p {margin-top:0px;margin-bottom:0px;} 'sleep 10' would be enough.
Then you say the simulator to create the checkpoint, i.e., 'm5 checkpoint'. Over here it creates the checkpoint.
When you restore from the checkpoint the same script is there and the next command is there to be executed, i.e., 'm5 exit'.
i.e., your script says that after the checkpoint restore do exit. Which should take seconds :)
--------- ������ ������ ---------
������������: Peng Wei <peng.wei.prc@gmx.com>
������������ : gem5-users@gem5.org
������: 2013��� 12��� 30��� ���������, 21��� 02��� 57��� +0900
������: [gem5-users] Problem about Create Checkpoint
Hi there,I tried to create a checkpoint in gem5 ARM full system mode, but it had been created successfully but cannot be used correctly.
I used default CPU mode, 4 core, no caches, and use an .rCS file like this:
sleep 600
/sbin/m5 checkpoint
/sbin/m5 exit
I would like to wait 10 minutes to let the system fully booted and then create checkpoint and close m5.
However, when I tried to restore the system from that checkpoint, it always was finished in several seconds and showed that "exit due to m5_exit instruction received" no matter which CPU mode or which rCS file I used.
Could anybody help me fix it? I suspend that it is likely that my script used to create the checkpoint was problematic, but it is so easy that I don't know what is the problem.
Thanks very much.
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
p {font-size:10pt ! important;font-family:������,'������',gulim,tahoma,sans-serif ! important;}
<!-- __Hanmail-sig-Start__ --><!-- __Hanmail-sig-End__ --><html><img src="http://wwl1453.hanmail.net:4280/@from=sarilian&rcpt=gem5%2Dusers%40gem5%2Eorg&msgid=%3C20131230211635%2EHM%2E0000000000000ze%40sarilian%2Ewwl1453%2Ehanmail%2Enet%3E">Hi Tod,
Thank you for your reply. However, I didn't catch your point very well.
Firstly, I ran a command in order to create a checkpoint:
./build/ARM/gem5.opt configs/example/fs.py --num-cpus=4
--kernel=vmlinux.smp.ics.arm.asimbench.2.6.35
--disk-image=ARMv7a-ICS-Android.SMP.Asimbench-v3.img --mem-size=256MB
--script=../gem5-stable/system_images/asimbench_boot_scripts/mixture.rcS
The mixture.rCS includes the "sleep 600; /skin/m5 checkpoint; /skin/m5 exit"
This step ran correctly and created a checkpoint.
The problem is the second step:
./build/ARM/gem5.opt configs/example/fs.py --num-cpus=4
--kernel=vmlinux.smp.ics.arm.asimbench.2.6.35
--disk-image=ARMv7a-ICS-Android.SMP.Asimbench-v3.img --mem-size=256MB
--script=../gem5-stable/system_images/asimbench_boot_scripts/adobe.rCS -r
1
My objective is to restore from the checkpoint I created and run new rCS
file. However, no matter what the new rCS file is, the execution of gem5
finishes in seconds.
Could you briefly explain what the mechanism of checkpoint restore is, and
what will happen when I use one rCS file (mixture.rCS) to create a
checkpoint and restore it to run another rCS file (adobe.rCS)?
Thank very much
On Mon, Dec 30, 2013 at 4:16 AM, Tod <sarilian(a)hanmail.net> wrote:
'sleep 10' would be enough.
Then you say the simulator to create the checkpoint, i.e., 'm5
checkpoint'. Over here it creates the checkpoint.
When you restore from the checkpoint the same script is there and the next
command is there to be executed, i.e., 'm5 exit'.
i.e., your script says that after the checkpoint restore do exit. Which
should take seconds :)
--------- 원본 메일 ---------
보낸사람: Peng Wei <peng.wei.prc(a)gmx.com>
받는사람 : gem5-users(a)gem5.org
날짜: 2013년 12월 30일 월요일, 21시 02분 57초 +0900
제목: [gem5-users] Problem about Create Checkpoint
Hi there,
I tried to create a checkpoint in gem5 ARM full system mode, but it had
been created successfully but cannot be used correctly.
I used default CPU mode, 4 core, no caches, and use an .rCS file like this:
sleep 600
/sbin/m5 checkpoint
/sbin/m5 exit
I would like to wait 10 minutes to let the system fully booted and then
create checkpoint and close m5.
However, when I tried to restore the system from that checkpoint, it
always was finished in several seconds and showed that "exit due to m5_exit
instruction received" no matter which CPU mode or which rCS file I used.
Could anybody help me fix it? I suspend that it is likely that my script
used to create the checkpoint was problematic, but it is so easy that I
don't know what is the problem.
Thanks very much.
gem5-users mailing list
gem5-users(a)gem5.orghttp://hanmail/mail/MailComposeFrame.daum?TO=gem5-users(a)gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
gem5-users mailing list
gem5-users(a)gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
Hi Peng,
In the AsimBench (renamed as Moby now) image provided by me, there
is a rcS file named arm_ckpt_asim.rcS. You can use this file to create the
checkpoint.
Best regards,
Yongbing Huang
From: gem5-users-bounces(a)gem5.org [mailto:gem5-users-bounces(a)gem5.org] On
Behalf Of Peng Wei
Sent: Tuesday, December 31, 2013 1:09 PM
To: gem5 users mailing list
Subject: Re: [gem5-users] Problem about Create Checkpoint
Hi Tod,
Thank you for your reply. However, I didn't catch your point very well.
Firstly, I ran a command in order to create a checkpoint:
./build/ARM/gem5.opt configs/example/fs.py --num-cpus=4 --
kernel=vmlinux.smp.ics.arm.asimbench.2.6.35 --disk-image=ARMv7a-ICS-Android.
SMP.Asimbench-v3.img --mem-size=256MB --script=../gem5-
stable/system_images/asimbench_boot_scripts/mixture.rcS
The mixture.rCS includes the "sleep 600; /skin/m5 checkpoint; /skin/m5 exit"
This step ran correctly and created a checkpoint.
The problem is the second step:
./build/ARM/gem5.opt configs/example/fs.py --num-cpus=4 --
kernel=vmlinux.smp.ics.arm.asimbench.2.6.35 --disk-image=ARMv7a-ICS-Android.
SMP.Asimbench-v3.img --mem-size=256MB --script=../gem5-
stable/system_images/asimbench_boot_scripts/adobe.rCS -r 1
My objective is to restore from the checkpoint I created and run new rCS
file. However, no matter what the new rCS file is, the execution of gem5
finishes in seconds.
Could you briefly explain what the mechanism of checkpoint restore is, and
what will happen when I use one rCS file (mixture.rCS) to create a
checkpoint and restore it to run another rCS file (adobe.rCS)?
Thank very much
On Mon, Dec 30, 2013 at 4:16 AM, Tod <sarilian(a)hanmail.net> wrote:
'sleep 10' would be enough.
Then you say the simulator to create the checkpoint, i.e., 'm5 checkpoint'.
Over here it creates the checkpoint.
When you restore from the checkpoint the same script is there and the next
command is there to be executed, i.e., 'm5 exit'.
i.e., your script says that after the checkpoint restore do exit. Which
should take seconds :)
--------- 원본 메일 ---------
보낸사람: Peng Wei <peng.wei.prc(a)gmx.com>
받는사람 : gem5-users(a)gem5.org
날짜: 2013년 12월 30일 월요일, 21시 02분 57초 +0900
제목: [gem5-users] Problem about Create Checkpoint
Hi there,
I tried to create a checkpoint in gem5 ARM full system mode, but it had
been created successfully but cannot be used correctly.
I used default CPU mode, 4 core, no caches, and use an .rCS file like this:
sleep 600
/sbin/m5 checkpoint
/sbin/m5 exit
I would like to wait 10 minutes to let the system fully booted and then
create checkpoint and close m5.
However, when I tried to restore the system from that checkpoint, it always
was finished in several seconds and showed that "exit due to m5_exit
instruction received" no matter which CPU mode or which rCS file I used.
Could anybody help me fix it? I suspend that it is likely that my script
used to create the checkpoint was problematic, but it is so easy that I
don't know what is the problem.
Thanks very much.
gem5-users mailing list
gem5-users(a)gem5.org <http://hanmail/mail/MailComposeFrame.daum?TO=gem5-
users(a)gem5.org>
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
<http://wwl1453.hanmail.net:4280/@from=sarilian&rcpt=gem5-
users(a)gem5.org&msgid=%3C20131230211635.HM.0000000000000ze(a)sarilian.wwl1453.h
anmail.net%3E>
gem5-users mailing list
gem5-users(a)gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
Hi Yongbing,
Thanks very much. However, the problem for me now is still the huge
differences between the gem5 results and development board results, such as
MPKI of I/DCache. Did you try to find the reason why there are so huge
differences between the simulator and the real data? Thanks very much.
On Mon, Dec 30, 2013 at 10:15 PM, huangyongbing <huangyongbing(a)ncic.ac.cn>wrote:
Hi Peng,
In the AsimBench (renamed as Moby now) image provided by me,
there is a rcS file named arm_ckpt_asim.rcS. You can use this file to
create the checkpoint.
Best regards,
Yongbing Huang
From: gem5-users-bounces(a)gem5.org [mailto:gem5-users-bounces(a)gem5.org] *On
Behalf Of *Peng Wei
Sent: Tuesday, December 31, 2013 1:09 PM
To: gem5 users mailing list
Subject: Re: [gem5-users] Problem about Create Checkpoint
Hi Tod,
Thank you for your reply. However, I didn't catch your point very well.
Firstly, I ran a command in order to create a checkpoint:
./build/ARM/gem5.opt configs/example/fs.py --num-cpus=4
--kernel=vmlinux.smp.ics.arm.asimbench.2.6.35
--disk-image=ARMv7a-ICS-Android.SMP.Asimbench-v3.img --mem-size=256MB
--script=../gem5-stable/system_images/asimbench_boot_scripts/mixture.rcS
The mixture.rCS includes the "sleep 600; /skin/m5 checkpoint; /skin/m5
exit"
This step ran correctly and created a checkpoint.
The problem is the second step:
./build/ARM/gem5.opt configs/example/fs.py --num-cpus=4
--kernel=vmlinux.smp.ics.arm.asimbench.2.6.35
--disk-image=ARMv7a-ICS-Android.SMP.Asimbench-v3.img --mem-size=256MB
--script=../gem5-stable/system_images/asimbench_boot_scripts/adobe.rCS
-r 1
My objective is to restore from the checkpoint I created and run new rCS
file. However, no matter what the new rCS file is, the execution of gem5
finishes in seconds.
Could you briefly explain what the mechanism of checkpoint restore is,
and what will happen when I use one rCS file (mixture.rCS) to create a
checkpoint and restore it to run another rCS file (adobe.rCS)?
Thank very much
On Mon, Dec 30, 2013 at 4:16 AM, Tod <sarilian(a)hanmail.net> wrote:
'sleep 10' would be enough.
Then you say the simulator to create the checkpoint, i.e., 'm5
checkpoint'. Over here it creates the checkpoint.
When you restore from the checkpoint the same script is there and the next
command is there to be executed, i.e., 'm5 exit'.
i.e., your script says that after the checkpoint restore do exit. Which
should take seconds :)
--------- 원본 메일 ---------
보낸사람: Peng Wei <peng.wei.prc(a)gmx.com>
받는사람 : gem5-users(a)gem5.org
날짜: 2013년 12월 30일 월요일, 21시 02분 57초 +0900
제목: [gem5-users] Problem about Create Checkpoint
Hi there,
I tried to create a checkpoint in gem5 ARM full system mode, but it had
been created successfully but cannot be used correctly.
I used default CPU mode, 4 core, no caches, and use an .rCS file like this:
sleep 600
/sbin/m5 checkpoint
/sbin/m5 exit
I would like to wait 10 minutes to let the system fully booted and then
create checkpoint and close m5.
However, when I tried to restore the system from that checkpoint, it
always was finished in several seconds and showed that "exit due to m5_exit
instruction received" no matter which CPU mode or which rCS file I used.
Could anybody help me fix it? I suspend that it is likely that my script
used to create the checkpoint was problematic, but it is so easy that I
don't know what is the problem.
Thanks very much.
gem5-users mailing list
gem5-users(a)gem5.orghttp://hanmail/mail/MailComposeFrame.daum?TO=gem5-users(a)gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
gem5-users mailing list
gem5-users(a)gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
gem5-users mailing list
gem5-users(a)gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users