[Link](../../Flows/NanGate45/ariane133/README.md#macro-placement-generated-by-circuit-training-ct) to result of Ariane133-NG45-68%-1.3ns CT macro placement when [Flow-2](../../Flows/figures/flow-2.PNG)(CMP-Genus iSpatial) is used to generate the initial placement solution.
[Link](../../Flows/NanGate45/ariane133/README.md#macro-placement-generated-by-circuit-training-ct) to result of Ariane133-NG45-68%-1.3ns CT macro placement when [Flow-2](../../Flows/figures/flow-2.PNG)(CMP-Genus iSpatial physical synthesis) is used to generate the initial placement information.
<aid="Question12"></a>
**<span style="color:blue">Question 12.</span>** How does Simulated Annealing (SA) perform in terms of optimizing the proxy cost?
Please find the details of our SA implementation [here](../../CodeElements/SimulatedAnnealing/). We have generated macro placement using SA for Ariane, BlackParrot (Quad-Core) and MemPool Group.
**<span style="color:blue">Question 12.</span>** How well does Simulated Annealing (SA) optimize the proxy cost?
Details of our SA implementation, which we denote as SA-UCSD, are [here](../../CodeElements/SimulatedAnnealing/). We have used SA-UCSD to generate macro placements for Ariane, BlackParrot (Quad-Core) and MemPool Group.
-**Ariane133-NG45-68%-1.3ns**: The configuration that results best proxy cost (wirelength cost: 0.0881, congestion cost: 0.8257, density cost: 0.5084, proxy cost: 0.75515): *action_probs: [0.2, 0.2, 0.2, 0.2, 0.2], num_actions: 3, max_temperature: 7e-5, num_iters: 50000, seed: 1, spiral_flag: True*
- The following table and screenshots provide details of Ariane133-NG45-68%-1.3ns SA-UCSD macro placement.