| Both sides previous revisionPrevious revisionNext revision | Previous revision |
| start [2023/05/10 09:33] – [Learning Materials] juang | start [2023/08/16 14:15] (current) – [Learning Materials] ewent |
|---|
| ==== Livestream ==== | ==== Livestream ==== |
| [[https://www.youtube.com/live/GIb5EgSrWk0?feature=share|YouTube livestream]] | [[https://www.youtube.com/live/GIb5EgSrWk0?feature=share|YouTube livestream]] |
| | {{youtube>GIb5EgSrWk0?large}} |
| |
| ==== Organizers ==== | ==== Organizers ==== |
| ===== Agenda (June 18, 2023) ===== | ===== Agenda (June 18, 2023) ===== |
| |
| ==== Lectures (tentative) ==== | ==== Lectures ==== |
| * 1. Introduction: PIM as a paradigm to overcome the data movement bottleneck. | * 8:55am-9:00am, Dr Juan Gómez Luna, “Welcome & Agenda”. |
| * 2. PIM taxonomy: PNM (processing near memory) and PUM (processing using memory). | * 9:00am-10:20am, Prof. Onur Mutlu, “Memory-centric Computing: Introduction to PIM as a Paradigm to Overcome the Data Movement Bottleneck”. |
| * 3. General-purpose PNM: UPMEM PIM. | * PIM taxonomy: PNM (processing near memory) and PUM (processing using memory). |
| * 4. PNM for neural networks: Samsung HBM-PIM, SK Hynix AiM. | * 10:20am-11:00pm, Dr. Juan Gómez Luna, “Processing-Near-Memory: Real PNM”. |
| * 5. PNM for recommender systems: Samsung AxDIMM, Alibaba PNM. | * PNM prototypes: Samsung HBM-PIM, SK Hynix AiM, Samsung AxDIMM, Alibaba HB-PNM. |
| * 6. Programming recommendations for general-purpose PNM. | * UPMEM PIM: Architecture and Programming. |
| * 7. General-purpose PNM for ML workloads, sparse linear algebra, bioinformatics… | |
| * 8. Adoption issues: How to enable PIM? Need for high-level PIM programming. | |
| * 9. What’s next? PUM prototypes, frameworks and programming for modern workloads. | |
| |
| ==== Hands-on Labs (tentative) ==== | * Coffee break (11:00am-11:20am) |
| * 1. Benchmarking of the UPMEM PIM system. | |
| * 2. Accelerating real-world workloads with the UPMEM PIM system: Common parallel patterns (reduction, prefix sum, histogramming, sorting, etc.), sparse matrix computation (e.g., SpMV), dynamic programming for sequence alignment, training of machine learning workloads, etc. | * 11:20am-11:50am, Prof. Izzat El Hajj (AUB), “High-throughput Sequence Alignment using Real Processing-in-Memory Systems”. |
| * 3. If time permits: In-memory copy, random number generation and bitwise operations with PiDRAM. | * 11:50am-12:30pm, Dr. Christina Giannoula (UofT), “SparseP: Towards Efficient Sparse Matrix Vector Multiplication on Real Processing-In-Memory Systems”. |
| | |
| | * Lunch break (12:30pm-2:00pm) |
| | |
| | * 2:00pm-2:45pm, Dr. Sukhan Lee (Samsung), “Introducing Real-world HBM-PIM Powered System for Memory-bound Applications”. |
| | * 2:45pm-3:30pm, Dr. Juan Gómez Luna/Ataberk Olgun, “Processing-Using-Memory and PUM Prototypes: Ambit/SIMDRAM, PiDRAM”. |
| | |
| | * Coffee break (3:30pm-4pm) |
| | |
| | * 4:00pm-4:40pm, Dr. Juan Gómez Luna, “Accelerating Modern Workloads on a General-purpose PIM System”. |
| | * 4:40pm-5:20pm, Dr. Juan Gómez Luna, “Adoption Issues: How to Enable PIM?” |
| | * 5:20pm-5:30pm, Dr. Juan Gómez Luna, “Introduction/Preparation for Hands-on labs”. |
| | * Optional - Hands-on Lab: Programming and Understanding a Real PIM Architecture. |
| | |
| | ==== Tutorial Materials ==== |
| | |
| | ^ Time ^ Speaker ^ Title ^ Materials ^ |
| | | 8:55am-9:00am | Dr. Juan Gómez Luna | Welcome & Agenda |{{realpimtutorial-isca23-agenda-juan-slides.pdf|(PDF)}} {{realpimtutorial-isca23-agenda-juan-slides.pptx|(PPT)}}| |
| | | 9:00am-10:20am | Prof. Onur Mutlu | Memory-Centric Computing |{{RealPIMTutorial-ISCA-Intro-onur-slides-18-June-2023.pdf|(PDF)}} {{RealPIMTutorial-ISCA-Intro-onur-slides-18-June-2023.pptx|(PPT)}}| |
| | | 10:20am-11:00am | Dr. Juan Gómez Luna | Processing-Near-Memory: Real PNM Architectures / Programming General-purpose PIM |{{realpimtutorial-isca23-processingnearmemory-juan-slides.pdf|(PDF)}} {{realpimtutorial-isca23-processingnearmemory-juan-slides.pptx|(PPT)}}| |
| | | 11:20am-11:50am | Prof. Izzat El Hajj | High-throughput Sequence Alignment using Real Processing-in-Memory Systems |{{AIM_ISCA2023_Real_PIM_Tutorial.pdf|(PDF)}} {{AIM_ISCA2023_Real_PIM_Tutorial.pptx|(PPT)}}| |
| | | 11:50am-12:30pm | Dr. Christina Giannoula | SparseP: Towards Efficient Sparse Matrix Vector Multiplication for Real Processing-In-Memory Systems |{{SparseP_sigmetrics22_full_clean.pdf|(PDF)}} {{SparseP_sigmetrics22_full_clean.pptx|(PPT)}}| |
| | | 2:00pm-2:45pm | Dr. Sukhan Lee | Introducing Real-world HBM-PIM Powered System for Memory-bound Applications |{{Introducing+Real-world+HBM-PIM+Powered+System+for+Memory-bound+Applications.pdf|(PDF)}}| |
| | | 2:45pm-3:30pm | Dr. Juan Gómez Luna / Ataberk Olgun | Processing-Using-Memory: Exploiting the Analog Operational Properties of Memory Components / PUM Prototypes: PiDRAM |{{realpimtutorial-isca23-processingusingmemory-juan-slides.pdf|(PDF)}} {{realpimtutorial-isca23-processingusingmemory-juan-slides.pptx|(PPT)}} {{realpimtutorial-isca23-pidram-ataberk-slides.pdf|(PDF)}} {{realpimtutorial-isca23-pidram-ataberk-slides.pptx|(PPT)}}| |
| | | 4:00pm-4:40pm | Dr. Juan Gómez Luna | Accelerating Modern Workloads on a General-purpose PIM System |{{realpimtutorial-isca23-modernworkloads-juan-slides.pdf|(PDF)}} {{realpimtutorial-isca23-modernworkloads-juan-slides.pptx|(PPT)}}| |
| | | 4:40pm-5:20pm | Dr. Juan Gómez Luna | Adoption Issues: How to Enable PIM? |{{realpimtutorial-isca23-enablingpim-juan-slides.pdf|(PDF)}} {{realpimtutorial-isca23-enablingpim-juan-slides.pptx|(PPT)}}| |
| | | 5:20pm-5:30pm | Dr. Juan Gómez Luna | Hands-on Lab: Programming and Understanding a Real Processing-in-Memory Architecture |{{lab_isca2023.pdf|(Handout)}} \\ {{realpimtutorial-isca23-lab-juan-slides.pdf|(PDF)}} {{realpimtutorial-isca23-lab-juan-slides.pptx|(PPT)}}| |
| |
| ==== Learning Materials ==== | ==== Learning Materials ==== |
| |
| === More Learning Materials === | === More Learning Materials === |
| * Mutlu O., Memory-Centric Computing (Keynote Talk at the Thoughtworks Engineering for Research Symposium (E4R), February 2022): | * Mutlu O., Memory-Centric Computing (IMACAW Keynote Talk at DAC 2023), July 2023: |
| * [[https://people.inf.ethz.ch/omutlu/pub/onur-ThoughtWorks-e4r-Keynote-MemoryCentricComputing-February-19-2022.pdf | PDF]] [[https://people.inf.ethz.ch/omutlu/pub/onur-ThoughtWorks-e4r-Keynote-MemoryCentricComputing-February-19-2022.pptx | PPT]] [[https://www.youtube.com/watch?v=r605d_C4WiY | Video]] | * [[https://people.inf.ethz.ch/omutlu/pub/onur-IMACAW-Keynote-MemoryCentricComputing-9-July-2023.pdf | PDF]] [[https://people.inf.ethz.ch/omutlu/pub/onur-IMACAW-Keynote-MemoryCentricComputing-9-July-2023.pdf | PPT]] [[https://www.youtube.com/watch?v=npr5zMEkANQ | Video]] |
| * Processing-in-memory: A workload-driven perspective (summary paper about recent research in PIM): | * Processing-in-memory: A workload-driven perspective (summary paper about recent research in PIM): |
| * [[https://people.inf.ethz.ch/omutlu/pub/processing-in-memory_workload-driven-perspective_IBMjrd19.pdf | PDF]] | * [[https://people.inf.ethz.ch/omutlu/pub/processing-in-memory_workload-driven-perspective_IBMjrd19.pdf | PDF]] |
