Rem@IMSSS (02/12/86)
Hmmm, if 5 minutes time on a Cray costs less than 10 million dollars, and it will prevent loss of another 2000 million dollar orbiter, then it sounds like a good idea, providing NASA knows what software to run on it to make effective use of it. Anybody have Cray cost figures? -------
ecl@mtgzy.UUCP (e.c.leeper) (02/14/86)
> Hmmm, if 5 minutes time on a Cray costs less than 10 million dollars, > and it will prevent loss of another 2000 million dollar orbiter, > then it sounds like a good idea, providing NASA knows what software ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > to run on it to make effective use of it. Anybody have Cray cost figures? ^^^^^^^^^^^^ And if wishes were horses, then beggars would ride... Before we start looking for cost figures, let us know where the software is. Evelyn C. Leeper ...ihnp4!mtgzz!ecl (or ihnp4!mtgzy!ecl)
paul@hp-lsd.UUCP (Paul D. Bame) (02/15/86)
Except for the small (:-)) datacomm problem, maybe NASA could buy some
time on their own Cray (2?) at AMES.
--Paul Bame
UUCP: {hplabs,ihnp4!hpfcla}!hp-lsd!paul
CSNET: hp-lsd!paul@hp-labs.csnet
ARPA: hp-lsd!paul%hp-labs@csnet-relay.arpawaynekn@tekig5.UUCP (Wayne Knapp) (02/15/86)
> Hmmm, if 5 minutes time on a Cray costs less than 10 million dollars, > and it will prevent loss of another 2000 million dollar orbiter, > then it sounds like a good idea, providing NASA knows what software > to run on it to make effective use of it. Anybody have Cray cost figures? > ------- You can buy a Cray for less than 10 million. 5 minutes time shouldn't be more than a $1000.
davidsen@steinmetz.UUCP (Davidsen) (02/28/86)
In article <493@tekig5.UUCP> waynekn@tekig5.UUCP (Wayne Knapp) writes: >> Hmmm, if 5 minutes time on a Cray costs less than 10 million dollars, >> and it will prevent loss of another 2000 million dollar orbiter, >> then it sounds like a good idea, providing NASA knows what software >> to run on it to make effective use of it. Anybody have Cray cost figures? >> ------- > >You can buy a Cray for less than 10 million. 5 minutes time shouldn't be >more than a $1000. The following figures are reasonably available and checkable: a new Cray2 costs about $16 million. Adding some standard figures for installing and maintaining the environment (water cooled), operator salaries, etc, the cost is about $25 million for five years, or about $571/hr. The problem is, that if you divide the cost by the number of hour of launch the rate gets pretty high, and there are not a lot of people who want to tie up their machine to let NASA use it a few times a year. The issue of cost is specious! A Cray (1, 2, or XMP) is not a realtime machine and would probably be totally unsuited to such monitoring. It is the perfect engine to run simulations before launches, and to look a data after a launch and find posible problems, but I have seen no evidence that it is intended for or suited to realtime monitoring. Finally, the accountability question: given the problems in aborting a shuttle launch during the SRB phase, would we want a machine to decide to try it? Please take all answers to that question to mod.risks, not here. Having done a little realtime work myself (including designing and writing a multiprocessor R/T operating system), I feel that having redundant small systems doing the monitoring is both cheaper, easier to impliment, and more reliable. The problem is knowing what things to monitor, and how to recover from a failure. The intrinsic problem is that the shuttle is (at least) one step away from a practical solution to getting things into space. Many posible alternatives have been suggested in this group, indicating that others agree with me that the next launch vehicle must have (a) lower cost, (b) shorter turnaround, (c) higher payload percentage, and (d) better failure modes. Hopefully the shuttle will be allowed to continue until that "next step" is taken. +================================================================+ All figures in this posting are public information. all opinions are my own. -- -bill davidsen seismo!rochester!steinmetz!--\ / \ ihnp4! unirot ------------->---> crdos1!davidsen \ / chinet! ---------------------/ (davidsen@ge-crd.ARPA) "It seemed like a good idea at the time..."