1 / 15

CSE 691: Energy-Efficient Computing Lecture 7 SMARTS: custom-made systems

CSE 691: Energy-Efficient Computing Lecture 7 SMARTS: custom-made systems. Anshul Gandhi 1307, CS building anshul@cs.stonybrook.edu. memcache. Benchmark competitions. Benchmark competitions. fawn paper. Data-intensive workloads. Scan-bound large sequential examples? problems?.

shana-berry
Download Presentation

CSE 691: Energy-Efficient Computing Lecture 7 SMARTS: custom-made systems

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. CSE 691: Energy-Efficient ComputingLecture 7SMARTS: custom-made systems Anshul Gandhi 1307, CS building anshul@cs.stonybrook.edu

  2. memcache

  3. Benchmark competitions

  4. Benchmark competitions

  5. fawn paper

  6. Data-intensive workloads • Scan-bound • large • sequential • examples? • problems? • Seek-bound • small • random • examples? • problems?

  7. Why FAWN? • Memory wall (??) • Increased CPU power consumption • DVFS is limited • Modern CPUs operate close to Vmin • Constant leakage current • Peak power and data center density

  8. FAWN results

  9. Processor scaling trends • Moore’s law (observation) • # transistors/chip ↑ 2X/2yr (how?) • Frequency ↑ as transistor size ↓ (max 9GHz) • Leakage current/power ↑ as transistor size ↓ • Heat ↑ as frequency ↑

  10. Processor scaling trends • Dennard scaling (observation) • Transistor power (V+I) ↓ as transistor size ↓ • + Moore’s law = perf/watt ↑ 2X/2yr (how?) • Not true now due to leakage current • So we did multicore!

  11. heteromates paper

  12. Dark silicon • Cannot power on all of the CPU • Result of: • Success of Moore’s law • Failure of Dennardscaling • Turbo boost

  13. Main ideas • Heterogeneous cores • For different performance requirements • Energy considerations (battery vs. plugged) • Thermal considerations • Dark silicon

  14. Main ideas

  15. Main ideas

More Related