RAD – 出版物
AMD 研究與開發部門 (RAD) 致力於關鍵科研成果之發表，尤為重視有同儕審查機制的會議及期刊。
本頁提供了 RAD 近年來多項出版物的連結。
2024
- AI-Based Approaches in Network Security（以 AI 為基礎的網路安全性方法）– AI4Good 2024
- T3：Transparent Tracking & Triggering for Fine-grained Overlap of Compute and Collectives（實現運算與集體通訊細粒疊合的透明追蹤與觸發機制）– ASPLOS 2024
- Integrating FPGA and GPU Acceleration to OpenMP Distributed Computing（將 FPGA 與顯示卡加速整合至 OpenMP 分散式運算）– FPL 2024
- Turn-based Spatiotemporal Coherence for GPUs（顯示卡輪流式時空一致性）– HiPEAC 2024
- Networking Technologies for Handling AI Workloads（處理 AI 工作負載的網路技術）– ISC 2024
- Sustainable Computing at Scale（大規模永續運算）– MODSIM 2024
2023
- Spectrum Usage and Occupancy Monitoring: Challenges and Software-Defined Radio Solutions（頻譜使用與佔用率監測：挑戰與軟體定義無線電解決方案）– IIIE WCNC 2023
- Improving DNN Throughput Via Intelligent Concurrent GEMM Executions（透過智慧化並行 GEMM 執行提升 DNN 輸送量）– arXiv 2023
- The Next Era for Chiplet Innovation（小晶片創新的新時代）– DATE 2023
- Leveraging MLIR to Design for AI Engines（運用 MLIR 進行 AI 引擎設計）– FCCM 2023
- Reducing Internode Communication Using FPGA-Accelerated Neural Network Surrogate Models（使用以 FPGA 加速的神經網路代理模型降低節點間通訊負擔）– FIRE 2023
- Navigating the Future Landscape of System-On-Chip Technology（掌握系統單晶片技術的未來發展）– IEEE SOCC 2023
- Tale of Two Cs: Computation vs. Communication Scaling for Future Transformers on Future Hardware（兩個 C 的故事：運算擴充與通訊擴充之比較，以未來硬體上的未來 Transformer 模型為核心）– IISWC 2023
- SPARTA: Spatial Acceleration for Efficient and Scalable Horizontal Diffusion Weather Stencil Computation（SPARTA：用於高效率且可擴充水平擴散天氣模板運算的空間加速）– ICS 2023
- Introduction to the AMD Versal ACAP Adaptable Intelligent Engine and to its Programming Model（AMD Versal ACAP 自行調適智慧引擎及其程式設計模型介紹）– SC 2023
- Innovative Approaches to AI with Adaptive Computing（以自適應運算推動 AI 的創新方法）– SPL 2023
2022
- Demystifying BERT: System Design Implications（解構 BERT：對系統設計的影響）– IISWC 2022
- A Case for Fine-grain Coherence Specialization in Heterogeneous Systems（異質系統細粒一致性專用化案例研究） - TACO
- Virtual Coset Coding for Encrypted Non-Volatile Memories with Multi-Level Cells（適用於多層單元加密非揮發性記憶體的虛擬陪集編碼技術）– HPCA 2022
- Data Convection: A GPU-Driven Case Study for Thermal-Aware Data Placement in 3D DRAMs（資料對流：3D DRAM 熱感知資料配置案例研究，以顯示卡工作負載為例）– SIGMETRICS 2022
- Cloak: Tolerating Non-Volatile Cache Read Latency（Cloak：容忍非揮發性快取讀取延遲）– ICS 2022
- Uncertainty Quantification Methods for ML-based Surrogate Models of Scientific Applications（用於科學應用之機器學習代理模型的不確定性量化方法）– NeurIPS 2022
- Eager Memory Cryptography in Caches（快取中的預先記憶體加密）– MICRO 2022
- Athena: An Early-Fetch Architecture To Reduce On-Chip Page Walk Latencies（Athena：用於降低晶片上分頁查詢延遲的提前擷取架構）– PACT 2022
- Improving Energy Efficiency of Permissioned Blockchains Using FPGAs（使用 FPGA 提升許可制區塊鏈能效）– ICPADS 2022
2021
- Analyzing and Leveraging Decoupled L1 Caches in GPUs（分析並活用顯示卡中解耦的 L1 快取）– HPCA 2021
- Deadline-Aware Offloading for High-Throughput Accelerators（具截止期限感知能力的高輸送量加速器卸載機制）– HPCA 2021
- Understanding Chiplets Today to Anticipate Future Integration Opportunities and Limits（從目前的小晶片出發，窺探未來的整合機會與限制）– DATE 2021
- Systems-on-Chip with Strong Ordering（論採強順序規則的系統單晶片） - TACO
- Pioneering Chiplet Technology and Design for AMD EPYC™ and Ryzen™ Processor Families（為 AMD EPYC™ 與 Ryzen™ 處理器系列產品開創小晶片技術與設計）– ISCA 2021（產業議程）
- Quantifying Server Memory Frequency Margin and Using it to Improve Performance in HPC Systems（量化伺服器記憶體頻率餘量，並將其用於提升高效能運算系統效能）- ISCA 2021
- Interconnect Modeling for Homogeneous and Heterogeneous Multiprocessors（同質與異質多處理器的互連模擬）– Springer（書籍章節）
- Increasing GPU Translation Reach by Leveraging Under-Utilized On-Chip Resources（利用未充分使用的晶片上資源擴大顯示卡轉譯範圍）- MICRO 2021
- DUB: Dynamic Underclocking and Bypassing in Network-on-Chip for Heterogeneous GPU Workloads（DUB：適用於異質顯示卡工作負載之晶片上網路動態降頻與旁路技術）- NOCS 2021
- A New Era of Tailored Computing（特製化運算的新時代）- VLSI Symposium 2021（短論文）
- Efficient Cache Utilization via Model-aware Data Placement for Recommendation Models（透過模型感知資料配置提升推薦模型的快取使用效率）– MEMSYS 2021
- Virtual Coset Coding for Encrypted Non-Volatile Memories with Multi-Level Cells（適用於多層單元加密非揮發性記憶體的虛擬陪集編碼技術）– HPCA 2022
- Using neural networks to reduce communication in numerical solution of partial differential equations（使用神經網路降低偏微分方程式數值解的通訊需求）– NEURIPS 2021
- Using physics-informed regularization to improve extrapolation capabilities of neural networks（利用物理定律正規化提升神經網路的外推能力）– NEURIPS 2021
2020
- Kite: A Family of Heterogeneous Interposer Topologies Enabled via Accurate Interconnect Modeling（Kite：透過精準互連模擬實現的一系列異質中介層拓撲法）– DAC 2020
- SeqPoint: Identifying Representative Iterations of Sequence-based Neural Networks（SeqPoint：辨識序列式神經網路具代表性的迭代階段）– ISPASS 2020
- Improving the Utilization of Micro-operation Caches in x86 Processors（提升 x86 處理器微運算快取的使用率）– MICRO 2020
- Centaur: A Novel Architecture for Reliable, Low-Wear,High-Density 3D NAND Storage（Centaur：用於高可靠、低抹寫、高密度 3D NAND 儲存裝置的創新架構）– SIGMETRICS 2020
- （分析並運用顯示卡的共用 L1 快取）– PACT 2020
- PreFAM: Understanding the Impact of Prefetching in Fabric-Attached Memory Architectures（PreFAM：理解預先擷取在網狀附接式記憶體架構中的影響）– MEMSYS 2020
- CFDNet: a deep learning-based accelerator for fluid simulations（CFDNet：用於流體模擬採用深度學習的流體模擬加速技術）– ICS 2020
- Optimizing of Intercache Traffic Entanglement in Tagless Caches With Tiling Opportunities（在無標籤快取中利用分塊機會最佳化快取間流量糾纏）– TCAD 2020
- Optimizing of Intercache Traffic Entanglement in Tagless Caches With Tiling Opportunities（在無標籤快取中利用分塊機會最佳化快取間流量糾纏）– CASES 2020
- Independent Forward Progress of Work-groups（工作群組的獨立前向推進）– ISCA 2020
- Experiences with ML-Driven Design: A NoC Case Study（機器學習驅動設計的實務經驗：一份 NoC 案例研究）– HPCA 2020
- GPU Initiated OpenSHMEM: Correct and Efficient Intra-Kernel Networking for dGPUs（由顯示卡啟動的 OpenSHMEM：正確且高效率的獨立顯示卡內核中網路通訊）– PPoPP 2020
- Centaur: A Novel Architecture for Reliable, Low-Wear,High-Density 3D NAND Storage（Centaur：用於高可靠、低抹寫、高密度 3D NAND 儲存裝置的創新架構）– SIGMETRICS 2020
- DSM: A Case for Hardware-Assisted Merging of DRAM Rows with Same Content（DSM：以硬體輔助方式合併具相同內容之 DRAM 列的案例）– SIGMETRICS 2020