sapiogram

6 ms·

181.
by sapiogram 2y ago
Does anyone know what they mean by "set-theoretic types"? I'm a PL nerd, but I've never heard this term before.
182.
by sapiogram 2y ago
Given how long arm64 took them, my guess is "3 years after major cloud vendors start offering risc-v VMs".
183.
by sapiogram 2y ago
4c/8t gives more throughput than 4c/4t.
184.
by sapiogram 2y ago
> there is hardly a reason to keep hyperthreading around. Performance is still a reason. Anecdote: I have a pet project that involves searching for chess puzzles, and hyperthreading improves throughput 22%. Not massive, but definitely no
185.
by sapiogram 2y ago
Wouldn't it be even more frustrating if your opponent was constantly picking up pieces when thinking?
186.
by sapiogram 2y ago
Earlier this years, Github would routinely say "2 years ago" for comments made 11 months ago.
187.
by sapiogram 2y ago
What's RISC about adding more registers?
188.
by sapiogram 2y ago
For large (>10m) telescopes, it doesn't matter much, because no one is building rockets with big enough fairings. Even if the cost of launching to LEO was $0, in-orbit unfolding/assembly is a bottomless hole of engineering comp
189.
by sapiogram 2y ago
The largest filled-aperture radio telescope in the world has a 500-meter diameter, and reportedly only cost $180 million. For comparison, the International Space Station costs $3 billion per year just to operate. Doing stuff in space is ex
190.
by sapiogram 2y ago
They will survey the entire sky every week or so. At the end, you can create a nice 60 fps timelapse of many objects in the southern sky.
191.
by sapiogram 2y ago
Note that rustc will not emit this instruction by default, to support older CPUs. Without `target-cpu=haswell` or similar, the operation will be quite slow, but still O(1) of course.
192.
by sapiogram 2y ago
Thank you for your response. I think your final remark was unnecessary though, I think less people asking stupid questions just leads to more confidently incorrect non-questions, which is not an improvement.
193.
by sapiogram 2y ago
> You are still dependent on foreign inputs that are dependent on the Dollar or Yuan payments For Russia in particular, do we have some idea of how true this actually is? Russia lacks a lot of things, but natural resources is generally n
194.
by sapiogram 2y ago
> 1) the memory speeds are increasing and approach those of the CPU ― that's definitely not true today; What does "speed" mean here? Latency? Bandwidth?
195.
by sapiogram 2y ago
I'm not familiar with ocaml, but the LEA thing can reasonably be fixed by a better compiler, no? Is there no LLVM backend for ocaml?
196.
by sapiogram 2y ago
That's an entirely different concept imo. Overloading in Rust would be to have separate `Vec::new()` and `Vec::new(capacity: usize)` functions, which is not allowed. Case in point: Java does allow exactly this, while also allowing cl
197.
by sapiogram 2y ago
That can't really be the main reason, Rust is doing just fine without it.
198.
by sapiogram 2y ago
> I don't see why did you label Aliaksandr Valialkin, the author, an "expert". I mean, he's no dummy but what exactly makes him an expert on optimizing Go code? I was trying to convey the meaning of "far more exp
199.
by sapiogram 2y ago
The Go code is already hyper-optimized by experts, just not by the blog authors, so you don't read about it here. As someone who has tried to write high-performance Go code on occasion, I can assure you that a ton of digging would have
200.
by sapiogram 2y ago
I haven't dug into the code, but it's clear that the Golang library has had a ton of optimization work put into it by very knowledgeable people. Techniques like object pooling are highly error prone, and certainly not "out of
201.
by sapiogram 2y ago
> I am guessing it’s the Rust stuff that caused the performance hit From Github's code stats, Zed is 97.3% Rust, so Rust is almost certainly both the problem and the solution here.
202.
by sapiogram 2y ago
Nope, and I don't think Google will publish concrete numbers anytime soon, if ever.
203.
by sapiogram 2y ago
> lower emissions Do you know of any cloud providers that publish actual data for this? Preferably verifiable, but I'll take anything at this point.
204.
by sapiogram 2y ago
Why?
205.
by sapiogram 2y ago
Javascript is even better, by not having multithreading at all. I am not joking, Go is much worse than Java and C#, but Javascript and Rust are the only mainstream languages where I've seen non-experts reliably write correct concurrent
206.
by sapiogram 2y ago
> Please. Stop posting about async/await here. I agree with everything else you said, but that's all the more reason to post it imo. The misconceptions aren't going away just because you don't hear about them, and eve
207.
by sapiogram 2y ago
I think your theory makes a lot of sense if you reverse the causality: Mammals have evolved to not engage in cannibalism to avoid prions. Other diseases as well, of course.
208.
by sapiogram 2y ago
Well, names are useful for the same reason people's names are useful. The rest just kinda happens naturally, I think.
209.
by sapiogram 2y ago
Is there a writeup somewhere by someone with a slightly stronger command of English?
210.
by sapiogram 3y ago
https://youtu.be/THJP_HB5HEk?si=gmdl6vIBVREc5rpA Link for the lazy

More ›