Inexpensive Monoprice Tablet

Back in April of 2014 I bought myself a birthday gift, a $50 digital drawing tablet from Monoprice. I read reviews stating that it worked well, and noted that Linux drivers existed thanks to the excellent DIGImend Project(old wiki here). The tablet as I later learned, is a rebranded Huion h610, and seems relatively well regarded. Compared to similarly priced Wacom tablets it was a great value. When it arrived, I expected to do some tinkering to get it working, but six months later I realized the magnitude of this undertaking. Following months of hard work attempting to get my Monoprice graphics tablet working well on Ubuntu 12.04, I had it mostly succeeded. (A dramatic recounting of this story will be told another time) Unfortunately for me, a bug in the version 12.04 included forced me to either upgrade to a newer release or slowly descend into the hell that is maintaining complex software and its dependencies outside the package manager. Because I’m lazy I chose to upgrade to 14.04. The tablet is now reporting motion events in its full native resolution, pressure events are mostly correct, and even the buttons work. Unfortunately, though the key IDs reported are mind bogglingly stupid, such as Alt+F4, Ctrl+G, Ctrl+C with predictably annoying effects. I intend to address these issues in the coming months, but for the moment, I want to ensure the main feature, the pen and tablet, are functioning correctly. Before I dive into the problem and how I solved it, I want to explain the XRandR multi-monitor model. xrestrict itself directly mentions CRTCs and Screens, rather than intuitive end-user concepts like “Monitors” so it is useful to understand them.

XRandR Model

XRandR setups typically have a single object called a Screen, which is a virtual rectangular region in which windows exist. X11 may have multiple Screen objects, but long time X users will note that windows may not cross from one Screen to another, limiting the usefulness of having multiple X Screens. As a result, modern X11 configurations have this single virtual Screen. The Screen is then divided into one or more rectangular regions called CRTCs. These CRTCs are then displayed on zero or more outputs, which represent connections to monitors and may be rotated or scaled to display as the user desires.

XRandR Screen Model

XRandR’s model of the world.

Back to the Tablet

I was very nearly satisfied with my tablet’s functionality, with one significant caveat due to my multi-monitor setup. If I positioned my pen in the lower left of my tablet, the cursor appeared in the lower left corner of my left monitor. If I positioned my pen in the upper right of my tablet, the cursor appeared in the upper right of my right monitor. This is the obvious behavior for the tablet, and I initially didn’t think anything of it. There was one problem however, if I drew a circle on the tablet, it appeared as a wide oval on screen. The issue was in how tablet coordinates were being translated to screen coordinates. In the X axis, coordinates ranged from 0 to 40,000, which were then mapped to the entire width of my X screen. As I have a dual monitor setup, the width of my X screen is actually the sum of the width of my two monitors: 3280 pixels. In the Y axis however, 25,000 units on my tablet, was mapped to the y axis of my screen: 1050 pixels. The aspect ratio of my X screen was approximately 3:1, while the aspect ratio of my tablet is 16:10. Directly mapping one to the other without regard for aspect ratio produced a distortion of nearly 2:1 from tablet coordinates to on-screen coordinates. This distortion is why a perfect circle drawn on the tablet, even with a guide, appeared as an oval on screen.

To remedy this I needed to provide a different mapping. Handily, XInput2 devices provide just the thing, the “Coordinate Transformation Matrix” which alters the pointer coordinates. If you alter this “Coordinate Transformation Matrix”, you can introduce scaling, skew, rotation and translation to the coordinates calculated by I initially calculated a “Coordinate Transformation Matrix” using a simple Python script, to restrict the pointer to my right screen. This worked, though I found I still had significant difficulty drawing. I attribute this to the “Coordinate Transformation Matrix” still being imperfect. It mapped tablet coordinates 40000×25000 to my right screen which is 1600×900. Though much closer to the correct aspect ratio, it still caused approximately an 11% distortion, which I believe to be significant.

Default Naive Mapping of Pointer Coordinates

Default Mapping of Pointer Coordinates

Because I have two monitors, stretching the tablet’s limited resolution over my entire screen seemed wasteful, as it would reduce effective resolution in the X axis by half. Furthermore, regardless of 12.2x greater precision than the screen, I wasn’t certain of the accuracy of the tablet, and the jagged lines I was drawing made me wonder if the tablet’s position readings might be noisy/jittery. If this was the case, then stretching the axes two times would double the size of these inaccuracies (not to mention any inaccuracies my unskilled hand might have). So I decided that I should instead restrict the tablet’s range to simply cover one of my monitors. But this introduced the problem that should I ever want to switch monitors, it would be quite a pain (the name xrestrict describes this behavior, though it may be used to cover an area larger than the X Screen now). I would have to recalculate this matrix, which I had by now realized was more complicated than simply chopping the screen up. I could rather trivially calculate these two matrices myself and store them somewhere, but it occurred to me that others may have similar issues, and resolved to make a reusable utility. This decision was re-enforced by the thought that this problem likely affects more people than myself. In fact, my hasty dump to github and this blog post are a direct result of discovering others with similar problems to mine, and a hope that xrestrict could help them.

Enter xrestrict

xrestrict‘s core functionality is to calculate the correct “Coordinate Transformation Matrix” to map a tablet device’s effective area to a portion of your screen without distortion. Originally, xrestrict needed to be told the XID of the pointer device. From there it discovers the X and Y axis bounds of that pointer device. It also discovers the size of the X Screen, which may be larger than any single monitor on multi-monitor systems. In fact, typically it is large enough to contain all monitors side by side, except when monitors are mirrored. I want to ensure it’s useful for as many people as possible, regardless of linux experience.

Improved Scaling With xrestrict Preserving Aspect Ratio

Basic Usage

The current recommended usage is to invoke

xrestrict -I 

This will prompt you to use your tablet device to select the CRTC you wish to restrict your tablet to. Use your tablet cursor to click somewhere on the screen on the monitor you want to use.

Alternatively, if you want your tablet to be able to access all monitors, but you want to correct the aspect ratio, you can invoke

xrestrict -I --full 

This will “restrict” you tablet to the entire virtual X Screen. The only downside to this is, unless your tablet’s aspect ratio matches your Screen’s aspect (in which case you never needed xrestrict to begin with!) some portion of your tablet’s surface will be mapped to portions of the Screen which are not displayed on any monitor. This means that portions of your tablet will be effectively useless. If you want to avoid this there are alternative options to control how your tablet’s surface is mapped. View xrestrict’s usage for more information.

Pending Problems

Due to feature “creep”, xrestrict is no longer a descriptive name, as it does more than restrict a tablet device to a particular monitor. I’d hate to keep a non-descriptive name like that for historical reasons when that “history” is only a few weeks old. The hid-huion drivers also report having an X rotational axis and a Z axis, which causes a small amount of confusion for X clients. Gimp, for instance, seems blissfully unaware that X tells it specifically what axis is the pressure axis, and instead uses a numbering scheme. This scheme causes one of the mis-reported axes (Z if I remember correctly) to be mapped to pressure for Gimp’s paint tool by default. I suppose this is a Gimp bug as the information describing the type of each axis is available, but my tablet also shouldn’t be reporting it has axes it actually doesn’t. For the time being I’ve worked around this by configuring the tablet in Gimp, which allowed me to easily remap the correct axis to pressure. At some point I may explore correcting this in the kernel driver, but I’m not particularly comfortable making modifications which “work for me” on my hardware, on a driver which is intended to be generic for Huion tablets. I don’t have the hardware to test it on all possible combinations, though perhaps through the DIGImend project I can get some regression testing.

The hid-huion drivers I modified still report painfully stupid buttons. I can either remap them in, or attempt to remap them using udev’s hwdb. For now I’m planning to use hwdb so it can potentially be compatible with Wayland in the future. My tablet appears as three identically named devices in, which makes it difficult for the user to identify the correct device id at a glance. For this reason, requiring the user to find the XID of their tablet is a poor user interface (even for a command line application!). Between my initial draft of this article and its publishing, I’ve added the “interactive” mode to xrestrict, which eliminates the need to discover the XID manually. Another user friendly feature I want is to provide a persistent identifier to the user’s tablet device. By providing the identifier, the user could for instance use interactive discovery once to determine their tablet, and from then on have a small button which invokes xrestrict --device-id-file=~/.config/xrestrict/mytablet which automagically restricts their tablet to CRTC 0.


  1. xrestrict
  2. DIGImend Project
  3. DIGImend Old Wiki
Inexpensive Monoprice Tablet

Expect this Inquisition!

One of the things I’ve been working on lately, aside from my resumé and everything else, is a little object inspector for Python objects. I plan on using it to debug PyPy translation test failures, as the object trees are often complex, and I think this will help me.

A screenshot of the inspector in action
A screenshot of the inspector in action

I’ve put the code up on github at the basic usage is to call inspect_object(my_object) or inspect_dictionary(my_dict) to inspect either an object or dictionary. This specifies a root node to start browsing at. You’ll need a bleeding edge version of urwid that has the treetools module

Expect this Inquisition!

Pythonic Pumpkin Carvings

PyPy and the JVM

I admit I’ve been lazy lately, but I have been working on a few things, the most interesting of which, is translating PyPy to JVM. Antonio Cuni did a lot of work getting PyPy to translate to the .NET platform, and to allow JIT generation for that platform. As a mixed blessing, his code has remained in a branch that is massively out of sync with trunk (here). The good news is this allowed me to very quickly revive JVM support in his branch, the patch is somewhere on, I’ll track it down later. Antonio expressed to me that he would like to port his changes by hand to trunk, rather than attempting any sort of merge. Because of that, I’ve focused my attention on trunk (here). The PyPy translator is extremely powerful, and consequently, extremely complex. I feel I’ve wrapped my head around a significant portion of it, and I was able to address one major issue preventing translating trunk to .NET and the JVM. There’s plenty more to go, but I’m still working.


I’ve slowed down a lot, but I recently added the support code for array broadcasting, which is essential for proper handling of arrays, after that, there should be a solid enough foundation to implement most of NumPy in pure python on top of micronumpy arrays, and through profiling, implement some in RPython. I may try my hand at improving performance before I finish broadcasting support, I’m not sure.

All Hallows Eve

My pumpkin with the Python logo

Not that this excuses my sloth, but here’s what I carved today :-). I’d never tried to carve a pumpkin to be semi-transparent before, but it turned out quite well actually, better than my pumpkins usually do…

Pythonic Pumpkin Carvings

Performance Update

As promised, I haven’t just dropped micronumpy, I’m continuing to work on it. As of September 10th, 2010 micronumpy takes 1.28736s per iteration over the convolve benchmark, and NumPy on CPython takes 1.87520s per iteration. This is about a 31.3% speedup over NumPy, I didn’t record the exact numbers near the end of the SoC but I believe I’ve made things slower still… On the bright side, I’m passing more tests than ever, and support slicing correctly. On the downside, I have no idea why it’s slower, I eliminated a whole loop from the calculation, I expected at least a moderate gain in performance… I’m investigating now, so I’m keeping this short.

Performance Update

When All is Said and Done

In the Beginning

Back when I was young and naïve at the beginning of the summer, I proposed to continue the work that a few PyPy developers and I had worked on, a reimplementation of NumPy in RPython. The project holds a lot of promise, as PyPy can generate a JIT compiler for itself and its components written in RPython. With a NumPy array written in RPython, the PyPy JIT can see inside of it and from that can make far more optimizations than it could otherwise. Since the PyPy JIT is especially good at optimizing CPU/computationally expensive code, bringing the two together could go a long way to bridge the gap between Python performance, and statically compiled languages.

As luck would have it, my project was categorized by the Python Software Foundation as a NumPy project, rather than a PyPy project, whose developers I’d been bugging and asking questions for some time. I soon came into contact with Stéfan van der Walt, a member of the NumPy Steering Committee. After consulting with him and the NumPy mailing list, it was decided that most people would not find a super fast NumPy array very useful by itself. For it to matter to most people, it would need to be able to do everything that the existing NumPy array does, and someone brought up the point that there is a great deal of C and Cython code already written which interacts with NumPy arrays, and it’s important that my project would allow these things.

So my project ballooned to a huge size, and I thought I could handle it all. The new burden of full compatibility was to be attacked by porting NumPy to PyPy and providing an easy interface for switching to and from NumPy and micronumpy arrays. Unfortunately, this pursuit wasn’t very fruitful, as PyPy’s CPyExt isn’t yet equipped to handle the demands of a module as all encompassing as NumPy. I spent a fair amount of time simply implementing symbols to satisfy the dependencies of NumPy. I made some significant changes to NumPy which are currently sitting in my git repository on github. I don’t know what the future holds for them unfortunately (If the NumPy refactor is completed soon enough, I may be able to sidestep CPyExt which will be faster anyways).


Around midterms I had micronumpy arrays working reasonably well enough that they could run the convolve benchmark, and handily beat NumPy arrays (twice as fast is fairly impressive). However, the point is to demonstrate that the JIT can speed up code to near compiled code, theoretically removing the need to rewrite large portions of python code in C or Cython. By this time, it was becoming clear that getting NumPy to work with PyPy was not going to happen over the summer. I’ve adjusted my expectations, NumPy working on PyPy is still on my TODO list but won’t be completed this summer. This might be for the better anyways, as NumPy is being refactored to be less Python (and therefore CPython) centric, as a result, in the near future I may be able to completely avoid CPyExt and use RPython’s foreign function interface to call NumPy code directly.

The Final Stretch

One of the beautiful things about PyPy’s JIT is that it’s generated, not hard coded, so I didn’t have do to anything in order to have micronumpy be JITed. Unfortunately, in the past three days or so, I’ve discovered that my code no longer works with the JIT. I’ve done all I can to figure out what’s wrong, and I can’t fix it on my own. Diving into the JIT in the last 24 hours of the summer of code surely won’t bear any fruit. I’ve put up my distress signal on the mailing list, and hopefully this issue can be resolved in time to provide some awesome benchmark results. If not, at least I can get this resolved in the next couple of weeks and then move on to the other things I want to fix.

EDIT: Thanks to the help of the core PyPy developers we determined that the problem was that arrays allocated with the flavor “raw” were being accepted. Apparently these arrays still have length fields, by using rffi.CArray I was able to instruct PyPy to construct an array without a stored length field.

I’d also like to add that in the final hours, we added support for the NumPy __array_interface__ so that as soon as NumPy is working on PyPy, NumPy can take micronumpy arrays and do all sorts of useful things with them, and then when you need speed for simpler operations, you can pass your NumPy arrays to micronumpy (this side of the transaction hasn’t been implemented yet).

The End

So here we are at the end of the Summer of Code, and my project isn’t where I wanted it to be. Specifically, given my addition of slice support, performance has dropped to around 50% faster than NumPy, even farther from my goal, so that’s my top priority to address in the coming weeks. In my previous blog post I outlined what my plans are for the future (as I don’t like leaving things undone). Basically it comes down to:

  • Optimizing!
  • Minor compatibility fixes
  • Bridging NumPy and PyPy completely


I’d just like to thank Google very quickly, and specifically Carol Smith, who has done a great job of managing the Google Summer of Code this year. I thoroughly enjoyed the program, and would love to do it again given the chance. I’ve learned a lot about writing software, dealing with deadlines, and time management (which is a skill I’ve let atrophy…) this summer. And thanks to you who’ve taken interest in my project. If you want to check back occasionally, the summer may be over, but my project isn’t, and I’ll be sure to brag about benchmark results as soon as they’re more favorable :-).

I’d also like to thank my mentor, Stéfan van der Walt for his help throughout my project, for being supportive and understanding when unexpected problems occurred and set us back. And I’d like to thank Maciej FijaŃkowski for his support from the PyPy side. The rest of the PyPy developers have all been helpful at some point or another, so thanks to them too.

When All is Said and Done

The End is Nigh!


We’re already past the suggested pencils down date for the Google Summer of Code, and I’m certainly paying my penance for my previous sloth. Just last night I got the test suite passing again, after several hours of hacking. Advanced indexing is nearly done, which is wonderful. I currently have one slice-related issue, which I’ll hopefully be resolving in the next couple of hours.

The Next Week

As soon as I have this indexing done, it’ll be time to optimize. Maciej was kind enough to show me how to get tracing information, so that I can produce the most JIT friendly code I can. I probably will spend the next five days working on tuning that. The original goal, of course, was to be near Cython speeds using the JIT, and we were nowhere near that on the first pass (though twice as fast as CPython and normal NumPy). Unfortunately, with the addition of the advanced slicing, I may have made Cython speeds harder to achieve. Hopefully the JIT will be ok with my first pass with slicing, however I’m prepared to revert to “dumb slicing” for the end of the GSoC and resume advanced indexing support after it is over. I’d feel bad about that though, as that’s been my major stumbling block these past weeks. In the next few hours I need to make sure everything translates so that I have something to show to Stéfan tonight.


My biggest regret from the summer of code, is that I haven’t succeeded in porting NumPy to PyPy. This is something I hope to address in my free time this coming semester. This will require extensive work on CPyExt which is a complicated beast.

Additionally, I want to make sure that micronumpy is as useful as it can be, and that’s something that should be pretty easily accomplished in my free time. This will include implementing basic math operations, and some ufuncs. I may make a first pass at everything with naïve implementations written in applevel Python, then progressively optimize things. Depending on the progress for the refactoring of NumPy, I might be able to plug in some NumPy code for fast implementations of some things which would be great.

Back to work with me,
The End is Nigh!

Camping and NumPy


So I’ve been back from camping for around a week, and it definitely derailed my train of thought… Subsequently, I went to Reno, Nevada with my girlfriend to meet up with one of her friends. We stayed at a casino in a nice room for 30 USD which is pretty awesome. But enough about me…

I’m starting to worry a bit about my progress, the past two trips have put me behind (I’ve only been gone a cumulative five days, but the interim days were mostly unproductive as well. I tend to code straight through the night if I’m on a roll, because even the eight hours I would sleep might throw off my current train of thought, so this traveling has been unhelpful)


What I have been working on, is advanced array indexing in micronumpy. I’ve pretty much broken indexing for the moment, but out of this should come slicing and ellipses support, because we don’t all use simple indexing. I’m afraid that with this significantly more complicated indexing scheme, is going to come alot of overhead, so it may set back performance, we’ll see. I’ve tried my best to put the common case first (single dimensional index handled first, then simple multi-dimensional indices). I’m actually not sure how much PyPy will be able to optimize out via JIT compilation, since dynamic types become static for the JIT’s purposes. I may find that the extra complexity is irrelevant to the JIT-ed code.

There’s more to say, but I should get back to work 🙂

That’s all for the moment.

Camping and NumPy