Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Does memory leak? (1995)

groups.google.com

Post reply on HN
Print view View original

131–140 of 289 posts

Re: Does memory leak? (1995)

#131

Post by dahart »

alex_young wrote: post #124
dahart wrote: post #115

Earlier quoted context omitted.

I had a family friend who worked on missiles and drones and other defense systems. He was really one of my dad’s running buddies, and he was a super nice guy, had 4 kids, went to church, etc. One day, I believe during the Iraq occupation, maybe ~12 or 13 years ago, I asked him very directly how he felt about working on these killing machines and whether it bothered him. He smiled and asked if I’d rather have the war…

Costa Rica hasn’t had a standing military since 1948. They are in one of the most politically unstable parts of the world and do just fine without worry of invasion. The US hasn’t been attacked militarily on its own soil in the modern era. The US military monopoly hasn’t prevented horrific attacks such as 9/11 executed by groups claiming to be motivated by our foreign military campaigns. I think there is a valid ques…

Of course there is a valid question about the morals of war technology. You are absolutely right about that, and I am not even remotely suggesting otherwise. Like I said, I don’t think I would ever choose to work on it.

There’s a vast chasm in between right and wrong though. There can be understanding of others’ perspectives, regardless of my personal judgement. And there is also a valid question and tightly related question here about the morals of mitigating damage during a military conflict, especially if the mitigation prevents innocent deaths. If there’s a hard moral line between doctors and cooks and drivers and snipers and drone programmers, I don’t know exactly where it lies. Doctors are generally considered morally good, even war doctors, but if we are at war, it’s certainly better to prevent an injury than to treat one.

The best goal in my opinion is no war.

Re: Does memory leak? (1995)

#132

Post by monocasa »

derefr wrote: post #68

Erlang has a parameter called initial_heap_size. Each new actor-process in Erlang gets its own isolated heap, for which it does its own garbage-collection on its own execution thread. This initial_heap_size parameter determines how large each newly-spawned actor’s heap will be. Why would you tune it? Because, if you set it high enough, then for all your short-lived actors, memory allocation will become a no-op (= bum…

So I kind of disagree with the idea that arenas are all about deallocation at once. There's other contexts where you have separate arenas but don't plan on deallocating in blocks, mainly around when you have memory with different underlying semantics. "This block of memory is faster, but not coherent and needs to be manually flushed for DMA", "this block of memory is fastest but just not DMA capable at all", "there's only 2k of this memory, but it's basically cache", "this memory is large, fairly slow, but can do whatever", "this block of memory is non volatile, but memory mapped", etc.

I'd say that arenas are kind of a superset of both what you and I are talking about.

Re: Does memory leak? (1995)

#133

Post by 3pt14159 »

alex_young wrote: post #124
dahart wrote: post #115

Earlier quoted context omitted.

I had a family friend who worked on missiles and drones and other defense systems. He was really one of my dad’s running buddies, and he was a super nice guy, had 4 kids, went to church, etc. One day, I believe during the Iraq occupation, maybe ~12 or 13 years ago, I asked him very directly how he felt about working on these killing machines and whether it bothered him. He smiled and asked if I’d rather have the war…

Costa Rica hasn’t had a standing military since 1948. They are in one of the most politically unstable parts of the world and do just fine without worry of invasion. The US hasn’t been attacked militarily on its own soil in the modern era. The US military monopoly hasn’t prevented horrific attacks such as 9/11 executed by groups claiming to be motivated by our foreign military campaigns. I think there is a valid ques…

It's a valid question, but realistically if Costa Rica were invaded a number of countries would step in to help them. I love Costa Rica, it's one of the most beautiful countries I've been to and I do appreciate the political statement their making, but at the same time they're in a pretty unique situation.

As for the ethics of working on weapons, I think there is a lot of grey when it comes to software. It tends to centralize wealth, since once you get it right it works for everyone. It tends to be dual use, because a hardened OS can be used for both banks and tanks. Even developments in AI are worrying because they're so clearly applicable to the military.

Would I work on a nuclear bomb? No. Would I work on software that does a better job of, say, facial recognition to lessen the likelihood of a predator drone killing an innocent civilian? Maybe. It's not an all or nothing thing.

Re: Does memory leak? (1995)

#134

Post by Tuna-Fish »

needusername wrote: post #125
derefr wrote: post #68

Erlang has a parameter called initial_heap_size. Each new actor-process in Erlang gets its own isolated heap, for which it does its own garbage-collection on its own execution thread. This initial_heap_size parameter determines how large each newly-spawned actor’s heap will be. Why would you tune it? Because, if you set it high enough, then for all your short-lived actors, memory allocation will become a no-op (= bum…

> memory allocation will become a no-op (= bump allocation) No, that's a cache miss.

No, as memory is allocated linearly the cpu prefetchers will most likely keep the heap in cache.

Re: Does memory leak? (1995)

#135

Post by amelius »

giu wrote: post #85
jakeinspace wrote: post #69

As somebody working on embedded software for aerospace, I'm surprised this missile system even had dynamic memory allocation. My entire organization keeps flight-critical code fully statically allocated.

I'm always fascinated about software running on hardware-restricted systems like planes, space shuttles, and so on. Where can someone (i.e., in my case a software engineer who's working with Kotlin but has used C++ in his past) read more about modern approaches to writing embedded software for such systems? I'm asking for one because I'm curious by nature and additionally because I simply take the garbage collector f…

Just read docs that were written in the 70s, before the advent of garbage collection.

Re: Does memory leak? (1995)

#136

Post by fancyfredbot »

enriquto wrote: post #122
GordonS wrote: post #42

A bit OT, but I wonder how I'd feel if I was offered a job working on software for missiles. I'm sure the technical challenge would be immensely interesting, and I could tell myself that I cared more about accuracy and correctness than other potential hires... but from a moral standpoint, I don't think I could bring myself to do it. I realise of course that the military uses all sorts of software, including line of b…

> A bit OT, but I wonder how I'd feel if I was offered a job working on software for missiles. Unless you are an extreme pacifist (which is a perfectly reasonable thing to be), you'll acknowledge the legitimate need for the existence of an army in your country. In that case, the army better be equipped by missiles than by youths carrying bayonets. Then, there's nothing wrong in providing these missiles with technolog…

I suppose we have the 2008 crisis to thank for creating a popular view that finance is an entirely morally corrupt industry. Perhaps it's not surprising given the role "fancy financial instruments" played there. All the same, it strikes me as strange to find moving risk around to be more morally difficult than designing a missile - moving risk around is at least sometimes straightforwardly beneficial for everyone involved, a missile strike less so...

Re: Does memory leak? (1995)

#137

Post by GordonS »

acqq wrote: post #74
crawshaw wrote: post #7

This is an example of garbage collection being more CPU efficient than manual memory management. It has limited application, but there is a more common variant: let process exit clean up the heap. You can use an efficient bump allocator for `malloc` and make `free` a no-op.

There was also a variant of it with the hard drives: building Windows produced a huge amount of object files, so the trick used was to use a whole hard disk (or a partition) for that. Before the next rebuild, deleting all the files would took far more time than a "quick" reformatting of the whole hard disk, so the later was used. (I am unable to find a link that talks about that, however). In general, throwing away a…

That's quite a clever solution, I doubt I would have thought of that!

Windows has always been my daily drivers, and I really do like it. But I wish deleting lots of files would be much, much faster. You've got time to make a cup of coffee if you need to delete a node_modules folder...

Re: Does memory leak? (1995)

#138

Post by LucaSas »

This pops up again from time to time, I think what people should take away from this is that garbage collection is not just what you see in Java and other high level languages.

There are a lot of strategies to apply garbage collection and they are often used in low level systems too like per-frame temporary arenas in games or in short lived programs that just allocate and never free.

Re: Does memory leak? (1995)

#139

Post by hinkley »

monocasa wrote: post #132
derefr wrote: post #68

Erlang has a parameter called initial_heap_size. Each new actor-process in Erlang gets its own isolated heap, for which it does its own garbage-collection on its own execution thread. This initial_heap_size parameter determines how large each newly-spawned actor’s heap will be. Why would you tune it? Because, if you set it high enough, then for all your short-lived actors, memory allocation will become a no-op (= bum…

So I kind of disagree with the idea that arenas are all about deallocation at once. There's other contexts where you have separate arenas but don't plan on deallocating in blocks, mainly around when you have memory with different underlying semantics. "This block of memory is faster, but not coherent and needs to be manually flushed for DMA", "this block of memory is fastest but just not DMA capable at all", "there's…

I can’t remember the last time I read C code, but I do recall a particular time when I was reading a library that had been written with a great deal of attention to reliability. The first thing it did was allocate enough memory for the shutdown operations. That way on a malloc() failure, it could still do a a completely orderly shutdown. Or never start in the first place.

From that standpoint, you could also categorize arenas on a priority basis. This one is for recovery operations, this one for normal operation, and whatever is left for low priority tasks.

Re: Does memory leak? (1995)

#140

Post by MaxBarraclough »

monocasa wrote: post #132
derefr wrote: post #68

Erlang has a parameter called initial_heap_size. Each new actor-process in Erlang gets its own isolated heap, for which it does its own garbage-collection on its own execution thread. This initial_heap_size parameter determines how large each newly-spawned actor’s heap will be. Why would you tune it? Because, if you set it high enough, then for all your short-lived actors, memory allocation will become a no-op (= bum…

So I kind of disagree with the idea that arenas are all about deallocation at once. There's other contexts where you have separate arenas but don't plan on deallocating in blocks, mainly around when you have memory with different underlying semantics. "This block of memory is faster, but not coherent and needs to be manually flushed for DMA", "this block of memory is fastest but just not DMA capable at all", "there's…

Those aren't arenas. I'm inclined to agree with Wikipedia's definition, which does emphasise deallocation all at once:

> A region, also called a zone, arena, area, or memory context, is a collection of allocated objects that can be efficiently deallocated all at once.

Post reply on HN