Raspberry Pi Zero - the $5 Computer - Computerphile

formatting link

I was trying to compare this computer to the computer power of the space program. The best I had was that it was more powerful.

He wanted to compare it to the computer power on the ground. I didn't really have much to go on.

How does the the Pi 0 compare to NASA's desktop at the time?

Reply to
Seymore4Head
Loading thread data ...

formatting link

formatting link

Reply to
Seymore4Head

What is a NASA desktop? Don't they use the same computers on the ground as everyone else? Also, what time are you talking about? The rPi is now. Do you mean desktops now? NASA is still operating you know.

--

Rick
Reply to
rickman

Not an answer to the above.

I seem to recall that when desktop computers started to appear, the computers NASA was sending into space were about 15 years behind the times. This was because they had to be built of radiation-hardened components, and radiation-hardened versions of the microprocessor chips simply didn't exist.

--
Alan Adams, from Northamptonshire 
alan@adamshome.org.uk 
 Click to see the full signature
Reply to
Alan Adams

I think radiation resistance is only needed on unmanned missions. They sent marginally reasonable machines on manned missions. I believe the landing computer on the moon missions had 64 kB or less.

--

Rick
Reply to
rickman

Radiation hardening isn't really an issue in LEO unless a decent-sized CME arrives while its in orbit, but a moon shot is a long way out beyond the Earth's magnetic field, so anything going to the moon or further must be radiation hardened because (a) they are exposed to high energy solar and interstellar radiation and (b) because they are going to spend several years bathed in it.

I think the same goes for anything orbiting outside the Van Allen belts, parked at a Lagrange point or sent to another planet, i.e. all the geostationary comms birds, to say nothing of Hubble, SOHO, Planck, LISA, and of course the Voyagers and New Horizons.

The GPS/Glonass/Gallileo consellations also need to be hardened because their orbits intersect the van Allen belts, which trap and store high- energy particles.

--
martin@   | Martin Gregorie 
gregorie. | Essex, UK 
 Click to see the full signature
Reply to
Martin Gregorie

I will acknowledge a lack in this area, but if the equipment needs to be radiation hardened, how do you radiation harden the people? Are you suggesting we have sent people to the moon where they were exposed to radiation that would incapacitate the HP hand calculators they carried?

--

Rick
Reply to
rickman

On Thu, 03 Dec 2015 21:24:21 GMT, Alan Adams declaimed the following:

Heck... the 737 fleet flies on 68040 cores running 30MHz.

The NASA stuff in the 70s was probably flying stuff with less power than mil-std 1750a (a 1980 spec) style processors

formatting link
(as the link indicates, the Cassini and Mars Global Surveyor both use 1750A class processors)

The "supermini" of the early 80s, the VAX-11/780 was based off TTL

74181 ALU chips, if one trusts
formatting link
Today, you can emulate a microVAX cluster on a pair of R-Pis
formatting link
or IBM 4381
formatting link

{Hmmm, wonder how difficult it would be to duplicate those using the BeagleBone Black I have}

--
	Wulfraed                 Dennis Lee Bieber         AF6VN 
    wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/
Reply to
Dennis Lee Bieber

On Thu, 3 Dec 2015 18:25:08 -0500, rickman declaimed the following:

The main purpose of rad-hardening is to ensure the ([E]P)ROM (or now FLASH) doesn't encounter bit flips which could corrupt the program. These things aren't easily field-programmable to correct such errors. I wouldn't be surprised if half the CPU time was spent running CRC checks on the "permanent" program/data memory to detect errors, along with (non-destructive) RAM checks. If they copy from FLASH to RAM to run, the RAM likely has lots of ECC overhead, but a bit flip (once ECC reports it is making corrections) can be recovered by reloading the code -- so long as the FLASH has not become corrupted. If they run from FLASH, it's a more critical matter.

--
	Wulfraed                 Dennis Lee Bieber         AF6VN 
    wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/
Reply to
Dennis Lee Bieber

formatting link

--
Cheer, 
Chris
Reply to
Chris

I suspect it would be a case of using the same computer systems for the manned missions that had been successfully used for the unmanned ones. There isn't a lot to say against a high degree of caution.

--
Alan Adams, from Northamptonshire 
alan@adamshome.org.uk 
 Click to see the full signature
Reply to
Alan Adams

The 780 CPU was, if I recall correctly, 8 large boards of TTL, with the memory boards stacked in pairs either side. Our first one came with a massive 1.5 MB of RAM, later upgraded to 4M at a cost of 15K ukp.

The system disk was 100M.

--
Alan Adams, from Northamptonshire 
alan@adamshome.org.uk 
 Click to see the full signature
Reply to
Alan Adams

In the case of all the American and Russian orbital flights there wasn't a problem because all were in LEO and they didn't launch if the sun had burped. Similarly, all space stations have been in LEO and I think most/ all of them have either been fitted with a radiation-resistant area, or (more likely) have always got enough re-entry capsules on hand to let the crew bail out in the face of a massive CME heading our way.

The Apollo flights all initially went into LEO for one or two orbits before heading moonwards. The transfer to lunar orbit went through the Van Allen belt, but quite quickly, so their radiation dose remained in safe limits both going and returning. I don't know what they did about CMEs during the mission, but probably not much more than not launching if one seemed likely.

Since nobody else has been out of LEO since Apollo, thats all manned missions covered.

--
martin@   | Martin Gregorie 
gregorie. | Essex, UK 
 Click to see the full signature
Reply to
Martin Gregorie

ICL's 2966 was soft. Its microcode was executed by a 2MHz 6809 with, presumably a fair amount of peripheral hardware. I don't know how fast it could run its 1900 emulation microcode, but IIRC the 2900 microcode (the so-called native instruction set) ran at about 200K instructions/sec and was capable of running both sets microcode simultaneously - the only external sign of this was that you needed two operator consoles: one for George3 and one for VME/B.

I never saw the bigger 2972 and 2980 but was told they were hardware implementations (the 2972 was really MU5 hardware) and could only run the native instruction set.

I recently found out, rather to my surprise, that Fujitsu, who bought ICL, still supports VME/B, but under revised microcode running on a fairly ordinary PC - and most likely somewhat faster than any 3900 could do it and with more memory too. Why? Apparently some ex-2900 shops are still running applications developed on 2900s and this is the most cost- effective way of doing it.

--
martin@   | Martin Gregorie 
gregorie. | Essex, UK 
 Click to see the full signature
Reply to
Martin Gregorie

In comp.sys.raspberry-pi message , Fri, 4 Dec 2015 22:26:30, Martin Gregorie posted:

They *always* have enough Soyuz to take everyone down. I don't recall mention of a shielded area inside.

IIRC, James Mitchener, in his lengthy novel "Space" depicting a later Apollo-type mission, had an unexpected lethal CME or similar during a moon-drive; presumably he had done his "research" first.

--

 Merlyn Web Site <                       > - FAQish topics, acronyms, & links.
Reply to
Dr J R Stockton

My original questions should have been the original NASA space program and not desktops, but main frames.

The biggest example of change I had for my nephew was that my first hard drive was a external dual height 64 Meg. It was as large as a loaf of bread and as heavy as a brick. Sim cards hold 1000 times more and are the size of a thumbnail. It cost "used" 600 bucks.

Reply to
Seymore4Head

I still have the telex response from Prime computer where they confirmed the puchase of one 315MB drive, 2 MB memory and one 16(15 usable) async line cards for our Prime 2550. The total was NOPK 485 000; or around $83 000 at the time.

And the 315MB drive only gave around 272 mb with decent formatting.

-- mrr

Reply to
Morten Reistad

On Sun, 06 Dec 2015 13:51:06 -0500, Seymore4Head declaimed the following:

Well... So far as I recall, NASA was one of the places heavy on SDS/XDS Sigma systems (along with McDonnell-Douglas Automation, Lockheed MSD/Navy, and for some reason a slew of MI colleges -- I remember the time we had problems getting time at GVSC because Wayne State's had gone down and they were running registration using our Sigma 6)

Cheap... I could never afford it, but the 5M hard-drive for that came out for TRS-80 Mod=III sold for $5000... I was scrounging to get the $800 for floppy controller and first floppy.

--
	Wulfraed                 Dennis Lee Bieber         AF6VN 
    wlfraed@ix.netcom.com    HTTP://wlfraed.home.netcom.com/
Reply to
Dennis Lee Bieber

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.