dkirk@k.gp.cs.cmu.edu (Dave Kirk) (01/27/89)
I am looking for references to techniques for modelling memory access patterns and approximate cache hit (miss) rates. I'd like to reduce specific access strings to a few parameters which could be used to determine approximate cache hit (miss) rates. I know of at least two such techniques. A fractal model was proposed by Thiebaut (1987), and models a fully associative cache hit rate using four program parameters. Another approach was proposed by Strecker (1983) in which he modified a linear approach presented by Saltzer (1974). Strecker's approach derived two parameters by measuring the program miss rate in a large and small cache. These parameters can then be used to calculate the miss rate for any other cache size. Additional references have been hard to come by, and would be greatly appreciated. (replies can be sent to above account, or kirk@maxwell.ece.cmu.edu) --
jja@etana.tut.fi (Ahola Jari) (01/30/89)
From article <4122@pt.cs.cmu.edu>, by dkirk@k.gp.cs.cmu.edu (Dave Kirk): > I am looking for references to techniques for modelling memory access > patterns and approximate cache hit (miss) rates. I'd like to reduce > specific access strings to a few parameters which could be used to > determine approximate cache hit (miss) rates. > > Additional references have been hard to come by, and would be greatly > appreciated. > Best reference to my knowledge is Pohm & Agrawahl: High speed memory systems which covers many cache organisations and models for modelling their performance. -jja Jari 'jja' Ahola |Tampere University of Technology, Software Systems Lab Opiskelijankatu 16A12 |P.O. Box 527, 33101 Tampere, Finland 33720 Tampere, Finland|Tel (intl) 358 31 162708 (work)/358 31 174009 (home) Puh. 931-174009 |Net address: jja@tut (UUCP) AHOLA@FINTUTA (BITNET)