Ah. We're back to the days of Emacs' old `M-x psychoanalyze-pinhead`, then. (Psychoanalyze-pinhead ran the Eliza chat-bot and fed it bizarre quotations collected from the Zippy the Pinhead comics.)
At least for the CPU/GPU split, llama.cpp recently added a `--fit` parameter (might default to on now?) that pairs with a `--fitc CONTEXTSIZE` parameter. That new feature will automatically look at your available VRAM and try to figure out a good CPU/GPU split for large models that leaves enough room for the context size that you request.
Hardware floating point was rare before the 486 DX and Pentiums. Not to mention that Integer<->FP conversion was slow. And division of any kind has always been slow. So you'd see a lot of fixed-point math approximations with power-of-two divisors so that you can shift-right.
I think the point the GP was trying to make is that the GitHub UI ought to be able to allow you to submit a branch with multiple well-organized commits and review each commit separately with its own PR. The curation of the commits that you'd do for stacked PRs could just as easily be done with commits on a single branch; some of us don't just toss random WIP and fixup commits on a branch and leave it to GitHub to squash at the end. I.e., it's the GitHub UI rather than Git that has been lacking.
(FWIW, I'm dealing with this sort of thing at work right now - working on a complex branch, rewriting history to keep it as a sequence of clean testable and reviewable commits, with a plan to split them out to individual PRs when I finish.)
> I think the point the GP was trying to make is that the GitHub UI ought to be able to allow you to submit a branch with multiple well-organized commits and review each commit separately with its own PR.
That's what this feature is, conceptually. In practice, it does seem slightly more cumbersome due to the fact that they're building it on top of the existing, branch-based PR system, but if you want to keep it to one commit, you can (and that's how I've been working with PRs for a while now regardless, honestly).
They confirmed in other comments here that you don't have to use the CLI, just like you don't have to use gh in general to make pull requests, it's just that they think the experience is nicer with it. This is largely a forge-side UI change.
> I think the point the GP was trying to make is that the GitHub UI ought to be able to allow you to submit a branch with multiple well-organized commits and review each commit separately with its own PR
So the point he's trying to make is that Gituhub UI should support Stacked PRs but call them something else because he doesn't like the name?
Cave Johnson here. I'll be honest, we're throwing science at the wall here to see what sticks. No idea what it'll do. Probably nothing. Best-case scenario, you might get some superpowers.
I've been using the 1M window at work through our enterprise plan as I'm beginning to adopt AI in my development workflow (via Cline). It seems to have been holding up pretty well until about 700k+. Sometimes it would continue to do okay past that, sometimes it started getting a bit dumb around there.
(Note that I'm using it in more of a hands-on pair-programming mode, and not in a fully-automated vibecoding mode.)
My take as a graphics programmer is that angles are perfectly fine as inputs. Bring 'em! And we'll use the trig to turn those into matrices/quaternions/whatever to do the linear algebra. Not a problem.
I'm a trig-avoider too, but see it more as about not wiggling back and forth. You don't want to be computing angle -> linear algebra -> angle -> linear algebra... (I.e., once you've computed derived values from angles, you can usually stay in the derived values realm.)
Pro-tip I once learned from Eric Haines (https://erich.realtimerendering.com/) at a conference: angles should be represented in degrees until you have to convert them to radians to do the trig. That way, user-friendly angles like 90, 45, 30, 60, 180 are all exact and you can add and subtract and multiply them without floating-point drift. I.e., 90.0f is exactly representable in FP32, pi/2 is not. 1000 full revolutions of 360.0f degrees is exact, 1000 full revolutions of float(2*pi) is not.
Hah. I think we're and the author of both articles on the same page about this. (I had to review my implementations to be sure). I'm a fan of all angles are radians for consistency, and it's more intuitive to me. I.e. a full rot is τ. 1/2 rot is 1/2 τ etc. Pi is standard but makes me do extra mental math, and degrees has the risk of mixing up units, and doesn't have that neat rotation mapping.
Very good tip about the degrees mapping neatly to fp... I had not considered that in my reasoning.
If you want consistency, you should measure all angles in cycles, not in radians.
Degrees are better than radians, but usually they lead to more complications than using consistently only cycles as the unit of measure for angles (i.e. to plenty of unnecessary multiplications or divisions, the only advantage of degrees of being able to express exactly the angle of 30 degrees and its multiples is not worth in comparison with the disadvantages).
The use of radians introduces additional rounding errors that can be great at each trigonometric function evaluation, and it also wastes time. When the angles are measured in cycles, the reduction of the input range for the function arguments is done exactly and very fast (by just taking the fractional part), unlike with the case when angles are measured in radians.
The use of radians is useful only for certain problems that are solved symbolically with pen on paper, because the use of radians removes the proportionality constant from the integration and derivation formulae for trigonometric function. However this is a mistake, because those formulae are applied seldom, while the use of radians does not eliminate the proportionality constant (2*Pi), but it moves the constant into each function evaluation, with much worse overhead.
Because of this, even in the 19th century, when the use of radians became widespread for symbolic computations, whenever they did numeric computations, not symbolic, the same authors used sexagesimal degrees, not radians.
The use of radians with digital computers has always been a mistake, caused by people who have been taught in school to use radians, because there they were doing mostly symbolic computations, not numeric, and they have passed this habit to computer programs, without ever questioning whether this is the appropriate method for numeric computations.
Unfortunately, IEEE Std 754 contains a huge mistake, which has been followed by some standard libraries for programming languages.
As an alternative to the trigonometric functions with arguments measured in radians, it recommends a set of functions with arguments measured in half-cycles: sinPi, cosPi, atanPi, atan2Pi and so on.
I do not who is guilty for this, because I have never ever encountered a case when you want to measure angles in half-cycles. There are cases when it would be more convenient to measure angles in right angles (i.e. quarters of a cycle), but half-cycles are always worse than both cycles and right angles. An example where measuring angles in cycles is optimal is when you deal with Fourier series or Fourier transforms. When the unit is the cycle that deletes a proportionality constant from the Fourier formulae, and that constant is always present when any other unit is used, e.g. the radian, the degree or the half-cycle.
Due to this mistake in the standard, it is more likely to find a standard library that includes these functions with angles measured in half-cycles than a library with the corresponding functions for cycles. Half-cycles are still better than radians, by producing more accurate results and being faster, but it may be hard to avoid some scalings by two. However, usually it is not necessary to do a scaling at every invocation, but there are chances that the scalings can be moved outside of loops.
Such functions written for angles measured in half-cycles can be easily modified to work with arguments measured in cycles, but if one does not want to touch a standard library, they may be used as they are.
When one uses consistently the cycle as the unit of angle, which is consistent with measuring frequencies in Hertz, i.e. cycle per second, instead of measuring them in radian per second, one must pay attention to the fact that a lot of formulae from most handbooks of physics are incorrect. Despite the claim that those formulae are written in a form that is independent of the system of units, this claim is false because many formulae are written in a form that is valid only when the unit of angle is the radian.
For example, all formulae for quantities related to rotation movements, as they are written in modern handbooks contain the "radius". This use of the "radius" creates a wrong mental model of the rotational quantities, both for students and also even for many experienced physicists.
In reality, in all those formulae, e.g. in the definition of the angular momentum, in order to obtain the correct formulae one must replace the "radius" with the inverse of the curvature of the trajectory of the movement. Thus the angular momentum is not the product of the linear momentum by the radius, but it is the ratio between the linear momentum and the curvature.
Then one must use the correct definition for the curvature. Most handbooks define the curvature as the inverse of the radius. This is a wrong definition, which is based on the non-explicit assumption that angles are measured in radians.
The correct definition of the curvature is as the ratio between rotation angle and length, for the movement, i.e. more precisely it is the derivative of the rotation angle as a function of the length of the curve on which something moves. When angles are measured in radians, the curvature is the inverse of the radius. When angles are measured in cycles, the curvature is the inverse of the perimeter. Thus with angles measured in cycles the angular momentum is defined as the product between the linear momentum and perimeter. Similarly for the other rotational quantities, like angular velocity and acceleration, moment of inertia and so on.
Great's great info on curvature! Also don't see the use of the half cycle, although obviously see use of cycle. (As was thinking in terms of them anyway, just with a tau term to convert to radians.)
Author here. I have a JavaScript port of my automated test suite (https://github.com/a-e-k/canvas_ity/blob/main/test/test.html) that I used to compare my library against browser <canvas> implementations. I was surprised by all of the browser quirks that I found!
But compiling to WASM and running side-by-side on that page is definitely something that I've thought about to make the comparison easier. (For now, I just have my test suite write out PNGs and compare them in an image viewer split-screen with the browser.)
Author here. There's no AI-generated code in this. But yes, security hardening this has not been a priority of mine (though I do have some ideas about fuzz testing it), so for now - like with many small libraries of this nature - it's convenient but best used only with trusted inputs if that's a concern.
Author here. No vibe-coding, all human-written. Are you thinking of my use of GitHub emoji on the section headings in the README? I just found they helped my eye pick out the headings a little more easily and I'd seen some other nice READMEs at the time do that sort of thing when I went looking for examples to pattern it off of. I swear I'd had no idea it would become an LLM thing!
Or better yet, pitting Eliza vs. Parry (https://logic.stanford.edu/complaw/readings/elizaandparry.pd...), where Parry was meant to simulate a paranoid schizophrenic. That was 1973, more than 50 years ago.
Everything old is new again.
reply