hedley@imagen.UUCP (Hedley Rainnie) (08/12/87)
I enclose here a previous posting that I use for processor
comparisons. I also claim that if those who bellyache about
their computing speed actually spent the bellyaching time doing
thier computation then these arguments would be moot.
Can you perform the computations you want in a timely fasion?
if no : look into alternatives.
Inserted text begins:
From: paf@oblio.UUCP (Paul Fronberg)
Date: 20-Sep-86 13:29 EDT
Subject: Benchmarks in August IEEE Micro
In the August issue of IEEE Micro there is a very interesting article
concerning benchmarking 32-bit microprocessors. The following table is
abstracted from page 57. Numbers are time in seconds. (N=no cache enabled;
C=cache enabled). This table reflects the results for dynamic memory.
MHz E F H I K
80286 (10) 4.89 13.63 6.59 11.20 19.39
80386 (16) 3.57 5.16 3.63 6.86 6.20
68000 (8) 13.73 14.61 8.79 12.08 16.59
68020 N (16) 8.02 5.55 3.84 5.65 4.78
68020 C (16) 3.84 2.47 2.14 2.75 3.02
32032 (10) 12.52 13.07 6.21 8.57 13.07
32100 N (18) 16.81 8.84 5.05 8.57 9.17
32100 C (18) 6.75 4.29 2.74 3.63 4.45
Benchmarks are EDN 16 benchmarks modified for 32 bits. The benchmarks were
coded in assembly code for each processor.
The following EDN programs were used.
Test E is a character-string search routine.
Test F is a bit test, set, and reset routine.
Test H is a linked-list insertion routine.
Test I is a quicksort routine.
Test K is a bit-matrix transposition routine.
Inserted text ends.
Cheers Hedley
{decwrl|sun}!imagen!hedleydbt@faccs.UUCP (Data base translator ) (08/20/87)
Another thing to keep in mind is the age of the 68020 vs. 80386. The way the Intel processor releases "leapfrog" with the Motorola processor releases, one should keep in mind that the Motorola counterpart to the 80386 would probably be the 68030 which is just going into production and will begin appearing in Suns, Macs, Apollos, etc. after the first of the year. Looking at it this way, the '020 not only compares well with the '386; it blows the socks off its Intel counterpart, the '286. And the '030? I can't wait.
sho@tybalt.caltech.edu (Sho Kuwamoto) (08/23/87)
In article <116@faccs.UUCP> dbt@faccs.UUCP (Data base translator ) writes: > >Looking at it this way, the '020 not only compares well with the '386; it >blows the socks off its Intel counterpart, the '286. And the '030? I can't >wait. Does anyone know what Apple plans to do with the 030, if anything? I imagine third party 030 cards will become available for the MacII and maybe the SE, but they may not utilize the full instruction set. Specifically, does Apple plan to put the 030 on a slightly revised version of the MacII (Call it the MacII+ with Applesoft Basic instead of integer basic), do they plan to put out a wholly new machine, or do they not plan to use it at all (doubtful)? Also, has anyone heard anything about NeXT? I'm not really worried, but I might possibly get a Mac II, and it would really suck if something came and blew it out of the water. I admit, the possibility is miniscule, but we are talking a heavy cash investment here. -Sho
cbenda@unccvax.UUCP (carl m benda) (08/23/87)
In article <116@faccs.UUCP>, dbt@faccs.UUCP (Data base translator ) writes: > Looking at it this way, the '020 not only compares well with the '386; it > blows the socks off its Intel counterpart, the '286. And the '030? I can't > wait. It really angers many people to see blind statements such as yours about the 68020 'blowing the socks off the '286'... In case you hadn't been reading the trade journals. AMD has recently been shipping 16Mhz versions of the 80286 which incidently contains, on the chip, memory management capabilities. Something that the 68020 MUST go off chip for, namely, the 68851 MMU. I believe most net readers would appreciate documented numbers, BEFORE you make any more of these irritable BLIND statements. /Carl ...decvax!mcnc!unnccvax!cbenda
jww@sdcsvax.UCSD.EDU (Joel West) (08/24/87)
The 68030 from Apple is probably late 1988. Sorry... I might point out, Sun might consider 20,000 68030 chips a large order. Apple would consider that a month's worth of sales, so there are limitations on how close to the cutting edge it can get.
chi@tybalt.caltech.edu (Curt Hagenlocher) (08/24/87)
In article <799@unccvax.UUCP> cbenda@unccvax.UUCP (carl m benda) writes: >the trade journals. AMD has recently been shipping 16Mhz versions of the >80286 which incidently contains, on the chip, memory management capabilities. >Something that the 68020 MUST go off chip for, namely, the 68851 MMU. I Memory management aside, it's pretty ridiculous claiming that the 80286, with a mere 16 bit data bus, can do as well as or better than the 68020, which has a full 32 bit bus. The fact that the 80286 can be obtained in a 16Mhz version doesn't change these facts. (I believe that the 68020 is avail- able in a 20- or even 24- Mhz model, if it's clock speed you want to compare.) It's pretty senseless comparing the two because of this, even though the 68020 came out only shortly after the Intel chip. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Curt Hagenlocher truC rehcolnegaH chi@tybalt.caltech.edu ude.hcetlac.tlabyt@ihc chi@citromeo.caltech.bitnet tentib.hcetlac.oemortic@ihc seismo!cit-vax!tybalt.caltech.edu!chi ihc!ude.hcetlac.tlabyt!xav-tic!omsies ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Curt Hagenlocher truC rehcolnegaH chi@tybalt.caltech.edu ude.hcetlac.tlabyt@ihc chi@citromeo.caltech.bitnet tentib.hcetlac.oemortic@ihc seismo!cit-vax!tybalt.caltech.edu!chi ihc!ude.hcetlac.tlabyt!xav-tic!omsies "If it didn't kank, it wouldn't be me""em eb t`ndluow ti ,knak t`ndid ti fI" ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
root@sbcs.UUCP (Root) (08/24/87)
> to the 80386 would probably be the 68030 which is just going into production > and will begin appearing in Suns, Macs, Apollos, etc. after the first of the > year. Perhaps I am mistaken, but my local Motorola Field Engrs told me that the '030 will not be in production until the first of the year. They claimed that the first sampling was scheduled for end of summer. This would seem to say that '030 systems will not really be showing up until late next spring; there is no telling where 80386 (and its offspring) will be by then.. As for showing up in new industry iron, I thought Sun went with their Sparc RISC chip (with promises to upgrade the existing Sun-3 line to 68030 "sometime in the future"), and Apollo was going AMD29000 - thus just jumping ahead to Supercomputing workstations rather than bothering about the next incremental speed change in 680?0. Rick Spanbauer SUNY/Stony Brook
jww@sdcsvax.UCSD.EDU (Joel West) (08/24/87)
The Sun-3/260 uses a 25 MHz 68020. I don't believe there is a part
between the 16 MHz and 25 Mhz models.
--
Joel West (c/o UCSD)
Palomar Software, Inc., P.O. Box 2635, Vista, CA 92083
{ucbvax,ihnp4}!sdcsvax!jww jww@sdcsvax.ucsd.edu
or ihnp4!crash!palomar!joel joel@palomar.cts.comjwhitnel@csib.UUCP (Jerry Whitnell) (08/24/87)
In article <799@unccvax.UUCP| cbenda@unccvax.UUCP (carl m benda) writes: |In article <116@faccs.UUCP>, dbt@faccs.UUCP (Data base translator ) writes: |> Looking at it this way, the '020 not only compares well with the '386; it |> blows the socks off its Intel counterpart, the '286. And the '030? I can't |> wait. | |It really angers many people to see blind statements such as yours about the |68020 'blowing the socks off the '286'... In case you hadn't been reading |the trade journals. AMD has recently been shipping 16Mhz versions of the |80286 which incidently contains, on the chip, memory management capabilities. |Something that the 68020 MUST go off chip for, namely, the 68851 MMU. I |believe most net readers would appreciate documented numbers, BEFORE you |make any more of these irritable BLIND statements. From IEEE Micro, June 1987, "Evaluating Microprocessors" by John C. McCallum and Tat-Seng Chua. From a table on page 74 (edited to show only 286 and the 020): 286/287 M68020 16 Mhz 16 MhZ Scientific 0.14 0.43 Commerical 0.48 1.44 Systems 0.70 1.50 General 0.31 0.93 Numbers are in a measurement unit call pMIPS (millions of psudeo-instruction per seconds). See the article for details of the psudeo-instructions used, but the basic idea is they are common macro-level instructions for various common operators. One problem is that they are all memory-to-memory, which hurts the '020 compared to the 286 because of the extra register resources in the 020. To repeat, the 020 BLOWS THE SOCKS OFF the 286 by a factor of 2-1 (at least). | |/Carl |...decvax!mcnc!unnccvax!cbenda Jerry Whitnell It's a damn poor mind that can only Communication Solutions, Inc. think of one way to spell a word. -- Andrew Jackson
lalonde@nicmad.UUCP (John Lalonde) (08/25/87)
In article <3709@sdcsvax.UCSD.EDU> jww@sdcsvax.UCSD.EDU (Joel West) writes: >The Sun-3/260 uses a 25 MHz 68020. I don't believe there is a part >between the 16 MHz and 25 Mhz models. The Sun 3/60 uses a 20 MHz 68020. 20MHz is a standard clock frequency for all of Motorola's Gang of Three (68020,68881,68851). The clock frequencies available for the '020 family are 12.5 MHz, 16.67 MHz, 20 MHz, and 25 MHz. -- John LaLonde Systems Engineering Group Nicolet Instrument Corporation uucp: {ihnp4,seismo,decvax,harvard}!uwvax!nicmad!lalonde
jwhitnel@csib.UUCP (Jerry Whitnell) (08/25/87)
In article <1178@csib.UUCP> jwhitnel@csib.UUCP (Jerry Whitnell) writes: In article <799@unccvax.UUCP| cbenda@unccvax.UUCP (carl m benda) writes: |In article <116@faccs.UUCP>, dbt@faccs.UUCP (Data base translator ) writes: |> Looking at it this way, the '020 not only compares well with the '386; it |> blows the socks off its Intel counterpart, the '286. And the '030? I can't |> wait. | |It really angers many people to see blind statements such as yours about the |68020 'blowing the socks off the '286'... In case you hadn't been reading |the trade journals. AMD has recently been shipping 16Mhz versions of the |80286 which incidently contains, on the chip, memory management capabilities. |Something that the 68020 MUST go off chip for, namely, the 68851 MMU. I |believe most net readers would appreciate documented numbers, BEFORE you |make any more of these irritable BLIND statements. Here is another set of numbers also showing the 020 is faster then a 286 and a 386. From IEEE Micro Vol 6 No. 4. "A Benchmark Comparison of 32-bit Microprocessors." These are based on the EDN 16-bit bench marks, modified for 32 bit operation. The table is copied from Table 3 in the article and includes only the ratio column for dynamic memory. The 68000 was measured at 8 Mhz and the 286 at 10 Mhz, so I multiplied by the appropiate ratios to get numbers for 16 Mhz. The ratios are the time processor X took divided by the time a 68020 took. So a 286 to 2.45 times as long to complete the tests as did the 020 at 16 Mhz. 80286 2.45 68000 2.31 80386 1.79 All numbers at 16 Mhz 68020 N 1.96 68020 C 1.00 The N and C are No cache and Cache. Conclusion: The 020 blows the socks of the 286 and does a good job aginst the 386 as well. You milage may differ. | |/Carl |...decvax!mcnc!unnccvax!cbenda Jerry Whitnell It's a damn poor mind that can only Communication Solutions, Inc. think of one way to spell a word. -- Andrew Jackson
alan@pdn.UUCP (Alan Lovejoy) (08/26/87)
In article <3725@cit-vax.Caltech.Edu> sho@tybalt.caltech.edu.UUCP (Sho Kuwamoto) writes: >In article <116@faccs.UUCP> dbt@faccs.UUCP (Data base translator ) writes: >Does anyone know what Apple plans to do with the 030, if anything? I >imagine third party 030 cards will become available for the MacII and >maybe the SE, but they may not utilize the full instruction set. The '030 differs from the '020 as follows: 1) It has a 256 byte data cache. 2) It has a SUBSET of the 68851 MMU on-chip. 3) It can do burst-mode memory transfers. 4) It can make more efficient use of static colum ram. 5) It takes only one clock cycle to access the instruction and/or data caches (instead of 2). 6) It takes only 2 clock cycles to access ram (instead of 3). 7) It has two 32-bit buses, internally and externally, one for addresses and one for data, that operate in parallel ("Harvard Architecture"). Note that only #2 makes any difference to software, and that Apple is going in the direction of making the 68851 standard equipment anyway. There is almost nothing that needs to be done to '020 software in order to take full advantage of the '030. This is Motorola's most compatible CPU yet. Alan "If President Reagan is telling the truth, why do they have to keep cutting off part of his nose?" Lovejoy
breck@aimt.UUCP (Robert Breckinridge Beatie) (08/26/87)
In article <1178@csib.UUCP>, jwhitnel@csib.UUCP (Jerry Whitnell) writes: > From IEEE Micro, June 1987, "Evaluating Microprocessors" by John C. McCallum > and Tat-Seng Chua. From a table on page 74 (edited to show only 286 and the > 020): > > 286/287 M68020 > 16 Mhz 16 MhZ > Scientific 0.14 0.43 > Commerical 0.48 1.44 > Systems 0.70 1.50 > General 0.31 0.93 > I notice that the 286 column is labeled as 286/287. Does this mean that the 286 also had a numeric co-processor? Since the 68020 column isn't labeled 68020/68881 does that mean that the 020 did NOT have a numeric co-processor? If so, then this is amazing. The 020 without floating point support beats a 286 with floating point support at SCIENTIFIC applications and does it by better than 3-1!!! This seems too good to be true. Could some kind soul post either confirmation or rufutation of my assumptions. I don't have easy access to the article in question. Even if the 020 did have a numeric co-processor, it looks as if, "yes Virginia the 020 does beat the pants of the 286." -- Breck Beatie uunet!aimt!breck
alan@pdn.UUCP (Alan Lovejoy) (08/28/87)
In article <509@sbcs.UUCP> root@sbcs.UUCP writes: >As for showing up in new industry iron, I thought Sun went with their >Sparc RISC chip (with promises to upgrade the existing Sun-3 line to >68030 "sometime in the future"), and Apollo was going AMD29000 - >thus just jumping ahead to Supercomputing workstations rather than >bothering about the next incremental speed change in 680?0. Ahem. 68020 * 4 is an "incremental" speed change? The 68030 at 30 Mhz more than matches the SPARC/29000/MIPS crowd in actual performance, and I wouldn't want to bet against Motorola on the 68040, which will probably be out before the '486 (Intel claims 1990 as the ETA for this). Also, my sources tell me that "first silicon" of the 68030 was last April, and general availability at 16 Mhz is 4th Qrtr 87, about the same time Intel finally gets to production of the '386 at 20 Mhz. Given the history of CPU clock speeds, it's a good bet Motorola will be producing 30 Mhz '030's before Intel produces a single 30 Mhz 'x86. Motorola can now achieve significant performance improvements in the '030 simply by increasing the cache sizes and translation lookaside buffer size--a rather minimal design change. Don't be surprised if a "68031" (or whatever) is announced in about a year. --Alan@pdn
stiber@CS.UCLA.EDU (08/28/87)
As far as I know, Intel is still having problems with bugs in their
80386s. Does anyone out there know if they have fixed chips in
production quantities? If not, Intel might not have much of a time
advantage over Motorola (assuming Moto. avoids such a faux pas).
"Faith is good, but skepticism is better"
- Giuseppe Verdi
Michael Stiber
Email: US Mail:
stiber@cs.ucla.edu UCLA Computer Science Dept.
...{ucbvax,ihpn4}!ucla-cs!stiber Machine Perception Laboratory
3564 Boelter Hall
Los Angeles, CA 90024jwhitnel@csib.UUCP (Jerry Whitnell) (08/28/87)
In article <80@aimt.UUCP| breck@aimt.UUCP (Robert Breckinridge Beatie) writes: |In article <1178@csib.UUCP>, jwhitnel@csib.UUCP (Jerry Whitnell) writes: | |> 286/287 M68020 |> 16 Mhz 16 MhZ |> Scientific 0.14 0.43 |> Commerical 0.48 1.44 |> Systems 0.70 1.50 |> General 0.31 0.93 |> | |I notice that the 286 column is labeled as 286/287. Does this mean that |the 286 also had a numeric co-processor? Since the 68020 column isn't |labeled 68020/68881 does that mean that the 020 did NOT have a numeric |co-processor? If so, then this is amazing. The 020 without floating point |support beats a 286 with floating point support at SCIENTIFIC applications |and does it by better than 3-1!!! This seems too good to be true. Could |some kind soul post either confirmation or rufutation of my assumptions. I |don't have easy access to the article in question. According to the instruction sequences given in the article, the 80286 did have a 287, whereas floating point for the 020 says "estimated from Motorola software". The 020 is actually slower doing floating point then the 287 (big surpise :-)), but the instruction mixes choosen are such that floatting point does not have a strong influence. For the Scientific mix, it is < 15%, for the others, all less the 3%. | |Even if the 020 did have a numeric co-processor, it looks as if, "yes Virginia |the 020 does beat the pants of the 286." |-- |Breck Beatie |uunet!aimt!breck Jerry Whitnell It's a damn poor mind that can only Communication Solutions, Inc. think of one way to spell a word. -- Andrew Jackson
jjg@walden.UUCP (John Grana) (09/05/87)
In article <3728@cit-vax.Caltech.Edu> chi@tybalt.caltech.edu (Curt Hagenlocher) writes: > >...which has a full 32 bit bus. The fact that the 80286 can be obtained in a >16Mhz version doesn't change these facts. (I believe that the 68020 is avail- >able in a 20- or even 24- Mhz model, if it's clock speed you want to compare.) >It's pretty senseless comparing the two because of this, even though the >68020 came out only shortly after the Intel chip. > How true. We (Perf. Tech.) recently completed a project that involved using 4 68020's running at 25 Mhz, 0 wait states. We used very fast static ram (each 020 has a 64K fast buffer, all 4 'share' a slower 8K buffer and all 4 have access to a 32 bit VMEbus and 64 bit VSBbus. The first time I ran the EDN benchmarks (against an older 16 Mhz 68020 design) I had to re-check the code/timings! john grana ..!rochester!cci632!walden!jjg
barnett@vdsvax.steinmetz.UUCP (Bruce G Barnett) (09/10/87)
In article <1179@pdn.UUCP> alan@pdn.UUCP (0000-Alan Lovejoy) writes: |In article <509@sbcs.UUCP> root@sbcs.UUCP writes: |>As for showing up in new industry iron, I thought Sun went with their |>Sparc RISC chip (with promises to upgrade the existing Sun-3 line to |>68030 "sometime in the future"), ... |>thus just jumping ahead to Supercomputing workstations rather than |>bothering about the next incremental speed change in 680?0. | |Ahem. 68020 * 4 is an "incremental" speed change? The 68030 at 30 Mhz |more than matches the SPARC/29000/MIPS crowd in actual performance, ^^^^^^^ ^^^^^^ My figures are: 68020 @ 25MHz = 4 MIPS (Sun 3/200) SPARC = 10 MIPS (Sun 4) I believe you are saying: 68030 @ 30MHz = 16 MIPS ??? I can only assume you literally have a 68031 chip NOW!? No fair jumping generations in a comparison. Otherwise I would like you to show documentation that the 68030 is a 16 MIP CPU. -- Bruce G. Barnett <barnett@ge-crd.ARPA> <barnett@steinmetz.UUCP> uunet!steinmetz!barnett
dsc@izimbra.CSS.GOV (David S. Comay) (09/11/87)
In article <2508@vdsvax.steinmetz.UUCP> barnett@steinmetz.UUCP (Bruce G Barnett) writes: >My figures are: > > 68020 @ 25MHz = 4 MIPS (Sun 3/200) > SPARC = 10 MIPS (Sun 4) check the recent articles in comp.arch for figures that dispute sun's ten mips claim for the sun 4. it actually is more like six to seven mips on real programs. also, although the comparisions between the low end 68020 workstations are interesting, they seem to ignore some of my concerns: 1. i want to run affordable software in a windowing environment. there are few programs in that class available on the sun or for that matter, the apollo. there are quite a few programs including pagemaker, more, ready set go, etc that are extremely useful and resonably affordable (at least compared to the per cpu charges for things like interleaf) available today on the macintosh ii. 2. i'm not personally interested in running aux on a macintosh ii. if i wanted a decent inexpensive unix box, then i would consider something like a 3/50 or a 3/60 or a 80386 based system or perhaps a 3b1 (hmm, i did say decent, right? :-) if i purchased a macintosh ii, it's because i want to run macintosh software and not unix software. 3. the macintosh ii is expandable while the sun 3/50 is not. if you want to compare hardware and their costs, a comparision between a sun 3/160 and a macintosh ii would be more useful. disclaimer: i like unix, i've used unix for about eight years and it has many strong points. i like the macintosh, i've used it for a couple of years and it also has many strong points. there are places in my work for both of them. i don't own stock in either company and am more certainly not a spokesperson for anyone (except perhaps for the levi stubbs' appreciation society. :-) sigh, dsc `see across the view. see the sky ripped open. see the rain through the gaping wound, pounding the women and children who run into the arms of america'
phssra@emory.UUCP (09/12/87)
In article <46807@seismo.CSS.GOV> dsc@izimbra.CSS.GOV (David S. Comay) writes: >In article <2508@vdsvax.steinmetz.UUCP> barnett@steinmetz.UUCP (Bruce G Barnett) writes: > > 1. i want to run affordable software in a windowing > environment. there are few programs in that class available on > the sun or for that matter, the apollo. there are quite a few > programs including pagemaker, more, ready set go, etc that are > extremely useful and resonably affordable (at least compared to > the per cpu charges for things like interleaf) available today > on the macintosh ii. This is certainly true, but that will hopefully change as more programs are developed with interfaces for UNIX-based window systems such as X and NeWS (e.g. TeX). The type of software one wishes to run is definitely an important factor. For more traditional programming, though, you have to keep in mind that UNIX boxes like Sun workstations come with a lot of standard software, such as editors, compilers, and debuggers. Buying these for the Mac adds up very quickly. > 2. i'm not personally interested in running aux on a macintosh > ii. if i wanted a decent inexpensive unix box, then i would > consider something like a 3/50 or a 3/60 or a 80386 based > system or perhaps a 3b1 (hmm, i did say decent, right? :-) if i > purchased a macintosh ii, it's because i want to run macintosh > software and not unix software. I agree; if you equip a Mac II with all of the standard equipment that comes with a Sun 3/52 (MMU, large disk, large screen, tape drive, ethernet, UNIX), you will end up paying a fair amount more for the Mac. > > 3. the macintosh ii is expandable while the sun 3/50 is not. > if you want to compare hardware and their costs, a comparision > between a sun 3/160 and a macintosh ii would be more useful. This depends on whether you need to expand it or not. Many people will be satisfied with ethernet, disk and tape drive as the only peripherals. One other advantage of the Mac II is that it provides color relatively cheaply (it is unavailable on the low-end Sun 3/50). * Scott Anderson * ** gatech!emoryu1!phssra * * * ** phssra@emoryu2.{bitnet,csnet} * * * * * ** * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
alan@pdn.UUCP (09/12/87)
In article <2508@vdsvax.steinmetz.UUCP> barnett@steinmetz.UUCP (Bruce G Barnett) writes:
/My figures are:
/
/ 68020 @ 25MHz = 4 MIPS (Sun 3/200)
/ SPARC = 10 MIPS (Sun 4)
/
/I believe you are saying:
/
/ 68030 @ 30MHz = 16 MIPS ???
/
/I can only assume you literally have a 68031 chip NOW!?
/No fair jumping generations in a comparison.
/
/Otherwise I would like you to show documentation that the 68030 is a
/16 MIP CPU.
First of all, ANY chip could be a "16 MIP" CPU: "MIP" means "Million
Instructions Per ...". Per what? Per Minute? Oh! You mean "Per Second!"
Well, that's MIPS. The VAX 780 is generaly rated as 1 MIPS (NOT
MIP!!!).
Secondly, Sun (and various other sources) like to quote MIPS figures
as if the machine in question were a VAX. VAXen get more work done
per instruction than most micropocessors (even m68030's) do. According
to Motorola, the 68020 at 25MHz does 5-6 native MIPS, and the 68030
is (according to Motorola) "at least twice as fast as a 68020" at the
same clock cycle speed. Therefore, at thirty MHz, it should run 4 times
as fast as a 68020 at 16 MHz -- provided the rest of the hardware in the
system with the '030 in it is designed to take advantage of the '030's
ability to do "burst mode" memory transfers, parallel address and data
bus activity, and the improved bus interface logic. Simply plopping
a '030 into a system designed for a '020 probably won't improve
things more than about 1.5x (at the same clock speed).
Several weeks ago, benchmark figures comparing the Sun-3 to the Sun-4
were posted in comp.sys.sun. These showed the Sun-4 to be about twice
the speed of the Sun-3 (25Mhz 68020). The Sun-3 is NOT the fastest
68020 system, by the way. Masscomp has a 68020 system that does almost 6000
Dhrystones/second (using the rather slow portable C compiler), while the
best the Sun-3 can do is somewhat over 4000.
Edge Computer has their own proprietary CMOS VLSI implementation of the
68010 that does 11 VAX-equivalent MIPS. The cost is comparable to what
the RISC companies are charging for their CPUs. (It's a board-level
product only that includes a generous amount of fast RAM and an operating
system license).
--alan@pdn