EENG 449bG/CPSC 439bG Computer Systems Lecture 17 Memory Hierarchy Design Part I. April 7, 2005 Prof. Andreas Savvides Spring 2005 http://www.eng.yale.edu/courses/2005s/eeng449b. µProc 60%/yr. 1000. CPU. “Moore’s Law”. 100. Processor-Memory Performance Gap: (grows 50% / year).
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.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.
April 7, 2005
Prof. Andreas Savvides
Performance Gap:(grows 50% / year)
2000Who Cares About the Memory Hierarchy?
1st Alpha: 340 ns/5.0 ns = 68 clks x 2 or 136
2nd Alpha: 266 ns/3.3 ns = 80 clks x 4 or 320
3rd Alpha: 180 ns/1.7 ns =108 clks x 6 or 648
Processor % Area %Transistors
Disk, Tape, etc.
Selects the set
Selects the desired data from the block
Compared against for a hit
While in cache
Cache-2Unified vs Split Caches
AMATHarvard=75%x(1+0.64%x50)+25%x(1+6.47%x50) = 2.05
Performance with cache misses
Four main categories of optimizations
- multilevel caches, critical word first, read miss before write miss, merging write buffers and victim caches
- larger block size, larger cache size, higher associativity, way prediction and pseudoassociativity and computer optimizations
2. Reduce the miss penalty or miss rate via parallelism
- non-blocking caches, hardware prefetching and compiler prefetching
3. Reduce the time to hit in the cache
- small and simple caches, avoiding address translation, pipelined cache access
3) Change Compiler:
Which of 3Cs is obviously affected?
MissesLarger Block Size (fixed size&assoc)
What else drives up block size?
Flaws: for fixed block size
Good: insight => invention