benchmark suites like SPEC, PARSEC, SPLASH etc. Currently we can invoke
garnet under uniform, tornado and bit complement traffic patterns. How can
Post by Tushar Krishna**
Hi Pavan,
Virtual networks carry different coherence message types to avoid
protocol level deadlocks in gem5.
As a simplest example, vnet 0 might carry requests and vnet 1
might carry responses.
Thus each vnet has its own set of buffers to avoid deadlocks.
Each vnet itself can have one or more virtual channels (VCs) for
performance reasons (to avoid head of line blocking etc).
VCs within vnets that carry control messages (typically 1-flit
wide) are 1 buffer deep.
VCs within vnets that carry data messages (typically 5-flit wide)
are typically more than 1 buffer deep.
- Tushar
Hi Tushar,
Thanks for the tip!! Now, I am able to run the code without it
being aborted. I am also able to see some network related statistics in the
ruby.stats file.
I am trying to understand how the router architecture has been
modeled within garnet. There are two fields which I am not able to
understand as to what do they mean.
1. m_virtual_networks -> number of virtual networks
2. m_vc_per_vnet -> number of vcs per virtual network
What does number of virtual networks mean?
Thanks,
Pavan
Hi Pavan,
I don't think its been pushed yet as I didnt see any email either.
If you are in a hurry, perhaps you can just add the change
ruby_port.access_phys_mem = False
i += 1
system.system_port = system.ruby._sys_port_proxy.port
But remember to remove it when you do a hg pull -u later when the
patch gets pushed in.
cheers,
Tushar
Hi Tushar,
I did not receive any emails regarding the patch for the issue
with GARNET. I was just wondering may be I subscribed to a different list
and the patch is already out. Has the patch been approved and already out
in the latest version of gems code?
Thanks a lot!!
Thanks,
Pavan
No emails related to code reviews and changes only come on the dev list.
Hi Tushar,
I am currently building gem5 binary with MOESI_hammer. I will try
to invoke garnet with that binary as you suggested after it is done. As per
the patch, can I subscribe to the gem5-dev list to receive the emails
regarding it or do those emails appear on gem5-users list too?
Thanks,
Pavan
On Fri, Jan 27, 2012 at 11:57 AM, Tushar Krishna <
Post by Tushar KrishnaHi Pavan,
I saw Nilay approve the patch on the gem5-dev mailing list today
so just wait for the author of the patch (Andreas) to push it in. You will
see an email on the gem5-dev list when he does that.
The Network_test protocol is tied to the ruby_network_test.py script.
It basically allows you to perform network-only simulations.
By running a different protocol, I meant building MOESI_hammer
or MOESI_CMP_directory for instance, and running it with
ruby_random_test.py, ruby_se.py and so on.
Basically garnet itself can be plugged in when running any protocol.
If you do want to run garnet-only simulations, I would say wait
for the patch to be pushed in and then perform a hg pull -u and run the
same command you ran earlier.
ruby_network_test.py seems to have broken recently due to some
other changes.
cheers,
Tushar
Hi Tushar,
Thank you for your quick reply. Can you please tell me how can
I get access to this patch you are suggesting? And also you said that with
any other protocol garnet would work fine. I did not really understand the
statement. Does changing the protocol mean running a different python
script instead of ruby_network_test.py or changing the topology?
Thanks,
Pavan
Yeah the warnings are always there but they are not a problem.
The specific error with regard to the system port are specific to the ruby
network tester which the patch fixes.
If you run garnet with any other protocol, it should work fine.
- Tushar
I've been running Garnet Networks for a while in GEM5 and I've
always had those warnings. They do not seem to interfere with the overall
results (not the ones I'm using anyway)
Also, the warnings don't appear in flexible pipeline Garnets.
Have you tried that to see what happens?
Your command line ran just fine over here both in the debug and
opt versions (provided the missing space after --fixed-pkts was a typo)
If you still have no success, I'd do what Tushar said and wait
for the patch to be pushed.
Regards,
Alexandre Yasuo Yamamoto
On Fri, Jan 27, 2012 at 11:17 AM, Tushar Krishna <
Post by Tushar KrishnaHi Pavan,
As you probably noticed, a patch has just been approved on the
gem5-dev list to fix this issue...
Keep a look out on that list for the patch to be pushed, and
then try it ..
cheers,
Tushar
Hello,
I am trying to invoke garnet module in gem5.
As given in the documentation I built gem5 binary using
*scons build/ALPHA_SE_Network_Test/gem5.debug*
After that, I executed the following command
*./build/ALPHA_SE_Network_test/gem5.debug configs/example/ruby_network_test.py --num-cpus=16 --num-dirs=16 --topology=Mesh --mesh-rows=4 --sim-cycles=1000 --injectionrate=0.01 --synthetic=0 --fixed-pkts--maxpackets=1 --garnet-network=fixed*
I get a lot of warning messages that say
*warning: add_child('cls') : child 'credit_links0 credit_links1' already has parent*
and then a panic message like the following is displayed
*panic: System port on system is not connected.* (I checked out this message is printed from system.cc file in sim directory which is under src directory)*Program aborted at cycle 0.
Aborted.*
I have tried reading through the documentation as well as the Frequently Asked Questions to find out a solution for it. I was not able to.
Any help/advice on solving this issue is greatly appreciated.
Thanks for your time.
Thanks,
Pavan
_______________________________________________
_______________________________________________
gem5-users mailing list
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
_______________________________________________
gem5-users mailing list
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
_______________________________________________
gem5-users mailing list
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
_______________________________________________
gem5-users mailing list
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
_______________________________________________
gem5-users mailing list
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
_______________________________________________
gem5-users mailing list
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users