Ali,
Thank you for responding.
Basically what I have set up now is a separate branch monitor program
running on a second core; I want to compare this against instrumentation
on one core. My thread context idea was to have the monitor running in a
separate thread context. Then it would be simil
Hello
When building Gem5, I met with someting strange, hope someone can help.
my building cmd is
scons build/ALPHA/gem5.fast
I am not going to use ruby here. To make sure the changes I made to
example/fs.py is really compiled, I added a wrong statement to the fs.py
file
from m5.internal.sta
I think you should always choose the one which has long execution time,
The difference between times are significant.
Usually the longer one matches the time in the terminal.
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailma
Hi,
I had exactly the same problem, and since some benchmarks take a really
long time with detailed ruby networks, I have to stop them before the
end, so sometimes I have 2 dumps and sometimes 1
So, how do I understand which of the stats contains useful information
for me? I just look at the C
Hi, Malek
Thanks for the reply. You are right, I am using precompiled image of
UT-Austin Parsec. So I should ignore the first round dumped stats,and use
the second one? However, sometimes, I get only one dump sections, sometimes
I got two sections, do you know what is causing the difference? I gu
Hui,
I think you should always get 2 sections of stats. If not, then it
means the benchmark did not exit correctly at the end (check m5term to
see if there was a segfault or something).
Those binaries called m5_dumpresetstats() so it dumps the stats then
resets those variables, so the 2nd set sho
Hi
I have a simulation that dumped the stats twice, does anyone know why?
Here is the simulation cmd:
./build/ALPHA/gem5.fast --stats-file=stats.txt configs/example/fs.py
--kernel=./ALPHA/parsec_alpha/vmlinux_2.6.27-gcc_4.3.4
--disk-image=./linux-parsec-2-1-m5-with-test-inputs.img -n 16 -r 1
Hi Hui,
The first section corresponds to stats because of the psuedoInst
dumpstats call that binary makes and the 2nd stats dump corresponds to
the stats dump that occurs at the end of the simulation (looks like
you are using the precompiled images from UT-Austin).
Malek
On Mon, Mar 25, 2013 at
Thanks for your response Lu. Actually, an alternative is working for
me now. I switched to x86, and now I am able to run my own benchmarks
(mapreduce) on four O3 cores with Ruby in FS mode (gem5 revision
9321). The only problem is with larger systems: since we cannot take
checkpoints, and booting O
Hi all,
I would like to know the relationship between an application binary and
gem5 in FS mode. Is the binary directly fed to kernel? While running the
binary file, can I still access information about threads through gem5? Or
once run in fs mode gem5 can only provide trace of CPUs (about interac
Hi all,
I would like to know the relationship between an application binary and
gem5 in FS mode. Is the binary directly fed to kernel? While running the
binary file, can I still access information about threads through gem5? Or
once run in fs mode gem5 can only provide trace of CPUs (about interac
Hi,
So you mean that the size of screen is determined by the Linux
kernel, not the gem5. I have tried to find out the root source of
configuring the lcd, but failed.
Are there detailed description about the parameters?
Best regards,
Yongbing Huang
From: gem5-user
The code you're pointing to specifies the max size for the lcd controller. If
you want to change the size of the vnc window you need to change the size of
screen that you're simulating (e.g. pass different parameters to the kernel on
boot).
Ali
On Mar 25, 2013, at 3:10 AM, "huangyongbing" wr
Hi David,
It's certainly possible, but it's going to take some work. My random idea would
be to change how all the branch instructions are defined, so the stash their
real target address somewhere and jump to your handler. Other than that
manually instrumenting the program (or doing it with som
Hi all,
To reduce simulation time, Simpoint approach is used in my simulation.
However, I am not aware of how to perform simpoint on multi-programmed
workloads.
For example, suppose a workload consists of 429.mcf plus 470.lbm. I have
got the simpoint for each of the two applications using Pi
Hi all,
I also try to run 0xbench on gem5. However, when I start the
application, it always says that "when an app was designed for a smaller
screen, a zoom control will appear by the clock" and stops here. So how to
handle it?
Thanks.
Best regards,
Yongbing Huang
Hi all,
I want to adjust the window size of VNC. On default, it is 1024*768
defined in src/dev/arm/pl111.hh. I simply set the width and height variables
in Pl111.write() function, 800*600 for example. However, the output of
vncviewer is not correct. Have I missed other things? What sho
17 matches
Mail list logo