No that's not as simple because each router performs a vc allocation and 
assumes that the next input port has vcs_per_vnet x num_vnet number of VCs.

If you want to experiment with different number of VCs in each input port, then 
you should update the vc_select function in the switch allocator and use a 
different VC number for each router. It's somewhat of a hack. A more robust 
technique might be to initialize different number of VCs in each router -- but 
a lot of the code is written assuming same number of VCs. So it's easier to 
change the vc select function and make sure some VCs are never assigned 
depending on the router id (effectively modeling fewer VCs).
Take a look at the wakeup function in 
SwitchAllocator.cc<http://SwitchAllocator.cc>

Cheers,
Tushar


On Aug 19, 2017, at 4:41 AM, matild breo 
<matild.b...@gmail.com<mailto:matild.b...@gmail.com>> wrote:

Hi Tushar

Thank you very much. if I want to implement different number of vcs in 
different links, can I use vcs_per_vnet with different value in topology file? 
for example as a link is created, I assign the vcs_per_vnet?
like bellow:
                    int_links.append(IntLink(link_id=link_count,
                                             src_node=routers[east_out],
                                             dst_node=routers[west_in],
                                             src_outport="East",
                                             dst_inport="West",
                                             latency = link_latency,
                                             weight=1))
would be correct that I assign vcs_per_vnet in this function?

Thanks

On Sat, Aug 19, 2017 at 10:19 AM, matild breo 
<matild.b...@gmail.com<mailto:matild.b...@gmail.com>> wrote:
Hi every body
I have some questions:

1- what is default link width in garnet 2 in bytes?

2- if the ni_flit_size is 16 byte, does it mean 2 flit pass from every link in 
one cycle (for example link width is 128bit)?

3-what does mean vcs_per_vnet?

4- how can determine vcs per link??

Thank you

_______________________________________________
gem5-users mailing list
gem5-users@gem5.org<mailto:gem5-users@gem5.org>
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to