Featured Post

This year, for Christmas, my children* got a Valve Steam Deck under the Christmas tree.  It's a pretty cool device that looks a little  like a monsterous Nintendo Switch, but it can run an impressive subset of the Steam video game catalog, games mostly designed to run on Windows PCs.  It manages this by sporting a custom x86_64 processor by AMD and running a customized version of Arch Linux that uses Wine via Valve's Proton tool.  The key point here, is that it is a tiny x86_64 compatible computer running Linux. So, of course, I needed to install Maple on it. So, I just paired a bluetooth keyboard, rebooted it into desktop mode and with a few small trick, bam, Maple on the Steam Deck:

There were a few small hiccups that required some work. I had absolutely no problems getting the Maple installer onto the device via a USB drive and no problems running it. I only ran into problems durring license activation:

Fortunately, I talked to our crack technical support team and they were able to identify this as a problem with Arch Linux not having full LSB 3.0 support installed by default. The process for fixing that is documented on the Arch Linux Wiki and involves just installing the ld-lsb package via pacman -- with the small additional wrinkle that you need to take the Steam Deck operating system out of 'read-only' mode in order to do that. But once that was done, I had a full version of Maple running well (albeit at 1280x800 resolution on a 7" display).

Since this device is designed for gaming, I was curious how fast it is compared to some other machines I work on. I chose an arbitrary benchmark of exactly solving a random linear system with integer coefficients.

N := 400;
A := LinearAlgebra:-RandomMatrix(N, N):
b := LinearAlgebra:-RandomVector(N):
v := [seq(cat(v__, i), i = 1 .. N)]:
sys := LinearAlgebra:-GenerateEquations(A, v, b):
CodeTools:-Usage(SolveTools:-LinearSolvers:-Rational(`~`[lhs - rhs](sys), v, dense = false)):

which it solves in decent time:

For comparison, this is 30% faster than the 32 core Xeon e5 workstation I do most of my work on, and only 5% slower than my notebook computer with an 8th gen Intel i7.  Not bad for a toy! (please don't make me sad by telling me how much faster this is on a Mac M1 or M2)

Let me know in the comments if you have other benchmarks you want me to run on the Steam Deck. Also, please let me know if you manage to get your employer to buy you a Steam Deck to do scientific computing.


*Okay, maybe it was a gift for me. Shhhh, don't tell.

Featured Post

Last week, one of our Maple Learn developers, Valerie McKay-Crites, published a Maple Learn document, based on the very popular Maple application by Highschool Teacher, Jason Schattman called "Just Move It Over There, Dear!".

In the Maple application, Schattman explains the math behind moving a rectangular sofa down a hallway with a 90-degree turn. In the 3D Moving Sofa Problem Estimate, Valerie uses Schattman’s math to determine the largest rectangular sofa that can be taken down a flight of stairs and down a hallway with a 90-degree turn. Both applications reminded me of how interesting the Moving Sofa Problem is, which inspired me to write a blog post about it!

If you’ve ever been tasked with moving a rectangular sofa around a 90-degree turn, you might wonder:

What is the largest sofa that can make the move?



Description automatically generated with medium confidence


Following these steps as outlined in Schattman’s "Just Move It Over There, Dear!", will guarantee that the sofa will make the turn:

  1. Measure the width of the hallway (h)
  2. Measure the length (L) and width (w) of the sofa.
  3. If L + 2w is comfortably less than triple the width of the hall, you'll make it!

When we work out the math exactly, we see that if the sofa's length plus double its width is less than 2*h*sqrt(2), the sofa will make the turn!


Chart, line chart

Description automatically generated


This problem is easy if we only consider rectangular sofas, however, the problem becomes significantly more complex if we consider sofas of different shapes and areas. In mathematics, this problem is known as the Moving Sofa Problem, and it is unsolved. If we look at a hallway with a 90-degree turn and legs of width 1 m (i.e. h = 1 above), the largest known sofa that can make the turn is Gerver’s Sofa which has an area of 2.2195 m2, this area is known as the Sofa Constant. Gerver’s Sofa, created in 1992, was constructed with 18 curve sections:


Description automatically generated


Check out this GIF of the sofa moving through the turn. It provides some insight into why Gerver’s sofa is such an interesting shape:

What is fascinating is that no mathematician has yet to prove that Gerver’s sofa is the sofa with the largest area capable of making the 90-degree turn.

The Moving Sofa Problem, is a great example of how math is embedded in our everyday lives. So, don’t stop being curious about the math around you as it can be fascinating and sometimes unproven!

If you are curious to learn more about the moving sofa problem check out this video by Numberphile, featuring Dan Romik from UC Davis: https://www.youtube.com/watch?v=rXfKWIZQIo4&t=1s

how to solve this in maple?

Maple asked by wijdan 10 Today

How to Plot this ?

Maple 18 asked by Saha 60 Today

ODE with complex variable

Maple asked by arashghgoo... 50 January 29