You have to copy the compiled binary into disk image.
1- mount your current disk image
2- copy binaries into it
3- unmount it
4- run the simulation
On Fri, Jun 19, 2015 at 12:18 PM, Junaid Shuja
wrote:
> Hi,
> During full system emulation for ARM, how can I push a compiled C file so
> that it i
Dear list!
I am simulating (ruby, fs) a 4 core ALPHA system with Mesh interconnect
(simple).
The system runs bodytrack benchmark.
Looking at NoC statistics, I'm seeing percent_links_utilized of all routers
to be always less than 1 (that is 1% !).
First I guess it might be fraction (rather than pe
Did you run them all in x86, multi thread with ruby?
I guess you should've edit the source code ...
I'd appreciate if you post the disk image :)
On Wed, Oct 31, 2012 at 5:34 PM, Runjie Zhang wrote:
> I have successfully ran Parsec with recent version(changeset:01c8c5ff2c3b)
> of Gem5 for both AL
Dear All,
I think doxygen-generated documents of source code is located in
http://www.gem5.org/docs/index.html
However, It's outdated. Is there a newer version somewhere online?
I know I can generate it myself! I want to know if this is available online.
Best
Seyed Hossein
__
you can run a multi-programmed workload in a multi-processor in SE mode.
you may compile your multi-threaded program with m5threads. This can
be run is SE mode.
remember: m5threads is not a complete alternative to pthreads. Some
features are not implemented.
if you want a native pthreads applicat
Thanks!
On Mon, Nov 5, 2012 at 6:41 PM, Ali Saidi wrote:
> It's been updated to the latest version.
>
>
>
> Ali
>
>
>
> On 04.11.2012 01:37, Hossein Nikoonia wrote:
>
> Dear All,
>
> I think doxygen-generated documents of source code is located in
>
Or you can use more swap!
On Dec 28, 2012, at 5:30 PM, Nilay Vaish wrote:
> On Thu, 27 Dec 2012, ali bagherian wrote:
>
>> Hi,
>>
>> I tried to compile Gem5. but I got following error:
>>
>> scons: Building targets ...
>> [ CXX] X86/arch/x86/generated/decoder.cc -> .o
>> virtual memory
I did this in Lion without a problem. Maybe you have multiple versions of
gcc and that is confusing ...
On Sat, Mar 2, 2013 at 10:34 PM, Anthony Gutierrez wrote:
> Hm... I've seen the previous posts from you and Ali about compiling on OS
> X. When I started, I tried installing swig/hg/scons/pyth
ITY or FITNESS FOR A PARTICULAR PURPOSE.
>
> Anthony Gutierrez
> http://web.eecs.umich.edu/~atgutier
>
>
> On Sat, Mar 2, 2013 at 10:38 PM, Hossein Nikoonia wrote:
>
>> I did this in Lion without a problem. Maybe you have multiple versions of
>> gcc and that is confusing
Hi,
I am preparing a patch that enables gem5 to run VMs like a hypervisor.
I might be able to help if you want to do that...
However, as andreas noted it is not possible to simulate different ISAs
Sent from my iPad
On Jun 6, 2013, at 12:03 PM, Andreas Sandberg wrote:
> Hi Mazen,
>
> On 0
openmp and pthreads are not supported in SE mode.
you may use m5threads or FS mode.
On Thu, Aug 8, 2013 at 11:25 AM, Fatemeh Ali Akbari
wrote:
> Hello,
> I want to change a c++ simple code into multithread one. So I use OpenMP
> for it. I just put "#pragma omp parallel for" before "for" in C++.
It is very hard to run a full hypervisor in gem5 since the simulator is not a
complete environment.
However, if you want to run virtual machines on gem5 to have a virtualized
environment, i might be able to help.
I have a patch for that
Its similar to Virtual-GEMS .
Sent from my iPad
> On
gt; Thanks,
> Binh
>
> On 9/29/13 2:24 PM, Hossein Nikoonia wrote:
>
> It is very hard to run a full hypervisor in gem5 since the simulator is not a
> complete environment.
> However, if you want to run virtual machines on gem5 to have a virtualized
> environment, i might
; Thanks a bunch!
> Binh
>> On 9/30/13 12:37 AM, Hossein Nikoonia wrote:
>> Hi,
>>
>> See Virtual-GEMS paper. My patch is similar to this.
>>
>> http://www-mount.ece.umn.edu/~jjyi/MoBS/2009/program/02D-Garcia-Guirado.pdf
>>
>> For a brief overview
-virtualization-support.README inside. And if needed, don't
hesitate to contact me :)
Best
Seyed Hossein
On Mon, Sep 30, 2013 at 6:03 PM, Hossein Nikoonia wrote:
> I'am happy to see you are interested! I'll contact you soon :)
>
> Sent from my iPad
>
> On Sep 30, 2
Dear List,
I run a system with 2 CPUs to run two concurrent parsec benchmarks. The
problem is that I see a memory controller bandwidth of 6.3 GBps. But the
peakBW is 4.2 GBps !
system.mem_ctrls.bw_total::total 6311068945
# Total bandwidth to/from this memory (bytes/s)
system.mem_ctrls.
e does sound quite
> large. Could some Ruby ninja out there verify that this is right? Also
> beware that the “CPU memory”, I.e. The one baked into Ruby, does not care
> about the “—men-type” on the command line.
>
> Andreas
>
> From: Hossein Nikoonia
> Reply-To: gem5
Dear list,
any idea on this?
On Wed, Nov 6, 2013 at 9:20 AM, Hossein Nikoonia wrote:
> Thanks Andreas ...
>
> I also guess this bandwidth is quite big ! ... I will also try with the
> latest source code ...
>
> To get things more complicated, I also repeat the above experien
Running the above experience with 4 CPU and 4 concurrent benchmarks (ALPHA,
Ruby, MOESI_CMP_token) results in 12.1 GBps memory controller bandwidth! I
guess this is too big for 1 GHz Ruby ... Isn't it?
On Sun, Nov 10, 2013 at 5:59 AM, Hossein Nikoonia wrote:
> Dear list,
>
> an
Ruby has a built in memory
> controller that scales one way or another with the network size.
>
> The “classic” SimpleDRAM controller behaves as I described earlier, with the
> peak bandwidth referring to the interface.
>
> Andreas
>
> From: Hossein Nikoonia mailto:nikoo...@g
Dear list,
The problem is that I do not see memory contention in gem5 :( ... let me
explain
I write two simple C++ codes:
- work: which accesses memory very frequently and each access produces a L1
and L2 miss
- sleep: which is a simple busy wait.
[Source codes bellow]
I run two very simple expe
Hi,
Yes. Hit ratio is almost zero. system.ruby.l2_cntrl0.L2cache.demand_hits is
59 while system.ruby.l2_cntrl0.L2cache.demand_accesses is 264194.
I forgot to mention that I have compiled those programs with -O3.
Any idea?
On Sat, Dec 7, 2013 at 5:50 PM, Mahmood Naderan wrote:
> Hi
> >- work: wh
Dear List,
I am running an experiment with gem5 in X86 + SE + Classic Memory.
The system has two cpus and runs two workloads. The workload is very
simple. It accesses memory in each cycle and (almost) each memory access
misses in L2 (miss ratio = 0.85 as reported by
system.l2.overall_miss_rate
Thanks...
On Sun, Dec 8, 2013 at 12:20 PM, Amin Farmahini wrote:
> MSHR, non blocking cache, memory level parallelism.
> On Dec 8, 2013 1:13 AM, "Hossein Nikoonia" wrote:
>
>> Dear List,
>>
>> I am running an experiment with gem5 in X86 + SE + Classic M
Dear list,
I have a question about validity of latency values in MOESI_CMP_token. L1
latency is set to 2, and L2 latency is set to 5. However, since the default
cpu frequency is 2x ruby frequency, it means that L1 responses (hit) in 4
cycles of cpu.
I guess this is not true in real systems. L1 sh
Dear list,
The default buffer_size in src/mem/ruby/slicc_interface/Controller.py is
defined zero which means infinite. Is it a correct configuration?
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
take a look at src/mem and src/mem/ruby
you should also decide about your memory type: Classic or Ruby.
On Wed, Dec 18, 2013 at 2:58 AM, Sanem Arslan wrote:
> Hi all,
>
> I have a question about cache protection. I want to apply parity or ECC on
> L1/L2 caches. Should I enable or disable ECC on
- telnet should not be related to that "disappearance".
- Its good to do /bin/sync before umount-ing the disk image. Did you ever
umount it?
- Remember that gem5 does not write on disk image file (by default... I
don't know how to make it write into the disk image) so if your code made a
file, it w
I couldn't get at what step I should sync.
>
>
>
>
> On Tue, Dec 24, 2013 at 4:56 AM, Hossein Nikoonia wrote:
>
>> - telnet should not be related to that "disappearance".
>> - Its good to do /bin/sync before umount-ing the disk image. Did you ever
>>
what about binding to 127.0.0.1 instead of 0.0.0.0 ?
On Sat, Jan 4, 2014 at 5:12 PM, Zohreh Naghibi wrote:
> Hi all
> I have implemented UDP socket programming in c code. It sends packets (1
> KB) via UDP port. The code works correctly in my own linux but NOT in
> gem5's Linux. Here's my output
I mean binding the server to 127.0.0.1
Are you doing that?
On Sat, Jan 4, 2014 at 5:24 PM, Zohreh Naghibi wrote:
> Thanks for your quick reply..
> I have tested it but I've got the same error.
>
>
> On Sat, Jan 4, 2014 at 5:50 AM, Hossein Nikoonia wrote:
>
>> wh
I guess the one that is shown in cpuinfo is fake... the system is already
configured with the cache you've specified in command line arguments.
Cache experts! Am I right?
On Tue, Jan 14, 2014 at 3:16 PM, Ahmad Hassan wrote:
>
> Also 'cat /proc/cpuinfo' shows the wrong cache size 1024KB where th
-d is also useful when you want to run multiple concurrent simulations on a
machine :)
On Sat, Feb 8, 2014 at 2:41 PM, Steve Reinhardt wrote:
> You can use the '-d' option to put all of the output files in a different
> directory for each run.
>
> Steve
>
>
>
> On Fri, Feb 7, 2014 at 9:13 PM, T
Maybe the problem is with your --mem-size ... 512MB is good to go
On Tue, Feb 18, 2014 at 5:32 AM, Rodrigo Reynolds Ramírez <
rodrigo.r...@hotmail.com> wrote:
> Hello everyone,
>
> I don't know why when I define a memory type different from simple_mem I
> get a Seg Fault. For example with --mem-
here is the file... It should be available in utexas website as well.
On Tue, Mar 4, 2014 at 10:46 AM, Alberto Martinez wrote:
> Hello,
> I am trying to run the Parsec benchmarks with Gem5. I followed the
> instructions listed in
>
> http://www.m5sim.org/PARSEC_benchmarks
>
> And the disk images
use --caches
On Wed, Mar 5, 2014 at 1:01 AM, Joseph Noor wrote:
> Hello,
>
> I am trying to run PARSEC benchmarks while varying the cache size and view
> the effect on simulation.
>
> For some reason, I am getting the same results no matter what cache size I
> am setting.
>
> My command line:
>
this is the inputsets.txt file :)
On Sun, Mar 9, 2014 at 2:58 PM, Alberto Martinez Guerrero <
alber...@gmail.com> wrote:
> Hossein Nikoonia gmail.com> writes:
>
> >
> >
> > here is the file... It should be available in utexas website as well.
> >
>
List,
I want to run multi-core ARM in FS mode. I guess number of cores are
limited to 4 (-n parameter of fs.py). Am I right?
If so, is it a logical problem? Is there any way to go around this?
Best
Seyed Hossein
___
gem5-users mailing list
gem5-users@ge
List,
I did compiled Gem5 in 2012-02-27 (=modification date of gem5.opt file)
successfully. However, when I "hg pull"ed today, I got
scons: Reading SConscript files ...
Error: Don't know what compiler options to use for your compiler.
Please fix SConstruct and src/SConscript and try again.
I've modified the wiki to reflect this ...
http://gem5.org/Introduction#Building
On Wed, Mar 7, 2012 at 12:03 AM, Malogro wrote:
> You have to declare the compiler as being GCC. Use this line:
>
> scons CC=gcc CXX=g++ build...etc
>
> This should work. Also, make sure you have all of the require
e found if it's a standard compiler
> (e.g. gcc) and in a normal location (e.g. /usr/bin).
>
> Ali
>
> On Mar 4, 2012, at 7:58 AM, Hossein Nikoonia wrote:
>
>> List,
>>
>> I did compiled Gem5 in 2012-02-27 (=modification date of gem5.opt file)
>> succ
Hi
Thats great to have kernel v3.2 !
Could you please kindly post your config and also your exact kernel version?
Did you modify kernel source code or you just config-ed it the right way?
Thank you very much in advance
Seyed Hossein
On Sun, Apr 15, 2012 at 11:12 PM, Jason Clemons wrote:
> Hey,
Did you
hg pull
hg update
?
You might have modified FSConfig.py (which probably you did this) and
hg might not be able to merge your changes with the recent version of
FSConfig ..
On Tue, Apr 17, 2012 at 12:08 PM, Hamid Reza Khaleghzadeh
wrote:
> Hello dear Mahmood,
>
> It worked. But I have a
ate both
> modified and not modified files by hg update?
>
> On 4/17/12, Hossein Nikoonia wrote:
>> Did you
>>
>> hg pull
>> hg update
>> ?
>>
>> You might have modified FSConfig.py (which probably you did this) and
>> hg might not be
Does http://gem5.org/Ubuntu_Disk_Image_for_ARM_Full_System help?
On Thu, Apr 19, 2012 at 12:29 PM, Ira Ray Jenkins
wrote:
> There are some provided images for Android and BBench using
> Gingerbread. Does anyone have any later/newer images, or can anyone
> direct me on how to build my own images?
Hi,
I guess you are interested to change the scheduling algorithm of OS so
you should go for modifying linux kernel ...
If you want to test your scheduling algorithm, you can either
- use performance measurement tools for your real system to see cache misses ...
- or use Gem5 to test your algo fo
nel of the used linux platform (e.g. ubantu
> ,fedora,...) as i am using MAC OS?
>
> Thank you in advance for your time.
> Thanks
> Best Regards
> Wael Amr
>
> On Sun, Apr 22, 2012 at 6:23 AM, Hossein Nikoonia
> wrote:
>>
>> Hi,
>>
>> I guess you are int
simple because scones is not installed!
what is the output of the following command? "whereis scons"
On Sat, Apr 28, 2012 at 9:51 PM, wael Amr wrote:
> Hi All,
>
> I have installed all the dependencies (i.e. scons,swig,python) using
> Macports for building and running gem5.
>
> Then i tried to b
Hey All + Nilay!
I'm going to compile m5thread for X86. I 'hg clone'd it. in tests dir,
i've changed CC and CPP to point to gcc and g++. but when i make:
In file included from ../pthread.c:48:0:
../profiling_hooks.h:37:6: error: #error "M5 profiling hooks not
implemented for your architecture, wr
Linux kernel is another big project. See kernel.org
Sent from my iPad
On May 11, 2012, at 2:25 PM, wael Amr wrote:
> Hello All,
>
> As i need to use gem5 for assigning and scheduling tasks (threads) among
> multicore platform
> I notified from this email
> http://www.mail-archive.com/gem5-us
take a look at graphviz, using a little bit of scripting, you might be able
to generated the desired FSM.
On Thu, Jun 14, 2012 at 7:27 AM, Nilay Vaish wrote:
> On Thu, 14 Jun 2012, GE ZHIGUO wrote:
>
> Hi, Is there any tool to convert SLICC description to FSM diagram
>> directly.
>> Thanks!
>>
List,
suppose you have configured (caches parameters and memory size) gem5 to
model a real X86 CPU. running a benchmark gives us some statistics. Running
this on the real CPU will also gives us some stats.
Is it possible/logical/acceptable (in computer architecture research
community) to compare t
are you sure you can run a multi-threaded benchmark (pthread or omp) in x86
SE mode?
for FS mode, you need to have a linux kernel compiled with the ability to
handle 256 cores
On Mon, Jun 25, 2012 at 9:38 AM, gem5 gem5 wrote:
> Hi all,
>
> I need to run some simulation with about 256 cores. So
List,
I'm (in the beginning of) going deep into Ruby memory model. However, I
cannot understand the bellow things:
- Why ruby is a part of 'system'? could it be child of 'root'? so 'root'
have two childs: 'system' (containing cpu and other things) and 'ruby'. is
it logical and/or possible?
- wher
rnel with the ability to handle 256 cores?
>
> Thanks!
>
> Jinzhu
>
>
> On Mon, Jun 25, 2012 at 1:46 AM, Hossein Nikoonia wrote:
>
>> are you sure you can run a multi-threaded benchmark (pthread or omp) in
>> x86 SE mode?
>>
>> for FS mode, you need to
pthread version of parsec2.1 on linux not gem5. Most of them
> support 256 threads very well. Does that mean linux kernel can handle 256
> cores well?
>
> On Mon, Jun 25, 2012 at 2:08 AM, Hossein Nikoonia wrote:
>
>> you should config the kernel with CONFIG_NR_CPUS options
o/gem5-stable/util/m5 -o
> $(TARGET)"
>
> Another one I found which might work is canneal. The same way to modify its
> makefile.
> On Mon, Jun 25, 2012 at 5:20 AM, Hossein Nikoonia wrote:
> i know linux can have 256 or more pthreads running but i don't k
Hi,
there should be better solutions. you might be able to define a new stat to
collect needed statistics.
see tutorials available on gem5 website…
Best
Seyed Hossein
On Jul 3, 2012, at 2:40 AM, Bojun Ma wrote:
> Hi,All
> I'd like to get the simulation result every few cycles in order to
t;
>
> On Tue, Jul 3, 2012 at 12:10 AM, Hossein Nikoonia wrote:
> Hi,
>
> there should be better solutions. you might be able to define a new stat to
> collect needed statistics.
> see tutorials available on gem5 website…
>
> Best
> Seyed Hossein
>
> On J
Yes, it is possible
see fs.py and included .py s.
you should create 3 (or more) "system" and make them child of "root"
when simulating one system, "root" has only one child. see m5out/config.ini.
Best
Seyed Hossein
On Tue, Jul 17, 2012 at 9:41 AM, Anirudh Sivaraman wrote:
> Hi
>
> I was wonder
I think their name in m5out/config.ini that is the same as the ones in
m5out/stats.text or ruby.stats should do it for you
On Aug 17, 2012, at 2:31 AM, Mahshid Sedghi wrote:
> Hello all,
>
> I need to print out the identifier of the memory controller module in order
> to distinguish the statis
This might help:
http://gem5.org/M5ops
You may try to complete it by reading source codes ... :D
On Wed, Aug 22, 2012 at 11:21 AM, Yi Liu wrote:
> Hi All
> I found there are several interesting m5ops such as "m5 pin
> [args ...]" which unfortunately has no document.
> Can someone give some d
see util/m5/m5.c and other files in that directory.
they use special ops that are not available in real cpus! they are only
available within gem5!
source code of them inside cpu source codes might also help!
On Wed, Aug 22, 2012 at 12:09 PM, Yi Liu wrote:
> 2012/8/22 Hossein Nikoo
see gem5.org
On Sat, Oct 13, 2012 at 3:07 PM, Bagus Budy wrote:
> Hi all
>
> I am a new comer here.
>
> I will hope help from every one about gem5 uses. I want to know, how can
> start gem5?
>
> Budy
>
> ___
> gem5-users mailing list
> gem5-users@gem
also see Parsec 3.
It is still in beta; it includes splash2 and splash2x.
Best
Seyed Hossein
On Thu, Oct 25, 2012 at 7:24 AM, GE ZHIGUO wrote:
> You might want to try this.
>
> http://www.capsl.udel.edu/splash/
>
> ** **
>
> Regards,
>
> ** **
>
> ** **
>
> *From:* gem5-users-boun
gem5 uses interconnection network within a "system" (i.e.
CPU+OS+Caches+DirecrotyControllers+ ...)
I guess what you want is a TCP/IP-network-connected systems. This is also
supported; at least for two systems. For more than that, you have to add
software-simulated swtich/routers between.
remember!
you can also use util/term/m5term 3456
you have to "make" it first :)
On Oct 29, 2012, at 4:16 PM, Payne, Benjamin wrote:
> Hi Marko,
>
> I am a new user to gem5, but I also experienced this confusion running full
> system mode. Once you launch the full system mode, you open a new terminal on
> From: gem5-users-boun...@gem5.org [mailto:gem5-users-boun...@gem5.org] On
> Behalf Of Hossein Nikoonia
> Sent: Friday, October 26, 2012 11:57 PM
> To: gem5 users mailing list
> Subject: Re: [gem5-users] cluster simulation
>
> gem5 uses interconnection network within a "system
This is up and running !
On Oct 29, 2012, at 4:24 PM, Payne, Benjamin wrote:
> Hello,
>
> I am unable to access the gem5 web pages. Is this true for anyone else?
>
> C:\Users\bpayne>ping gem5.org
>
> Pinging gem5.org [141.212.106.70] with 32 bytes of data:
> Request timed out.
> Request timed
69 matches
Mail list logo
