Sayak Kundu

Sayak Kundu

Auto-generated profile, based on publicly available information

Sayak Kundu

University of California, San Diego, La Jolla, CA, USA

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 March 2023
Verified
Conference Paper
Open access
CCChung-Kuan Cheng
Chung-Kuan Cheng
AKAndrew B. Kahng
Andrew B. Kahng
SKSayak Kundu
Sayak Kundu
YWYucheng Wang
Yucheng Wang
ZWZhiang Wang
Zhiang Wang
We provide open, transparent implementation and assessment of Google Brain's deep reinforcement learning approach to macro placement (Nature) and its Circuit Training (CT) implementation in GitHub. We implement in open-source key "blackbox" elements of CT, and clarify discrepancies between CT and Nature. New testcases on open enablements are developed and released. We assess CT alongside multiple alternative macro placers, with all evaluation flows and related scripts public in GitHub. Our experiments also encompass academic mixed-size placement benchmarks, as well as ablation and stability studies. We comment on the impact of Nature and CT, as well as directions for future research.

12 June 2024
Verified
Conference Paper
Open access
AKAndrew B. Kahng
Andrew B. Kahng
SKSayak Kundu
Sayak Kundu
STShreyas Thumathy
Shreyas Thumathy
Multi-bit flip-flop clustering is a well-studied optimization problem in physical design: carefully merging multiple single-bit flip-flops into a single multi-bit flip-flop can decrease the total power consumption in a clock distribution network due to lower clock power and routed clock wirelength. We propose a pointset decomposition heuristic that in conjunction with capacitated k-means [4] enables a scalable, divide-and-conquer flow for multi-bit flip-flop clustering. Our flow produces high-quality flip-flop clustering and placement solutions with respect to total power consumption, area, timing, and wirelength metrics evaluated after the post-routing optimization (PRO) stage of P&R. We test our flow on five designs of varying input size (0.5K to 64K clusterable single-bit flip-flops) implemented using the ASAP7 7nm research enablement [3] [9]. Empirical results show that our new flow is competitive with current state-of-the-art flows. Compared to MeanShift [2], we achieve 6.18% (resp. 1.90%) maximum (resp. average) reduction in total power consumption, along with improved total negative slack and wirelength. Compared to FlopTray [4], we achieve a 400 × speedup on larger designs such as VGA (17K single-bit flip-flops), but with an average 1.12% degradation in total power consumption.

23 June 2024
Verified
Conference Paper
Open access
AKAndrew Kahng
Andrew Kahng
SKSeokhyeong Kang
Seokhyeong Kang
SKSayak Kundu
Sayak Kundu
KMKyungjun Min
Kyungjun Min
SPSeonghyeon Park
Seonghyeon Park
BPBodhisatta Pramanik
Bodhisatta Pramanik
Today's place-and-route (P&R) flows are increasingly challenged by complexity and scale of modern designs. Often, heuristics must trade off between turnaround time and quality of PPA outcomes. This paper presents a clustered placement methodology that improves both turnaround time and final-routed solution quality. Our PPA-aware clustering considers timing, power and logical hierarchy during netlist clustering, effectively reducing problem size and accelerating global placement runtime while improving post-route PPA metrics. Additionally, our machine learning (ML)-accelerated virtualized P&R methodology predicts the best cluster shapes (i.e., aspect ratios and utilizations) to use in P&R of the clustered netlist. With the open-source OpenROAD tool, our methods achieve up to 47% (average: 36%) global placement runtime improvement with similar half-perimeter wirelength (HPWL) and 90% (29%) improvement in post-route total negative slack (TNS). With the commercial Cadence Innovus tool, our methods achieve up to 3.92% (1%) improvement in power and 99% (49%) improvement in TNS.

15 December 2025 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Open access
CCChung-Kuan Cheng
Chung-Kuan Cheng
AKAndrew B. Kahng
Andrew B. Kahng
SKSayak Kundu
Sayak Kundu
YWYucheng Wang
Yucheng Wang
ZWZhiang Wang
Zhiang Wang
We provide an improved assessment of Google Brain’s deep reinforcement learning approach to macro placement [29] and its updated Circuit Training (CT) implementation in GitHub [53]. A stronger simulated annealing (SA) baseline leverages the “go-with-the-winners” metaheuristic [3] and a multi-threading implementation. We develop and release new public benchmarks in sub-10nm technology: LEF/DEF for Google’s 7nm TSMC Ariane protobuf and scaled variants, as well as testcases implemented in the open-source ASAP7 7nm research enablement. We evaluate from-scratch training and fine-tuning results for the latest “AlphaChip” release of Circuit Training, alongside multiple alternative macro placers. We also study the recently-published pre-training guidance in [53]. A commercial place-and-route tool is used to provide “true reward” post-route power, performance and area metrics. All data, evaluation flows and related scripts are publicly available in the MacroPlacement GitHub repository [63]. Our study affords insights into reproducibility and reporting in the research literature, and points out still-missing confirmations (e.g., of CT’s scalability and pre-training methodology) that remain open questions for the research community.


9 July 2024 in ACM Transactions on Design Automation of Electronic Systems
Verified
Journal Article
Open access
HEHadi Esmaeilzadeh
Hadi Esmaeilzadeh
SGSoroush Ghodrati
Soroush Ghodrati
AKAndrew Kahng
Andrew Kahng
JKJoon Kyung Kim
Joon Kyung Kim
SKSean Kinzer
Sean Kinzer
ZZZiqing Zeng
Ziqing Zeng
Parameterizable machine learning (ML) accelerators are the product of recent breakthroughs in ML. To fully enable their design space exploration (DSE), we propose a physical-design-driven, learning-based prediction framework for hardware-accelerated deep neural network (DNN) and non-DNN ML algorithms. It adopts a unified approach that combines power, performance, and area (PPA) analysis with frontend performance simulation, thereby achieving a realistic estimation of both backend PPA and system metrics such as runtime and energy. In addition, our framework includes a fully automated DSE technique, which optimizes backend and system metrics through an automated search of architectural and backend parameters. Experimental studies show that our approach consistently predicts backend PPA and system metrics with an average 7% or less prediction error for the ASIC implementation of two deep learning accelerator platforms, VTA and VeriGOOD-ML, in both a commercial 12 nm process and a research-oriented 45 nm process.