A fascinating reminder that “progress” in computing is not always a straight line. As AI makes data movement and energy efficiency increasingly important, some of ENIAC’s old ideas about physically mapping computation start to look surprisingly modern again.
Thank you! FPGAs could be made to be much closer to the ENIAC concept for sure, but also to simulate the EDVAC one. I think they give you the equivalent of the logic units and wires to build either. However, you have a great point that programming them is conceptually just as difficult or more so than the ENIAC version. It takes much longer than it takes to compile a C program!
A fascinating reminder that “progress” in computing is not always a straight line. As AI makes data movement and energy efficiency increasingly important, some of ENIAC’s old ideas about physically mapping computation start to look surprisingly modern again.
Great article. I think CGRAs might be able to remove the abstraction of data movement while still keeping themselves sufficiently programmable.
Nice read. Wonder how FPGAs could factor into this?
Thank you! FPGAs could be made to be much closer to the ENIAC concept for sure, but also to simulate the EDVAC one. I think they give you the equivalent of the logic units and wires to build either. However, you have a great point that programming them is conceptually just as difficult or more so than the ENIAC version. It takes much longer than it takes to compile a C program!