Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ICNT Bound #345

Open
Evane5cence opened this issue Nov 4, 2024 · 2 comments
Open

ICNT Bound #345

Evane5cence opened this issue Nov 4, 2024 · 2 comments

Comments

@Evane5cence
Copy link

Hi,
I found that the performance of ICNT performance is low, making my application bounded by it.
May I ask how can I tune the config file to increase the ICNT performance?
Thank you!

@Evane5cence
Copy link
Author

----------------------------Interconnect-DETAILS--------------------------------
Req_Network_injected_packets_num = 2872744
Req_Network_cycles = 262598
Req_Network_injected_packets_per_cycle =      10.9397 
Req_Network_conflicts_per_cycle =      64.3919
Req_Network_conflicts_per_cycle_util =      65.1456
Req_Bank_Level_Parallism =      11.0677
Req_Network_in_buffer_full_per_cycle =       0.0000
Req_Network_in_buffer_avg_util =     250.5793
Req_Network_out_buffer_full_per_cycle =       0.8221
Req_Network_out_buffer_avg_util =      11.4200

Reply_Network_injected_packets_num = 2872744
Reply_Network_cycles = 262598
Reply_Network_injected_packets_per_cycle =       10.9397
Reply_Network_conflicts_per_cycle =       13.6370
Reply_Network_conflicts_per_cycle_util =      13.9928
Reply_Bank_Level_Parallism =      11.2252
Reply_Network_in_buffer_full_per_cycle =       0.0000
Reply_Network_in_buffer_avg_util =       1.4637
Reply_Network_out_buffer_full_per_cycle =       0.0000
Reply_Network_out_buffer_avg_util =       0.1013
----------------------------END-of-Interconnect-DETAILS-------------------------

It seems that the Req_Network_conflicts_per_cycle is too high, and becomes the bottleneck.
How should I solve this?

@JRPan
Copy link
Collaborator

JRPan commented Nov 7, 2024

You can try different scheduling algorithms. I believe default uses iSLIP_Advance()

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants