MSX vs Atari 800 Vs ZX spectrum/BBC nicro

[carlsson]

BASIC V2 in the VIC-20 and C64 actually is a cut down version of BASIC V4 in the CBM 4032/8032 series, not the old BASIC V2 in the CBM 2001/3032 series. Given that Microsoft implanted the infamous Easter egg in V4, which also can be seen in both other 6502 versions as well as 6809 versions of Microsoft BASIC, it is an indication that Commodore actually had Microsoft updating BASIC for them, just that they already had paid a flat fee. Whether updates were included in that fee or cost extra, I'm not sure but the key point is that Microsoft were not paid for BASIC per computer sold but a flat fee. Somehow that agreement only applied to 8-bit computers though so with the Commodore Amiga, they had to come up with a new agreement and include Microsoft on the credits, which also appears on the C128 for similar reasons I think.

 

So yes, Commodore did have BASIC updated at least up to 1980, just that they never introduced additional commands. BASIC V4 on the 4032/8032 has those commands for CATALOG, HEADER, DLOAD etc which may look pretty but frankly added only a little bit of value vs the additional ROM cost. To be able to display the directory on screen without losing the program in memory probably was the biggest feature to ask for, but then I know many other systems have trouble with doing that using built in commands. By loading third party routines that are resident in memory, it is no biggie.

[artrag]

About msx, do you know that msx computers were used on the Russian MIR space station ?

 

https://youtu.be/nLjtW-FwWM0?t=753

Edited May 6, 2017 by artrag

[Rybags]

Probably thanks to the tech embargo of the 80s... plenty of better candidates around so they probably took what they could easily get.

[emkay]

Probably thanks to the tech embargo of the 80s... plenty of better candidates around so they probably took what they could easily get.

Sorry, no. Depending on the used CPU, there only have been comparable machines.

Not all computers have been build aiming towards gamers with just a TV set at home. From a Developer's view, the MSX machines were very powerful, small and efficient, thus very handy ...

[artrag]

That Msx was the most robust solution for sure.

[JamesD]

That Msx was the most robust solution for sure.

That SONY would have certainly been more suited to the task than other off the shelf 8 bit systems Russia would have had access to.

Really, most 8 bit systems did not have a version suited to a task like that, they were aimed at the home market.

I think JVC had some nice professional models as well at that time.

Russia had PDP-11, Z80, and 8080 clones. I wonder why they chose the MSX over any of those.

Maybe a foreign Z80 machine was chosen for lower power usage.

 

The US space shuttle supposedly used 8086 based machines in the ground computers that performed tests (NASA is looking to buy spares in bulk), and OS-9 was used on space shuttle computers, which means they were based on the 6809. These would have been custom built computers rather than an off the shelf design.

OS-9 was supposedly still in use when the Challenger blew up in 1986, which is the year that SONY was released.

I have to wonder if anyone at NASA figured out the 6309 could be dropped in place of the 6809 and it could run faster and draw less power.

[carlsson]

How much less power would the onboard computer draw if it has a Hitachi CPU instead of the Motorola one? A few watts? Faster might not be that important. Validated and affirmed it is a working solution probably is, so it might be a far longer and more complex process than just a drop in CPU change.

[JamesD]

How much less power would the onboard computer draw if it has a Hitachi CPU instead of the Motorola one? A few watts? Faster might not be that important. Validated and affirmed it is a working solution probably is, so it might be a far longer and more complex process than just a drop in CPU change.

I don't know exactly how much less power would be involved, but if they have several computers on board it could add up.

The 6309 has a sleep mode, so on a system that waits for interrupts, it could save even more power than just by using a newer die process.

The machine could sleep till the next interrupt.

When combined with the faster speed, it could potentially finish what it's doing faster and spend more time asleep.

Also, with the added 16 and 32 bit support, the code might be smaller and they wouldn't need as much ROM/RAM space.

Or one CPU with a little more RAM could monitor more systems and require fewer computers.

 

Validation would definitely be a concern here.

Switching CPUs is usually a redo from scratch.

 

*edit*

Keep in mind more power requires more weight in batteries, solar panels, fuel for fuel cells, fuel to enter orbit....

Edited May 8, 2017 by JamesD

[Van]

How much less power would the onboard computer draw if it has a Hitachi CPU instead of the Motorola one? A few watts? Faster might not be that important. Validated and affirmed it is a working solution probably is, so it might be a far longer and more complex process than just a drop in CPU change.

Actually the shuttle had 4 identical systems, IBM AP-101 CPUs, running in parallel with a 5th AP-101 system comparing and error checking the results, Self -test could detect 95% of hardware and software errors and redundancy would handle the last 5%. So not an off the shelf Rotamola CPU: AP-101s are a custom build of the AP-1 Flight computer used on aircraft such as the B-52, tried and true designs. These builds had custom apt specific microcode and were largely built with a mix of MSI and LSI TTL chips. There were separate registers for the FP and Fixed point operations, 16b and 32b. Instruction were either 16b or 32b wide and the core memory was 104KW shared storage. The systems ran at 480K IPS. Basically a flying mainframe.

Once the software was certified and approved for manned flight it would take years to change it and re-certify.

Yogi

[JamesD]

Actually the shuttle had 4 identical systems, IBM AP-101 CPUs, running in parallel with a 5th AP-101 system comparing and error checking the results, Self -test could detect 95% of hardware and software errors and redundancy would handle the last 5%. So not an off the shelf Rotamola CPU: AP-101s are a custom build of the AP-1 Flight computer used on aircraft such as the B-52, tried and true designs. These builds had custom apt specific microcode and were largely built with a mix of MSI and LSI TTL chips. There were separate registers for the FP and Fixed point operations, 16b and 32b. Instruction were either 16b or 32b wide and the core memory was 104KW shared storage. The systems ran at 480K IPS. Basically a flying mainframe.

Once the software was certified and approved for manned flight it would take years to change it and re-certify.

Yogi

So the OS-9 story is bs even though I've read it in numerous places. Ah the internet.

Custom CPUs with floating point are certainly consistent with the space program's history.

[Van]

So the OS-9 story is bs even though I've read it in numerous places. Ah the internet.

Custom CPUs with floating point are certainly consistent with the space program's history.

Yea, if it's on the interweb it must be true. More then likely the OS-9 design shares so many similarities with the design of the Flight Computer's modular OS that assumptions were made.

 

NASA, IBM and Rockwell started development of the Pri. Avionics Software System, PASS, in the '70s. At the base is the Flight Computer OS, FCOS, and on top of that are the apps for navigation and control, guidance, systems management and such. Overall it was a priority-driven RTOS built to specific NASA requirements, so not really an off the shelf soft, but I'm sure the lessons learned were incorporated into commercial products.

 

It was written in HAL/S, a high level language for aerospace; only a small amount of the FCOS was written in Assembly. The development systems in the Software Production Facility at Johnson Space Flight Center at first were 5 IBM 360/75s and a 3 AP-101 system and software 6 DOF simulator. By the early '80s they switched to 2 IBM 3033 and kept the 360s as backup. Software Labs at Cambridge, Johnson, Goddard, Kennedy SFCs and Rockwell International were all networked into the SPF. There was at least a 2 year turn around to add functions and software fixes to go through the verification to deployment system.

 

By '83 NASA did allow HP-41C calculators and GRiD Compass 8086 laptops on board for non-critical tasks. In the '90s they updated the AP-101s to F series which used MC68000s and 256K Word semiconductor battery-backed RAM (the earlier AP-101s used ferro Core memories), and the programming language shifted to C. So they went from '60 era computers to late '80s tech with the upgrade. Maybe the newer AP specs COULD run OS-9 but to my knowledge the PASS was ported over to the newer HW.

Yogi

[JamesD]

Yea, if it's on the interweb it must be true. More then likely the OS-9 design shares so many similarities with the design of the Flight Computer's modular OS that assumptions were made.

 

NASA, IBM and Rockwell started development of the Pri. Avionics Software System, PASS, in the '70s. At the base is the Flight Computer OS, FCOS, and on top of that are the apps for navigation and control, guidance, systems management and such. Overall it was a priority-driven RTOS built to specific NASA requirements, so not really an off the shelf soft, but I'm sure the lessons learned were incorporated into commercial products.

 

It was written in HAL/S, a high level language for aerospace; only a small amount of the FCOS was written in Assembly. The development systems in the Software Production Facility at Johnson Space Flight Center at first were 5 IBM 360/75s and a 3 AP-101 system and software 6 DOF simulator. By the early '80s they switched to 2 IBM 3033 and kept the 360s as backup. Software Labs at Cambridge, Johnson, Goddard, Kennedy SFCs and Rockwell International were all networked into the SPF. There was at least a 2 year turn around to add functions and software fixes to go through the verification to deployment system.

 

By '83 NASA did allow HP-41C calculators and GRiD Compass 8086 laptops on board for non-critical tasks. In the '90s they updated the AP-101s to F series which used MC68000s and 256K Word semiconductor battery-backed RAM (the earlier AP-101s used ferro Core memories), and the programming language shifted to C. So they went from '60 era computers to late '80s tech with the upgrade. Maybe the newer AP specs COULD run OS-9 but to my knowledge the PASS was ported over to the newer HW.

Yogi

I suppose it's possible a 6809 was used in a non-flight computer role on some NASA mission, but it sounds like it could just be urban legend.

FWIW, if you put a large amount of external ROM filled with math tables on an 8 bit CPU, it could be faster than the Apollo guidance computer even without internal floating point math support. At 2 MHz it probably takes fewer cycles to calculate and set a page of ROM than the Apollo computer's floating point unit took for operations.

But that doesn't sound like a NASA solution does it? Why spend $1000 when you can spend $100,000,000?

A cheaper system might make more sense for something like Skylab, but that predated 8 bit microprocessors.

There was a version of OS-9 for the 68000, but it was called OS-9000.

 

That's not the only NASA stories I've read that might be urban legend.

Another involved a machine that logged data from a space probe using UCSD Pascal or maybe Apple Fortran (based on the same virtual machine), and data was lost when they exceeded the maximum number of files. I always thought that was running on something like a mainframe or PDP-11 though.

Yet another involved using an Amiga for something at NASA. I think that was supposedly another data logging system, and it was a cowboy solution to a problem rather than an official project.

NASA used the 8086 for some sort of ground systems as NASA was trying to purchase those. But exactly what it was used for may be in question.

 

I ran across several articles stating the Orion spacecraft computers are based on a single core Power-PC.

Given the number of articles that back that up, I'd say that has to be legit.

 

Edited May 8, 2017 by JamesD

[Van]

I suppose it's possible a 6809 was used in a non-flight computer role on some NASA mission, but it sounds like it could just be urban legend.

FWIW, if you put a large amount of external ROM filled with math tables on an 8 bit CPU, it could be faster than the Apollo guidance computer even without internal floating point math support. At 2 MHz it probably takes fewer cycles to calculate and set a page of ROM than the Apollo computer's floating point unit took for operations.

But that doesn't sound like a NASA solution does it? Why spend $1000 when you can spend $100,000,000?

A cheaper system might make more sense for something like Skylab, but that predated 8 bit microprocessors.

There was a version of OS-9 for the 68000, but it was called OS-9000.

 

Hi JamesD,

I'm sure that back in the '60s when they began designing the Apollo craft they would have loved to have had a ROM chip. They used a ferro CORE system call rope memory (also called LOL, Little Old Ladies Memory. MIT hired women to build the COREs because they had needle work experience), kind of like weaving the program. Each row wire for an address would be passed thru a bead for a 1 or bypass it for a 0 output on the bit wire, each bead could handle 64 row wires. So with 16 beads you would have a 16x64 ROM. The ferro CORE RAM could actively magnetize beads to store bits. The breakthrough back then was building the CPU with RTL ICs, the transistor density was only about 100 per chip, I think.

 

Here is an interesting link that lists CPUs on space missions, one of the main criteria was/is rad hardened chips. http://www.cpushack.com/space-craft-cpu.html It lists 8086s were used on the shuttle, and upgraded to 80386 but they may be referring to the display or engine sub-systems as NASA docs lists the MC68000 as the upgrade with the AP-101S.

 

Can't find much info on the AP-101 hardware design as the IBM System/4 Pi series (The AP-101 is the top model in this line) are/were used in a lot of DoD projects from the B52 to the B1 and specifics are classified, I think, aside from they are similar to the IBM 360 mainframe design. The 4 Pi family name is a play on words, referring to steradians of a full sphere and it's relation to the 360 degrees of a circle.

 

But yea, it's possible that they used 6809s in one of the many subsystems. Not sure about OS-9 as the whole mind set for avionics systems is to keep the software as simple as possible for reliability and once HW and SW is certified for flight, 'don't change what works'. So they tend to choose tech that has a long track record in the field.

 

 

That's not the only NASA stories I've read that might be urban legend.

Another involved a machine that logged data from a space probe using UCSD Pascal or maybe Apple Fortran (based on the same virtual machine), and data was lost when they exceeded the maximum number of files. I always thought that was running on something like a mainframe or PDP-11 though.

Yet another involved using an Amiga for something at NASA. I think that was supposedly another data logging system, and it was a cowboy solution to a problem rather than an official project.

NASA used the 8086 for some sort of ground systems as NASA was trying to purchase those. But exactly what it was used for may be in question.

 

I ran across several articles stating the Orion spacecraft computers are based on a single core Power-PC.

Given the number of articles that back that up, I'd say that has to be legit.

 

 

Oh I don't doubt your Amiga story, as I know first hand that there was allot of discretion in the labs and engineers and scientist loved their 'toys'. I rescued a TRS 80 model 100 and TRS model III from the dumpster when I was helping a friend clean out a building at Goddard that had become a 'storage attic'.

 

Yea the 8086 was used in the SRB engine systems and at one point NASA was buying NOS chip lots off of Ebay to maintain stock till the updated design was delivered. This was around 2000 I think.

And you're right about the Power PC, again they are going with proven HW that offers better rad harden robustness. But this is really the state of modern avionics; the In Flight Entertainment system is the most powerful computer on modern jets

Yogi

[JamesD]

...

Can't find much info on the AP-101 hardware design as the IBM System/4 Pi series (The AP-101 is the top model in this line) are/were used in a lot of DoD projects from the B52 to the B1 and specifics are classified, I think, aside from they are similar to the IBM 360 mainframe design. The 4 Pi family name is a play on words, referring to steradians of a full sphere and it's relation to the 360 degrees of a circle.

...

There was a page dedicated to it when I did a search. It was more of a high level explanation that explained the evolution of the systems that have been used.

[bugbear]

 

I have to wonder if anyone at NASA figured out the 6309 could be dropped in place of the 6809 and it could run faster and draw less power.

Probably not - NASA is just packed with ignorant, stupid people. :-)

 

BugBear

[Stormtrooper of Death]

The Tandy 100 uses a 2.4 Mhz Intel 80C85 cpu. program storage was on ram chip. so, no movable parts.

[2600problems]

please keep to the post guys. and stormtrooper, yes i would

[artrag]

About Amiga and msx2 this is for plain msx2 in screen 8 (256 colors)

[emkay]

About Amiga and msx2 this is for plain msx2 in screen 8 (256 colors)

Well, it's a shame to see that Atari had such a powerful "before the time" computer line back in 1979. If you see the advancing of the MSX hardware and do a timeline simulation, the Amiga should have been there in 1983 and in 1990 some spectacular Supercomputer.... and inbetween...

But Atari sold the same 1979's hardware even in the 1990s. The ridiculous ST not to mention... eventually got a foot in the MIDI sector, but surely was a step back in history.

 

Comparing the MSX2 with the Amiga is on the graphics line hard to compare 256 colours against 64 colours out of 4096 plus multi level parallax... and so on. You have to know that the AMIGA could run such a game in a separated thread and Window on the Workbench.

[2600problems]

how does the Amiga stand up against DOS?

[2600problems]

was the MSX capable of 3D? I mean, not literal but did it at least go the altered beast route?

[+Stephen]

With it's 4MHz Z-80, it was capable of 4D.

[artrag]

was the MSX capable of 3D? I mean, not literal but did it at least go the altered beast route?

 

Yes but do not expect miracles

[artrag]

Naturally if you move from full 3d to raycasting things go smoother

[emkay]

Hehe... MSX has the 16 Bit memory commands.... the Atari has the line blitter, to speed graphics up.