cbhl

no profile record

投稿

We tried Go's experimental Green Tea garbage collector and it didn't help perf

2 ポイント投稿者 cbhl9 か月前0 コメント

コメント

22 日前議論

It still sucks to be on the receiving end of genuine surprise.

I found the most helpful reframing is to replace the words and emotions with ones that encourage learning and question-asking. For example you can try being excited instead of surprised, or say something like, "that's a great question, let's figure it out together."

Going through the Fermi estimation in the xkcd comic Ten Thousand also helped me to be a lot less genuinely surprised when someone didn't know something: https://xkcd.com/1053/

27 日前議論

I remember watching Gary Bernhardt give this talk live at the Canadian Undergraduate Software Engineering Conference (CUSEC) back in 2014. PNaCl had just come out the year prior, and Google was using it to cross-compile, run, and sandbox OpenSSH and RDP clients inside of Chrome and ChromeOS, and the Mozilla/Firefox folks counter-proposed asm.js as a response.

At the time I just thought it was funny. Now I find it surprising how much of these ideas ended up sticking around.

2 か月前議論

> still had to use their human words to prompt the AI

In that case, I would like to know what prompt they used so that I can learn to use the AI tools better.

Like, we've had "let me google that for you" since 2008...

3 か月前議論

Looks like the docs start here: https://developer.android.com/tools/agents/android-skills#us...

There's a link to a repo or you can use the CLI tool to init the skills

3 か月前議論

To be fair they "only" have a 20 Mbps laser uplink from the capsule to Earth, and that's shared between all the systems and uploading images/video.

4 か月前議論

it doesn't give a sense of how much the Neo moves the needle or grows the pie though. Mac and iPad are each 8% of Apple's business. iPhone is 50%.

for Apple to focus on fixing things in macOS again (as opposed to just backporting stuff from their other lines) Mac would need to grow by a lot.

8 か月前議論

Shouldn't it be the other way around if the population is increasing? Every minute one is born = 1440 born/day, every minute and a sixteenth ~= 1335 dead/day for a net population increase of 105/day.

9 か月前議論

https://archive.is/qb2h8

10 か月前議論

The last time I played with NFTs it was about $50 USD just in processing fees to mint the token; not sure if that has changed since then. So it's not necessarily the case that the author is making a bunch of money off of this project; they might actually be just passing along costs.

10 か月前議論

Assuming that the boxes are 13.25oz/18.25oz, looks like an updated recipe could be:

- 2 boxes cake mix

- 3 eggs (rounding up from ~2.9 eggs)

- 1/2 cup neutral oil (rounding up from ~0.48c)

YMMV

11 か月前議論

Huh.

https://kaisercougarconnection.com/2784/news/musical-trains-...

My impression is that all of the Yamanote line stations are above ground -- I'd have expected it to be possible to have "one button plays the right sound at each station" if you used a standard phone's GPS to figure out which station you were at.

3 年前議論

The author, Fabrice Bellard, is well-known for his work on QEMU (among other things), but I'd treat the rest of the posts as "idea sharing" at an early stage. For example, while BPG (Better Portable Graphics) didn't take off, the core idea (image formats based on video codecs) can be found in WebP (VP8, VP9) and AVIF (AV1) and HEIF (HEVC/H.265).

For this, I'd look at comparable work in the audio codec space, including Lyra (https://arxiv.org/abs/2102.09660) and SoundStream (https://arxiv.org/abs/2107.03312).

For text compression, I think it would be novel if you could get a lossy but semantically equivalent decompression that was resilient to the exact hardware used for inference. I don't think that's what happened here, though, given the requirement for "exact same GPU and model".

4 年前議論

If I recall correctly, there was also a lot of controversy over the personal views of some folks in ENS leadership. If someone does analysis of these users I'd consider whether their tweets indicate a trend of being on a particular side of said controversy.

9 年前議論

In that case, then you can to go back to

    int mid = (low + high) / 2;

but you'd have to constrain low and high to be in the range [-2^30, 2^30 - 1] so as to not overflow a (assuming 32-bit) signed integer. (And you're probably using 64-bit floating point, or similar, which has other fiddly bits.)

However, the article specifically talks about merge sort in arrays, so low (your "a") is always >= 0. The discovery was "oh, we went decades before someone had to sort an array with more than 2^31 elements". Frankly, it's akin to when we ran out of IPv4 addresses -- when it was originally built, there was some range over which the computation was defined to be correct, and we later discovered that we wanted to do computation outside that range.

9 年前議論

IIRC, the git client is deprecated; the mercurial one is meant to replace it for the use cases where you would want a DVCS client interfacing with Piper in the first place.

10 年前議論

If each half of the controller is similar to a Wii Remote (+-pad and two or three buttons and an accelerometer/gyroscope) then it'll be plenty to play a lot of Wii U games.

12 年前議論

There is no reason why couldn't it be made out of more functions. Absolutely none. In fact, you can fork the repository and do the refactoring yourself, right now.

If I had to guess, the functions are tied pretty closely to how the author is parsing the file; "next" (next token), "expr" (expression), "stmt" (statement), and "main".

As for the project being called "C in 4 functions"; at best, I'd argue that's just a linkbait-y title since it's not actually C (it's a subset). I don't have a problem with the code _per se_; just the title.

12 年前議論

I would be worried about handling the not-context-free parts of the language: http://trevorjim.com/c-and-cplusplus-are-not-context-free/

How would you handle these?

12 年前議論

I actually found the code surprisingly easy to read; "tk" stands for "token", "ty" for "type", and so forth.

It's worth noting that compilers don't pop out of thin air -- you have to start with something simple in order to compile a more complicated compiler. Bootstrapping your own self-hosting compiler is a useful academic exercise, and you should try it sometime if you haven't already: http://en.wikipedia.org/wiki/Bootstrapping_(compilers)

12 年前議論

Judging from the comments in c4.cpp, it probably only supports enough of a subset to compile itself.

Granted, while building a parser that can parse (let alone compiling) the full C language is nontrivial, any undergrad should be able to build a parser and compiler for a sufficiently simple subset of it. (In my undergrad, we used this subset to build a "compiler" in second year: https://www.student.cs.uwaterloo.ca/~cs241/wlp4/WLP4.html)