Dataset
Data Sheet 1_Architecture and engineering trade-offs of a heterogeneous spaceborne supercomputing node for future onboard intelligent computing.docx
| Τίτλος: | Data Sheet 1_Architecture and engineering trade-offs of a heterogeneous spaceborne supercomputing node for future onboard intelligent computing.docx |
|---|---|
| Συγγραφείς: | Jianing Rao, Wenjie Zhao, Shuangxue Han, Wanying Zhang, Miao Ma, Wenyuan Jia, Tianzi Wang, Guodong Yin, Yan Zhu, Junshe An, Lianguo Wang |
| Έτος έκδοσης: | 2026 |
| Συλλογή: | Frontiers: Figshare |
| Θεματικοί όροι: | Cosmology, aerospace realtime intelligent data processing, earth observation, high performance computer (HPC), modular design, space based supercomputing node, synthetic aperture radar (SAR) |
| Περιγραφή: | The rapid growth of onboard data generation in Earth observation and deep-space missions is driving increasing demand for high-throughput and intelligent onboard processing. Conventional onboard computing platforms based on CPU or CPU–FPGA architectures often provide insufficient computational flexibility for emerging workloads such as real-time Synthetic Aperture Radar (SAR) imaging and AI-assisted telemetry analytics. This work presents a heterogeneous and fault-tolerant architectural framework for future spaceborne intelligent computing systems. Onboard workloads including SAR processing and intelligent telemetry analysis are analyzed to derive preliminary system-level requirements. Based on these requirements, a modular computing node integrating CPU, GPU, NPU, and FPGA resources through PCIe-based interconnection is proposed. An asymmetric redundancy strategy combining high-performance heterogeneous computing units and conservative space-grade backup units is introduced to balance computational capability and mission survivability. Processor benchmark analysis, workload mapping strategies, radiation-aware architectural considerations, and thermal feasibility analysis are discussed. Rather than claiming a fully validated flight-qualified implementation, this work focuses on engineering trade-offs and architectural feasibility for future onboard heterogeneous supercomputing systems supporting high-autonomy space missions. |
| Τύπος εγγράφου: | dataset |
| Γλώσσα: | unknown |
| DOI: | 10.3389/fspas.2026.1832412.s001 |
| Διαθεσιμότητα: | https://doi.org/10.3389/fspas.2026.1832412.s001 https://figshare.com/articles/dataset/Data_Sheet_1_Architecture_and_engineering_trade-offs_of_a_heterogeneous_spaceborne_supercomputing_node_for_future_onboard_intelligent_computing_docx/33001607 |
| Rights: | CC BY 4.0 |
| Αριθμός Καταχώρησης: | edsbas.A7228E00 |
| Βάση Δεδομένων: | BASE |
| DOI: | 10.3389/fspas.2026.1832412.s001 |
|---|