CA7 - 1 Chapter 7 Large and Fast: Exploiting Memory...

Info iconThis preview shows pages 1–13. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1 Chapter 7 Large and Fast: Exploiting Memory Hierarchy The Big Picture: Where are we now? The Big Picture: Where are we now? Datapath Memory Processor Input Output Control Next: Focus on the memory system Memory Processor Input Output Control So far: Focus on processor datapath and control Datapath 3 A DRAM Cell A DRAM Cell Word-line Bit-line Word-line Bit-line Word-line Bit-line ++++- - - - 1 Memory Gap: Since 1980, CPU has Memory Gap: Since 1980, CPU has outpaced DRAM ... outpaced DRAM ... CPU 60% per yr 2X in 1.5 yrs DRAM 9% per yr 2X in 10 yrs 1 DRAM CPU Performance (1/latency) 1 1 1 9 8 2 1 9 9 Year Gap grew 50% per year 5 Are Memory Designers Doing Are Memory Designers Doing Their Job? Their Job? 1980 1982 1984 1986 1988 1990 1992 1994 50 100 150 200 250 1 Mbit 64 Mbit 16 Mbit 4 Mbit 256 Kbit 64 Kbit Speed (ns) Review: Two inverters store a bit Review: Two inverters store a bit The other elements in a memory circuit control reading and writing 16-transistor circuit. Most transistors implement read/write semantics 1 1 1 Example: Flip-Flop D Q CLK For use in arrays: Static RAM (SRAM) For use in arrays: Static RAM (SRAM) cell cell Writing a bit Drive bit lines with new data and activate word line 1 1 bit bit word Reading a bit Activate word line let cell drive bit lines. 1 1 1 bit bit word 1 Putting it all together: an SRAM array Putting it all together: an SRAM array Write Driver Write Driver Write Driver Write Driver Parallel Data I/O Lines Add muxes to select subset of bits Memory Gap: Since 1980, CPU has Memory Gap: Since 1980, CPU has outpaced DRAM ... outpaced DRAM ... CPU 60% per yr 2X in 1.5 yrs DRAM 9% per yr 2X in 10 yrs 1 DRAM CPU Performance (1/latency) 1 1 1 9 8 2 1 9 9 Year Gap grew 50% per year Q. How do architects address this gap? A. Put smaller, faster cache memories between CPU and DRAM. Create a memory hierarchy. 10 10 Structure of a Memory Hierarchy Structure of a Memory Hierarchy CPU Level 1 Level 2 Level N . Size of the memory at each level Increasing distance from the CPU In access time 11 11 Locality Locality Temporal locality (locality in time): If an item is referenced, it will tend to be referenced again. Spatial locality (locality in space): If an item is referenced, items whose addresses are close by will tend to be referenced soon. 12 12 Locality Locality Spatial locality Spatial locality Solution: move data located near a Solution: move data located near a recently accessed data closer to the recently accessed data closer to the processor processor...
View Full Document

This note was uploaded on 01/26/2011 for the course CSCI 4203 taught by Professor Weichunghsu during the Fall '05 term at Minnesota.

Page1 / 33

CA7 - 1 Chapter 7 Large and Fast: Exploiting Memory...

This preview shows document pages 1 - 13. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online