In the case of Fortran in France, nothing to do with the language. Just an homonym…
In the case of Wales, it’s in Cardiff which has an university, although those roads seem far away of the university.
1 Like
Cray-3 Supercomputer Systems (1993)
There are cool pictures of Cray machines. Pages 14 and 16 mention that there were Cray compilers for C, Fortran 77, and a subset of Fortran 90. Cray is now part of Hewlett Packard Enterprise. Current compiler
1 Like
Cray went through phases of being independent, part of SGI, independent again, and, most recently, part of HPE. Throughout, the “Cray” Fortran compiler has continued to evolve with new standards and added support for things like GPU’s and the ARM-SVE processors made by Fujitsu. The current compiler version is essentially F2018 compliant. The coarray parallel model has been part of the compiler since the T3E days in the 1990’s.
As pmk noted, the Cray C-90 was a reliable, though expensive, machine that had a human-understandable hardware instruction set. That tradition continued with subsequent vector machines, such as the T90 (which added IEEE floating point hardware), the X1/X1E (added distributed memory parallelism) and the X2, the last of the line, and my favorite in terms of network integration and easy-to-understand hardware instructions.
1 Like
Fortran may be the only language to have an asteroid but Algol has a star (beta Persei according to Norton’s Star Atlas, which says its name means the Demon Star because it’s variable). But the star’s name appeared centuries before the language did.
3 Likes
Funny. Also, I scored 68.
1 Like
I scored 65 (over 70 point). I lost 5 points in question 3…
Python and Fortran share something: both are divine.
- Python was Gaia’s son: Python (mythology) - Wikipedia
- Fortran’s father is Backus, and Bacchus is the roman name of Dionysus, a son of Zeus and a god of many things. One of his epithets was Dimetor Διμητωρ (“Twice-Born”) and refers to Dionysus’ two births (Fortran I and Fortran 90?). He was also said to be young and old, “the older and the younger of all gods”.
1 Like
It was largely due to the efforts of thousands of Real Programmers working for NASA that our boys got to the moon and back
To be specific, searching the NASA Technical Report Service for the earliest Fortran documents gives results like this:
Astronomical integration code - fortran solution of space mechanics - computer program(1963)
Fortran computer program for gas-liquid partition chromatography data processing(1963)
The documents list the codes.
Do you remember the days when there was not only search engines but also directories?
There was the Open Directory Project (ODP) also known as DMOZ. It is still alive (or reborn) and is now known as Curlie, and this is the Fortran directory:
1 Like
Guy Steele Jr. designed an language, now discontinued, intended for HPC. It was called Fortress, implying a secure replacement for Fortran.
Sometimes, I feel Fortran is like a grandad blessed with elixir of life. Destined to see death of many of his children.
2 Likes
Like C, every week a new language pretends to replace it. But C is still at the top of the TIOBE index.
CYBERNETIC LANDSCAPES (1971): Fortran program for displaying a virtual reality environment on an LDS-1.
3 Likes
@interkosmos , thanks for sharing. I was also impressed by the text SWI.V1.CybLandscp3.311213.pdf, linking that program to poetry. And the graphics reminds me of the Vectrex game console.
If someone is interested to try to port that program to gtk-fortran, I could help with the GTK part. We are missing the definitions of the subroutines, but for the drawing ones we can guess what they do:
CALL POS(IX(N),0,IZ(N))
...
CALL DRAWTR(0,IY(N),0)
...
CALL DRAWT(-131071,0,131071)
I have found an animated gif here:
and an interview here:
2 Likes
@vmagnin: The rendering is probably the easiest part. But without documentation on the LDS-1 and its FORTRAN IV library, figuring out the inner workings of subroutines like DTABLE() or MAKMT2() is quite difficult.
1 Like
From The art of solving problems with Monte Carlo simulations
Generate uniform deviates until their sum exceeds one. The average number of deviates needed is Euler’s number, e. Proof.
program main
! simulate Euler's number, e
implicit none
integer, parameter :: nsim = 10000000, niter = 10
integer :: i,iter,j,nsum(nsim)
real :: xsum,xran
call random_seed()
do iter=1,niter
do i=1,nsim
xsum = 0.0
j = 0
do
j = j + 1
call random_number(xran)
xsum = xsum + xran
if (xsum > 1.0) exit
end do
nsum(i) = j
end do
print*,sum(nsum)/real(nsim)
end do
print*,exp(1.0)," true"
end program main
output:
2.71822739
2.71828914
2.71854210
2.71806741
2.71793246
2.71829391
2.71816230
2.71832871
2.71852446
2.71853471
2.71828175 true
1 Like
That reminds me of Buffon’s needle, where you can use a Monte Carlo simulation to determine the value of pi. (Tricky part: do not use cos() and sin(), as that implicitly means you have to use pi before you have estimated it)
1 Like
Yes, that’s in the cited link: The art of solving problems with Monte Carlo simulations | Gabriel Carvalho
(first example)
I am playing with that for one week, and will soon make a post.
Note that the precision is like \frac{1}{\sqrt{N}}, so it’s a very inefficient algorithm to compute \pi…