than an academic difference in speed.

-- 
B.Walter                http://www.bwct.de      http://www.fizon.de
bernd@bwct.de           info@bwct.de            support@fizon.de



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20080317180315.GA72551>