CompuServe Thread

#"Super-HAM' mode?

29 messages in this thread
#49500From: Ethan SolomitaMay 20, 1989 7:14 PM
Hold on. I think the most important consideration regarding a possible bus speed is not to make it 14.28 MHz, or any other arbitrary speed. If its possible, It would be best to make it flexible (I'm not an engineer). The complaints we have about the current bus is that it is too slow. In 2 years I ASSURE YOU that a 14 MHz bus will be laughable in a business situation. Make the bus flexible, of course having a maximum limit. Or maybe make a A3014, A3020, A3028 and an A3050, each having a different bus speed. That might be too complicated, but if the bus is limited to 14.28 MHz I fear the worst. — Ethan
#49577From: Kelly JordanMay 21, 1989 8:03 AM
Ethan, if you start going much higher than the 14.28 MHz, things start to get very expensive. Right now, the chip bus (What I was referring to in the previous message) uses 150 nS DRAMs. I don't know what the actual timing requirements are for the chip bus, but if we assume they are in the area of 160 nS, then if the chip bus speed is doubled, this 160 nS would be 80 nS, then to get any kind of leaway for timing slop, the DRAMs would have to be 65-70 nS devices. They aint cheap, and wuldn't be for years to come. And when you start looking at larger quanitites of chip ram (hopefully come standard with 2 megs, expandle to 16 megs), at a much higher cost, the base machine will have to be priced about the same as a MacII. Ethan, maybe you should look at what having a 32 bit chip bus at 14.32 Mhz will give you. You will have four times the bandwidth over the existing bus, so the chips should be able to read four times the data they do now in the same time. One of the new modes in the new chip set will be 640x400x2 non-interlace with some bus contention. With four times the bandwidth, this mode should be extrapolate to 640x400x8 (256 colours on screen) with the same degree of bus contention. (Thats 256k bytes) Or if you go back to the standard modes, ie. 640x200x4, with some contention, we could easily have 640x200x8 with no contention. (Actually, just going to a 32 bit bus would give you this mode with the same amount of contention) (I really doubt any new chip set would have <continued>
#49578From: Kelly JordanMay 21, 1989 8:03 AM
<continuation> more than 256 colour registers). And could we ever have some fine looking HAM modes! Ethan, if you want an expansion bus going faster than 14.32MHz, your really going to have to pay some big bucks for high speed memory boards! 73, Kelly
#49629From: Ethan SolomitaMay 21, 1989 5:15 PM
Is there any way to make the bus speed flexible, i.e. runs at the CPUs speed? Also, what is the bus speed of Mac IIs, IBM's MCA and Compaqs. How about the new EISA standard? — Ethan
#49894From: Kelly JordanMay 23, 1989 5:06 AM
Ethan, which bus? Chip bus – No, and Why? The best thing to do with the chip bus is to make it go as fast as possible at an affordable cost. Right now, to me, that would be 32 bits, with twice the clock speed we have now. Expansion Bus – No, and Why? You have to build/design the expansion bus to a certain speed. All the control circuitry on the cards must work at the maximum specified speed. If you have memory cards, you would want them to work at the maximum speed, otherwise they are slowing down the machine. Peripheral cards, like serial and parallel ports can run at a slower speed, and this is done with the DTACK control line for 68020/000 type peripheral devices. This is done when some part on the board can't do it's job in the normal bus access time, so this time is stretched a couple of clock cycles. An example of this is when using slow EPROMS. The highest bus speeds I have seen are in the 25MHz range for the Motorola chips. This is on a VME bus system, and this stuff aint cheap! If they made a machine that ran at these speeds, who would be able to afford it? You would probably be bad mouthing them for bringing a machine to market at a price nobody can afford. <continued>
#49895From: Kelly JordanMay 23, 1989 5:06 AM
<continuation> I don't know much about Macs, archaic technology, or clones (in that order). Which new standard is the EISA standard? There are more standards out there than we can shake a stick at! IMHO, with what little I have seen on the VME bus, it should be the basis for a high power Amiga system (25+MHz), modified to allow autoconfig, and what ever else is required for the Amiga specific stuff. This bus even allows true multiprocessing! Now if it just wasn't so darn expensive! 73, Kelly
#49924From: Black Belt SystemsMay 23, 1989 12:13 PM
Kelly, I have laid hands upon a 33 Mhz 68030 running with no waits whatsoever, and Motorola is promising another 10% or so before the next 1/4… I wouldn't go TOO far towards saying 4x 14 Mhz is near any current limits, and certainly not any future ones. Just a thought. –Ben-t(/ex
#50241From: Kelly JordanMay 25, 1989 4:53 AM
Ben, who is going to make an (affordable) expansion bus that will operate at those speeds? That's what I think the limit is. Actuall, if they want to go to CPUs of that speed range, then there should be a CPU board that has a capacity for up to 8 or 16 Meg of ram, with the MMU/FPU (I know, already integral to the 68040, just need the FPU for the 68030) all on the same board. Kind of like a super-duper A2620. Then, the rest of the machine can have a slower bus, 32 bits of course. Of course, if they are going to that extreme to get the speed, then the CPU board should be a two slot/piggy back affair, with the second board being a vary high speed custom chip set with memory for the video, etc. Hmmm…. A machine built like that might just be affordable! 73, Kelly
#49961From: Ethan SolomitaMay 23, 1989 6:18 PM
EISA is the new standard being produced by the Gang of Seven (Compaq, AST, Epson, AT&T, …) to compete with MCA. What is VME? — Ethan
#50102From: Rick StevensMay 24, 1989 2:59 AM
Ethan, VME is the bus used in various Sun workstations and was originally developed by Motorola for all 680×0-based systems. It started out as the "VersaBus". Later on, it acquired the name VME (I can't remember what it stands for) and has evolved into a more involved bus called VXI (VME eXtended for Instrumentation). Rick Stevens/MicroBotics
#50207From: Don Curtis/SYSOPMay 25, 1989 12:10 AM
Rick, Memory says that VME stands for "Virtual Machine Extension"…however memory is often faulty…so that could easily be wrong. Don
#50235From: Rick StevensMay 25, 1989 2:55 AM
Sounds good to me, Don. I really can't remember what it was. I seem to recall that it was something like "Versabus Microprocessor Extended", but that's probably wrong. (The memory is the second thing to go. I can't remember what the first is.) [O O] Rick "Duh-hhh!" Stevens ? V
#50712From: Michael St. LaurentMay 27, 1989 7:59 PM
Hmmm… How does that Bob Hope joke go? First you forget names Next you forget faces Then you forget to zip up Finally you forget to zip down! Michael St. Laurent
#50783From: Rick StevensMay 28, 1989 2:50 AM
Har!
#50243From: Kelly JordanMay 25, 1989 4:54 AM
Rick , don't correct me in my reply to Ethan! Your right, I forgot it started as "VersaBus", and this is the one the DIN standard is based on, and VME is Motorola's extansion to that. VXI for instrumentation, you say. Might have to check into that one. 73, Kelly
#50751From: Rick StevensMay 28, 1989 12:45 AM
Oops! Messed up there. Yup, VXI is (as Arte Johnson would say) "velly intelesting!" Rick Stevens/MicroBotics
#50364From: Ethan SolomitaMay 25, 1989 11:01 PM
Is VME the bus, then, that we use? Does your message imply that the same bus is on all 680×0 machines, including Macs and Suns? — Ethan
#50759From: Rick StevensMay 28, 1989 12:47 AM
I said that Motorola designed it with the intent that it would be the standard. It didn't become THE standard, just one of many. But, no, Macs and Amoebas use their own busses, but VME _IS_ used in Suns. Heck, the PC bus was to be THE bus standard for clones, but Toshiba, Zenith, Tandy and others came up with their own busses for clones. RS-232C is a standard, but darn few manufacturers use it properly. The same is true for whole bunches of standards. Perhaps it should be used thus in the computer world: "standard (i.e. kinda common)" Rick Stevens/MicroBotics
#50841From: Kevin DarlingMay 28, 1989 5:06 PM
I forgot which computer type said it, but 'twas something like: "Standards are wonderful. There are so many of them". He was moaning about the RS232 "standard" at the time, I think.
#51327From: Steve PietrowiczMay 31, 1989 6:44 AM
That "computer type" is Andrew Tannenbaum, and that was from his "Computer Networks" book. I like that quote too. 🙂 Steve
#50883From: Vic WagnerMay 28, 1989 8:09 PM
Rick, One should keep firmly in mind that the RS in RS-232C stands for "Recommended Standard".
#50242From: Kelly JordanMay 25, 1989 4:54 AM
Ethan, okay, you can keep the EISA standard, since it is a piece of archaic garbage, while MCA is just newer archaic garbage <GRIN>. VME came out of europe, and is a DIN standard, though motorola (and others) have added to the standards. Motorola is into it pretty big. The bus uses 96 pin DIN (not pcb) connectors, while Mororola's version of VME allows the use of two busses to be accessed by one board, one local, one system wide, or at least I think that is the way it works – not to sure. You will probably find these systems being used in reasearch, industry, probably a million other places. There are a number of multi-tasking O/S available for 680×0 VME systems. Like I said before, the only thing keeping me from having one here to fool with is the price! 73, Kelly
#50769From: Rick StevensMay 28, 1989 12:49 AM
Kelly, Sun Microsystems uses VME busses in the Sun 3/xxx line. Also in the SPARCStation, I think. (Larry?) Rick Stevens/MicroBotics
#50960From: John DraperMay 29, 1989 12:42 AM
Rick, Aside from the 3/50, 3/60, and SPARCStation 1, we use the VME bus, with a few extensions (private bus specs). -larry
#50036From: Ray LewisMay 23, 1989 10:18 PM
Kelly: EISA is an alternate bus system to microchannel being developed by a consortium of clone makers, led by Compaq. They have no faith in microchannel, and are leading and insurection against BigBlue. If successful, they hope to convince the business world that computing has grown up, and you can have Fords, Chevys, Cadillacs, Rolls Royces, Fiats, Sterlings, etc., that do not all have to be clones of the model T. Even tho they are still not at the advanced level of the Amiga, let's hope they succeed. We need to the world to understand that desk-top does not equal IBM/clone any more than Frigidaire = refrigerator. ?|^] Ray Lewis
#50244From: Kelly JordanMay 25, 1989 4:54 AM
Ray, I was aware of this group (groupies?), just didn't associate the EISA to them. Wouldn't you know it, 1.5 years ago I bought a Ford truck (4×4, Diesel), after being a died in the wool Chevy/GMC truck owner. Took a good look at those new GM's, and said no way. Actually the first time I saw the new GM trucks in a mall, I swore it was a compact! Rice boxes should be banned! 73, Kelly
#50101From: Rick StevensMay 24, 1989 2:58 AM
Kelly, ISA is "Industry Standard Archetecture" or the AT-bus. It was created by Compaq as an alternative to the MicroChannel bus used on PS/2s. Isn't it wonderful where one manufacturer can declare itself as a "standard" to which everyone else should conform? Especially a kludgy archetecture as either the AT or MicroChannel bus? [O O] Rick (washing mouth out with soap) Stevens ? V
#50245From: Kelly JordanMay 25, 1989 4:54 AM
It's only a standard if you recognize it as one. I don't, so to me, it doesn't exist <GRIN>! 73, Kelly
#50752From: Rick StevensMay 28, 1989 12:45 AM
Well, I recognize it a a standard. Not a good standard, but a standard. Ah, well. Rick Stevens/MicroBotics