neighorn@qiclab.UUCP (05/11/87)
In a previous article, Michel Bruneau writes: >It is a well know fact that the 8086-8087 combinaison beats easily the >80286-80287 setup in term of numbercrunching speed (at least for the >Fortran programs I use!) .. Now that the 80386 is out, is there a >math-coprocessor that will come with it , and can we expect real output >from it (i.e. speeding up 20 times or more, like the 8087 did with the 8086)? While it is true the combinatation of an 8086/8087 surpasses the number crunching output of the 286/287 pair operating at similar clock speeds, this is not always the case. Higher speed 286/287 combo's (ie 10-12Mhz) are faster than the slower speed 86/87 pair. The speed differential is due mainly to the 80287 accessing memory through the 80286 due to protection schemes, where the 8087 does not have this requirement. Early results with the 16Mhz 80386/80387 show the numerical throughput of the 80387 to be 5-6 times that of the 80287. This is good news to me, for I am one of those lucky individuals with a 10Mhz 80287 on an emulator board stuck in the 80387 socket. f their 80386 machine. Another problem is the complete lack of software to take advantage of the extra functionality on the 80*87 chips. MSDOS, in all its glory, will never do much more than make the higher end math coprocessors run 8087 code. Only unix and xenix take advantage of the 80287's added instructions. Who knows what Microsoft's 'soon-to-be-released' OS/2 will do with the chip. Another case of unused potential! :-) -- Steven C. Neighorn tektronix!{psu-cs,reed}!qiclab!neighorn Portland Public Schools "Where we train young Star Fighters to defend the (503) 249-2000 ext 337 frontier against Xur and the Ko-dan Armada"
toma@tekgvs.TEK.COM (Thomas Almy) (05/12/87)
In article <447@qiclab.UUCP> neighorn@qiclab.UUCP (Steven C. Neighorn) writes: > >Another problem is the complete lack of software to take advantage of the >extra functionality on the 80*87 chips. MSDOS, in all its glory, will never >do much more than make the higher end math coprocessors run 8087 code. >Only unix and xenix take advantage of the 80287's added instructions. Not true, Steve. The Metaware High C 386 compiler emits inline 80387 code for trig functions. It it runs under MS-DOS! Their 286 version emits different code for 287s than for 87s. On the other hand protected mode 286 code is beyond MSDOS except for VDISK type use. But we know that the 286 is just a Stop-Gap measure for the REAL processor, the 80386! Tom Almy (*NIX -- I hate it but I use it twice a day.)
neighorn@qiclab.UUCP (05/15/87)
In article <2279@tekgvs.TEK.COM> toma@tekgvs.UUCP (Thomas Almy) writes: > <comments about msdos only supporting 8087 code> > >Not true, Steve. The Metaware High C 386 compiler emits inline 80387 code for >trig functions. It it runs under MS-DOS! Their 286 version emits different >code for 287s than for 87s. On the other hand protected mode 286 code is >beyond MSDOS except for VDISK type use. But we know that the 286 is just a >Stop-Gap measure for the REAL processor, the 80386! > >Tom Almy >(*NIX -- I hate it but I use it twice a day.) Ok, you got me on this one. I forgot about them high-falutin fancy-dan C compilers that actually know what kind of ship they are running on.... and more importantly, take advantage of it. Now if software producers would start compiling their C code with compilers such as the Metaware High C compiler, maybe we could start using these advanced chips the way they were meant to be used. Of course that would leave out zillions of 8086/8088 machines, and that doesn't go well with the marketing department. I know that compatibility with older machines is important. I have heard all those lines. I guess that's why unix is running on my machine 80% of the time - no 640k memory barriers, no funny disk limitations, etc. A real operating system as opposed to a control program and device driver handler. :-) -- Steven C. Neighorn tektronix!{psu-cs,reed}!qiclab!neighorn Portland Public Schools "Where we train young Star Fighters to defend the (503) 249-2000 ext 337 frontier against Xur and the Ko-dan Armada"