1 / 11

Power-Aware Hardware Prefetching

Power-Aware Hardware Prefetching. Yao Guo. Prefetching Energy Consumption. L2 + Memory Access Energy very minor energy impact. Leakeage Energy reduced due to performance improvement L1 Access energy increased significantly. Prefetching Energy Consumption. Cache Way-Prediction.

makya
Download Presentation

Power-Aware Hardware Prefetching

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. Power-Aware Hardware Prefetching Yao Guo

  2. Prefetching Energy Consumption • L2 + Memory Access Energy • very minor energy impact. • Leakeage Energy • reduced due to performance improvement • L1 Access energy • increased significantly.

  3. Prefetching Energy Consumption

  4. Cache Way-Prediction • ISLPED’99, Inoue et al. • MRU (Most Recently Used Way) • Remember the MRU way for each cache set.

  5. Prefetching Tag Buffer (PTB) • Use a small hardware buffer (PTB) to remember the most recently prefetched cache tags. • No prefetching if data is found in PTB. • Use LRU algorithm when the buffer is full. • Energy cost is small. • Access the buffer only during prefetching • Buffer is much smaller compared to MRU.

  6. How does PTB work? Prefetching Request • A prefetching request is dropped if found in the PTB. PTB Hit Prefetching Dropped Miss DL1

  7. Predictor Hit Rate

  8. # of prefetches dropped due to misprediction

More Related