Just to show you how far we've come -- NASA CEA

For a bit of nostalgia, I set a goal to recompile and run sample cases of the NASA Chemical Equilibrium Program written in fortran 2 ( I believe) . One ref is below. This program was a bulwark of all propulsion efforts on Apollo project. I used it around 1965.

NASA TN D-1737 (1963)

The program was about 1/2 a box of cards and used to have to be run on an IBM 7094 that took up a room; really a floor. You used to get paper output back the next day. This program needed a Thermo tape of needed thermodynamic data so you also had to hand in a 12" magnetic data tape and tell the data center people what tape unit to mount the tape on.

I found a copy of the source code through: NASA software. On my desktop pc, I recompiled the code in gfortran using code-blocks ide. The code is 6000 lines. It compiled in literally a second or two with no mods needed. I ran the benchmark set of sample cases The run took about 2 seconds again, literally. I got the output in a 158 kb file; not an inch of paper. Not one issue with compilation or execution.

I thought it might be fun to just sit back and see how far all this computing stuff has come over a lifetime. Also, I thought it would be good to let younger people get an idea of how things used to be.

tom kosvic

5 Likes

Nice, thank you for sharing!
Compilation without error in 2026 is quite remarkable… They must have used AI-assisted development :thinking:

These days you can use the open source cantera for doing chemical calcs, it’s Python-based and quite a heavy install.
Although completely irrelevant, it would be interesting to compare the computational resources needed for the same equilibrium calculation… again to show how technologies have evolved…

This is a fantastic bit of computing history. Going from an IBM 7094 that occupied an entire room, punched cards and magnetic tapes, to recompiling 6,000 lines of Fortran on a desktop in seconds is an incredible comparison.

What really stands out is that the original code compiled and ran without modification. It’s a great reminder of how much computing hardware has advanced while well-written scientific software can remain surprisingly durable. Thanks for sharing this piece of nostalgia!

1 Like

The funny thing is that “AI” is that old with a previous bubble in the 60s. Compiling 1963 code today without issue is impressive. I was impressed when I found that Gnat 2022 compiles Booch components written in Ada 83 without changes other than the file extensions from .1.ada and .2.ada to .ads for spec and .adb for bodies.

2 Likes