jdudeck@polyslo.CalPoly.EDU (John R. Dudeck) (02/06/90)
In article <25c86533:4479.2comp.sys.ibm.pc;1@vpnet.UUCP> akcs.amparsonjr@vpnet.UUCP (Anthony M. Parson, Jr.) writes: >will say an 80386 (25 mhz) without a coprocessor, is STILL faster by far >than the 8088/8087 combination. Moral: get a fast 80286 or 80386, and >possibly save the cost of buying a math coprocessor. Of course, it depends I haven't done any precise comparisons, but I watched a friend run Generic CADD 2.0 on his 8mhz luggable XT with a 4 mhz 8087, and then tried running the same thing on a Premium/286 10 mhz with no math chip. There was no contest. The XT won easily. Now I have a 386/20, and I haven't had a chance to do any comparisons. I also have a math chip. But I would be willing to bet my money on the XT against a 386 with no math chip. My 386 has the option of EITHER an 80287 at 8 mhz (costs under $200) or an 80387 at 20mhz (big bucks). Needless to say, I chose the '287, since I am a poor student and not doing big time number cruching or lots of CAD. The motherboard is an X'Golden. I think some other motherboards give this same option. When buying a budget system it is definitely a point to consider if you are going to be wanting a math chip. -- John Dudeck "You want to read the code closely..." jdudeck@Polyslo.CalPoly.Edu -- C. Staley, in OS course, teaching ESL: 62013975 Tel: 805-545-9549 Tanenbaum's MINIX operating system.
baer@uwovax.uwo.ca (02/06/90)
In article <25ce13f4.5fd6@polyslo.CalPoly.EDU>, jdudeck@polyslo.CalPoly.EDU (John R. Dudeck) writes: > In article <25c86533:4479.2comp.sys.ibm.pc;1@vpnet.UUCP> akcs.amparsonjr@vpnet.UUCP (Anthony M. Parson, Jr.) writes: >>will say an 80386 (25 mhz) without a coprocessor, is STILL faster by far >>than the 8088/8087 combination. Moral: get a fast 80286 or 80386, and >>possibly save the cost of buying a math coprocessor. Of course, it depends > > I haven't done any precise comparisons, but I watched a friend run Generic > CADD 2.0 on his 8mhz luggable XT with a 4 mhz 8087, and then tried running > the same thing on a Premium/286 10 mhz with no math chip. There was no > contest. The XT won easily. Now I have a 386/20, and I haven't had a chance > to do any comparisons. I also have a math chip. But I would be willing to > bet my money on the XT against a 386 with no math chip. > We did some timings on computationally intensive programs, with relatively little I/O, on three machines: 1) my 286 with an 8mHz math chip, 2) a grad student's machine, a 386, 25 mHz, with cache, no math chip, 3) the department's 386, 20 mHz, no cache, with a 387 chip (20mHz). (BTW, all machines have disk drives running at 28ms, but as I said, the programs involved relatively little I/O). The 286/287 combination "timed in" at about the same (perhaps 10% slower) as the 386/cache/no chip configuration. The 387/no cache/math chip configuration was 4-5 times faster. I understand that the Intel 387 chip is considerably faster than the 287 chip (even if we control for speed differences). This seems to bear this out. I have no timings on a 386sx-387sx combination, but I understand the 387sx chip is considerably faster than an equivalent 287 chip. [This implies that while a 16mHz 386sx machine may not be that much faster for most operations than a plain vanilla 12mHz 286, and perhaps even a bit slower than a 16mHz 286, if it has a 387sx chip it should be *much* faster than a 286-287 -- perhaps other newsgroup contributors might care to comment on this]. --------------------------------------------------------------------- Douglas Baer, Assoc. Prof., Dept. of Sociology, University of Western Ontario, London, Canada N6A 5C2 Internet: BAER@UWO.CA Bitnet: BAER@UWOVAX Telephone: [home] (519)-657-4799 *most reliable number [office] (519)-661-3859 [leave messages, M-F 9-4pm EST] (519)-661-3606
boyd@hpavla.AVO.HP.COM (Bill Boyd) (02/07/90)
About doing math-intensive operations with and without a coprocessor: I adapted a program for displaying the Mandelbrot set from an old issue of Turbo Technix. It would take 20 minutes to run on my XT with 8087. I bought a 16-MHz 80386SX machine without a math coprocessor. It took 53 (!) minutes to run the same program! I bought an 80387SX, and now it takes 3 minutes to complete. I say the math coprocessor makes a HUGE difference if the program is math-intensive enough! Bill Boyd
chasm@attctc.Dallas.TX.US (Charles Marslett) (02/07/90)
In article <4911.25ce0291@uwovax.uwo.ca>, baer@uwovax.uwo.ca writes: :: In article <25ce13f4.5fd6@polyslo.CalPoly.EDU>, jdudeck@polyslo.CalPoly.EDU (John R. Dudeck) writes: :: We did some timings on computationally intensive programs, with relatively :: . . . The 387/no cache/math chip configuration was 4-5 times :: faster. I understand that the Intel 387 chip is considerably faster than :: the 287 chip (even if we control for speed differences). This seems to :: bear this out. I have no timings on a 386sx-387sx combination, but :: I understand the 387sx chip is considerably faster than an equivalent :: 287 chip. [This implies that while a 16mHz 386sx machine may not be :: that much faster for most operations than a plain vanilla 12mHz 286, :: and perhaps even a bit slower than a 16mHz 286, if it has a 387sx chip :: it should be *much* faster than a 286-287 -- perhaps other newsgroup :: contributors might care to comment on this]. The 387sx is rumored to be pretty much a 387 with a 16-bit inteface glued on. The IIT 287 is essentially that (just a FAST 387 with a 16-bit interface -- ok, so it's just 10% faster ;^). So they still sit about even so long as you run DOS code. What's more, the IIT chip seems to be a bit cheaper when you can find it. :: --------------------------------------------------------------------- :: Douglas Baer, Assoc. Prof., Dept. of Sociology, :: University of Western Ontario, London, Canada N6A 5C2 Charles Marslett chasm@attctc.dallas.tx.us
sbnicol@rose.waterloo.edu (Scott Nicol) (02/08/90)
In article <8370003@hpavla.AVO.HP.COM> boyd@hpavla.AVO.HP.COM (Bill Boyd) writes: >About doing math-intensive operations with and without a coprocessor: > >I adapted a program for displaying the Mandelbrot set from an old issue >of Turbo Technix. It would take 20 minutes to run on my XT with 8087. >I bought a 16-MHz 80386SX machine without a math coprocessor. It took >53 (!) minutes to run the same program! I bought an 80387SX, and now >it takes 3 minutes to complete. About five years ago, as a school project, a friend and I wrote a program that accepted a function in three variables (x, y, and z), and plotted a (accurate) 3D plot of the function. The machine we were using was a (brand new at the time) Compaq Deskpro (8 Mhz 8086), with an 8 Mhz 8087. Using an inefficient C compiler (C86), the program could plot a reasonably complex function in a minute. without the coprocessor, we estimated it would take 2 hours (we didn't run it for 2 hours, we ran it for 5 minutes and estimated how far along it was). Last year, I discovered the source code lying around on one of my old disks. I ported it to Turbo C (2.0), and compiled it (with 286 code generation) on an IBM PS/2 model 70 (25Mhz 8. The same function that took a minute on my old Deskpro took about 10 minutes on the 386. After that, I took the program to a 33Mhz Compaq Deskpro (with a coprocessor and a real fancy EGA), it took less than 5 seconds. So, if software is numerically intensive, and if it is smart enough to recognize the coprocessor, then it can make a huge difference. Of course, if the program is I/O bound, the '387 will hardly make a dent. - Scott Email: sbnicol@rose.waterloo.edu -or- ...!uunet!watmath!rose!sbnicol SlowMail: 546 FallingBrook Dr., Waterloo, Ont., Canada, N2L 4N4. Phone: (519) 725-1980
besler@egrunix.UUCP (Brent Besler) (02/09/90)
Since the issue of coprocessors is being brought up, I noticed in the latest PC magazine some test of the Cyrix 25 Mhz 80387 replacement vs. the Intel part were done. They didn't find much of a speed increase for the price. The increae was about 15% for their Autocad benchmark. The Cyrix is about twice the price of the Intel part. It is CMOS, so it will run cooler and probably last longer. I am curious to know what results people on the net have had with the Cyrix or IIT replacement chips for the 80387. Brent H. Besler