jqj@cornell.UUCP (J Q Johnson) (01/17/86)
In article <182@intelca.UUCP> glen@intelca.UUCP (Glen Shires) writes: >> Please note, and tell all your friends, that Norton's SI test results >> show NO relationship to the actual overall machine performance. >> >I agree SI is weighted, but what is a realistic system performance test? >One application may be spreadsheet recalculation which is numerics intensive. >Another may be a compile with multiple libraries which is disk intensive >(Or a compile with RAMDISK which is string-operation intensive). > >Maybe the best test is to simply time your favorite application. > Yes, the best test is clearly application-specific. But if you want to decide which PC-clone to buy you generally don't have the chance to try out your particular application on dozens of alternative systems. So some standard performance benchmark is needed, even if we agree to take its results with a grain (or even a shakerfull) of salt. A cpu/compiler performance test that is becoming standard is the Dhrystone. It could be used as a measure of PC-compatible cpu performance if a standard compiler (perhaps Lattice C?) were agreed upon as the comparator. Note that the Dhrystone benchmark is a "typical" (empirically determined for a large set of high-level-language programs) mix of integer, string, flow-control, and procedure calling. It does not test IO performance at all.