I reproduce.
There was already a patch to port to 7.2.0 at
5c41076bd7610d03431fd0dd89bd0fdc7f30d6bd but the migration wasn't complete.
One workaround that does not require modifying the code is:
sudo apt-get install g++-6 gcc-6
CC=gcc-6 CXX=g++-6 scons
From
https://github.com/gem5/gem5/issues/16
fbe63074e3a8128bdbe1a5e8f6509c565a3abbd4
Boot works if I remove `--cpu-type=DerivO3CPU`.
Possible same: https://www.mail-archive.com/gem5-users@gem5.org/msg14062.html
but `DeriveO3CPU` not mentioned so I'm not sure.
Automated setup to reproduce:
https://
Hello Sahana,
If you don't absolutely need Ubuntu, I also have highly automated Buildroot
setup at:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/6b099ebd82c81a093922cda6ab5385e86238ec54#gem5
From: gem5-users on behalf of Sahana Prasad
Sen
at you did.
From: Sahana Prasad
Sent: Friday, March 2, 2018 12:10:13 PM
To: gem5 users mailing list; Ciro Santilli
Cc: Sepulveda Florez, Martha Johanna
Subject: Re: [gem5-users] Gem5 full system simulation on ARM
Thanks for this Ciro.
I will have a look and try it
One possibility is to instrument the benchmark with gem5 magic
instructions (the same instructions that m5 dumpstats ends up
calling).
Parsec in particular already has instrumentation stubs called
__parsec_roi_begin(), which you can implement for gem5 as shown at:
https://github.com/arm-university
m5_roi_begin are to be defined and placed inside the program that you
are benchmarking, not in gem5 itself (although it would be a good idea
to put those stubs as a header file in gem5 if not already done?).
There is an example of that at:
-
https://github.com/arm-university/arm-gem5-rsk/blob/aa
Yes, if you don't modify the dtb, the Linux kernel won't see the
change in the number of CPUs change, this can be verified with cat
/proc/cpuinfo.
You can modify the dtb directly by first converting it to dts human
readable form:
https://stackoverflow.com/questions/14000736/tool-to-visualize-the-
core counts:
https://github.com/gem5/gem5/blob/a66fe6a8c36c9ab49cb3a35065bfc645d51036c8/system/arm/dt/Makefile#L48
So hacking up that Makefile is another good option to get your dtbs.
On Tue, Mar 6, 2018 at 10:55 AM, Ciro Santilli wrote:
> Yes, if you don't modify the dtb, the Linux kernel w
What is your gem5 commit?
I had that problem but it went away after
https://gem5-review.googlesource.com/c/public/gem5/+/7301 apparently.
Here is a fully working setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/6b099ebd82c81a093922cda6ab5385e86238ec54#gem5
, use "-a x86_64".
On Wed, Mar 7, 2018 at 6:48 AM, Boyang Xu <6172...@gmail.com> wrote:
>
> Hi, all
>
> I am supposed to simulate an application scenario by gem5. There is a server
> with Linux OS, and I initialize an Android VM or JVM by this server. and then
> some Android or Java benchmarks will be executed in V
https://stackoverflow.com/questions/49261801
With QEMU, I can use either use `-initrd '${images_dir}/rootfs.cpio`
for the initrd, or pass the initramfs image directly to `-kernel
Image`.
But if I try the `initramfs` image with gem5 `fs.py --kernel Image` it
fails with:
fatal: Could not load
https://stackoverflow.com/questions/49268571/how-to-use-9p-virtio-mounts-in-gem5
I want to mount folders from host without networking on a Linux guest,
much as explained for QEMU at:
https://superuser.com/questions/628169/how-to-share-a-directory-with-the-host-without-networking-in-qemu
I have tr
https://stackoverflow.com/questions/49296092
On QEMU, I can use `-S -s`, which waits for GDB to connect before
starting the simulation
How to do the same thing on gem5, hopefully with the `fs.py` script?
By the time I swap terminals and connect, the simulation has already
gone past `start_kernel
https://stackoverflow.com/questions/48941494
For example, I want to be able to do on guest:
nc -l -p 45455
and then on host:
echo asdf | nc localhost 45455
and have the guest receive the `asdf`.
My initial goal with this is to setup `gdbserver`, but this could also
be used for fun th
On Wed, Mar 21, 2018 at 10:12 AM, Timon Evenblij
wrote:
> Hi all,
>
> A quick question: Is it possible to restore from a checkpoint with a
> different runscript then used during the first boot?
> This way, I could run a lot of benchmarks without losing time waiting for
> gem5 to boot.
> I know I c
I have successfully run PARSEC with this setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/29908dffb2726b0819fdf95240ae513dddb8af48#parsec-benchmark
and never encountered those problems.
Please also provide more detailed reproduction steps if you can.
On Fri, Mar 23, 2018 at 1
https://github.com/gem5/gem5/issues/22
Has anyone ever made this work?
68af229490fc811aebddf68b3e2e09e63a5fa475, gdb 7.11.1 built by Buildroot
An extremely similar aarch64 setup works for QEMU, and x86 and arm
both work for gem5, just aarch64 does not.
gem5 command:
```
M5_PATH='/home/ciro/bak
eems to be working, but I want to be sure I don't
> run into issues because of this.
>
> Regards
>
> Timon
>
> On Wed, Mar 21, 2018 at 4:47 PM, Ciro Santilli
> wrote:
>>
>> On Wed, Mar 21, 2018 at 10:12 AM, Timon Evenblij
>> wrote:
>> > Hi al
https://stackoverflow.com/questions/49516399
I expect `m5 readfile` will read a file from the host, and `m5
execfile` will execute a file from host, but the documentation is
scarce and I couldn't get them to work.
Maybe the path is determined by `readfile=` in `m5out/config.ini`, but
I don't know
If the init process is exits or is killed, the kernel panics, and
shows the message:
Kernel panic - not syncing: Attempted to kill init! exitcode
you can it out with a minimal init that just prints hello world and
exits, e.g. along:
https://unix.stackexchange.com/questions/122717/custom-linux-dis
This is likely a problem with the interpreter.
Either check:
- the shebang line #! of FFT if it is a script
- if it is an executable, the dynamic loader with readelf -a, look for
a line of type: "Requesting program interpreter:
/lib64/ld-linux-x86-64.so.2"
and make sure the program exists.
I al
How did you cross compile PARSEC exactly? It requires patches to do so
AFAIK.
I have most of splash2x working at: https://github.com/
cirosantilli/linux-kernel-module-cheat/tree/52358543846d7d9524c79908638951
d8a9bbdb64#parsec-benchmark (except some documented segfaults), and the
patches are at: h
Hello Xin,
Here is a setup that just works:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/52358543846d7d9524c79908638951d8a9bbdb64#gem5-getting-started
In particular:
- it uses this kernel config file:
https://github.com/cirosantilli/linux-kernel-module-cheat/blob/52358543846d7d
Have you tried gem5.opt --debug-flags=Exec as explained at
http://www.gem5.org/Trace_Based_Debugging ?
I can't test it on SE because I can't get SE working at all:
https://stackoverflow.com/questions/48959349/how-to-solve-fatal-kernel-too-old-when-running-gem5-in-syscall-emulation-se-m
, how did y
at happening?
>
> As for your problem, I have never encountered it, but it might be because
> I'm running Ubuntu 16.04 at the moment and I haven't updated gem5 since
> commit 9651c0278d93bf5c4620e198baac975f0c84e9a0.
>
> On Sat, Mar 31, 2018 at 6:43 AM, Ciro Santilli
>
I build in 10 minutes with this native machine:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/1f5baf5371f0670f7724f53ddb7793701ed84b9e#p51
Ubuntu 17.10 near master.
It happened to me a few times that the build became very slow, but I could
not understand why and it went away when
on-2.7.13/lib/libpython2.7.so.
> 1.0(PyEval_EvalCodeEx+0x830)[0x7f4f631d5ea0]
> ../../common/software/el6/python-2.7.13/lib/libpython2.7.so.
> 1.0(PyEval_EvalFrameEx+0x5aba)[0x7f4f631d1cea]
> ../../common/software/el6/python-2.7.13/lib/libpython2.7.so.
> 1.0(PyEval_EvalCodeEx+0x830)[0
Both KVM and DerivO3CPU without restore were broken for me without
checkpoints when I last tried them:
- https://github.com/gem5/gem5/issues/16
- https://www.mail-archive.com/gem5-users@gem5.org/msg15048.html
On Thu, Apr 5, 2018 at 4:12 AM, Da Zhang wrote:
> Hey guys,
>
> My gem5 real simulati
My setup https://github.com/cirosantilli/linux-kernel-module-cheat/tree/
32920fd25d6473934a89c48c56f4e12d03abde85#gem5-getting-started correctly
boots x86_64 with 4 cores (./run -c 4), and `cat /proc/cpuinfo` shows them.
I haven't however used your exact same build / run options: RUBY=True
PROTOCO
went forward in my boot flow.
>
>
>
> Regards,
>
> Vitorio.
>
>
>
> *From:* gem5-users [mailto:gem5-users-boun...@gem5.org] *On Behalf Of *Ciro
> Santilli
> *Sent:* Wednesday, April 4, 2018 2:39 PM
> *To:* gem5 users mailing list
> *Subject:* Re: [gem5-user
attempt using a Hello world and once I removed the
> –smt parameter it worked.
>
>
>
> Any taugths anybody ?
>
>
>
>
>
> *From:* gem5-users [mailto:gem5-users-boun...@gem5.org] *On Behalf Of *Ciro
> Santilli
> *Sent:* Wednesday, April 4, 2018 2:39 PM
> *To:* gem
nvironment it is working, I found how to add my own binaries,
> however I try to execute them it keeps saying it is not found, despite it
> shows up on ls.
>
>
>
> Regards,
>
> Vitorio
>
>
>
>
>
> From: gem5-users [mailto:gem5-users-boun...@gem5.org] On Behal
gt;
> Thanks for all the help so far.
> Rebuilding the image and will test it again in a few minutes.
>
>
>
> -Original Message-
> From: gem5-users [mailto:gem5-users-boun...@gem5.org] On Behalf Of Ciro
> Santilli
> Sent: Tuesday, April 10, 2018 11:06 AM
> To: gem5 user
Does not need to be running on top of a full Linux, could be just some
bare-metal setup that jumps to userland and then breaks memory
separation.
This could be a fun educational project, and would increase gem5's fame :-)
___
gem5-users mailing list
gem5
ool. I am planning on
> writing up a blog post on this, but I need to find the time.
>
> Cheers,
> Jason
>
> On Wed, Apr 11, 2018 at 6:07 AM Ciro Santilli
> wrote:
>>
>> Does not need to be running on top of a full Linux, could be just some
>> bare-metal setup th
Related: https://www.mail-archive.com/gem5-users@gem5.org/msg01431.html
On Thu, Apr 12, 2018 at 7:31 PM, Xin Zhan wrote:
> Hi all,
>
> I know the gem5ops like dumpstats [delay [period]] can be used in FS mode to
> periodically dump stats, but anyone has any idea about how to do it in SE
> mode?
>
What is your gem5 version, full gem5 CLI, and how did you get the images?
Here is a working setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/7b0bd10c0b83cf88ef5676d54d1c927c3ea57436#number-of-cores
On Fri, Apr 13, 2018 at 9:40 PM, Haiyang Han
wrote:
> Hi all,
>
> I am trying
Forgot to say: one possibility is to GDB step debug the Linux kernel:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/7b0bd10c0b83cf88ef5676d54d1c927c3ea57436#gem5-gdb
On Sat, Apr 14, 2018 at 4:24 AM, Ciro Santilli wrote:
> What is your gem5 version, full gem5 CLI, and how did
On Tue, Apr 17, 2018 at 8:10 AM, Singh, G. wrote:
> Hi All,
>
>
>
> I am trying to run a simple multi-threaded program on ARM processor with 4
> cpu's and 4 threads.
>
> This program works perfectly for DerivO3CPU but the same program on timing
> model gives the errors given
>
> below. Actually I
2cache --l2_size=2kB
>
> But if I use DerivO3CPU it works fine.
>
> Best
> Gagan
>
> -----Original Message-
> From: gem5-users [mailto:gem5-users-boun...@gem5.org] On Behalf Of Ciro
> Santilli
> Sent: Tuesday, 17 April 2018 10:04
> To: gem5 users mailing list
> Subj
Hi Liu,
I don't know the cause of your problem, but I have the following recommendation:
Can you try it on gem5 master to see if it is still present? It would
make it much more likely that people will know the solution to the
problem, or bother to reproduce it.
On Thu, Apr 19, 2018 at 11:45 AM,
master is the git name for the most recent version.
You can obtain it with:
git clone https://gem5.googlesource.com/public/gem5
as explained at: http://gem5.org/Download
I highly recommend that you become familiar with git if you are not
already, it will greatly help your software projects.
On
On Thu, Apr 19, 2018 at 5:03 PM, Bryce Patel wrote:
> Hello,
>
> I am very new to GEM5 and I am wondering how best to simulate thread
> migration on an asymmetric multicore system. I was thinking I could take a
> checkpoint partway through execution using the checkpoint m5op, then use the
> switch
Here is an alternative setup that just works in a single command:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/2852fe1989a6f1ab546e9a4fa88724423b3949f5#gem5
On Thu, Apr 19, 2018 at 8:28 PM, Mandrita wrote:
> Hello all,
>
> I am beginner in gem5. I am using ARM archtecture. When
a syscall.
> Thanks. Bryce Patel
>
> On Thu, Apr 19, 2018 at 1:08 PM, Ciro Santilli
> wrote:
>>
>> On Thu, Apr 19, 2018 at 5:03 PM, Bryce Patel
>> wrote:
>> > Hello,
>> >
>> > I am very new to GEM5 and I am wondering how best to simulate thread
l Message-
> From: gem5-users [mailto:gem5-users-boun...@gem5.org] On Behalf Of Ciro
> Santilli
> Sent: Tuesday, 17 April 2018 14:41
> To: gem5 users mailing list
> Subject: Re: [gem5-users] Multi-threading on ARM timing CPU
>
> On Tue, Apr 17, 2018 at 10:15 AM, Singh, G. wro
koverflow.com/questions/48959349/how-to-solve-fatal-kernel-too-old-when-running-gem5-in-syscall-emulation-se-m
> Best
> Gagan
>
> -Original Message-
> From: gem5-users [mailto:gem5-users-boun...@gem5.org] On Behalf Of Ciro
> Santilli
> Sent: Friday, 20 April 2018 00:2
Ping :-)
On Sun, Mar 25, 2018 at 11:52 PM, Boris Shingarov
wrote:
> I thought Paul and I fixed this back in September.
> I'll investigate in more detail tomorrow.
>
> -"gem5-users" wrote: -
> To: gem5 users mailing list
> From: Ciro Santilli
> Sent
In case you can't get that setup working, here is an alternative
working Parsec setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/14965a40d27c8d9d1ff5b023ace827b288a024ef#parsec-benchmark
On Mon, Apr 23, 2018 at 10:23 AM, Raman Arora wrote:
> Hi All,
>
> I am trying to compile
Reading current count from inactive timer.
> fatal: Kernel is mapped to invalid location (not memory). kernelStart
> 0x(fc31) - kernelEnd 0x(fc899860)
> 0xfc31:0xfc0000899860
> Memory Usage: 696048 KBytes
>
> Raman
>
> On Mon, Apr 23, 2018 at
https://askubuntu.com/questions/904685/cross-compilation-for-arm-error-no-such-file-or-directory/1027651#1027651
On Tue, Apr 24, 2018 at 3:41 AM, Bryce Patel wrote:
> Hello,
>
> I am having encountering an issue with a cross compiled Hello World program.
> I used gcc-arm-linux-gnueabihf to cross
I will be trying to reproduce this soon I hope, but in the meantime,
can you check things like:
- does /proc/cpuinfo show multiple cpus?
- do kernel boot messages show multiple cpus being brought up?
- does the problem happen for other content, or just PARSEC? E.g. you
could try to install and run
The optional ARM Linux kernel patches claim to support up to 255
cores: https://www.mail-archive.com/gem5-dev@gem5.org/msg24593.html so
I expect that this has been tested and works, although I haven't tried
it myself yet.
If that does not work, please send an email explaining the problem and
we'll
Here is another working v4.16 setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/7c16eb6ec1c98ac605a9324f37a0d13a8349e756#9p
On Fri, Apr 27, 2018 at 5:52 PM, Jochen Hollmann wrote:
> So far I have been using the default linux kernel
> x86_64-vmlinux-2.6.22.9.smp, however I have
I don't know what causes that problem, but here is a working setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/0b9157c91912685b0c59773126e1c1fefc7eea84#gem5
After build, run:
./run -a aarch64 -g -c 4
and it boots correctly, and cat /proc/cpuinfo shows 4 cpus.
The scripts th
Have you tried include/gem5/m5ops.h ? Found with "git ls-files | grep
m5ops.h", and it is included by m5.c.
Alternatively, this other ARM PARSEC tutorial just defines the
functions directly:
https://github.com/arm-university/arm-gem5-rsk/blob/aa3b51b175a0f3b6e75c9c856092ae0c8f2a7cdc/parsec_patches
Is it possible to put all of gem5 compilation output, including
intermediate files, into a specified directory?
By default it all goes under build/, is it possible to change that?
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/m
http://gem5.org/wiki/images/0/0e/ASPLOS2017_gem5_tutorial.pdf page 97
contains some interesting notes.
Also have a look at the wiki pages:
- http://gem5.org/CPU_Models
- http://www.gem5.org/O3CPU
Not really enough info I feel though, when I understand all of those
really well I will try to make
; CPU implementer: 0x41
> CPU architecture: 8
> CPU variant: 0x0
> CPU part: 0xc0f
> CPU revision: 0
>
> Is it intended for the ARMv8 big.LITTLE info to output the same CPU part
> when it is running O3_ARM_v7a_3 and Minor CPUs?
>
> Thanks,
> Peter de Groot
>
Hi Muhammad, you are more likely to get answers by following these bug
report recommendations:
https://github.com/cirosantilli-work/gem5-issues/issues/7
On Fri, May 4, 2018 at 8:32 AM, Muhammad Avais wrote:
> Dear All,
>
> Can anyone suggest what this error message indicates
>
> read error, exit
Hopefully with fs.py, but not necessarily.
For example, I have some x86 BIOS example that draw a line on the
screen on QEMU, and I'd like to see that work on gem5 too.
Also interested in ARM.
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.
What do you mean by poor results?
My experiments with changes of cache size and CPU type gave results
which seem consistent:
-
https://github.com/cirosantilli/linux-kernel-module-cheat/blob/237b7603b7972142a99aab1a609aef12522d0ef7/gem5-bench-cache#L39
-
https://stackoverflow.com/questions/49011
Hi Elena, is this different from:
http://m5sim.org/cgi-bin/mailman/private/gem5-users/2018-May/031456.html
?
Please provide more details: what do you want to achieve? What stat do
you want to read and why?
On Sun, May 6, 2018 at 10:31 PM, Woo L.L. wrote:
> Hi all,
>
> I would like to access the
de the disk image.
> Thanks.
>
> Regards
> Elena
>
> On 06/05/2018, 11:05 PM, "gem5-users on behalf of Ciro Santilli"
> wrote:
>
>>Hi Elena, is this different from:
>>http://m5sim.org/cgi-bin/mailman/private/gem5-users/2018-May/031456.html
>>?
&g
ds,
> Elena
>
>> On 7 May 2018, at 12:23 AM, Ciro Santilli wrote:
>>
>> Do you want to access a host file (stats.txt) from a Linux guest?
>>
>> If that is the case, can you either use `m5 readfile` (maybe not since
>> taken by your rc script, and only wor
Hi Pawan,
Remember to always reply to both me and the gem5-users list to share
the knowledge, I only noticed this on you latest message. Proposal for
a more modern issue tracker at:
https://www.mail-archive.com/gem5-dev@gem5.org/msg24951.html
> AttributeError: object 'System' has no attribute 'sw
Alternative setup that makes everything just work in a single command:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/04f13937f4a68f44c5d302954892eac8a79bade5#parsec-benchmark-without-parsecmgmt
On Wed, May 9, 2018 at 4:13 AM, Raman Arora wrote:
> Hi All,
>
> I am trying to run th
Hi Pawan, isn't it the situation here the same for checkpoints and
simpoints? https://www.mail-archive.com/gem5-users@gem5.org/msg15457.html
On Wed, May 9, 2018 at 4:24 AM, Pawan Joshi wrote:
> Hello
> Is it not possible to use the checkpoints generated using an Atomic simple
> CPU to restore the
As mentioned at:
https://www.mail-archive.com/gem5-dev@gem5.org/msg24593.html you need
to:
- use the ARM linux kernel fork from:
https://gem5.googlesource.com/arm/linux/+/refs/heads/gem5/v4.15 in
particular the GICv2 extension script commit
- hack up gem5:
- fs.py to set `root.system.realview.gi
le. I actually meant the
> system hangs with 16 cores.
> Do I still need to set root.system.realview.gic.gem5_extensions to True in
> fs.py?
> Is the GICv2 extension only compatible with Linux kernel 4.15? Currently I'm
> using 4.4.
>
> Thanks,
> Haiyang
>
> On W
I'm not sure if this helps, but if you do `gem5.opt --debug-flags=Exec` it does
dump uops with format:
25007000: system.cpu T0 : @start_kernel: stp
25007000: system.cpu T0 : @start_kernel.0 : addxi_uop ureg0, sp,
#-112 : IntAlu : D=0xff8008913f90
25007500: system.cpu T0 : @start_ker
The init=/some/path linux kernel command line must match an executable
present in the disk image. Have you passed the init= option?
BTW, your email shows on my spam, I think you haven't configured your
email client to authenticate correctly, ensure that you can send
emails across two test accounts
gt; when I use Qemu to create disk, I have "/" and "/boot", no swap and others.
>
> On Thursday, May 10, 2018, 10:32:25 AM PDT, Ciro Santilli
> wrote:
>
>
> I don't think so: a quick test with QEMU shows that this message
> happens when the path to init
Immediately we suspect that the file is auto-generated during build.
So we do a shell find in the output directory:
```
find . -name OpClass.hh
```
This leads us to: ./build/ARM/enums/OpClass.hh
Open the file, look for interesting unique identifiers, and git grep
that on main source.
So:
```
I don't think those patches are enough to run all benchmarks, and I
have never seen a set of patches that just worked for all benchmarks so far.
I have part of parsec benchmarks working at:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/d1f676a680aca90e525f14d90605393a9742d09b#pars
inaro arch64 image or i add the patch for arch64 i did
> not understand well ??
> because I need to copy parsec 3.0 in the linaro image to run with gem5.
>
> thanks;
>
>
> 2018-05-12 17:28 GMT+02:00 Ciro Santilli :
>>
>> I don't think those patches are enough to
well ??
>> because I need to copy parsec 3.0 in the linaro image to run with gem5.
>>
>> thanks;
>>
>>
>> 2018-05-12 17:28 GMT+02:00 Ciro Santilli :
>>>
>>> I don't think those patches are enough to run all benchmarks, and I
>>>
What is the full gem5 command line cli and gem5 version?
Here is a working setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/99382df5ff1c1933be1371d060a221195608a9a5#gem5
On Wed, May 16, 2018 at 2:13 AM, Ryan Wang wrote:
> Hi all,
> I'm able to boot a kernel with the given im
Got it half working for ARM: https://stackoverflow.com/a/50364864/9160762
On Sun, May 6, 2018 at 3:42 PM, Ciro Santilli wrote:
> Hopefully with fs.py, but not necessarily.
>
> For example, I have some x86 BIOS example that draw a line on the
> screen on QEMU, and I'd like to
https://github.com/cirosantilli-work/gem5-issues/issues/18
I had both aarch64 and arm working with --dtb-file using:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/0678bd82ac61f0decce3480fef798bd5edbfd0c3#gem5
But now I've learnt about the awesome --generate-dtb option, which
woul
read RealView I/O at offset 0x60 that doesn't exist
warn: Tried to read RealView I/O at offset 0x48 that doesn't exist
warn: Tried to write RVIO at offset 0xa8 (data 0) that doesn't exist
Exiting @ tick 18446744073709551615 because simulate() limit reached
```
On Thu, May 17, 2018 a
Are you affiliated with an university or enterprise?
What is your end goal in using gem5? Does it justify the cloud
investment in your project?
How long does your build / Linux kernel boot take? Here are some
references:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/58face41ac84
Hi commerce, I'm maintaining a fork of parsec at:
https://github.com/cirosantilli/parsec-benchmark
I don't have all benchmarks working, but I documented how everything
fails, and if you get something working, please send a pull request
there as well.
On Wed, May 23, 2018 at 12:51 PM, commerce _co
a cloud to do the sumulation quickly, please if you have any idea help
> me and thank you.
>
> 2018-05-21 16:24 GMT+02:00 Ciro Santilli :
>>
>> Are you affiliated with an university or enterprise?
>>
>> What is your end goal in using gem5? Does it justify the cloud
Hi commerce, I don't know what is your setup, all I can say is that I
have a few parsec benchmarks working at on my setup:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/fe01ade54d0563b5dc8d8a7826d46b153d9a850a#parsec-benchmark
You will also notice that parsecmgmt scripts are too s
https://stackoverflow.com/questions/5054/how-to-run-a-dynamically-linked-executable-syscall-emulation-mode-se-py-in-gem5
After
https://stackoverflow.com/questions/48959349/how-to-solve-fatal-kernel-too-old-when-running-gem5-in-syscall-emulation-se-m
I managed to run a statically linked hello
; running on an ARM native machine) or if you could use QEMU user mode to get
> dynamically linked executables to work on a non-native machine. Both of
> these cases would likely require some re-working of the code.
>
> Cheers,
> Jason
>
> On Sat, May 26, 2018 at 11:56 AM Ciro Santil
Likely incompatible compiler using wrong dynamic loader, do "file
hello", see "interpreter /some/path", and check if "/some/path" is
present on guest.
On Thu, May 31, 2018 at 9:33 PM, commerce _com wrote:
> Hi all,
>
> i need to run hello.c in an architecture armv8 big.LITTLE
> I compile hello_wo
;>
>> but I did not understand how to solve this problem; Please.
>>
>>
>> com_.
>>
>> 2018-05-31 23:02 GMT+02:00 Ciro Santilli :
>>>
>>> Likely incompatible compiler using wrong dynamic loader, do "file
>>> hello", see
t has no relation with how the binary was
>>> compiled.
>>>
>>> My intuition is that the directory is in fact "/root/parsec". It's just a
>>> guess based on the absence of home directory for root user in my Ubuntu
>>> (although I don't know which
I think bootrscript is just an arbitrary script that your must create
yourself to run the benchmark you want, or give you a shell.
It must be being used from the /init of that image to set the m5
readfile path, and then run it.
This kind of thing allows you to checkpoint at the end of boot, but
s
Sorry, I don't know of anything free specifically for gem5 :-(
On Mon, Jun 4, 2018 at 2:04 PM, commerce _com wrote:
> Hi Ciro,
> do you know of a free cloud to use gem5, because I always took a lot of time
> to run the kernel armv8 on the fs.big.LITTLE mode.
>
> Thanks,
>
> com_.
Cache sizes, at least, can be varied for sure, and do take effect:
https://github.com/cirosantilli/linux-kernel-module-cheat/tree/d1ad6d8f7a4811f22f83166deaeb66ff2c54937d#gem5-cache-size
I'd recommend trying out similar setups for other parameters you care
about and see if they take effect.
On Mo
Does it happen on a full clean build? What is your gem5 git revision?
What is your OS? What is your compiler version? :-)
On Mon, Jun 4, 2018 at 9:44 PM, Mao Ye wrote:
> Hi all,
>
> I have been using gem5 for a while without any problem. But today I met a
> problem with my new modification and it
On Tue, Jun 5, 2018 at 3:21 PM, Georgios S. Bousdras
wrote:
> Dear all,
>
> I wonder if It is possible to run the gem5 on a multi-core mode(like the
> scons build/ARM/gem5.opt -j “n”) in order to reduce the simulation time.
>
No: https://github.com/cirosantilli-work/gem5-issues/issues/15
> In ad
armv8_gem5_v1_big_little_4_4 means 4 big and 4 little.
If you run it without anything other than --little-cpus=4 --big-cpus=4
is likely to blow up.
Failed to come online means the kernel could not properly find the CPU
and won't use it at all.
On Fri, Jun 8, 2018 at 3:57 PM, Georgios S. Bousdras
How to build that program step by step? What are the versions of your
OS, compiler and gem5 revision? Are you able to run other shared libraries fine?
Are you wearing a tie? Use GDB step debugging
or tracing to find out at which instruction / point of code it is
blowing up.
On Mon, Jun 11, 2018 at
At https://www.mail-archive.com/gem5-users@gem5.org/msg15233.html
Jason mentioned the technique of attaching two disk images to restore
a checkpoint and run a different benchmark.
Is it possible to specify multiple disk images when using fs.py? From
the source code I don't think it is currently su
w.
> I will probably make a cleaner way of doing this in the near future.
>
> However, the same method did not apply to a custom full-system script I am
> working on, so I am still investigating.
>
> On Thu, Jun 14, 2018 at 7:23 PM Ciro Santilli
> wrote:
>
>> At https://
bootscript.rcS is the default m5 readfile file that fs_bigLITTLE.py sets at:
https://github.com/gem5/gem5/blob/05c4c2b566ce351ab217b2bd7035562aa7a76570/configs/example/arm/fs_bigLITTLE.py#L64
panic: could not open file is just the default message that gem5 gives when
that file is not found.
So,
1 - 100 of 312 matches
Mail list logo