TY - JOUR
T1 - GAMER
T2 - GPU-Accelerated Maze Routing
AU - Lin, Shiju
AU - Liu, Jinwei
AU - Young, Evangeline F.Y.
AU - Wong, Martin D. F.
N1 - Funding Information:
This work was supported in part by the Grant from the AI Chip Center for Emerging Smart Systems Ltd.
Publisher Copyright:
© 1982-2012 IEEE.
PY - 2023/2
Y1 - 2023/2
N2 - Maze routing is usually the most time-consuming step in global routing and detailed routing. A commonly used maze routing method is to start from one pin and iteratively connect the current route to the closest unconnected pin. This method reduces the maze routing problem to multiple multisource-multidestination shortest path problems. The shortest path problem in VLSI routing has: 1) rectilinear routing directions and 2) preferably small via usage. By utilizing these two characteristics, we propose a novel parallel algorithm called GAMER to accelerate the multisource-multidestination shortest path problem for VLSI routing. GAMER decomposes the shortest path search into alternating vertical and horizontal sweep operations, and two parallel algorithms are proposed to accelerate a sweep operation from O(n2) to O(log2n) on a grid graph of n×n. Several techniques of applying GAMER on irregular routing regions are also introduced. Experiments are conducted by integrating GAMER into the state-of-the-art academic global router CUGR. CUGR adopts a two-level maze routing scheme, including coarse-grained routing and fine-grained routing, and they can be accelerated by 19.85× and 2.59×, respectively, with GAMER, achieving an overall speedup of 2.7× without quality degradation.
AB - Maze routing is usually the most time-consuming step in global routing and detailed routing. A commonly used maze routing method is to start from one pin and iteratively connect the current route to the closest unconnected pin. This method reduces the maze routing problem to multiple multisource-multidestination shortest path problems. The shortest path problem in VLSI routing has: 1) rectilinear routing directions and 2) preferably small via usage. By utilizing these two characteristics, we propose a novel parallel algorithm called GAMER to accelerate the multisource-multidestination shortest path problem for VLSI routing. GAMER decomposes the shortest path search into alternating vertical and horizontal sweep operations, and two parallel algorithms are proposed to accelerate a sweep operation from O(n2) to O(log2n) on a grid graph of n×n. Several techniques of applying GAMER on irregular routing regions are also introduced. Experiments are conducted by integrating GAMER into the state-of-the-art academic global router CUGR. CUGR adopts a two-level maze routing scheme, including coarse-grained routing and fine-grained routing, and they can be accelerated by 19.85× and 2.59×, respectively, with GAMER, achieving an overall speedup of 2.7× without quality degradation.
KW - Electronic design automation (EDA)
KW - global routing
KW - graphics processing unit (GPU) acceleration
KW - maze routing
KW - physical design
UR - http://www.scopus.com/inward/record.url?scp=85132766278&partnerID=8YFLogxK
U2 - 10.1109/TCAD.2022.3184281
DO - 10.1109/TCAD.2022.3184281
M3 - Journal article
AN - SCOPUS:85132766278
SN - 0278-0070
VL - 42
SP - 583
EP - 593
JO - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
JF - IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
IS - 2
ER -