Placement and routing for 3D-FPGAs using reinforcement learning and support vector machines

Abstract

The primary advantage of using 3D-FPGA over 2D-FPGA is that the vertical stacking of active layers reduce the Manhattan distance between the components in 3D-FPGA than when placed on 2D-FPGA. This results in a considerable reduction in total interconnect length. Reduced wire length eventually leads to reduction in delay and hence improved performance and speed. Design of an efficient placement and routing algorithm for 3D-FPGA that fully exploits the above mentioned advantage is a problem of deep research and commercial interest. In this paper, an efficient placement and routing algorithm is proposed for 3D-FPGAs which yields better results in terms of total interconnect length and channel-width. The proposed algorithm employs two important techniques, namely, Reinforcement Learning (RL) and Support Vector Machines (SVMs), to perform the placement. The proposed algorithm is implemented and tested on standard benchmark circuits and the results obtained are encouraging. This is one of the very few instances where reinforcement learning is used for solving a problem in the area of VLSI.

Efficient Detailed Routing for FPGA Back-End Flow Using Reinforcement Learning

“…It is important to mention here that the aforementioned state-of-the-art work mostly uses ML or artificial intelligence to optimize the auto tuning of the parameters for FPGA backend flow. These frameworks obtain the desired results through extensive training of underlying ML algorithms, such as Support Vector Machine (SVM) [26], Bayesian Learning (BL) and Knowledge-Based Neural Networks (KBNN) [27].…”

Review of Machine Learning in Logic Synthesis

“…Reinforcement learning and LS: Reinforcement learning is a ML paradigm in contrast to other learning types, such as supervised and unsupervised learning. This concept is further studied and matured in the physical design area, being explored in processes of placement and routing, for example, in Goldie [23], Mirhoseini [22], and Manimegalai [57], just to cite a few examples. In contrast to the physical design, LS techniques based on Reinforcement Learning (RL) are only recently being experimented with.…”

Guiding Global Placement With Reinforcement Learning

“…[8] was able to learn a policy to explicitly place a smaller number of large macro cells before using a force based method to place the remaining cells. Other work such as [9] has studied using reinforcement learning for the assignment of logic elements to FPGA logic blocks. However, these differ from our work significantly as we investigate ways to directly improve the forcebased method used to place smaller standard cells.…”