mikem@uhccux.uhcc.hawaii.edu (Mike Morton) (02/15/89)
I'm curious about the use of the 68882 math chip by SANE in the SE/30. I
believe that the older '881 chip has some admitted precision problems for
some functions (transcendentals?). Apple, being the purists that they
are, didn't use the '881 to implement these particular functions, but
left the computation in software.
Does the '882 fix these inaccuracies? If so, does SANE in the SE/30 use
the hardware for more operations?
On a related note, the March MacUser mentions a benchmark which runs much
more slowly on an SE/30 than a Mac IIx, and that "Apple's engineers were
mystified" as to why (p. 182). Anyone know why?
-- Mike Morton // P.O. Box 11378, Honolulu, HI 96828, (808) 676-6966 HST
Internet: msm@ceta.ics.hawaii.edu
(anagrams): Mr. Machine Tool; Ethical Mormon; Chosen Immortal; etc.
P.S. On the subject of SE hardware, note p. 180 of the same issue, with a
diagram listing the "Video ROM" -- but no corresponding number in the
picture. Is this a joke or a funny typo? [Presumably the Video ROM is
where the developer's pictures are hidden...]holland@m2.csc.ti.com (Fred Hollander) (02/17/89)
In article <3263@uhccux.uhcc.hawaii.edu> mikem@uhccux.uhcc.hawaii.edu (Mike Morton) writes: > >On a related note, the March MacUser mentions a benchmark which runs much >more slowly on an SE/30 than a Mac IIx, and that "Apple's engineers were >mystified" as to why (p. 182). Anyone know why? MacWorld also published benchmarks. The strangest was the graphics benchmark that ran roughly 30% faster on the IIx despite the video RAM being on a 10MHz NuBus. Fred Hollander Computer Science Center Texas Instruments, Inc. hollander@ti.com The above statements are my own and not representative of Texas Instruments.
alexis@ccnysci.UUCP (Alexis Rosen) (02/19/89)
In article <3263@uhccux.uhcc.hawaii.edu> mikem@uhccux.uhcc.hawaii.edu (Mike Morton) writes: >I'm curious about the use of the 68882 math chip by SANE in the SE/30. I >believe that the older '881 chip has some admitted precision problems for >some functions (transcendentals?). Apple, being the purists that they >are, didn't use the '881 to implement these particular functions, but >left the computation in software. > >Does the '882 fix these inaccuracies? If so, does SANE in the SE/30 use >the hardware for more operations? The '882 is identical to the '881 in terms of precision. These 'inaccuracies' are simply motorola's chosen precision. (I don't remember for sure, but I think moto used 80-bit numbers for transcendentals while Apple's SANE uses 96 bit numbers). SANE has not changed in the /30. >On a related note, the March MacUser mentions a benchmark which runs much >more slowly on an SE/30 than a Mac IIx, and that "Apple's engineers were >mystified" as to why (p. 182). Anyone know why? I don't know why they printed that. Neither do they :-) In fact, it's no big mystery, and when I talked to them monday after the show they confirmed my suspicion that it had to do with the fact that video is handled very differently (in terms of the actual hardware, not the software) on the SE/30 than on the Mac II. Video writes take quite a bit longer. (Shoot me if I'm wrong. This was only four days after it was announced, and I think they made the same logical guess I did, but who knows?) Alexis Rosen alexis@ccnysci.uucp
siegel@endor.harvard.edu (Rich Siegel) (02/20/89)
In article <1289@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >The '882 is identical to the '881 in terms of precision. These 'inaccuracies' >are simply motorola's chosen precision. (I don't remember for sure, but I >think moto used 80-bit numbers for transcendentals while Apple's SANE uses >96 bit numbers). SANE has not changed in the /30. The 6888x uses a 96-bit data type to represent IEEE extended- precision, and SANE uses an 80-bit type. However, the extra 16 bits are set to zero, and are used, according to the 6888x manuals, "for longword alignment and future expansion". The number of bits affecting the precision is the same in either case: both use 80-bit calculation precision. Transcendentals done in hardware are much faster, but somewhat less accurate; a good compromise can be reached by using SANE for transcendentals and hardware for four-banger (+ - * /) functions. If you're not paranoid about a few digits on the end of your transcendental results, you can use the hardware transcendentals and get about two orders of magnitude speed increase. --Rich Rich Siegel Staff Software Developer THINK Technologies Division, Symantec Corp. Internet: siegel@endor.harvard.edu UUCP: ..harvard!endor!siegel Phone: (617) 275-4800 x305
holland@m2.csc.ti.com (Fred Hollander) (02/20/89)
In article <1289@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >In article <3263@uhccux.uhcc.hawaii.edu> mikem@uhccux.uhcc.hawaii.edu >(Mike Morton) writes: > >>On a related note, the March MacUser mentions a benchmark which runs much >>more slowly on an SE/30 than a Mac IIx, and that "Apple's engineers were >>mystified" as to why (p. 182). Anyone know why? > >I don't know why they printed that. Neither do they :-) In fact, it's no >big mystery, and when I talked to them monday after the show they confirmed >my suspicion that it had to do with the fact that video is handled very >differently (in terms of the actual hardware, not the software) on the SE/30 >than on the Mac II. Video writes take quite a bit longer. And Apple's engineers don't know about this !? >(Shoot me if I'm wrong. This was only four days after it was announced, and >I think they made the same logical guess I did, but who knows?) > >Alexis Rosen >alexis@ccnysci.uucp Fred Hollander Computer Science Center Texas Instruments, Inc. hollander@ti.com The above statements are my own and not representative of Texas Instruments.
pollock@usfvax2.EDU (Wayne Pollock) (02/21/89)
I've read that there exist INITs (from companies such as Levco) that bypass
the software SANE calls and pass the instructions directly to the chip.
The reported speedup is amazing. I have looked for a PD/Shareware version
of this on information services such as GEnie, but have found none. Why
hasn't some (in)SANE guru provide one?
Wayne Pollock (The MAD Scientist) pollock@usfvax2.usf.edu
Usenet: ...!{uflorida, codas}!usfvax2!pollock
GEnie: W.POLLOCKalexis@ccnysci.UUCP (Alexis Rosen) (02/21/89)
In article <70359@ti-csl.csc.ti.com> holland@m2.UUCP (Fred Hollander) writes: >In article <1289@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >>I don't know why they printed that. Neither do they :-) In fact, it's no >>big mystery, and when I talked to them monday after the show they confirmed >>my suspicion that it had to do with the fact that video is handled very >>differently (in terms of the actual hardware, not the software) on the SE/30 >>than on the Mac II. Video writes take quite a bit longer. > >And Apple's engineers don't know about this !? They do. That's why I said, "I don't know why they printed that." They did discuss it with Apple, and Apple confirmed that it was related to the Video stuff. There is something that I would like to know which has not been mentioned anywhere yet that I know of. How is the video implemented on the SE/30? Does it look like a Nubus Card, did they fake it another way, or what? Alexis Rosen alexis@ccnysci.uucp
jurjen@cwi.nl (Jurjen N.E. Bos) (02/22/89)
In article <1303@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >There is something that I would like to know which has not been mentioned >anywhere yet that I know of. How is the video implemented on the SE/30? >Does it look like a Nubus Card, did they fake it another way, or what? They do it the fastest way possible: dual-port RAM. (For less than wizards: this is some weird kind of memory where you can access two addresses exactly simultaneously: one for the scanning and one for the processor. It requires a lot of pins on a chip, so it is costly, but it works.) For more details, read the article on the SE/30 in the latest BYTE. -- -- Jurjen N.E. Bos (jurjen@cwi.nl)
holland@m2.csc.ti.com (Fred Hollander) (02/24/89)
In article <1303@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >In article <70359@ti-csl.csc.ti.com> holland@m2.UUCP (Fred Hollander) writes: >>In article <1289@ccnysci.UUCP> alexis@ccnysci.UUCP (Alexis Rosen) writes: >>>I don't know why they printed that. Neither do they :-) In fact, it's no >>>big mystery, and when I talked to them monday after the show they confirmed >>>my suspicion that it had to do with the fact that video is handled very >>>differently (in terms of the actual hardware, not the software) on the SE/30 >>>than on the Mac II. Video writes take quite a bit longer. >> >>And Apple's engineers don't know about this !? > >They do. That's why I said, "I don't know why they printed that." They did >discuss it with Apple, and Apple confirmed that it was related to the Video >stuff. I thought the "... they :) ..." referred to MacUser not knowing why they, themselves printed it - that they in fact knew the video was handled differently. Now I see that you meant Apple had explained to MacUser who apparently wasn't listening. >There is something that I would like to know which has not been mentioned >anywhere yet that I know of. How is the video implemented on the SE/30? >Does it look like a Nubus Card, did they fake it another way, or what? I'm not satisfied either, just knowing that it's different. Can anyone explain how (and why)? Is the internal video handled differently than a card in the direct slot? Fred Hollander Computer Science Center Texas Instruments, Inc. hollander@ti.com The above statements are my own and not representative of Texas Instruments.