Conduit 1

Conduit 1, also known by its systematic name BFx59H, is one of the earliest and most remarkable converters, discovered by Dave Buckingham on July 5 1996. Composed of a block and a snake (or eater 1 to make it Spartan), it transforms a B-heptomino into a clean Herschel with very good clearance in 59 generations, allowing easy connections to other conduits (see images below). It forms the final stage of many of the known composite conduits, including the majority of the original sixteen Herschel conduits. It is important because it was one of the key pieces in finding a method for construction of oscillators and guns of arbitrary large period (see omniperiodic).

Conduit 1
<html><div class="rle"><div class="codebox"><div style="display:none;"><code></html>x = 11, y = 5, rule = B3/S23 ob2o$2obo2$9b2o$9b2o! #C [[ THUMBSIZE 2 THEME 6 GRID GRIDMAJOR 0 SUPPRESS THUMBLAUNCH ]] #C [[ AUTOSTART ]] <nowiki></nowiki> <html></code></div></div><canvas width="200" height="300" style="margin-left:1px;"><noscript></html> <html></noscript></canvas></div></html>
View static image
Pattern type Conduit
Conduit type Converter
Input B-heptomino
Output Herschel
Number of cells 10
Bounding box 11×5
Output orientation Unturned, flipped
Step 59 ticks
Recovery time
(ignoring FNG if any) 		54 ticks
Minimum overclock period
(ignoring FNG if any) 		Unknown
Spartan? Yes
Discovered by David Buckingham
Year of discovery 1996
  • Generation 0 with a B-heptomino
  • Generation 59 with a Herschel
gollark: This sort of thing is very good at the particular task it's optimized for, but expensive (initial-cost-wise, it's easy to churn out more of them) and entirely unable to do anything else, unlike general-purpose CPUs/GPUs, which are also hilariously expensive in initial investment but can do basically anything and are reusable all over the place.
gollark: Fortunately, we have good cryptography now as export controls were stupid and didn't actually work.
gollark: Well, "very good" varies.
gollark: Also, you shouldn't avoid asking questions, but remember that AI things are hard, don't work like humans, and aren't magic but very good pattern-matchy algorithms.
gollark: So you can just have lots of things generating hashes in parallel.
This article is issued from Conwaylife. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.