On 06/27/2011 12:36 PM, Xiaofan Chen wrote:
On Mon, Jun 27, 2011 at 6:13 PM, Tom Schouten<t...@zwizwa.be> wrote:
Hi,
I'm trying to figure out why Flash uploads are apparently slow on the
AT91SAM7 target with an Olimex ARM-USB-OCD adapter. This is with a OpenOCD
from git of early this week.
These are the speeds I get with different JTAG clock settings:
1MHz -> 5KB/sec
2MHz -> 7KB/sec
6MHz -> 8KB/sec
Have you tried to enable dcc downloads and fast memory access?
Yes, this is with:
arm7_9 dcc_downloads enable
arm7_9 fast_memory_access enable
But the speed is probably about right.
http://comments.gmane.org/gmane.comp.debugging.openocd.devel/13635
Seems similar indeed.
Comparing this to something like SAM-ICE on Windows this seems really slow.
(The SAM-ICE device says max 720KB/secs, gdb reports 873KB/sec which is
probably a bit off but it's still noticably a lot faster than the Olimex.)
Is there a known reason why this difference is so large?
ARM-USB-OCD hardware?
Adaptive clocking? (not on ARM-USB-OCD, but supported on SAM-ICE).
Some other issue related to OpenOCD itself?
I think all three are related with probably OpenOCD contribute the
major portion.
Any suggestion on where I might to look in the code to fix this?
What I find strange is that the data rate doesn't go proportional to the
clock speed.
Maybe combination of communication delay and waiting for the ack when
sending out a page? Some other polling delay that's too large?
You can use your SAM-ICE with OpenOCD (enable J-link support) and
then compare the results.
I do not use AT91SAM7. Last time I tried to use LPC-2148 with J-Link
and the speed is about 44KB/sec.
http://permalink.gmane.org/gmane.comp.debugging.openocd.devel/13920
Thanks, good to know.
_______________________________________________
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
