mcguffey@muvms3.bitnet (Michael McGuffey) (12/06/90)
This is probably a "frequently asked question", but where may I find what used to be sent out as a listing of benchmarks for "all" computers/OS's. In particular, I'm looking for performance comparisons between the 80386(MS-DOS), 80386SX(MS-DOS), 68030(MacOS), and 68040(NexT unix). Both general and a floating point hardware numbers would be appreciated. Three platforms are being looked at for a Mathematics Lab using mostly Mathematica, but also a variety of home grown and/or publicly available software. These numbers would only be one of several criteria for selection. One vendor is claiming that the any program (Mathematica, Pagemaker, etc.) running on a 16MHz 386SX MSDOS system will outperform the same program running on a MacIIsi (20MHz 68030). I find this difficult to swallow, so I'm appealing to the net for any assistance. If anyone knows of some available comparisons, I would appreciate hearing of them. Thanks, -- michael ----------------------------------------------------------------------------- Michael McGuffey, Director BITNET: mcguffey@marshall Office of Institutional Research Internet: mcguffey%muvms3@wvnvms.wvnet.edu Marshall University Phone: 304/696-3212 Huntington, WV 25755 FAX: 304/696-3601 -----------------------------------------------------------------------------
kaleb@thyme.jpl.nasa.gov (Kaleb Keithley ) (12/07/90)
In article <31766@muvms3.bitnet> mcguffey@muvms3.bitnet (Michael McGuffey) writes: > > In particular, I'm looking for performance comparisons between >the 80386(MS-DOS), 80386SX(MS-DOS), 68030(MacOS), and 68040(NexT unix). >Both general and a floating point hardware numbers would be appreciated. For what it's worth, I got the following numbers from one of the Dhrystone benchmarks floating around: 386SX (16mhz) 386DX (20mhz) 486 (25mhz) Sun 4/330 SunOS cc ---- ---- ----- 27000 ESIX cc 4500 8000 24000 ----- gnu 1.37.1 5000 9000 28000 ----- -- Kaleb Keithley Jet Propulsion Labs kaleb@thyme.jpl.nasa.gov You can please all of the people some of the time,
minich@d.cs.okstate.edu (Robert Minich) (12/07/90)
by mcguffey@muvms3.bitnet (Michael McGuffey): > ...These numbers would only be one of several criteria for > selection. One vendor is claiming that the any program (Mathematica, > Pagemaker, etc.) running on a 16MHz 386SX MSDOS system will outperform > the same program running on a MacIIsi (20MHz 68030). I find this difficult The only _reasonable_ difference I could forsee is the DOS machine would have an FPU (usually don't). The IIsi doesn't come with an FPU but you do get one with either slot-thing-card (NuBus or 030 Direct). As far as ALL programs, the sales-slime is probably talking a lot of BS. The 386sx has a 16 bit data path while the 030 in the IIsi is using a 32 bit path. This should make an appreciable difference on cpu intensive programs. Also, I am under the impression that most DOS stuff is NOT written for a 32 bit instruction set, which would preclude running on XTs and ATs. If this DOS machine is to be running Windows, then I suspect there isn't a chance in hell of it keeping up. > -- michael > ----------------------------------------------------------------------------- > Michael McGuffey, Director BITNET: mcguffey@marshall > Office of Institutional Research Internet: mcguffey%muvms3@wvnvms.wvnet.edu > Marshall University Phone: 304/696-3212 > Huntington, WV 25755 FAX: 304/696-3601 > ----------------------------------------------------------------------------- -- |_ /| | Robert Minich | |\'o.O' | Oklahoma State University| "Get bent." |=(___)= | minich@d.cs.okstate.edu | -- Bart Simpson | U | - Ackphtth |