GUBBINS@RADC-TOPS20.ARPA (Gern) (05/15/85)
Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows away an IBM-AT (16-bit 80286 at 6 MHz). the progam: 5 TIME$="00:00:00" 10 FOR X=1 TO 10000 20 IF SQR(X)<>INT(SQR(X)) THEN 40 30 PRINT SQR(X), 40 NEXT X 45 PRINT TIME$ 50 END NOTE: This program also is an accuracy test, the correct results are the numbers from 1 to 100. The TRS-80, Apple, HP, C64 can't do it right (they miss about 20 numbers). The time for the 6 MHz IBM-AT is 1 min, 13 sec The time for the standard 5 MHz Z-100 is 1 min, 13 sec The time for a turboed 7.37 MHz Z-100 is 49 sec The time for the newly released 8MHz Z-100 would be better still. The Z-100 is faster, can do better color graphics, has a real keyboard, has a second 8-bit CPU for CP/M and is less than half the price of the IBM-AT. The cost of being better is not being competely IBM compatible. Cheers, Gern -------
sdyer@bbnccv.UUCP (Steve Dyer) (05/16/85)
> > Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows > away an IBM-AT (16-bit 80286 at 6 MHz). Were you using the same BASIC interpreter in both machines? The BASIC benchmark might be standard, but certainly not the interpreters. A number of PC clones seem to have slightly "tweaked" Microsoft BASICs relative to the PC. I think a "sieve" object module from a single compiler and run on both the Z100 and the PC/AT might give a better, less language dependent benchmark. That sounds like an interesting experiment given your promising, but hardly conclusive, results. -- /Steve Dyer {decvax,linus,ima,ihnp4}!bbncca!sdyer sdyer@bbnccv.ARPA
GUBBINS@RADC-TOPS20.ARPA (Gern) (05/16/85)
I wondered about that myself, but both BASICS are the same tweeked version of GW-BASIC. I still put it that if the IBM-AT is not allowed to use its 80287 (I don't know if it is a standard feature) or if the $200 8087 is put into the Z-100 - bringing them at software/hardware parity, that the stock Z-100 (5MHz) will run any program at the same speed (if not faster) than the IBM-AT. Also, a turboed Z-100 (with a Gernware or SDS 7.37MHz, a CDR 7.5MHz, or the Zenith newly released 8MHz Z-100 (or upgrade to an older Z-100)) will completely blow an IBM-AT away in speed, at half the cost with 640x512 pixels in 8 (colors - 36 if you tweek). I don't think an IBM-AT can beat that. Doing a little checking last night, yielded that the stock IBM-AT at 6MHz uses a memory wait state (so IBM can use cheap 5 MHz memory), so now how fast is the machine? Reports indicate that 1 Memory wait state slows down throughput about 20-30% depending on the application. The Z-100s don't have problems with their Winnies either... (-: I know I am taking vicious stabs at IBM, that one benchmark doesn't prove much, but I was surprised myself, as I thought that the AT WOULD be a lot faster than a Z-100. It is faster than a 4.77MHz IBM-PC. Cheers, Gern -------
ritzenth@bgsuvax.UUCP (ritzenth) (05/24/85)
After this short review, I've got to put my 2 cents worth in . . . GUBBINS@RADC-TOPS20.ARPA (Gern) wrote: >Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows >away an IBM-AT (16-bit 80286 at 6 MHz). > >the progam: > >5 TIME$="00:00:00" >10 FOR X=1 TO 10000 >20 IF SQR(X)<>INT(SQR(X)) THEN 40 >30 PRINT SQR(X), >40 NEXT X >45 PRINT TIME$ >50 END > >NOTE: This program also is an accuracy test, the correct results are >the numbers from 1 to 100. The TRS-80, Apple, HP, C64 can't do it >right (they miss about 20 numbers). > >The time for the 6 MHz IBM-AT is 1 min, 13 sec > >The time for the standard 5 MHz Z-100 is 1 min, 13 sec > >The time for a turboed 7.37 MHz Z-100 is 49 sec > >The time for the newly released 8MHz Z-100 would be better still. > >The Z-100 is faster, can do better color graphics, has a real keyboard, >has a second 8-bit CPU for CP/M and is less than half the price of the >IBM-AT. The cost of being better is not being competely IBM compatible. > >Cheers, >Gern O.K. You've spent the time to read the above, now for my 2 cents . . . I've been chided here for buying a not-so-IBM-Compatible, Sanyo 555-2. I've upgraded it so that it would be a little more compatible, adding the 256K and installing the new Video Board (VB). They now say that it is approx. 80% compatible. SO . . . how does it really compare to the IBM-PC? COMPUTER VERSION OUTPUT DIRECTED TO TIME ======== ======= ================== ==== IBM-PC 256K screen 3:19 Sanyo STANDARD screen 4:37 with Sanyo Basic Sanyo VIDEO BOARD screen 1:44 with GW-Basic Sanyo VIDEO BOARD printer 1:40 with GW-Basic Can anyone out there in net-land explain these results? How can a machine that is running on the same CPU, a supposedly slower clock, and approximately the same basic (at least all of the commands are the same) run twice as fast as it's competitor? The only difference between the 2 Sanyos is the Video Board (similar to the IBM PC video board). Also, why did the program run faster when the output was directed to the printer? When I studied operating systems, we were taught that the printer was the slowest of the output devices. Oh, yes, on BOTH machines, all 100 numbers were listed! THANKS Phil Ritzenthaler cbosgd!osu-eddie!bgsuvax!ritzenth
john@moncol.UUCP (John Ruschmeyer) (05/28/85)
>From: ritzenth@bgsuvax.UUCP (ritzenth) >Organization: Bowling Green State University, OH >Message-ID: <463@bgsuvax.UUCP> > >After this short review, I've got to put my 2 cents worth in . . . > >GUBBINS@RADC-TOPS20.ARPA (Gern) wrote: > >>Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows >>away an IBM-AT (16-bit 80286 at 6 MHz). >> >>the progam: >> >>5 TIME$="00:00:00" >>10 FOR X=1 TO 10000 >>20 IF SQR(X)<>INT(SQR(X)) THEN 40 >>30 PRINT SQR(X), >>40 NEXT X >>45 PRINT TIME$ >>50 END >> >>NOTE: This program also is an accuracy test, the correct results are >>the numbers from 1 to 100. The TRS-80, Apple, HP, C64 can't do it >>right (they miss about 20 numbers). >> >>The time for the 6 MHz IBM-AT is 1 min, 13 sec >>The time for the standard 5 MHz Z-100 is 1 min, 13 sec >>The time for a turboed 7.37 MHz Z-100 is 49 sec > >O.K. You've spent the time to read the above, now for my 2 cents . . . >I've been chided here for buying a not-so-IBM-Compatible, Sanyo 555-2. > >I've upgraded it so that it would be a little more compatible, adding the >256K and installing the new Video Board (VB). They now say that it is >approx. 80% compatible. SO . . . how does it really compare to the >IBM-PC? > >COMPUTER VERSION OUTPUT DIRECTED TO TIME >======== ======= ================== ==== >IBM-PC 256K screen 3:19 > >Sanyo VIDEO BOARD screen 1:44 > with GW-Basic > >Can anyone out there in net-land explain these results? How can a machine >that is running on the same CPU, a supposedly slower clock, and approximately >the same basic (at least all of the commands are the same) run twice as >fast as it's competitor? > >Oh, yes, on BOTH machines, all 100 numbers were listed! The above was edited a bit, but I have tried to maintain the gist. To the above results, let me add the following: COMPUTER VERSION OUTPUT DIRECTED TO TIME ======== ======= ================== ==== Tandy 1000 256k screen 1:38 AT&T 6300 640k screen 0:32 (8mhz 8086) At this point, the results seem to get too bizarre to be believable. What I fail to see is how the standard (I assume) IBM-PC can pull a time so far from that of the Tandy 1000 or Sanyo. If I read my spec sheets, the Tandy and the IBM are very close. (By the way, I ran the benchmark on a standard IBM-PC here and it really does take that long under BASICA.) The AT&T numbers seem to look like the "turboed" Z-100 numbers, but its very far from the PC AT numbers. Is there really that big a difference between an 8086 running at 8mhz and an 80286 running at 6mhz? -- Name: John Ruschmeyer US Mail: Monmouth College, W. Long Branch, NJ 07764 Phone: (201) 222-6600 x366 UUCP: ...!vax135!petsd!moncol!john ...!princeton!moncol!john ...!pesnta!moncol!john Silly Quote: I never wanted to be a barber. I wanted to be... a LUMBERJACK!
oaa@houxl.UUCP (O.ALEXANDER) (05/29/85)
>Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows >away an IBM-AT (16-bit 80286 at 6 MHz). > ... >NOTE: This program also is an accuracy test, the correct results are >the numbers from 1 to 100. The TRS-80, Apple, HP, C64 can't do it >right (they miss about 20 numbers). Their's can't, but my Atari 400 (1Mhz 6502) does it right. It just takes it's time (40:02)! Owen Alexander microsys!oaa
indra@utai.UUCP (Indra Laksono) (05/30/85)
ritzenth@bgsuvax wrote > > > >GUBBINS@RADC-TOPS20.ARPA (Gern) wrote: > > > >>Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows > >>away an IBM-AT (16-bit 80286 at 6 MHz). > >> > >>the progam: > >> > >>5 TIME$="00:00:00" > >>10 FOR X=1 TO 10000 > >>20 IF SQR(X)<>INT(SQR(X)) THEN 40 > >>30 PRINT SQR(X), > >>40 NEXT X > >>45 PRINT TIME$ > >>50 END > >> > >>NOTE: This program also is an accuracy test, the correct results are > >>the numbers from 1 to 100. The TRS-80, Apple, HP, C64 can't do it > >>right (they miss about 20 numbers). > >> > >>The time for the 6 MHz IBM-AT is 1 min, 13 sec > >>The time for the standard 5 MHz Z-100 is 1 min, 13 sec > >>The time for a turboed 7.37 MHz Z-100 is 49 sec > > > >O.K. You've spent the time to read the above, now for my 2 cents . . . > >I've been chided here for buying a not-so-IBM-Compatible, Sanyo 555-2. > > > >I've upgraded it so that it would be a little more compatible, adding the > >256K and installing the new Video Board (VB). They now say that it is > >approx. 80% compatible. SO . . . how does it really compare to the > >IBM-PC? > > > >COMPUTER VERSION OUTPUT DIRECTED TO TIME > >======== ======= ================== ==== > >IBM-PC 256K screen 3:19 > > > >Sanyo VIDEO BOARD screen 1:44 > > with GW-Basic > > > > > >Oh, yes, on BOTH machines, all 100 numbers were listed! > > The above was edited a bit, but I have tried to maintain the gist. To the > above results, let me add the following: > > COMPUTER VERSION OUTPUT DIRECTED TO TIME > ======== ======= ================== ==== > Tandy 1000 256k screen 1:38 > > AT&T 6300 640k screen 0:32 > (8mhz 8086) > *** REPLACE THIS LINE WITH SOMETHING FUNNY *** I really can't say about the Tandy 1000, but I do have a sanyo 555-2 with the video board. Ok, so the video is faster than the pc, but how do you account for all that snow in text programs ? The only exception to this is 1-2-3 which is as clear as,well, a monitor. dBase III, Supercalc, KEDIT snows so much, you'd think it's christmas. I have nothing to do with Sanyo other than that I happen to own a 555. {allegra cornell decvax ihnp4 linus utzoo}!utcsri!utai!indra
peterb@pbear.UUCP (05/30/85)
I just had to throw in my two cents worth:
Using an IBM PC/XT running PC/IX as its OS, the program to find those
numbers who square roots are integers ran with the following times:
Using the 8087 coprocessor:
Real 11.7 (seconds!)
User 4.7 (run time in program)
Sys 6.7 (since context switch unloads 8087 registers)
Using subroutines for sqrt and floor (instead of using 8087 instructions):
Real 54.4
User 45.1
Sys 7.0
This just shows that here you are testing basics against one another, and
the math implementations. Start doing benchmarks in assembler to get the hard
core speed operation out of them, and also lets you have an exact grasp
of what you are comparing.
Peter Barada
{ihnp4!inmet|{harvard|cca}!ima}!pbear!peterbgnome@olivee.UUCP (Gary Traveis) (06/01/85)
> > Doing a little checking last night, yielded that the stock IBM-AT at 6MHz > uses a memory wait state (so IBM can use cheap 5 MHz memory), so now > how fast is the machine? Reports indicate that 1 Memory wait state > slows down throughput about 20-30% depending on the application. > > The Z-100s don't have problems with their Winnies either... (-: > > I know I am taking vicious stabs at IBM, that one benchmark doesn't > prove much, but I was surprised myself, as I thought that the AT WOULD > be a lot faster than a Z-100. It is faster than a 4.77MHz IBM-PC. > > Cheers, > Gern > ------- Let's face it, when taking the PC-AT's price/performance into consideration, almost any similarly max-ed out system can run rings around the AT. "But when Cray comes out with a $7.95 pocket-sized unit with full graphics, they'll still ask 'Is it IBM compatible?'"
maxg@tekig4.UUCP (Max Guernsey) (06/06/85)
In article <327@moncol.UUCP> john@moncol.UUCP (John Ruschmeyer) writes: >>>Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows >>>away an IBM-AT (16-bit 80286 at 6 MHz). >>> >>>the progam: >>> >>>5 TIME$="00:00:00" >>>10 FOR X=1 TO 10000 >>>20 IF SQR(X)<>INT(SQR(X)) THEN 40 >>>30 PRINT SQR(X), >>>40 NEXT X >>>45 PRINT TIME$ >>>50 END ... >>>The time for the 6 MHz IBM-AT is 1 min, 13 sec ... >COMPUTER VERSION OUTPUT DIRECTED TO TIME >======== ======= ================== ==== ... >AT&T 6300 640k screen 0:32 >(8mhz 8086) Now for my $.02 worth. I have an AT at work and a 6300 at home and I tried out the benchmark just to see. It's true. The AT when booted with and unformatted C: and nothing in A: still "comes up" in "cassett basic" I believe. So the AT is probably still looking to ROM to find some of the routines. Would this explain the AT's turtle like behavior? Max Guernsey, jr.
maxg@tekig4.UUCP (Max Guernsey) (06/06/85)
In article <119@tekig4.UUCP> maxg@tekig4ernsey.UUCP (Max Guernsey) writes: >Now for my $.02 worth. I have an AT at work and a 6300 at home and I tried out the benchmark just to see for myself what happened. > >The AT when booted with an unformatted C: and nothing in A: still "comes up" >in "cassett basic", I believe. So the AT is probably still looking to ROM to | BASIC or BASICA on the >find some of the routines. > >Would this explain the AT's turtle like behavior? > Max Guernsey, jr. P.S. Does anybody have a good termcap that will work with the XTALK vt100 emulator.
maxg@tekig.UUCP (Max Guernsey) (06/08/85)
In article <327@moncol.UUCP> john@moncol.UUCP (John Ruschmeyer) writes: >>From: ritzenth@bgsuvax.UUCP (ritzenth) >>Organization: Bowling Green State University, OH >>Message-ID: <463@bgsuvax.UUCP> >> >>After this short review, I've got to put my 2 cents worth in . . . >> >>GUBBINS@RADC-TOPS20.ARPA (Gern) wrote: >> >>>Using a standard BASIC benchmark, the Z-100 (8/16-bit 8088) blows >>>away an IBM-AT (16-bit 80286 at 6 MHz). >>> >>>the progam: >>> >>>5 TIME$="00:00:00" >>>10 FOR X=1 TO 10000 >>>20 IF SQR(X)<>INT(SQR(X)) THEN 40 >>>30 PRINT SQR(X), >>>40 NEXT X >>>45 PRINT TIME$ >>>50 END >>> >>>NOTE: This program also is an accuracy test, the correct results are >>>the numbers from 1 to 100. The TRS-80, Apple, HP, C64 can't do it >>>right (they miss about 20 numbers). >>> >>>The time for the 6 MHz IBM-AT is 1 min, 13 sec >>>The time for the standard 5 MHz Z-100 is 1 min, 13 sec >>>The time for a turboed 7.37 MHz Z-100 is 49 sec >> >>O.K. You've spent the time to read the above, now for my 2 cents . . . >>I've been chided here for buying a not-so-IBM-Compatible, Sanyo 555-2. >> >>I've upgraded it so that it would be a little more compatible, adding the >>256K and installing the new Video Board (VB). They now say that it is >>approx. 80% compatible. SO . . . how does it really compare to the >>IBM-PC? >> >>COMPUTER VERSION OUTPUT DIRECTED TO TIME >>======== ======= ================== ==== >>IBM-PC 256K screen 3:19 >> >>Sanyo VIDEO BOARD screen 1:44 >> with GW-Basic >> >>Can anyone out there in net-land explain these results? How can a machine >>that is running on the same CPU, a supposedly slower clock, and approximately >>the same basic (at least all of the commands are the same) run twice as >>fast as it's competitor? >> >>Oh, yes, on BOTH machines, all 100 numbers were listed! > >The above was edited a bit, but I have tried to maintain the gist. To the >above results, let me add the following: > >COMPUTER VERSION OUTPUT DIRECTED TO TIME >======== ======= ================== ==== >Tandy 1000 256k screen 1:38 > >AT&T 6300 640k screen 0:32 >(8mhz 8086) > > >At this point, the results seem to get too bizarre to be believable. What I >fail to see is how the standard (I assume) IBM-PC can pull a time so far >from that of the Tandy 1000 or Sanyo. If I read my spec sheets, the Tandy >and the IBM are very close. (By the way, I ran the benchmark on a standard >IBM-PC here and it really does take that long under BASICA.) > >The AT&T numbers seem to look like the "turboed" Z-100 numbers, but its >very far from the PC AT numbers. Is there really that big a difference >between an 8086 running at 8mhz and an 80286 running at 6mhz? >-- >Name: John Ruschmeyer >US Mail: Monmouth College, W. Long Branch, NJ 07764 >Phone: (201) 222-6600 x366 >UUCP: ...!vax135!petsd!moncol!john ...!princeton!moncol!john > ...!pesnta!moncol!john >Silly Quote: > I never wanted to be a barber. > I wanted to be... a LUMBERJACK! Is it because the BASICA and BASIC on the AT still go to ROM for some of the subroutines? maxg@tekig5 PS has anybody got a good termcap for Crosstalk's vt100?
carlj@hp-pcd.UUCP (carlj) (06/08/85)
I benchmarked the same program on my Columbia (PClone) and measured
1:11. Are you sure than the quoted time was for an AT and not an
XT?
Carl Johnson
...!hplabs!hp-pcd!carlj