Closed VikasChidananda closed 1 year ago
Some specification on performance benchmarks:
run the benchmarks on CPU and GPU
Profiling to find bottleneck (if any)
Metrics for comparison: timesteps / unit time, updates / unit time
vary NN architecture linearly with increasing nodes(N) [learnable parameters M = AN + b]
[x] Open loop simulation (without control)
[x] Closed loop simulation (with reinforcement learning / classical controls?)
Ideas for configuring num_sources and num_loads:
num_sources
num_loads
num_source_i
num_loads_j
num_nodes
num_source
Common pitfalls while benchmarking on Linux machines
-RL benchmarks
Some specification on performance benchmarks:
run the benchmarks on CPU and GPU
Profiling to find bottleneck (if any)
Metrics for comparison: timesteps / unit time, updates / unit time
vary NN architecture linearly with increasing nodes(N) [learnable parameters M = AN + b]
[x] Open loop simulation (without control)
[x] Closed loop simulation (with reinforcement learning / classical controls?)
Ideas for configuring
num_sources
andnum_loads
:num_sources
=num_loads
num_source_i
,num_loads_j
) such thatnum_source_i
+num_loads_j
=num_nodes
: not requirednum_source
such thatnum_source
>num_loads
andnum_source
+num_loads
=num_nodes
: not required