The Last Starfighter
x01001x
the FX for "The Last Starfighter" motion picture? I know all visual
effects were done on a CRAY XMP.
Eugene Miya
There was a ACM/SIGGRAPH paper at the time.
All highly proprietary and would never run w/o Freon, or an emuator and
locating the old OS and compiler sources.
Hmmm, can't find it immediately in my DB, must have been in the SIGGRAPH
DB which I turned over to the org. and sat on the DECSRC/DECWRL server.
%A Gary Deimos
%Z Whitney/Deimos
%T Supercomputing for High Complexity Computer Graphics
%B High-Speed & Large-Scale Computing: A Panoramic View
%E Myron Ginsberg
%R SP-763
%I Society of Automotive Engineers, Inc.
%D 1988
%P 117-126
%K applications for current and future vector and parallel processors,
--
Fred Bradford
>In article <1186395575.6...@r34g2000hsd.googlegroups.com>,
>x01001x <xem...@softhome.net> wrote:
>>Is there any DETAILED information on what software was used to make
>>the FX for "The Last Starfighter" motion picture? I know all visual
>>effects were done on a CRAY XMP.
"The Last Starfighter" computer graphics, 30 minutes in the movie, was
produced on a home-grown Fortran program (starting on a Cray-1) and
completed on a Cray-XMP at Digital Productions in Los Angeles.
The program was well over 500,000 lines of Fortran and Cray assembly
language, all written by employees. Possibly, another 200,000 lines of
code were added later to produce the few seconds of film in the movie
"2010" showing the whirling of the clouds of the planet Jupiter.
(The assembly language was used in areas that needed higher speed than
the Fortran could produce. Mostly in areas of vectorization on inner
loops that untill almost 1987, Fortran just was not performing well.
Other unique code dealing with zero-lenth arrays could not be coded in
Fortran)
Without the income later from major Hollywood productions, the company
simply closed its doors in 1987. Its owners from Canada (Omnibus)
simply went home. No bankruptcy papers were ever filed.
Whitney and Demos (note the spelling) were the founders [see the SAE
reference below.].
-Fred Bradford (Los Angeles)
Eugene Miya
Fred Bradford <avp...@earthlink.net> wrote:
>The program was well over 500,000 lines of Fortran and Cray assembly
>language, all written by employees. Possibly, another 200,000 lines of
>code were added later to produce the few seconds of film in the movie
>"2010" showing the whirling of the clouds of the planet Jupiter.
>
>(The assembly language was used in areas that needed higher speed than
>the Fortran could produce. Mostly in areas of vectorization on inner
>loops that untill almost 1987, Fortran just was not performing well.
>Other unique code dealing with zero-lenth arrays could not be coded in
>Fortran)
>
>Without the income later from major Hollywood productions, the company
>simply closed its doors in 1987. Its owners from Canada (Omnibus)
>simply went home. No bankruptcy papers were ever filed.
Did anyone keep the CFT or CAL source?
Did you guys run COS or CTSS or write your own OS?
>Whitney and Demos (note the spelling) were the founders [see the SAE
Good you bit the typo.
>reference below.].
>
>-Fred Bradford (Los Angeles)
>
>
>>%A Gary Demos
>>%Z Whitney/Demos
>>%T Supercomputing for High Complexity Computer Graphics
>>%B High-Speed & Large-Scale Computing: A Panoramic View
>>%E Myron Ginsberg
>>%R SP-763
>>%I Society of Automotive Engineers, Inc.
>>%D 1988
>>%P 117-126
>>%K applications for current and future vector and parallel processors,
--
Silver Bells
wrote:
>In article <ebvuj3drrnmcr8jcn...@4ax.com>,
>Fred Bradford <avp...@earthlink.net> wrote:
>>The program was well over 500,000 lines of Fortran and Cray assembly
>>language, all written by employees. Possibly, another 200,000 lines of
>>code were added later to produce the few seconds of film in the movie
>>"2010" showing the whirling of the clouds of the planet Jupiter.
>>
>>(The assembly language was used in areas that needed higher speed than
>>the Fortran could produce. Mostly in areas of vectorization on inner
>>loops that untill almost 1987, Fortran just was not performing well.
>>Other unique code dealing with zero-lenth arrays could not be coded in
>>Fortran)
>>
>>Without the income later from major Hollywood productions, the company
>>simply closed its doors in 1987. Its owners from Canada (Omnibus)
>>simply went home. No bankruptcy papers were ever filed.
>
>
>Did anyone keep the CFT or CAL source?
Yes....
There is a copy that has some 34,000 lines of CFT but no CAL. (My
original estimate of 500,000 lines of code was a HUGE mistake. Sorry.)
This copy is dated June 6, 1984.
I don't know if it has any of the Jupiter code.
>Did you guys run COS or CTSS or write your own OS?
>
It all ran on COS only. There were numerous calls from CFT and CAL to
the operating system that are certainly incompatible with CTSS.
The successor/owner to Digital Productions went out of business in the
Spring of 1985. By that time, it was owned by a Canadian company. The
officers simply walked away, the Canadians flew back to Canada. No
bankruptcy was ever declared to my knowledge.
Thus the ownership (as well as the copyright) are unkown.
-Fred Bradford