rra202@uiucuxa.CSO.UIUC.EDU (09/25/86)
hi I'm from a Commodore - 128 kind of scene I ran a program on both the apple //e and a C-128 10 fort=1 to 500 20 a=3.14159 30 b=sqr(val(str$(a*a))) 40 nextt 50 end and the C-128 in the 80-col mode is almost 2x as fast by the way, the values of 'b' were almost identacle acually, the C-128 is almost (within 20%) as fast as an IBM PC.
wtm@neoucom.UUCP (Bill Mayhew) (09/27/86)
[Line eater, get me.] Hi, It's been a while since I was dabbling on a c128. I do remember, though, that under certain circumstances the cpu on the c128 is run at 2.04 MHz, rather than the "standard" 1.02 MHz. Don't quote me on this, but I think it was durring the CRT retrace times that the cpu is running faster on the c128. Naturally, Commodore wasn't first to cook the idea up. Remeber the ill fated Apple III ran its cpu faster in the CRT retrace time. Comparing the c128 and apple II against the IBM pc is interesting, as in the normal precision mode, the IBM basica only carries 6 -7 digits of accuracy while apple and commodore carry typically 10-11 digits. Scrolling text on the 80 col screen is where the speed of the apple IIe is really sapped since the poor cpu has to do handstands to move the text around in the oddball memory map. (I wrote a subroutine in machine language to scroll down to augment the built in code that scrolls up.) I've always wondered why a decent text memory map wasn't designed in to the Apple II at the outset, as even in 1976, it would have added less than $5 to the cost of the components in the machine. I saw a review of the new //GS on TV, and they were loading an image, and it came up with the familiar fill-in-between the lines fashion of the II. Perhaps this was just done for "effect" to disolve the screen. I sure was hoping the GS would have a mode with a contiguous bit map to make manipulations easier (and faster). Perhaps, the 128K quickdraw ROM in the GS will be a relief, and I won't have to do the dogwork of figuring out how to make the hardware dance nicely. --Bill Bill Mayhew Division of Basic Medical Sciences Northeastern Ohio Universities' College of Medicine Rootstown, Ohio 44272 USA 216-325-2511 (wtm@neoucom.UUCP)
steve@jplgodo.UUCP (Steve Schlaifer x43171 301/167) (09/29/86)
In article <284@neoucom.UUCP>, wtm@neoucom.UUCP (Bill Mayhew) writes: > > ...... > > I saw a review of the new //GS on TV, and they were loading an > image, and it came up with the familiar fill-in-between the lines > fashion of the II. Perhaps this was just done for "effect" to > disolve the screen. I sure was hoping the GS would have a mode > with a contiguous bit map to make manipulations easier (and > faster). Perhaps, the 128K quickdraw ROM in the GS will be a > relief, and I won't have to do the dogwork of figuring out how to > make the hardware dance nicely. I did not see the demo mentioned but, when running in ][ emulation mode, the //GS must emulate the bizarre mapping of the graphics onto the screen. When running in native mode, however, the super-high-res graphics is linearly mapped so that the next line on the screen is a lineworth of bits later in memory just like one would like. On the other hand, the QuickDraw II in ROM is reasonably fast for many purposes and unloads the programmer to do bigger and better things. -- ...smeagol\ Steve Schlaifer ......wlbr->!jplgodo!steve Advance Projects Group, Jet Propulsion Labs ....logico/ 4800 Oak Grove Drive, M/S 301/165F Pasadena, California, 91109 +1 818 354 3171
ranger@ecsvax.UUCP (Rick N. Fincher) (09/30/86)
> > hi > I'm from a Commodore - 128 kind of scene > I ran a program on both the apple //e > and a C-128 > > 10 fort=1 to 500 > 20 a=3.14159 > 30 b=sqr(val(str$(a*a))) > 40 nextt > 50 end > > and the C-128 in the 80-col mode is almost 2x as fast > by the way, the values of 'b' were almost identacle > acually, the C-128 is almost (within 20%) as fast as > an IBM PC. That makes sense because the C128 runs at 2mhz compared to 1 mhz standard for the apple //. Drop in an accelerator however and the Apple runs twice as fast as an IBM or the c128. Where do you plug in an accelerator on a C128 and why would you want to run bad software twice as fast? The benchmark you list is not totally hardware dependent. The quality and type of Basic will make a difference too. Rick
nakada@husc4.harvard.edu (paul nakada) (10/01/86)
In article <2084@ecsvax.UUCP> ranger@ecsvax.UUCP (Rick N. Fincher) writes: >> >> hi >> I'm from a Commodore - 128 kind of scene >> I ran a program on both the apple //e >> and a C-128 >> [ a program and other stuff ] > >That makes sense because the C128 runs at 2mhz compared to 1 mhz standard for >the apple //. Drop in an accelerator however and the Apple runs twice as fast >as an IBM or the c128. Where do you plug in an accelerator on a C128 and >why would you want to run bad software twice as fast? The benchmark you >list is not totally hardware dependent. The quality and type of Basic >will make a difference too. > >Rick isn't it true that although the // series microprocessor runs at 1mhz, the computer runs at an apparent speed of 0.5mhz because of the video/computer division of time.. Somehow, I don't really think the microprocessor spped is quite as important as the apparent or visible speed... Here, the Apple, usually through excellent programming, makes up much ground on the PC series.. Anyone else out there feel the same? Paul Nakada
daveh@cbmvax.cbm.UUCP (Dave Haynie) (10/02/86)
> Summary: cpu clock speed faster in c128 > > [Line eater, get me.] > > It's been a while since I was dabbling on a c128. I do remember, > though, that under certain circumstances the cpu on the c128 is > run at 2.04 MHz, rather than the "standard" 1.02 MHz. Don't quote me > on this, but I think it was durring the CRT retrace times that the > cpu is running faster on the c128. Its been awhile since I dabbled on the C128 too. But I was one of the design engineers on the thing. The 1.02 Mhz speed, same as the C64, is what you get if you want to use both displays on the '128 (40 and 80 column, which are distinct video controllers). The 40 column display chip, when active, will alos steal a few cycles from the processor for character data fetches and sprite display. However, if the 40 column display is not needed, the '128 will run at 2.04 MHz nearly all the time, and the 80 column display chip will function as normal. The 2.04 MHz clock will slow down for two things. RAM refresh is one of these, though its pretty infrequent. The other slowdown is for any access of an I/O chip. You'll get a slowdown statistically 50% of the time you access an I/O chip. The reason behind this is that some of the I/O chips require the 1.02 MHz clock as a time base. Thus, if the I/O request occurs out of sync with the 1.02 MHz clock, it must be stretched to sync. If the access is in sync, no stretch is required. And there are no video DMA's to slow the processor down in this mode; the 80 column chip manages its own memory. > > Bill Mayhew > Division of Basic Medical Sciences > Northeastern Ohio Universities' College of Medicine > Rootstown, Ohio 44272 USA 216-325-2511 > (wtm@neoucom.UUCP) -- ============================================================================ Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh These opinions are my own, though if you try them out, and decide that you really like them, a small donation would be appreciated.
ranger@ecsvax.UUCP (Rick N. Fincher) (10/02/86)
> [Line eater, get me.] > > Hi, > > > I saw a review of the new //GS on TV, and they were loading an > image, and it came up with the familiar fill-in-between the lines > fashion of the II. Perhaps this was just done for "effect" to > disolve the screen. I sure was hoping the GS would have a mode > with a contiguous bit map to make manipulations easier (and > faster). Perhaps, the 128K quickdraw ROM in the GS will be a > relief, and I won't have to do the dogwork of figuring out how to > make the hardware dance nicely. > > --Bill > > Bill Mayhew > Division of Basic Medical Sciences > Northeastern Ohio Universities' College of Medicine > Rootstown, Ohio 44272 USA 216-325-2511 > (wtm@neoucom.UUCP) The new graphics modes on the //gs have 32K of contiguous screen memory with 16 color tables. You won't have to do handstands quite so much for them. Rick
daveh@cbmvax.cbm.UUCP (Dave Haynie) (10/06/86)
> >> I'm from a Commodore - 128 kind of scene I ran a program on both the >> apple //e and a C-128... >> >> 10 fort=1 to 500 >> 20 a=3.14159 >> 30 b=sqr(val(str$(a*a))) >> 40 nextt >> 50 end >> >> and the C-128 in the 80-col mode is almost 2x as fast >> by the way, the values of 'b' were almost identacle >> acually, the C-128 is almost (within 20%) as fast as >> an IBM PC. > > That makes sense because the C128 runs at 2mhz compared to 1 mhz standard for > the apple //. Drop in an accelerator however and the Apple runs twice as fast > as an IBM or the c128. Can you get a 4.08 MHz accelerator for the Apple (which is what you'd need for twice the speed of the C128)? Last I heard no one's delivering 6502s in production quantity that are speced to run above 3MHz. > The benchmark you list is not totally hardware dependent. The quality and > type of Basic will make a difference too. Exactly. The C128 vs. Apple comparison is very valid. You'll have to use Applesoft BASIC (floating point). This and the C128 BASIC both use the old Microsoft floating point routines; the operations involved will be just about identical. If the BASIC on the PC is Microsoft than the comparison is very valid; if you're using BASICA on the PC, its not, for the BASICA interpreter uses 7 digit floats verses the Microsoft 11 digit floats. > Rick -- ============================================================================ Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh These opinions are my own, though if you try them out, and decide that you really like them, a small donation would be appreciated.
ranger@ecsvax.UUCP (Rick N. Fincher) (10/08/86)
> In article <2084@ecsvax.UUCP> ranger@ecsvax.UUCP (Rick N. Fincher) writes: > >> > > isn't it true that although the // series microprocessor runs at 1mhz, the > computer runs at an apparent speed of 0.5mhz because of the video/computer > division of time.. > > Somehow, I don't really think the microprocessor spped is quite as important > as the apparent or visible speed... Here, the Apple, usually through > excellent programming, makes up much ground on the PC series.. > Anyone else out there feel the same? > > Paul Nakada Paul, memory refresh and video sharing does slow down the system from its full clock speed, but it doesn't cut it in half. I vaguely remember the ratio being 1 cycle in 50 or 60 being stretched. You may be thinkin - ng of the read cycles vs the write cycles, each only gets control of the bus for half the time. I agree, clock speed is not the only or even the best benchmark of performance. Even the best machine in the world world is useless without good software. Rick
ranger@ecsvax.UUCP (Rick N. Fincher) (10/09/86)
> > Can you get a 4.08 MHz accelerator for the Apple (which is what you'd need > for twice the speed of the C128)? Last I heard no one's delivering 6502s in > production quantity that are speced to run above 3MHz. Most accelerators for the Apple // run at 3.6 mhz because it is a di n even division of the 14mhz system clock. The 65C02 used is actually a 4mhz chip (available from Rockwell, GTE or WDC). I think that the Western Design Center is getting production yields on 6mhz 65802's which are pin compatible with the 65C02 and 65816's which are used in the //gs. > > > The benchmark you list is not totally hardware dependent. The quality and > > type of Basic will make a difference too. > > Exactly. The C128 vs. Apple comparison is very valid. You'll have to use > Applesoft BASIC (floating point). This and the C128 BASIC both use the > old Microsoft floating point routines; the operations involved will be just > about identical. If the BASIC on the PC is Microsoft than the comparison > is very valid; if you're using BASICA on the PC, its not, for the BASICA > interpreter uses 7 digit floats verses the Microsoft 11 digit floats. > > > Rick > -- > ============================================================================ > Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh > > These opinions are my own, though if you try them out, and decide > that you really like them, a small donation would be appreciated.
daveh@cbmvax.cbm.UUCP (Dave Haynie) (10/13/86)
> >> >> Can you get a 4.08 MHz accelerator for the Apple (which is what you'd need >> for twice the speed of the C128)? Last I heard no one's delivering 6502s in >> production quantity that are speced to run above 3MHz. > Most accelerators for the Apple // run at 3.6 mhz because it is a di n > even division of the 14mhz system clock. The 65C02 used is actually > a 4mhz chip (available from Rockwell, GTE or WDC). I think that the > Western Design Center is getting production yields on 6mhz 65802's > which are pin compatible with the 65C02 and 65816's which are used > in the //gs. Not quite. First of all, WDC doesn't make production silicon at all, I think they may even farm out much of their prototype work. GTE makes all of the 65SC816 parts; they've finally got the 4MHz part in reasonable quantities, though they're still expensive. I haven't heard anything positive on anything faster from them. Also, GTE only has 3MHz 65C02's in production quantities. They can get you 4MHz parts, but these are hand picked 3MHz parts. They're redesigning their 65C02 for real 4MHz operation, and expect to have that ready by early '87. WDC still doesn't have production quantities of any silicon product. I'm not sure how fast Rockwell parts are working (of course, they only make 65C02 and 6502), though Rockwell is in the process of de-emphasizing their 6502 lines in favor of 16 bit and greater chips (much to the approval of GTE, who's apparently firmly comitted to the 65C02 families as long as they're selling as well as they are). > Rick -- ============================================================================ Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh "Techno-Hippie, heathen, designing evil computers" These opinions are my own, though if you try them out, and decide that you really like them, a small donation would be appreciated.
ranger@ecsvax.UUCP (Rick N. Fincher) (10/16/86)
> > > >> > >> Can you get a 4.08 MHz accelerator for the Apple (which is what you'd need > >> for twice the speed of the C128)? Last I heard no one's delivering 6502s in > >> production quantity that are speced to run above 3MHz. > > Most accelerators for the Apple // run at 3.6 mhz because it is a di n > > even division of the 14mhz system clock. The 65C02 used is actually > > a 4mhz chip (available from Rockwell, GTE or WDC). I think that the > > Western Design Center is getting production yields on 6mhz 65802's > > which are pin compatible with the 65C02 and 65816's which are used > > in the //gs. > > > Not quite. First of all, WDC doesn't make production silicon at all, I > think they may even farm out much of their prototype work. GTE makes all > of the 65SC816 parts; they've finally got the 4MHz part in reasonable > quantities, though they're still expensive. I haven't heard anything > positive on anything faster from them. Also, GTE only has 3MHz 65C02's > in production quantities. They can get you 4MHz parts, but these are > hand picked 3MHz parts. They're redesigning their 65C02 for real 4MHz > operation, and expect to have that ready by early '87. WDC still doesn't > have production quantities of any silicon product. I'm not sure how fast > Rockwell parts are working (of course, they only make 65C02 and 6502), > though Rockwell is in the process of de-emphasizing their 6502 lines in > favor of 16 bit and greater chips (much to the approval of GTE, who's > apparently firmly comitted to the 65C02 families as long as they're selling > as well as they are). > > > Rick > -- > ============================================================================ > Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh > "Techno-Hippie, heathen, designing evil computers" > > These opinions are my own, though if you try them out, and decide > that you really like them, a small donation would be appreciated. I burned out the 4mhz 65C02 in my accelerator over a year ago and bought a replacement off the shelf for less than $12 (single unit) it was a Rockwell chip but the supplier said GTE was available too at the same price, but he didn't have any in stock. Bill Mensch said more than a year ago that the parts WDC was shipping were all 4mhz (even when 1 or 2 mhz were ordered), I don't know who manufactured these parts, but you are right, WDC sends all of their fabrication work out. Rick
daveh@cbmvax.cbm.UUCP (Dave Haynie) (10/20/86)
> > I burned out the 4mhz 65C02 in my accelerator over a year ago and bought > a replacement off the shelf for less than $12 (single unit) it was a > Rockwell chip but the supplier said GTE was available too at the same > price, but he didn't have any in stock. Bill Mensch said more than a > year ago that the parts WDC was shipping were all 4mhz (even when 1 or > 2 mhz were ordered), I don't know who manufactured these parts, but you > are right, WDC sends all of their fabrication work out. > > Rick $12 is a hand-tested price. A production 2MHz part is less than $1.00 in cost, that might be as much as $2.00 in single quantities. Production 4MHz 65C816 chips were started at around $8.00 a piece in large quantity (8 MHz 68000s are around $5.00 in quantity), though that price will certainly drop over the next year. About two years ago we go some parts directly through WDC, 65C816 chips, and they were labeled "4MHz", but they'd only run at around 500 KHz. The 4 MHz parts I have from GTE run fine at 2.04MHz, and I see no reason why they couldn't run at 4MHz. But these haven't been available for that long. I wouldn't expect to see 6MHz parts for awhile yet, except for an occasional hand-tested 4MHz part; OK for hobbying and high priced 3rd party add-ons, but nothing a large manufacturer is going to base a production computer on. Also, a system running at 4MHz will barely be able to use DRAMs. A 6MHz system will have to use static RAM, which will significantly increase the system price; you pay much more for memory than you do for the CPU chip. -- ============================================================================ Dave Haynie {caip,ihnp4,allegra,seismo}!cbmvax!daveh "Techno-Hippie, heathen, designing evil computers" These opinions are my own, though if you try them out, and decide that you really like them, a small donation would be appreciated.