How do you know ls or no command is running? Have you tried using the
trace flags to see if the detailed CPU is executing instructions?
What if you create a checkpoint before booting finishes? When you
restore from the checkpoint and switch to the detailed CPU does the
system finish booting?
If you run with a smaller switch time (say 2000 ticks) does that
change anything?
The only other thing I could guess is that you need to put a drain/
resume around the switching... Try something like this:
diff -r 209e5c11ccdd configs/common/Simulation.py
--- a/configs/common/Simulation.py Fri Aug 10 12:44:40 2007 -0400
+++ b/configs/common/Simulation.py Fri Aug 10 12:44:42 2007 -0400
@@ -139,7 +139,9 @@ def run(options, root, testsys, cpu_clas
if options.standard_switch:
exit_event = m5.simulate(options.warmup)
+ m5.drain(testsys)
m5.switchCpus(switch_cpu_list1)
+ m5.resume(testsys)
num_checkpoints = 0
exit_cause = ''
Ali
On Oct 14, 2007, at 7:46 AM, abc def wrote:
The problem with checkpointing detailed cpu is still
there.
What i am doing is following:
a) Run M5 b3 with atomic simple cpu.
b) take checkpoint after booting finishes.
c) Then trying to restore that checkpoint with
detailed cpu by this command:
./build/ALPHA_FS/m5.opt configs/example/fs.py -d
--caches --l2cache -s -w 200000000000 -r 1
It gets to the command prompt in console. But, ls does
not run at all. For that matter no other command is
running.
But if i dont use checkpoint and boot detailed cpu
from scratch, after booting commands seem to work.
--- Ali Saidi <[EMAIL PROTECTED]> escribió:
The detailed CPU is significantly slower than the
AtomicCPU, so ls
will probably eventually be run, but it's will take
a while. Normally
we checkpoint a certain amount of time after the
checkpoint
instruction after whatever benchmark of interest is
running.
We use the following command line to restore from
the checkpoint and
switch to a detailed CPU. I don't know exactly what
you're running so
I can't be sure if there is a problem or not.
./build/ALPHA_FS/m5.fast configs/example/fs.py -t
--caches --l2cache -
s -w 200000000000 -r 1 --checkpoint_dir=....
Where -s stands for standard switching (to a
detailed cpu), and -w
says warmup the caches/tlb/etc for 200000000000
ticks with a simple
cpu before switching.
Ali
On Sep 25, 2007, at 6:43 AM, abc def wrote:
Hi,
I am trying to create and restore checkpoints for
detailed CPU model in M5 v2.0b3.
But it is not restoring properly.
While taking checkpoint i am using AtomicSimpleCPU
and
then using "m5 checkpoint" from command line after
booting finishes.
Next while restoring, i am giving the checkpoint
with
--detailed switch.
It switches to timing mode and console goes to
command
prompt. But after that it doesn´t respond to any
command. For example if i restore from checkpoint
in
this way, in the console even "ls" is not
responding.
Am i doing it correctly?
______________________________________________________________________
______________
Sé un Mejor Amante del Cine
¿Quieres saber cómo? ¡Deja que otras personas te
ayuden!
http://advision.webevents.yahoo.com/reto/entretenimiento.html
_______________________________________________
m5-users mailing list
m5-users@m5sim.org
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
m5-users@m5sim.org
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
______________________________________________________________________
______________
Sé un Mejor Amante del Cine
¿Quieres saber cómo? ¡Deja que otras personas te ayuden!
http://advision.webevents.yahoo.com/reto/entretenimiento.html
_______________________________________________
m5-users mailing list
m5-users@m5sim.org
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
m5-users@m5sim.org
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
