Xiaqing Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xiaqing Li has a total of 63 co-authors. Here are the top co-authors by number of shared publications:
- Qi Guo
Institute of Computing Technology, Chinese Academy of Sciences
11 shared publications - Zidong Du
Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China
11 shared publications - Rui Zhang
Institute of Computing Technology, Chinese Academy of Sciences
7 shared publications - Yifan Hao
Institute of Computing Technology, Chinese Academy of Sciences
7 shared publications - Xing Hu
Institute of Computing Technology, Chinese Academy of Sciences
6 shared publications - Yongwei Zhao
Institute of Computing Technology, Chinese Academy of Sciences
6 shared publications
Latest Publications
- Open Access
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 489-503
Mosaic: Efficient CPU–GPU Coordinated Block-Sparse Attention for Long-Context Decoding
- 14 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 489-503
Open AccessMosaic: Efficient CPU–GPU Coordinated Block-Sparse Attention for Long-Context Decoding
- Open Access
- 3 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
- Vol. 32 (1) , 1-27
LE-Timing: Layout-Aware and Explainable Timing Prediction
Accurate and fast timing prediction at early design stages is crucial for achieving timing closure in very-large-scale integration (VLSI) design. Machine learning (ML) based approaches have been widely adopted for timing prediction to pursue both accuracy and speed. Unfortunately, these approaches fail to fully exploit the layout features, which can significantly impact the timing metrics, thus leading to great accuracy degradation. In this article, we propose LE-Timing, an end-to-end layout-aware timing prediction approach that can achieve better accuracy by leveraging important layout information. The key insight is that crucial information about coupling capacitance between routing wires can significantly impact timing metrics, and fully exploiting this layout information in the timing prediction model can substantially improve its accuracy. However, efficiently leveraging layout features while maintaining accuracy poses three key challenges: node-level layout-graph integration, domain experience utilization and training efficiency. To address these, we further propose three key technologies specially designed for LE-Timing correspondingly, including a layout-encoding net model, an explainable cell model, and an explicit-message-passing scheme. Experimental results on 16 open-source designs with an advanced 7-nm technology node demonstrate that LE-Timing achieves R 2 score 0.94 for arrival time prediction and lower error accumulation compared to the baselines.
- 3 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Design Automation of Electronic Systems
- Vol. 32 (1) , 1-27
Open AccessLE-Timing: Layout-Aware and Explainable Timing Prediction
Accurate and fast timing prediction at early design stages is crucial for achieving timing closure in very-large-scale integration (VLSI) design. Machine learning (ML) based approaches have been widely adopted for timing prediction to pursue both accuracy and speed. Unfortunately, these approaches fail to fully exploit the layout features, which can significantly impact the timing metrics, thus leading to great accuracy degradation. In this article, we propose LE-Timing, an end-to-end layout-aware timing prediction approach that can achieve better accuracy by leveraging important layout information. The key insight is that crucial information about coupling capacitance between routing wires can significantly impact timing metrics, and fully exploiting this layout information in the timing prediction model can substantially improve its accuracy. However, efficiently leveraging layout features while maintaining accuracy poses three key challenges: node-level layout-graph integration, domain experience utilization and training efficiency. To address these, we further propose three key technologies specially designed for LE-Timing correspondingly, including a layout-encoding net model, an explainable cell model, and an explicit-message-passing scheme. Experimental results on 16 open-source designs with an advanced 7-nm technology node demonstrate that LE-Timing achieves R 2 score 0.94 for arrival time prediction and lower error accumulation compared to the baselines.
- Open Access
- 27 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 381-391
Cross-Layer Performance Analysis of Single-GPU Large Language Model Inference
The performance of single-GPU LLM inference is characterized by strong cross-layer interactions spanning model architecture, runtime scheduling, operator execution, and GPU microarchitecture. Unfortunately, a unified understanding of single-GPU LLM inference bottlenecks is still lacking due to two limitations: the lack of traceable cross-layer attribution and the static characterization of KV-cache usage. To this end, we present a cross-layer analysis approach for single-GPU LLM inference that jointly characterizes latency and memory behavior. However, this requires addressing two challenges: cross-layer attribution and dynamic KV-cache characterization. To address the two challenges, we employ four-layer joint profiling for cross-layer attribution and introduce a logical KV computation metric for dynamic KV-cache characterization, respectively. Finally, we systematically characterize representative dense and Mixture-of-Experts (MoE) models under diverse workloads on a single A100 GPU, identify key bottlenecks across inference phases and architectures, and reveal opportunities for further optimization.
- 27 September 2026
- Published by Association for Computing Machinery (ACM)
- p. 381-391
Open AccessCross-Layer Performance Analysis of Single-GPU Large Language Model Inference
The performance of single-GPU LLM inference is characterized by strong cross-layer interactions spanning model architecture, runtime scheduling, operator execution, and GPU microarchitecture. Unfortunately, a unified understanding of single-GPU LLM inference bottlenecks is still lacking due to two limitations: the lack of traceable cross-layer attribution and the static characterization of KV-cache usage. To this end, we present a cross-layer analysis approach for single-GPU LLM inference that jointly characterizes latency and memory behavior. However, this requires addressing two challenges: cross-layer attribution and dynamic KV-cache characterization. To address the two challenges, we employ four-layer joint profiling for cross-layer attribution and introduce a logical KV computation metric for dynamic KV-cache characterization, respectively. Finally, we systematically characterize representative dense and Mixture-of-Experts (MoE) models under diverse workloads on a single A100 GPU, identify key bottlenecks across inference phases and architectures, and reveal opportunities for further optimization.
- 3 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (5) , 2362-2375
AGON: Automated Design Framework for Customizing Processors From ISA Documents
Customized processors are essential for domain-specific applications such as the Internet of Things (IoT) and multimedia embedded systems, yet their design often requires extensive expert intervention. Traditional approaches, including hardware design using encapsulated abstractions (e.g., Chisel) and high-level synthesis (HLS) from languages like C or SystemC, reduce some manual efforts but remain either costly or suboptimal. Recent explorations into leveraging large language models (LLMs) to generate register-transfer level (RTL) from natural language specifications have shown promise, but these methods still struggle with generating complex and high-performance processors, mainly due to the complicated low-level details in the RTL code. In this work, we introduce AGON, a novel framework designed to facilitate the development of customized processor RTL from instruction set architecture (ISA) documents using LLMs. The framework comprises two layers: a functional description layer and a hardware implementation layer. At the functional layer, AGON employs a nano-operator (nOP)-based intermediate representation (IR) that abstracts basic instruction operations, thereby reducing the semantic gap between natural language and the RTL code. This abstraction significantly shortens the descriptive code required for LLM generation, improving the generation accuracy in a single pass. At the hardware layer, AGON offers three abstraction levels (i.e., instruction, ISA, and processor) along with rule-based primitives to systematically lower the nOP-based IR into a fully optimized processor implementation. This decoupled design not only ensures correctness-by-construction but also enables automated, power, performance, and area (PPA)-aware performance optimization. We evaluate AGON by designing high-performance out-of-order (OoO) processors that correctly execute practical programs. Experimental results demonstrate that processors generated with AGON achieve an average speed of $4.51\times $ on specific tasks compared to expert-designed general-purpose central processing units (CPUs) while requiring minimal design effort.
- 3 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 45 (5) , 2362-2375
AGON: Automated Design Framework for Customizing Processors From ISA Documents
Customized processors are essential for domain-specific applications such as the Internet of Things (IoT) and multimedia embedded systems, yet their design often requires extensive expert intervention. Traditional approaches, including hardware design using encapsulated abstractions (e.g., Chisel) and high-level synthesis (HLS) from languages like C or SystemC, reduce some manual efforts but remain either costly or suboptimal. Recent explorations into leveraging large language models (LLMs) to generate register-transfer level (RTL) from natural language specifications have shown promise, but these methods still struggle with generating complex and high-performance processors, mainly due to the complicated low-level details in the RTL code. In this work, we introduce AGON, a novel framework designed to facilitate the development of customized processor RTL from instruction set architecture (ISA) documents using LLMs. The framework comprises two layers: a functional description layer and a hardware implementation layer. At the functional layer, AGON employs a nano-operator (nOP)-based intermediate representation (IR) that abstracts basic instruction operations, thereby reducing the semantic gap between natural language and the RTL code. This abstraction significantly shortens the descriptive code required for LLM generation, improving the generation accuracy in a single pass. At the hardware layer, AGON offers three abstraction levels (i.e., instruction, ISA, and processor) along with rule-based primitives to systematically lower the nOP-based IR into a fully optimized processor implementation. This decoupled design not only ensures correctness-by-construction but also enables automated, power, performance, and area (PPA)-aware performance optimization. We evaluate AGON by designing high-performance out-of-order (OoO) processors that correctly execute practical programs. Experimental results demonstrate that processors generated with AGON achieve an average speed of $4.51\times $ on specific tasks compared to expert-designed general-purpose central processing units (CPUs) while requiring minimal design effort.
- 9 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (12) , 4643-4656
VariPar: Variation-Aware Workload Partitioning in Chiplet-Based DNN Accelerators
Chiplet-based DNN accelerators have been extensively explored to save design and manufacturing costs. Previous works regard all chiplets as identical and employ uniform workload partitioning strategies. These workload partitioning strategies overlook various real-world factors that contribute to remarkable performance variations among chiplets, including manufacturing process variation, thermal condition, physical placement, and power supply condition. When considering these performance variations, a variation-aware workload partitioning can achieve superior performance. This article introduces VariPar, a systematic framework to employ variation-aware partitioning strategy in chiplet-based DNN accelerators. VariPar models performance variations for each chiplet and partition workloads accordingly. VariPar includes a simulator with multifactor variation modeling and a heuristic search engine to generate near-optimal partitioning within a reasonable time. Experiment results show that VariPar achieves $1.45{\times }$ performance and $1.82{\times }$ energy efficiency improvement on average when compared to uniform partitioning (UP) strategy.
- 9 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (12) , 4643-4656
VariPar: Variation-Aware Workload Partitioning in Chiplet-Based DNN Accelerators
Chiplet-based DNN accelerators have been extensively explored to save design and manufacturing costs. Previous works regard all chiplets as identical and employ uniform workload partitioning strategies. These workload partitioning strategies overlook various real-world factors that contribute to remarkable performance variations among chiplets, including manufacturing process variation, thermal condition, physical placement, and power supply condition. When considering these performance variations, a variation-aware workload partitioning can achieve superior performance. This article introduces VariPar, a systematic framework to employ variation-aware partitioning strategy in chiplet-based DNN accelerators. VariPar models performance variations for each chiplet and partition workloads accordingly. VariPar includes a simulator with multifactor variation modeling and a heuristic search engine to generate near-optimal partitioning within a reasonable time. Experiment results show that VariPar achieves $1.45{\times }$ performance and $1.82{\times }$ energy efficiency improvement on average when compared to uniform partitioning (UP) strategy.
- 20 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (10) , 3962-3975
SaaP: Rearchitect SoC-as-a-Processor to Orchestrate Hardware Heterogeneity
Due to the end of Moore’s Law and Dennard Scaling, Domain-Specific Accelerators (DSAs) have come to a Cambrian explosion. Especially when advancing into the intelligent era, more and more DSAs are integrated into System-on-Chips (SoCs) as intellectual property (IP) blocks to provide high performance and efficiency. Currently, IPs usually expose IP-dependent hardware interfaces, requiring SoCs to manage them as isolated devices with software running on the host CPU. However, such software-managed heterogeneity in CPU-centric SoCs leads to low IP utilization. This inefficiency arises from the dependence on software optimization, coupled with the control and data exchange overheads. To improve IP utilization of heterogeneous SoCs, in this article, we rearchitect the SoC as a processor (i.e., SaaP) to orchestrate hardware heterogeneity. SaaP features an orchestration pipeline where DSAs are integrated as execution units and managed directly by the hardware pipeline to conceal the hardware heterogeneity from software. Moreover, SaaP redesigns the register file and data paths to implement an IP-level data-forwarding mechanism, avoiding the costly control and data exchange in the CPU-centric execution model. Block data dependence among different DSAs is carefully resolved to exploit mixed-level parallelism and inter-IP data exchange. SaaP abstracts tasks as mixed-scale instructions, where each instruction can be mapped to different IPs. Experimental results show that compared against Xavier on six fully software-optimized benchmarks from different domains, SaaP-rearchitected Xavier achieves a $2.08{\times }$ speedup, with an 8.21% area reduction and only 2.98% increase in power consumption.
- 20 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
- Vol. 44 (10) , 3962-3975
SaaP: Rearchitect SoC-as-a-Processor to Orchestrate Hardware Heterogeneity
Due to the end of Moore’s Law and Dennard Scaling, Domain-Specific Accelerators (DSAs) have come to a Cambrian explosion. Especially when advancing into the intelligent era, more and more DSAs are integrated into System-on-Chips (SoCs) as intellectual property (IP) blocks to provide high performance and efficiency. Currently, IPs usually expose IP-dependent hardware interfaces, requiring SoCs to manage them as isolated devices with software running on the host CPU. However, such software-managed heterogeneity in CPU-centric SoCs leads to low IP utilization. This inefficiency arises from the dependence on software optimization, coupled with the control and data exchange overheads. To improve IP utilization of heterogeneous SoCs, in this article, we rearchitect the SoC as a processor (i.e., SaaP) to orchestrate hardware heterogeneity. SaaP features an orchestration pipeline where DSAs are integrated as execution units and managed directly by the hardware pipeline to conceal the hardware heterogeneity from software. Moreover, SaaP redesigns the register file and data paths to implement an IP-level data-forwarding mechanism, avoiding the costly control and data exchange in the CPU-centric execution model. Block data dependence among different DSAs is carefully resolved to exploit mixed-level parallelism and inter-IP data exchange. SaaP abstracts tasks as mixed-scale instructions, where each instruction can be mapped to different IPs. Experimental results show that compared against Xavier on six fully software-optimized benchmarks from different domains, SaaP-rearchitected Xavier achieves a $2.08{\times }$ speedup, with an 8.21% area reduction and only 2.98% increase in power consumption.
- Open Access
- 27 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-9
Explainable and Layout-Aware Timing Prediction
- 27 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-9
Open AccessExplainable and Layout-Aware Timing Prediction
- 8 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 74 (1) , 155-169
Efficient and Fast High-Performance Library Generation for Deep Learning Accelerators
The widespread adoption of deep learning accelerators (DLAs) underscores their pivotal role in improving the performance and energy efficiency of neural networks. To fully leverage the capabilities of these accelerators, exploration-based library generation approaches have been widely used to substantially reduce software development overhead. However, these approaches have been challenged by issues related to sub-optimal optimization results and excessive optimization overheads. In this paper, we propose Heron to generate high-performance libraries of DLAs in an efficient and fast way. The key is automatically enforcing massive constraints through the entire program generation process and guiding the exploration with an accurate pre-trained cost model. Heron represents the search space as a constrained satisfaction problem (CSP) and explores the space via evolving the CSPs. Thus, the sophisticated constraints of the search space are strictly preserved during the entire exploration process. The exploration algorithm has the flexibility to engage in space exploration using either online-trained models or pre-trained models. Experimental results demonstrate that Heron averagely achieves 2.71$\times$ speedup over three state-of-the-art automatic generation approaches. Also, compared to vendor-provided hand-tuned libraries, Heron achieves a 2.00$\times$ speedup on average. When employing a pre-trained model, Heron achieves 11.6$\times$ compilation time speedup, incurring a minor impact on execution time.- 8 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computers
- Vol. 74 (1) , 155-169
Efficient and Fast High-Performance Library Generation for Deep Learning Accelerators
The widespread adoption of deep learning accelerators (DLAs) underscores their pivotal role in improving the performance and energy efficiency of neural networks. To fully leverage the capabilities of these accelerators, exploration-based library generation approaches have been widely used to substantially reduce software development overhead. However, these approaches have been challenged by issues related to sub-optimal optimization results and excessive optimization overheads. In this paper, we propose Heron to generate high-performance libraries of DLAs in an efficient and fast way. The key is automatically enforcing massive constraints through the entire program generation process and guiding the exploration with an accurate pre-trained cost model. Heron represents the search space as a constrained satisfaction problem (CSP) and explores the space via evolving the CSPs. Thus, the sophisticated constraints of the search space are strictly preserved during the entire exploration process. The exploration algorithm has the flexibility to engage in space exploration using either online-trained models or pre-trained models. Experimental results demonstrate that Heron averagely achieves 2.71$\times$ speedup over three state-of-the-art automatic generation approaches. Also, compared to vendor-provided hand-tuned libraries, Heron achieves a 2.00$\times$ speedup on average. When employing a pre-trained model, Heron achieves 11.6$\times$ compilation time speedup, incurring a minor impact on execution time.- Open Access
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Revisiting Automatic Pipelining: Gate-level Forwarding and Speculation
Pipelining is a widely applied micro-architectural performance optimization and requires non-trivial designs for better execution throughput. The key to pipeline throughput optimization is to resolve data hazards caused by read-after-write (RAW) dependencies, which are traditionally tackled by forwarding and speculation to avoid pipeline stalls. However, existing approaches are conducted based on high-level dataflow analysis, with potential loss of optimization opportunities for lack of analysis of the netlist structures. We propose an efficient method to resolve RAW dependencies with low-level netlist analysis by gate-level forwarding and speculation. With a greedy search method to detect and resolve short-delay gate-level signal paths for forwarding and an approximate circuit synthesis method with formal verification for gate-level speculation, the method efficiently utilizes the gate-level information to further improve pipeline throughput. Even if the gate-level design space is vast, by assembling the logic gates in the short-delay signal paths in fewer adjacent pipeline stages, the forwarding paths and the speculator bypass invalid pipeline stages and thus conceal the latency of these paths. We conduct experiments on the widely-used ISCAS/EPFL benchmark circuits and a large-scale RISC-V CPU, and our approach can find better designs than human experts.
- 23 June 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-6
Open AccessRevisiting Automatic Pipelining: Gate-level Forwarding and Speculation
Pipelining is a widely applied micro-architectural performance optimization and requires non-trivial designs for better execution throughput. The key to pipeline throughput optimization is to resolve data hazards caused by read-after-write (RAW) dependencies, which are traditionally tackled by forwarding and speculation to avoid pipeline stalls. However, existing approaches are conducted based on high-level dataflow analysis, with potential loss of optimization opportunities for lack of analysis of the netlist structures. We propose an efficient method to resolve RAW dependencies with low-level netlist analysis by gate-level forwarding and speculation. With a greedy search method to detect and resolve short-delay gate-level signal paths for forwarding and an approximate circuit synthesis method with formal verification for gate-level speculation, the method efficiently utilizes the gate-level information to further improve pipeline throughput. Even if the gate-level design space is vast, by assembling the logic gates in the short-delay signal paths in fewer adjacent pipeline stages, the forwarding paths and the speculator bypass invalid pipeline stages and thus conceal the latency of these paths. We conduct experiments on the widely-used ISCAS/EPFL benchmark circuits and a large-scale RISC-V CPU, and our approach can find better designs than human experts.