Athas

26 ms·

241.
by Athas 7y ago
My K is weak, but I think that implementation is also putting the fixed point iteration on the outside, isn't it?
242.
by Athas 7y ago
People are so quick to reject K and APL-style languages for superficial reasons that they never get to the deep and interesting reasons! I am mostly familiar with APL, but I think the things I appreciate and dislike are about the same in K
243.
by Athas 7y ago
None of these are Standard ML! There are people who prefer SML for various reasons. Somewhat unexpectedly, there is in fact already a very good SML-to-browser solution, namely the unfortunately named SMLtoJs: https://github.com&
244.
by Athas 7y ago
I don't think this is true, or at least as simple a story as "Vulkan is faster". OpenCL is already a pretty good and modern API, and computation in Vulkan doesn't seem like it's the main point. It's my impres
245.
by Athas 7y ago
> Oh, actually: Stalin probably does. It's an R4RS compiler. Good luck getting it to compile, though. It took me a few hours and a lot of code changes to get it to compile half a decade ago (I wanted to compare the one I was writing
246.
by Athas 7y ago
Most of the APLs don't automatically parallelise array operations either, because it's hard to know automatically when it's worth it. This is exacerbated by the fact that most APLs don't have a compiler, so the granular
247.
by Athas 7y ago
> However, if you want to add new concepts into the IR, you should do it in the language that the IR is defined in. And isn't that what most users will want to do? The point of MLIR is to provide a framework for writing mid-level I
248.
by Athas 7y ago
> By the way, do you think futhark could support automatic differentiation? I don't see why not. It's similar to other languages that have been the subject of AD research: https://www.microsoft.com/en-us/r
249.
by Athas 7y ago
I'm the developer of a language that I think is in the target audience of MLIR: https://futhark-lang.org I'm still not completely certain MLIR addresses the difficulties I face, which is exactly to design the specific
250.
by Athas 7y ago
It's sometimes a good idea to do it that way in practice, but it's still a conflation of two conceptually distinct processes. I think it is a bad approach when teaching compiler implementation, as it means you avoid the extremely
251.
by Athas 7y ago
Old techniques are not necessarily more fundamental, they are just more primitive. (And often, but not always, more efficient.) In particular, old compiler texts will often do things like emit code directly from the parser, try to minimise
252.
by Athas 7y ago
By the taxonomy of the diagram, pure APL-style array programming is just first-order functional programming. Real array programming languages also introduce features from other branches of the tree (especially imperative programming and cl
253.
by Athas 7y ago
This does not save you if that null pointer is supposed to point to an array, and subsequent index operations move it out of the protected first page (or the first couple, I forget how modern Unix does it).
254.
by Athas 7y ago
Nothing is particularly simple about X11, especially in its modern incarnations. It's much simpler to have an API that maps cleanly to what the hardware is actually doing, which is what Wayland is. You can still prefer X11 for its fea
255.
by Athas 7y ago
I have added packages to Homebrew, and they install man pages when available. Homebrew provides a pretty simple mechanism for doing so. The problem is that many projects just prefer web documents to man pages, likely because the man page to
256.
by Athas 7y ago
I think all of the ACM conferences I contribute to have stopped actually printing physical proceedings. The consistent formatting is by a LaTeX document class, with an automated system that checks for bad fonts, text going into the margin,
257.
by Athas 7y ago
Co-dfns is interesting because I think it is the first compiler that has ever been written in a fully data-parallel style. But why does that make APL as a language particularly relevant, except that it was the implementation language?
258.
by Athas 7y ago
> How are you not? I research compilers and type systems. If the RAM dies while the compiler is running, you rerun the compiler on a new machine. A lot of computer science abstracts away the notion of hardware failure, because otherwis
259.
by Athas 7y ago
I still think it's interesting to reflect on the mindset that leads to this conclusion. As far as I have been able to determine (although I'm not in the aerospace field), the 737 MAX is procedurally identical to the NG, except
260.
by Athas 7y ago
A quite bright bachelor's student of mine did something similar for his undergraduate thesis[0]. Conceptually, he represented neural networks as functions from input to output, and connecting two networks required that the output of t
261.
by Athas 7y ago
If you have precise reference counts (which is often achievable in array languages because they don't have cycles), then you can dynamically do efficient in-place updates if the array you are updating has a count of 1. Uniqueness types
262.
by Athas 7y ago
> As I remember it now, this was taught as a bittersweet moment in our history. We united, we were brave and fought a stronger army but still we were defeated and our chief was captured and killed by the enemy. Isn't modern France m
263.
by Athas 7y ago
I'm always a sucker for parallel functional programming, but I could not find any details on how specifically it works (and importantly: performs) in Clio.
264.
by Athas 7y ago
It is definitely a problem that contributing to Rosetta Code is only possible through a clumsy WikiMedia interface. This makes it difficult to test and maintain the solutions for a given language. I'm not convinced a fork is the best
265.
by Athas 7y ago
Why is it important to ensure that the implementations are comparable? Why would someone contribute one that isn't? And if a language has to do something unusual because of specific quirks, I'd argue that's interesting by
266.
by Athas 7y ago
APEX? Does it still work?
267.
by Athas 7y ago
What did you find that depends on SaC?
268.
by Athas 7y ago
This is what I would recommend, but there are also other OpenCL stacks you could use. At least two from AMD (amdgpu-pro and ROCm), and and there is also one from the Mesa project (Clover).
269.
by Athas 7y ago
Yes, there is supposedly Tensorflow support in ROCm. I have never used it, so I cannot say how well it works.
270.
by Athas 7y ago
But CUDA C++ is a single-source programming language incorporating both CPU and GPU parts, so what is the one-to-one replacement for writing a kernel and launching it with the angle bracket syntax? Looking at the whitepaper[0] it looks lik

More ›