@dotstdy@mastodon.social titelbild
@dotstdy@mastodon.social avatar

dotstdy

@dotstdy@mastodon.social

phd in branch predicting, ceo of euro gaming

Dieses Profil is von einem föderierten Server und möglicherweise unvollständig. Auf der Original-Instanz anzeigen

fasterthanlime , an Random Englisch
@fasterthanlime@hachyderm.io avatar

I'm reading up about C++ move semantics and I don't understand why anybody in the history of humankind has ever tried defending this language.

dotstdy ,
@dotstdy@mastodon.social avatar

@fasterthanlime I find it to be a 50/50 whether the place people have typed out std::move actually makes sense as a move and works as a move. And then somebody changes a struct definition in a random unrelated way and you're no longer generating the automatic move implementation and you get a performance problem. So good.

dotstdy ,
@dotstdy@mastodon.social avatar

@fasterthanlime no it's not possible, because when you move from an object you always have to leave it in a valid state (since the destructor needs to be able to run) so it's not possible to statically block moving from an object twice.

foone , an Random Englisch
@foone@digipres.club avatar

You know the "one dimensional thinking" trope, where people run directly away from falling/rolling objects, rather than run a small distance to the side?

I just saw a letsplayer do that in a space game! With full 3D movement!

In deep space, they realize they've over-accelerated, and are heading towards a small satellite 13km away at 500m/s. They go "oh no, I'm going to crash into it again!" and slam on the retrorockets.
To slow down.

dotstdy ,
@dotstdy@mastodon.social avatar

@foone don't worry, bruce willis is ready

fasterthanlime , an Random Englisch
@fasterthanlime@hachyderm.io avatar

HOW MUCH??

dotstdy ,
@dotstdy@mastodon.social avatar

@fasterthanlime @paddi tfw your code is so large that firefox forgets how to render text by the end of it.

karolherbst , an Random Englisch
@karolherbst@chaos.social avatar

I wonder if Freedesktop.org as an organization has to grow into one actively funding projects, which need to be done.

But on the other side there isn't really enough money available (unless I'm mistaken) in order to fund people work on things full-time.

So I do wonder if FDO should just collect money from desktops and then hire people to work on things in a desktop agnostic way, where foundations of projects could only do that specific to theirs.

dotstdy ,
@dotstdy@mastodon.social avatar

@karolherbst i think it would be tough, because you'd need a lot more money that what is probably available. and also the politics of what that means by "desktop-agnostic" exactly. only paying desktops?

karolherbst , an Random Englisch
@karolherbst@chaos.social avatar

today: debugging a memory corruption which goes away with libasan...

dotstdy ,
@dotstdy@mastodon.social avatar

@karolherbst crashes on shutdown are just fast paths for exiting! It's a feature!

karolherbst , an Random Englisch
@karolherbst@chaos.social avatar

I'm considering building a home server which I kinda want to use as a private git forge, but maybe also distributed compiler and maybe other things.

And I want it passively cooled with a thermal budget of around 100W for the CPU.

But what I do wonder about is what CPU to choose. Do we have proper ARM boards now in the ITX form factor? Should I go with AMD or Intel? What are the power efficient options here and what option has the lower idle power consumption?

Anybody any experience with that?

dotstdy ,
@dotstdy@mastodon.social avatar

@karolherbst 100W is kinda a lot, so you should be able to get well under that. Modern AMD and Intel desktop parts both tend to idle quite high, AMD a bit higher, but generally speaking it's not hugely impactful either way. No idea about arm dev boars though. I wouldn't really expect that they're vastly different in practice.

dotstdy ,
@dotstdy@mastodon.social avatar

@karolherbst I would recommend seeking out actual numbers from people who have used that specific cooler with similar CPUs, my expectation is that unless you're in ideal conditions you might struggle to hit the advertised numbers :')

dotstdy ,
@dotstdy@mastodon.social avatar

@karolherbst (at least judging by the reviews of passive cooling solutions I've seen before)

dotstdy ,
@dotstdy@mastodon.social avatar

@karolherbst yeah if you're willing to underclock you can run lots of stuff, though the question does become a bit, why not buy a cheaper sku :P

foone , an Random Englisch
@foone@digipres.club avatar

Shout out to the DoorDash driver on a moped trying to do a mid intersection lane split to get ahead of the waymo self driving cab.
You have WAY more faith in Google's code than I do!

dotstdy ,
@dotstdy@mastodon.social avatar

@foone I knew I'd arrived in the tech capital of the world when I saw a waymo taxi stuck in the middle of the road due to some orange cones on the street, and the passenger seemingly abandoning ship into a lane full of traffic. The future is here.

fasterthanlime , an Random Englisch
@fasterthanlime@hachyderm.io avatar

Although I've known about it for years it's just now going click in my brain how big of a hack the syn crate is and how much I dislike it 🙃

https://lib.rs/crates/syn

dotstdy ,
@dotstdy@mastodon.social avatar

@fasterthanlime cranelift with optimizations is still much slower than debug mode, iirc. perhaps you could find a better balance than -O0 though, specifically for build scripts. even with LLVM.

  • Alle
  • Abonniert
  • Moderiert
  • Favoriten
  • •
  • random
  • haupteingang
  • Alle Magazine
    •