Re: [Gems-users] GARNET TOOL

[hb166307@xxxxxxxx]

I think I can explain my problem a little better now. 

In rubyconfig.defaults, I have:
g_endpoint_bandwidth: 10000
g_PROCS_PER_CHIP: 1
g_NUM_PROCESSORS: 1
g_NUM_L2_BANKS: 1
g_NUM_MEMORIES: 1
g_PRINT_TOPOLOGY: true
g_NETWORK_TOPOLOGY: FILE_SPECIFIED
g_CACHE_DESIGN : NUCA

I build the ruby module for MOESI_CMP_token and enter this directory

~/GEMS/simics/targets/MOESI_CMP_token$ ./simics -stall

simics> read-configuration Checkpoints/ftt-bench.check

[A python script runs these command which is in the same directory]

run_command('instruction-fetch-mode instruction-fetch-trace')
run_command('istc-disable')
run_command('dstc-disable')
run_command('stc-status')
run_command('cpu-switch-time 1')
run_command('load-module ruby')
run_command('ruby0.init')
run_command('continue') 

-------------------------------------------------------------------------------
SIMICS OUTPUT

[cpu0 info] Note that on this cpu, instruction-fetch-trace is implemented using
instruction-cache-access-trace with a suitable cache line size.
Turning I-STC off and flushing old data
Turning D-STC off and flushing old data
D-STC is currently *OFF*
I-STC is currently *OFF*
The switch time will change to 1 cycles (for CPU-0) once all processors have
synchronized.
successful installation of the ruby timing model.
Ruby Timing Mode
Creating event queue...
Creating event queue done
Creating system...
  Processors: 1

[This is a cout statement in Topoogy.C to see file name created]
Filename:
../../../ruby/network/simple/Network_Files/NUCA_Procs-1_ProcsPerChip-1_L2Banks-1_Memories-1.txt

failed assertion 'atoi((string_split(varStr, ':')).c_str()) ==
g_endpoint_bandwidth' at fn void Topology::makeFileSpecified() in
network/simple/Topology.C:540
failed assertion 'atoi((string_split(varStr, ':')).c_str()) ==
g_endpoint_bandwidth' at fn void Topology::makeFileSpecified() in
network/simple/Topology.C:540
At this point you might want to attach a debug to the running and get to the
crash site; otherwise press enter to continue
PID: 8421
-------------------------------------------------------------------------------

>From this error, I can not tell where g_endpoint_bandwith is coming from in
Topology.C, which might be why I am getting an assertion error, but I could be
wrong. There was NUCA_Procs-1_ProcsPerChip-1_L2Banks-1_Memories-1.txt in the
Network_Files Folder that was allready in there and when I remove the line
corresponding to 

bw_unit: ####

It no longer gives me the error. What effect is this going to have? Sorry for
not being clearer with my problem....

Quoting "Niket Agarwal (niketa@xxxxxxxxxxxxx)" <niketa@xxxxxxxxxxxxx>:

> 
> In the GEMS setup you can specify (either in the rubyconfig.defaults
> or while running) that you want to use the FILE_SPECIFIED topology
> type and the NUCA cache design with a particular number of
> processors, processors per chip, L2 banks and Memory banks. The
> topology file name is automatically calculated from that. 
> 
> You do not have to worry about the file name and folder. 
> 
> Niket
> 
> ----- Original Message -----
> From: hb166307@xxxxxxxx
> Date: Monday, September 29, 2008 11:14 am
> Subject: Re: [Gems-users] GARNET TOOL
> To: Gems Users <gems-users@xxxxxxxxxxx>
> 
> > >From what I understand the network description file I want to use
> 
> > should be in
> > GEMS/ruby/network/simple/Network_Files, but where to I specify the
> 
> > file name
> > with the of the file name in that folder that I want to use. I 
> > assume it would
> > be in the rubyconfig.defaults file. I could be wrong ...
> > 
> > Quoting Niket <niketa@xxxxxxxxxxxxx>:
> > 
> > > Yes, you have to specify the topology in the 
> > > GEMS/ruby/network/simple/Network_Files
> > > 
> > > You should name your file to be consistent with the number of L1,
> 
> > L2,> 
> > > Memory, etc. Look at other sample files provided in the same
> folder.
> > > 
> > > The ruby dump will have more network results like, average 
> > network 
> > > latency, link utilization, etc.
> > > 
> > > Cheers,
> > > Niket
> > > 
> > > hb166307@xxxxxxxx wrote:
> > > > Dear Users,
> > > >
> > > > I seem to missing the point when it comes to specifying a
> network
> > > file for the
> > > > Garnet tool. I have set the following
> > > >
> > > > g_GARNET_NETWORK true
> > > > g_DETAIL_NETWORK true
> > > > g_NETWORK_TOPOLOGY: FILE_SPECIFIED
> > > >
> > > > but am not sure where to you tell ruby to used a NETWORK_FILE
> from
> > > what I
> > > > assuming will be one of the files in the
> > > GEMS/ruby/network/simple/Network_Files
> > > > directory. When I get some help with this, does anyone know if
> a
> > > ruby stats dump
> > > > will have more network metric type information printed.
> > > >
> > > > Thanks in advance!
> > > > _______________________________________________
> > > > Gems-users mailing list
> > > > Gems-users@xxxxxxxxxxx
> > > > https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> > > > Use Google to search the GEMS Users mailing list by adding
> > > "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your
> search.
> > > >
> > > >   
> > > 
> > > _______________________________________________
> > > Gems-users mailing list
> > > Gems-users@xxxxxxxxxxx
> > > https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> > > Use Google to search the GEMS Users mailing list by adding
> > > "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your
> search.
> > > 
> > > 
> > _______________________________________________
> > Gems-users mailing list
> > Gems-users@xxxxxxxxxxx
> > https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> > Use Google to search the GEMS Users mailing list by adding 
> > "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your
> search.
> > 
> > 
> _______________________________________________
> Gems-users mailing list
> Gems-users@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/gems-users
> Use Google to search the GEMS Users mailing list by adding
> "site:https://lists.cs.wisc.edu/archive/gems-users/"; to your search.
> 
>