Implementation
1) nINTER, nLOCAL 대신 Round-Robin으로 선택하게 하고
Rx, Ry의 같은 방향으로 가는 port에 대해서만 prioritized arbitration구현
2) Pipeline stage
모두다 no VA stage
Baseline router: I, SA, ST, LT
Turning packet: SA/STx/INT, STy/LT
X->X: SA/STx/LT (1 cycle)
Y->Y: SA/STy/LT (1 cycle)
3) Network-only simulator
ex) a uniform random traffic
Evaluation
1) n value를 6개 사용한것과 4개 사용한것 비교
2) Count number: How many times is nX > n_max used
3) CPI & Average network latency
4) Baseline router의 Input Buffer: 2, 4로 실험
5) Network-only simulator에서
for the network-only, you dont' need VCs.. so 16x1 should be sufficient.
Pipeline (I/VA/SA/ST/LT 원래 그대로 실험)
@ BASELINE
/home/berebere/gems2/simics
buf가 2, 4일때 -> baseline_buf_4, baseline_buf_2
@ 4개 n value, Prioritized arbitration
/home/berebere/gems/simics
buf=2, inter_buf=4, n_max=4 일때 -> low_cost_max4
@ 6개 n value, Prioritized arbitration
/home/berebere/gems3/simics
buf=2, inter_buf=4, n_max=4 일때 -> low_cost_6n_max4
nX > n_max used 개수 비교
CPI, Average network latency 비교.
@ Baseline router: I, SA, ST, LT (flexible pipeline, 4 cycle)
/home/berebere/gems4/simics
buf가 2, 4일때 -> baseline_buf_4, baseline_buf_2
buf가 2이고 5stage (VA 포함) 일때 -> baseline_buf_2_5stage
@ Baseline router
pipeline: I/VA, SA, ST, LT (4 cycle)
/home/berebere/gems5/simics
buf가 2, 4일때 -> baseline_buf_4, baseline_buf_2
@ 4개 n value, Prioritized arbitration
Turning packet: SA/STx/INT, STy/LT
X->X or Y->Y : SA/(STx or STy)/LT (1 cycle)
@ Fixed-pipeline
Uniform Random traffic, 0.1~1.0까지 실험
buffer size = 16
@ Flexible-pipeline
Uniform Random traffic, 0.1~1.0까지 실험
buffer size = 16, pipeline_stage = 5
0.1 /home/berebere/gems_net_only/simics
0.2 /home/berebere/gems_net_only2/simics
0.3 /home/berebere/gems_net_only3/simics
0.4 /home/berebere/gems_net_only4/simics
0.5 /home/berebere/gems_net_only5/simics
0.6 /home/berebere/gems_net_only6/simics
0.7 /home/berebere/gems_net_only7/simics
0.8 /home/berebere/gems_net_only8/simics
0.9 /home/berebere/gems_net_only9/simics
