High-Level Synthesis (HLS) tools are widely adopted in FPGA-based domain-specific accelerator design. However, existing tools rely on fixed optimization strategies inherited from software compilations, limiting their effectiveness. Tailoring optimization strategies to specific designs requires deep semantic understanding, accurate hardware metric estimation, and advanced search algorithms - capabilities that current approaches lack.We propose DAPO, a design structure-aware pass ordering framework that extracts program semantics from control and data flow graphs, employs contrastive learning to generate rich embeddings, and leverages an analytical model for accurate hardware metric estimation. These components jointly guide a reinforcement learning agent to discover design-specific optimization strategies. Evaluations on standard HLS benchmarks demonstrate that our end-to-end flow delivers 1.67× speedup on pragma-free designs and a 2.36× speedup on designs with pragmas over Vitis HLS with comparable resource usage.

Coarse-Grained Reconfigurable Architectures (CGRAs) are a promising and versatile accelerator platform, offering a balance between the performance and efficiency of specialized accelerators and software programmability. However, their full potential is severely hindered by control flow in accelerated kernels, as control flow (e.g., loops, branches) is fundamentally incompatible with the parallel, data-driven CGRA fabric. Prior strategies to resolve this mismatch in CGRA kernel acceleration are either inefficient, sacrificing performance for generality, or lack generality due to the difficulty of adapting them across different execution models. Thus, a general and unified solution for efficient CGRA kernel acceleration remains elusive. This paper introduces NEURA, a unified and retargetable compilation framework that systematically resolves the control-dataflow mismatch in CGRAs. NEURA's core innovation is a novel, pure dataflow intermediate representation (IR) built on a predicated type system. In this IR, control contexts are embedded as a predicate within each data, making control an intrinsic property of data. This mechanism enables NEURA to systematically flatten complex control flow into a single unified dataflow graph. This unified representation decouples kernel representation from hardware, empowering NEURA to retarget diverse CGRAs with different execution models and microarchitectural features. When targeted to a high-performance spatio-temporal CGRA, NEURA delivers a 2.20x speedup on kernel benchmarks and up to 2.71x geometric mean speedup on real-world applications over state-of-the-art (SOTA) high-performance baselines. It also provides a competitive solution against the SOTA low-power CGRA when retargeted to a spatial-only CGRA. NEURA is open-source and available at https://github.com/coredac/neura.

Implementing HLS accelerators on large-scale multi-die FPGAs presents significant challenges. To address this, researchers have proposed High-Level Physical Synthesis (HLPS), which co-optimizes high-level synthesis and physical design to improve achievable frequency. However, existing HLPS techniques suffer from unstable and inconsistent quality of results (QoRs), largely due to the vast number of parameters that need to be selected by the user in an ad-hoc way. As a result, achieving satisfactory solutions still requires substantial manual effort and expertise in low-level circuit design.We propose a robust and practical design space exploration (DSE) framework that enhances the reliability and QoRs of HLPS by automating the iterative parameter tuning process. Informed by metrics extracted from physical implementation outcomes, the framework applies tailored heuristics to refine HLPS parameters, enabling consistent and automated timing closure. In evaluations with large-scale, real-world designs implemented on representative multi-die devices, our framework achieves an average frequency of 311.06 MHz, reaching 2.42× the frequency of the AMD Vitis/Vivado toolchain (128.48 MHz) and 1.67× that of the leading academic solutions (186.21 MHz).

Low cloud microphysical properties, including liquid water content (LWC), cloud effective radius (CER), and cloud optical thickness (COT), are fundamental to climate modeling and weather prediction. However, existing remote sensing techniques often rely on priori data and passive instrument observations, which limit the accuracy and consistency of retrievals, particularly in complex cloud regimes. In this study, we present an innovative radar-based algorithm that retrieve LWC, CER, and COT directly from CloudSat millimeterwavelength cloud profiling radar (CPR) observations. Unlike conventional approaches, our method derives physically consistent cloud properties with minimal reliance on ancillary data, thereby overcoming the inherent limitations of current algorithms. We demonstrate the efficacy of this method through detailed comparisons with established cloud products from CloudSat, MODIS, and AMSR2 sensors. The results show improved accuracy in LWC retrievals and a more consistent vertical distribution of CER and COT, particularly in challenging low marine boundary layer clouds. This novel algorithm offers a significant advancement in cloud remote sensing, facilitating more reliable cloud property retrievals for enhanced climate model simulations. The method is fully applicable for CloudSat’s 18-year record and the latest space radar mission of EarthCare, offering an essential tool to better capture the vertical complexity of clouds and to advance global cloud data quality for climate research.

The cloud droplets effective radius is a key feature that plays a critical role in influencing cloud microphysical processes and radiative effects. Accurate quantification of cloud effective radius (CER) is essential for advancing our understanding of cloud microphysics, refining cloud parameterization, and improving future climate prediction. Nonetheless, the accuracy of current CER retrieval algorithms, particularly relying on millimeter-wavelength cloud radar, is often largely affected by assumptions about the cloud droplet number concentration, inappropriate empirical coefficients, attenuated radar reflectivity, and limitations of other auxiliary instruments. In this study, we developed a novel CER retrieval algorithm for single-wavelength radar by leveraging the interconnections between CER, liquid water content (LWC), and cloud radar reflectivity. Unlike the previous studies, we first derive the LWC from a self-consistent method based on cloud liquid water mass absorption instead of empirical relationships. Subsequently, we correct the radar measured reflectivity attenuated by cloud water and perform sensitivity analysis to select an optimal parameter that minimizes the uncertainty associated with the given cloud droplet size distribution (DSD) assumption. Then, the CER is calculated from the retrieved LWC, corrected reflectivity, and the optimal parameter. We compared the frequency distribution, vertical structure, and error fraction of the retrieved CER with aircraft in situ measurements. Our results demonstrate higher consistency with in situ data compared to traditional empirical algorithms. Furthermore, the cloud optical thickness (COT) derived from the CER shows a much better agreement with Moderate Resolution Imaging Spectroradiometer (MODIS) products, which provides additional validation for the efficacy of our method in investigating cloud microphysical properties.

Multi-die FPGAs are widely adopted for large-scale accelerators, but optimizing high-level synthesis designs on these FPGAs faces two challenges. First, the delay caused by die-crossing nets creates an NP-hard floorplanning problem. Second, traditional directive optimization cannot consider resource constraints on each die or the timing issue incurred by the die-crossings. Furthermore, the high algorithmic complexity and the large scale lead to extended runtime for legalizing the floorplan of HLS designs under different directive configurations. To co-optimize the directives and floorplan of HLS designs on multi-die FPGAs, we formulate the co-search based on bin-packing variants and present two iterative optimization flows. The first (FADO 1.0) relies on a pre-built QoR library. It involves a greedy, latency-bottleneck-guided directive search, and an incremental floorplan legalization. Compared with a global floorplanning solution, it takes 693X~4925X shorter search time and achieves 1.16X~8.78X better design performance, measured in workload execution time. To remove the time-consuming QoR library generation, the second flow (FADO 2.0) integrates an analytical QoR model and redesigns the directive search to accelerate convergence. Through experiments on mixed dataflow and non-dataflow designs, compared with 1.0, FADO 2.0 further yields a 1.40X better design performance on average after implementation on the Alveo U250 FPGA.

Low-level clouds (LLC), mainly composed of liquid water droplets, cool the climate system by strongly reflecting solar radiation back to space, and thus play an important role in the Earth energy budget. However, the LLC properties and their radiative effects are poorly represented in climate models, leading to the largest source of uncertainty in the climate prediction. Liquid water content (LWC) is a key property of LLC determining cloud extinction characteristics and is a fundamental parameter in the radiative transfer model. To improve the understanding of LWC properties, algorithms have been proposed to retrieve LWC based on millimeter-wavelength radar. However, the traditional retrieval relies on pre-constructed empirical relationship between reflectivity and LWC and have noticeable limitations. Particularly, the retrieval uncertainty is strongly depended on the assumed particle size distribution, the existence of drizzle particle; and on the accuracy of reflectivity measurement. In this study, we develop a new self-consistent algorithm to retrieve LWC by constraining radar reflectivity factor and attenuation in the whole liquid cloud layer. A relationship between the radar measured reflectivity, LWC, and the intrinsic reflectivity is first constructed based on the radiative transfer theory under Rayleigh scattering regime. A nonlinear least-square regression technique is then applied to derive the optimal parameters in the retrieval equations to obtain the LWC. Comparison with the microwave radiometer (MWR) derived liquid water path (LWP) indicates that our proposed method retrieves more accurate LWC products than that from the traditional empirical algorithms.