Fixing The Only Thing That’s Slow About Grand Theft Auto V

March 2, 2021 by Jenny List 13 Comments

The driving, crime, and general mayhem game Grand Theft Auto V is something of a phenomenon that has lasted for the last seven or more years. Whether following the in-game missions, driving around like a hooligan for fun or performing crazy stunts, the depth of detail in its landscapes and the continual improvements to gameplay that have arrived over the years have assured it a massive following across multiple platforms. The game is not without its problems though, one of which is an unreasonably long loading time for its online version. This annoyed [T0st] to the extent that it was worth the effort of looking under the hood to find out where the problem lay.

It was evident that for PC users the effect varied depending on the hardware present. Furthermore AMD processors seemed worse-hit than Intel ones, and indeed they found an entire core maxed out by a couple of processes during the wait. Some diagnostics and disassembly led the trail to some string processing code which was identified as a JSON parser. This was not simply parsing the JSON but also performing a check for token uniqueness in an extremely inefficient manner, causing the whole process to be extremely slow. Sone nifty patching in a DLL containing a much more efficient function with a cache for unique values saved the day, and delivered an impressive 70% speed-up. It’s to be hoped that the game’s developers will take note, and a future GTA V update will deliver a fix.

Driving a car from a third-person viewpoint in a game like GTA V is a hoot. In real life though, not so much.

Thanks [Thanatos Erberus] for the tip.

Slick Web Oscilloscope Is Ready In A Flash (Literally)

February 22, 2021 by Tom Nardi 27 Comments

A bench oscilloscope is one of the most invaluable tools in the hardware hacker’s arsenal, but even the slimmest digital models are a bit large to be part of your everyday electronic carry. Sure you could throw one of those cheap pocket scopes in your bag, but what if there was an even easier way to take a peek at a few signals while you’re on the go?

For those who roam, the Arduino-web-oscilloscope project created by [David Buezas] is worth a close look. Using the Web Serial API built into recent versions of Google’s Chrome browser, this project allows you to pop open a software oscilloscope without installing anything locally. Whether it’s a public computer or that cheap Chromebook you keep around for emergencies, a valuable tool is just a few clicks away.

Flashing the MCU from the web interface.

Of course, there has to be some hardware involved. Despite what you might think given the name of the project, the code currently only supports the Logic Green LGT8F328P microcontroller. This cheap ATmega328P clone not only runs at 32 Mhz but according to [David], many operations can be done in fewer clock cycles than on the original 328P. In short it’s fast, and fast is good if you want more samples.

One of the best parts about this project is that a function to flash the firmware to the LGT8F328P is built right in the web interface. With the oscilloscope running in the browser, you just need to plug in a blank board, click the button to flash it, and start taking measurements. You could outfit a whole classroom or hackerspace with basic oscilloscopes in minutes, with a per-seat cost of just a few bucks.

Now as you might expect, there are some pretty hard limits on what you can realistically measure with this setup. For one thing, the board can’t handle anything higher than 5 volts. Even the cheapest oscilloscope kit is still going to be an upgrade, but the fact you can spin this up almost anywhere for the cost of a cheap MCU board makes it hard to complain about the results.

[Thanks to Bill for the tip.]

Open Source: It’s The Little Things

February 13, 2021 by Elliot Williams 108 Comments

I use open source software almost exclusively; at least on the desktop — the phone is another matter, sadly. And I do a lot of stuff with and on computers. Folks outside of the free software scene are still a little surprised when small programs are free to use and modify, but they’re downright skeptical when it comes to the big works of professional software. It’s one thing to write xeyes, but how about something to rival Photoshop, or Altium?

Of course, we all know the answer — mostly. None of the “big” software packages work exactly the same as their closed-source counterparts, often missing a few features here and gaining a few there, or following a different workflow. That’s OK, different closed-source programs work differently as well. I’m not here to argue that GIMP is better than Photoshop, but rather to point out what I really love about open software: it caters to the little guys and gals, the niche users, and the specialists. Or rather, it lets them cater to themselves.

I just started learning FreeCAD for a CNC milling project, and it’s awesome. I’ve used Fusion 360, and although FreeCAD isn’t “the same” as Fusion 360, it has most of the features that I need. But it’s the quirky features that set it apart.

The central workflow is to pick a “workbench” where specific tasks are carried out, and then you take your part to each bench, operate on it, and then move to the next one you need. But the critical bit here is that a good number of the workbenches are contributed to the open project by people who have had particular niche needs. For me, for instance, I’ve done most of my 3D modelling for 3D printing using OpenSCAD, which is kinda niche, but also the language that underpins Thingiverse’s customizer functionality. Does Fusion 360 seamlessly import my OpenSCAD work? Nope. Does FreeCAD? Yup, because some other nerd was in my shoes.

And then I started thinking of the other big free projects. Inkscape has plugins that let you create Gcode to drive CNC mills or strange plotters. Why? Because nerds love eggbots. GIMP has plugins for every imaginable image transformation — things that 99% of graphic artists will never use, and so Adobe has no incentive to incorporate.

Open source lets you scratch your own itch, and share your solution with others. The features of for-pay, closed-source software are driven by the masses: “is this a feature that enough of our customers want?” The features of open-source software are driven by the freaky ideas of nerds just like me. Vive la différence!

PyGame Celebrates 20 Years By Releasing PyGame 2.0

October 28, 2020 by Tom Nardi 7 Comments

Python is an absolutely fantastic language for tossing bits of data around and gluing different software components together. But eventually you may find yourself looking to make a program with an output a bit more advanced than the print() statement. Once you’ve crossed into the land of graphical Python programming, you’ll quickly find that the PyGame library is often recommended as a great way to start pushing pixels even if you’re not strictly making a game.

Today, the project is celebrating an incredible milestone: 20 years of helping Python developers turn their ideas into reality. Started by [Pete Shinners] in 2000 as a way to interface with Simple DirectMedia Layer (SDL), the project was quickly picked up by the community and morphed into a portable 2D/3D graphics library that lets developers deploy their code on everything from Android phones to desktop computers.

Things haven’t always gone smoothly for the open source library, and for awhile development had stalled out. But the current team has been making great progress, and decided today’s anniversary was the perfect time to officially roll out PyGame 2.0. With more than 3,300 changes committed since the team started working on their 2.0 branch in July of 2018, it’s a bit tough to summarize what’s new. Suffice to say, the library is more capable than ever and is ready to tackle everything from simple 2D art up to 4K GPU-accelerated applications.

If you haven’t given PyGame a try in awhile, don’t worry. The team has put special effort into making the library as backwards compatible as possible, so if you’ve got an old project kicking around that you haven’t touched in a decade, it should still run against the latest and greatest version. If you’ve never used it before, the team says they’ll soon be releasing new tutorials that show you how to get the most out of this new release.

Whether you’re putting together your own implementation of Conway’s “Game of Life” or creating the graphical front-end for your own Linux distribution, PyGame is a powerful tool to have in your collection. Our sincere congratulations to all PyGame developers, past and present, for making it to this auspicious occasion. We can’t wait to see what the next decade will bring.

[Thanks to deshipu for the tip.]

Hardware Vs Software: Fight!

October 10, 2020 by Elliot Williams 41 Comments

It’s one of the great cliches in the hacker world: the hardware type and the software type. You can tell which of these two you are quite easily. When a project is actually 20% done, but you think it’s 90% done, and you say to yourself “And the rest is a simple matter of software”, you’re a hardware type. Ask anyone who has read my code, and they’ll tell you, I’m a hardware type.

Along with my blindness to the difficulties of getting the code right, I’ve also admittedly got an underappreciation of what powers lie in the dark typing arts. But I am not too proud to tip my hat when I see an awesome application of the soft stuff. Case in point: this Go board sequencer that we ran last week. An overhead webcam parses players’ moves as they put black and white stones down while playing the game of Go, and turns this into music.

The pure software type will be saying “but there’s a webcam and a Go board”. And indeed, that’s true. There are physical elements to this project that anchor it in the shared reality of the two people playing. But a hardware project this isn’t; it’s OpenCV and Max/MSP that make it work.

For comparison, look at the complexity of this similar physical sequencer. It’s got a 16 x 16 array of LEDs and switches and a CNC milled, primed, and painted surface that’s the size of a twin bed. Sawdust and hand-soldering: that’s a hardware project.

What I love about the Go sequencer is that it uses software just right. The piece is still physical. It could have just as easily been a VR world, where the two people would interact with each other only inside their goggles. But somehow that’s not quite as human as putting stones on a wooden board, sitting across from, and maybe even looking at, your opponent. The players aren’t forced to think about the software. They don’t feel like they’re playing a video game.

But at the same time, the software side of things makes all of the horrible hardware problems go away. Nobody is soldering a rat’s nest of 169 switches. There’s a webcam plugged into the USB port of a laptop. There’s a deep simplicity there.

Should you always trade out arcade buttons for OpenCV? Absolutely not! But is it worth considering the soft side when doing it in hardware is just too, well, hard? I’m open.

Autodesk Announces Major Changes To Fusion 360 Personal Use License Terms

September 16, 2020 by Dan Maloney 315 Comments

Change is inevitable, and a part of life. But we’re told that nobody likes change. So logically, it seems we’ve proved nobody likes life. QED.

That may be a reach, but judging by the reaction of the Fusion 360 community to the announced changes to the personal use license, they’re pretty much hating life right now. The clear message from Autodesk is that Fusion 360 — the widely used suite of CAD and CAM software — will still offer a free-to-use non-commercial license for design and manufacturing work, with the inclusion of a few very big “buts” that may be deal-breakers for some people. The changes include:

Project storage is limited to 10 active and editable documents
Exports are now limited to a small number of file types. Thankfully this still includes STL files but alas, DXF, DWG, PDF exports are all gone
Perhaps most importantly to the makerverse, ~~STEP~~, SAT, and IGES file types can no longer be exported, the most common files for those who want to edit a design using different software.
2D drawings can now only be single sheet, and can only be printed or plotted
Rendering can now only be done locally, so leveraging cloud-based rendering is no longer possible
CAM support has been drastically cut back: no more multi-axis milling, probing, automatic tool changes, or rapid feeds, but support for 2, 2.5, and 3 axis remains
All support for simulation, generative design, and custom extensions has been removed

Most of these changes go into effect October 1, with the exception of the limit on active project files which goes into effect in January of 2021. We’d say that users of Fusion 360’s free personal use license would best be advised to export everything they might ever think they need design files for immediately — if you discover you need to export them in the future, you’ll need one of the other licenses to do so.

To be fair, it was pretty clear that changes to the personal use license were coming a while ago with the consolidation of paid-tier licenses almost a year ago, and the cloud-credit system that monetized rendering/simulation/generative design services happening on the Autodesk servers. Features removed from the free license in this week’s announcement remain in place for paid subscriptions as well as the educational and start-up license options.

The problem with these personal use licenses is that it’s easy to get used to them and think of them as de facto open-source licenses; changing the terms then ends up leaving a bad taste in everyone’s mouth. To their credit, Autodesk is offering a steep discount on the commercial license right now, which might take some of the sting out of the changes.

Update 09-25-2020: Autodesk has announced that STEP file export will remain in the free version of Fusion 360

Improving More Leaf Design Flaws

September 16, 2020 by Bryan Cockfield 19 Comments

[Daniel] was recently featured here for his work in improving the default charging mode for the Nissan Leaf electric vehicle when using the emergency/trickle charger included with the car. His work made it possible to reduce the amount of incoming power from the car, if the charging plug looked like it might not be able to handle the full 1.2 kW -3 kW that these cars draw when charging. Thanks to that work, he was able to create another upgrade for these entry-level EVs, this time addressing a major Leaf design flaw that is known as Rapidgate.

The problem that these cars have is that they still have passive thermal management for their batteries, unlike most of their competitors now. This was fine in the early ’10s when this car was one of the first all-electric cars to market, but now its design age is catching up with it. On long trips at highway speed with many rapid charges in a row the batteries can overheat easily. When this happens, the car’s charging controller will not allow the car to rapid charge any more and severely limits the charge rate even at the rapid charging stations. [Daniel] was able to tweak the charging software in order to limit the rapid charging by default, reducing it from 45 kW to 35 kW and saving a significant amount of heat during charging than is otherwise possible.

While we’d like to see Nissan actually address the design issues with their car designs while making these straighforward software changes (or at least giving Leaf owners the options that improve charging experiences) we are at least happy that there are now other electric vehicles in the market that have at least addressed the battery thermal management issues that are common with all EVs. If you do own a Leaf though, be sure to check out [Daniel]’s original project related to charging these cars.

Continue reading “Improving More Leaf Design Flaws” →