1 / 41

Veljko Milutinovi ć

Veljko Milutinovi ć. VLSI for SuperComputing: From Applications and Algorithms to Masks and Chips. 1 /40. World Top Foundries in VLSI for SuperComputing. #1?. 2 /40. World Top Foundries in VLSI for SuperComputing. #1? Qualcomm. 3 /40. World Top Foundries in VLSI for SuperComputing.

york
Download Presentation

Veljko Milutinovi ć

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Veljko Milutinović VLSI for SuperComputing: From Applications and Algorithms to Masks and Chips 1/40

  2. World Top Foundries in VLSI for SuperComputing • #1? 2/40

  3. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm 3/40

  4. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? 4/40

  5. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel 5/40

  6. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? 6/40

  7. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? Mubadala (AMD, ...) 7/40

  8. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? Mubadala (AMD, ...) • The Qualcomm VP/TD started from this course(first as my PhD student and later as the course TA) 8/40

  9. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? Mubadala (AMD, ...) • The Qualcomm VP/TD started from this course(first as my PhD student and later as the course TA) • An Intel VP/TD took the IR4RVL in-a-nut-shell(at my past IEEE/ACM-HICSS conference tutorial) 9/40

  10. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? Mubadala (AMD, ...) • The Qualcomm VP/TD started from this course(first as my PhD student and later as the course TA) • An Intel VP/TD took the IR4RVL in-a-nut-shell(at my past IEEE/ACM-HICSS conference tutorial) • Maybe, a next MubadalaSRB VP/TD comes from ETF  10/40

  11. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? Mubadala (AMD, ...) • The Qualcomm VP/TD started from this course(first as my PhD student and later as the course TA) • An Intel VP/TD took the IR4RVL in-a-nut-shell(at my past IEEE/ACM-HICSS conference tutorial) • Maybe, a next MubadalaSRB VP/TD comes from ETF or MF  11/40

  12. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? Mubadala (AMD, ...) • The Qualcomm VP/TD started from this course(first as my PhD student and later as the course TA) • An Intel VP/TD took the IR4RVL in-a-nut-shell(at my past IEEE/ACM-HICSS conference tutorial) • Maybe, a next MubadalaSRB VP/TD comes from ETF or MF  or FFH  12/40

  13. World Top Foundries in VLSI for SuperComputing • #1? Qualcomm • #2? Intel • #3? Mubadala (AMD, ...) • The Qualcomm VP/TD started from this course(first as my PhD student and later as the course TA) • An Intel VP/TD took the IR4RVL in-a-nut-shell(at my past IEEE/ACM-HICSS conference tutorial) • Maybe, a next MubadalaSRB VP/TD comes from ETF or MF  or FFH  or FON  13/40

  14. Who works here? 14/40

  15. The Holistic Foundry (R&DFab)in VLSI for SuperComputing • Phase#1: From Applications to AlgorithmsPhase#2: From Algorithms to MasksPhase#3: From Masks to Chips 15/40

  16. The Holistic Foundry (R&DFab)in VLSI for SuperComputing • Phase#1: From Applications to AlgorithmsPhase#2: From Algorithms to MasksPhase#3: From Masks to Chips • Verification is crucial in each one of these phases,and related teaching is done in coop with ELSYS! 16/40

  17. The Holistic Foundry (R&DFab)in VLSI for SuperComputing • Phase#1: From Applications to AlgorithmsPhase#2: From Algorithms to MasksPhase#3: From Masks to Chips • Verification is crucial in each one of these phases,and related teaching is done in coop with ELSYS! • Management issues of importance for an R&DFab are covered in the accompanying course: IR4USP (including 12 related homework assignments)! 17/40

  18. Contents: From Algorithms to Masks • Part#1: VLSI for ControlFlow SuperComputing • Part#2: VLSI for DataFlow SuperComputing • Part#3: VLSI for WirelessFlow SuperComputing 18/40

  19. VLSI for ControlFlow SuperComputing • ManyCore Systems: • VHDL vs Verilog (0.5 weeks) • Design and Programming of a 200MHz RISC Microprocessor(2.5 weeks) + HW#1 • MultiCore Systems: • MicroProcessor and MultiMicroProcessor Systems by Wiley (1 week) • Verification by Elsys (2 weeks) + Lab#1 19/40

  20. 20/40

  21. 21/40

  22. VLSI for DataFlow SuperComputing • FineGrain DataFlow: • Altera vs Xilinx (0.5 weeks) • Design and Programming of the 200MHz Maxeler Machine (3.5 weeks) + HW#2 • SystolicArray DataFlow: • Systolic Array Architectures (0.5 weeks) • DARPA Systolic Architectures (0.5 weeks) + Lab#2 22/40

  23. Advances in Computer Architecture (North Holland)by Veljko M. Milutinovic with a contribution from John Hennessy 23/40

  24. High-Level Language Computer Architecture (Elsevier Computer Science Press)by Veljko M. Milutinovic with a contribution from Michael Flynn 24/40

  25. VLSI for WirelessFlow SuperComputing • WSNs: Part#1 • Hardware (0.25 weeks) • Software (0.25 weeks) • WSNs: Part#2 • Systems (SUN+Slimmer) (0.25 weeks) • Applications (UbiComputing@WSN+DataMining@WSN) (0.25 weeks) 25/40

  26. 26/40

  27. opt.html 27/40

  28. VLSI for QuantumMechanical SuperComputing • Basics: • Hardware (0 weeks) • Software (0 weeks) • Advances: • Systems (0 weeks) • Applications (0 weeks) Optional class projects for extra points! 28/40

  29. European Projects • ESF: • RoMoL: Riding on Moore's Law (0 weeks) • HiPeac: Parallel Programming Models (0 weeks) • FP7/H20: • FP7: ProSense (0 weeks) • FP7: BalCon (0 weeks) 29/40

  30. Example Algorithms for Practical Implementations • Engineering: • Computer Engineering (0 weeks) • Financial Engineering (0 weeks) • Science: • Physical Chemistry (0 weeks) • Computer Science (0 weeks) 30/40

  31. SOME PREVIOUS OFFERINGS OF THIS TECH COURSE Purdue Stanford MIT Barcelona Siena Ljubljana Previous offerings of the related MGMT course: Purdue, Dartmouth, HarvardCNY, Barcelona, Pisa, Ljubljana 31/40

  32. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM 32/40

  33. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM 33/40

  34. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM 34/40

  35. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM (2)

  36. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM (2)

  37. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM (2)

  38. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM (3) 38/40

  39. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM (3) 39/40

  40. BOTTOM LINE: BRINGING ADVANCED INDUSTRIAL EXPERIENCE INTO THE CLASSROOM (3) Wireless Sensor Networks: ApplicationDesign and DataMining 40/40

More Related