Pictures of a working garbage collector
oilshell.org
oilshell.org
Watchpoints, when they're suitable, are an incredibly powerful way to query your program's runtime behaviour (When does this happen? Why does this value end up here?) instead of stepping through and printing things / logging stuff.
They also fast, provided you watch values suitable for the CPU's logic to handle them directly. They need to be small-ish and located in memory for hardware support to handle it, otherwise GDB needs to single step.
(the `-l` flag is also important here and even less well known - it tells GDB you really care about the memory address the expression is stored in, not the expression itself)
I'm wondering if it's possible to put some directives into the generated C++ code that would map its debug info directly back to source lines in the Python - I can't find any docs to confirm my feeling this should be possible.
Edit: Maybe this? https://gcc.gnu.org/onlinedocs/cpp/Line-Control.html
So I think that would be cool, but I don't usually use a debugger in Python (or even C++ -- I tend to write small programs using shell as the "REPL").
Oil is extremely test-driven and shell's stdin/stdout type interfaces means it's pretty easy to exhaustively test. Examples: https://www.oilshell.org/release/0.13.1/test/spec.wwz/survey...
The GC is the one case we have where that kind of testing/debugging doesn't suffice! So I started watching the CppCon videos on debuggers a couple years ago and found some good talks :)
In my experience rr's recording is around x2 slower for single threaded programs than running the program on its own.
I think featurewise they are in parity.
I heard that udb is like rr, but better on some fronts (except price and software freedom).
> gdb has had reversible debugging since release 7, in 2009. What does udb offer that it lacks?
GDB's built-in reversible debugging is cool (and it's helped raise awareness) but it doesn't scale well. We build on the same command set and serial protocol that GDB defined - UDB is GDB but with additional Python code hooking it up to our separate record/replay engine.
For UDB, I'd say we offer: 1. Performance & efficiency (orders of magnitude faster at runtime and lower in memory requirements). 2. Recordings can be saved to portable files (share with colleagues, receive from customers, etc). 3. Library API so applications can self-record with control of when to capture and save. 4. Wider support of modern software (proactively tracking modern CPU features, shared memory and device maps, etc). 5. Correctness (in the past we've found the reverse operations in GDB don't have as strong semantics as we'd hoped, though I'd also be happy to be wrong here)
FWIW, rr (https://rr-project.org/) also offers many similar benefits over GDB's built-in system (though not the library API in point 3) but with differences in what CPUs / systems are supported, ability to attach at runtime, etc.
If you're looking for an open source solution, I'd choose rr over GDB's built-in approach.
rr's recording works across system calls, thread context switches, etc, but gdb's doesn't.
rr's recording creates a persistent recording on disk that you can even move around between machines. This permits workflows that gdb's doesn't.
(Disclaimer: I work on rr)
I'm on the Undo mailing list as well -- nice to see that it's effective!
(Also should say that Clasp sounds very cool -- I'm a fan of anything that enables interop and reuse, rather than rebuilding the same thing in different languages, which may or may not be as good)
It succeeds because it only solves the GC problems that Oil has (and I mean this as a high complement).
Oil is single threaded, code runs in loops that are well understood, and the code is generated from a strongly typed subset of Python.
The GC then is run only between loop iterations, so there is no need for stack scanning. You never have to worry about a root being in a temporary (from the point of view of the C++ compiler) since there are just a few locals in the function running the loop, and any variables local to the loop are logically dead between iterations.
Since a goal of Oil is portability, not scanning registers and stacks is very important. Getting this right when the GC could be invoked at any allocation is potentially intractable with mypy semantics at least.
It's unusual because it's a precise collector in C++, and what I slowly realized is that that problem is basically impossible for any non-trivial software, without changing the C++ language itself :)
It seems like that hasn't happened, despite efforts over decades. I added this link about C++ GC support to the appendix, which also explains our unique constraints.
Garbage collection in the next C++ standard (Boehm 2009)
https://dl.acm.org/doi/abs/10.1145/1542431.1542437
http://www.oilshell.org/blog/2023/01/garbage-collector.html#...
---
The reason that precise GC can work for Oil is because it's a shell that links with extremely little 3rd-party code, and has relatively low perf requirements. We depend on libc and GNU readline, just like bash. And those libraries are basically old-school C functions which are easy to wrap with a GC.
(Also as Aidenn mentioned, shells use process-based concurrency, which means we don't have threads. The fact that it's mostly generated C++ code is also important, as mentioned in the post)
---
The funny thing is that one reason I started this project is because I worked with "big data" frameworks on clusters, but I found that you can do a lot on a single machine. (in spirit similar to the recent "Twitter on one machine post" https://news.ycombinator.com/item?id=34291191 )
I would just use shell scripts to saturate ~64 cores / 128 G of RAM, rather than dealing with slow schedulers and distributed file systems.
But garbage collectors and memory management are a main reason you can't use all of a machine from one process. There's just so much room for contention. Also the hardware was trending toward NUMA at the time, and probably is even more now, so processes make even more sense.
All of that is to say that I'm a little scared of multi-threaded GC ... especially when linking in lots of third party libraries.
And AFAIK heaps with tens or hundreds of gigabytes are still in the "not practical" range ... or they would take a huge amount of engineering effort
---
But of course there are many domains where you don't have embarrassingly parallel problems, and writing tight single- or multi-threaded code is the best solution.
Some more color here: https://old.reddit.com/r/oilshell/comments/109t7os/pictures_...
I wonder if Clasp has any support for multi-process programming? Beyond Unix pipes, you could also use shared memory and maybe some semaphores to synchronize access, and avoid copying. I think of that as sort of "inverting" the problem. Certain kinds of data like pointer-rich data is probably annoying to deal with in shared memory, but there are lots of representations for data and I imagine Lisps could take advantage of some of them, e.g. https://github.com/oilshell/oil/wiki/Compact-AST-Representat...
In any case, I would imagine embedding a C++ compiler at runtime does open up a lot more options!
Which effectively combines a `watch -l`, plus a reverse-continue but also monitors when the underlying memory was allocated / freed.
It's quite nice, basically "git blame" for your variables.
A thousand years from now when they're digging up texts and artifacts from this period, there's likely going to be a lot of confused academics.
Edit: this link posted by another commenter gives a nice overview of the tradeoffs between the methods and it has very nice graphs to boot https://spin.atomicobject.com/2014/09/03/visualizing-garbage...
That is, with ARC, the more garbage you generate, the more work will be spent on the GC vs normal program flow. In contrast, with tracing/copying, you can generate as much garbage as you want, without affecting GC time; but, the more memory you actually use, the more time you will spend in GC.
A is for Automatic, as in the compiler adds the calls needed to manage the reference counting.
If the language and compiler can guarantee a reference is strictly contained within a thread, it doesn't have to be atomic. Usually that is not the case, so atomic operations are used.
- your application is using close to 100% available RAM
- you don't care about the impact of all those extra writes clogging up CPU caches and bouncing cache lines back and forth across internal buses
- you can't afford to pay for a small amount of extra RAM
In that case reference counting sounds ideal. Good luck!
If you're on the server then just buy more RAM and use a GC, I agree.
But if you're making consumer software then buying more RAM isn't an option. Additionally, RAM costs battery life even when it isn't in use.
- ARC (upper case): Automatic Reference Counting (in Swift/ObjC)
- Arc (title case): Atomic reference counting (in Rust)
I assume OP intended the former.
Choosing to use automatic reference counting in a language is a design choice, with positives and negatives. Atomic reference counting is an implementation detail, a strict necessity whenever you can't be sure an object will only be referenced by one thread (in essence, non-atomic reference counting is just an optimization).
What do you mean by 'usage'? References don't change when reading or writing the data being reference counted, they change when being passed to a function or returned from one. In C++ it's rare that you would even use reference counting, since that essentially means you don't know the lifetime of your value. Most variables are going to be on the stack and the vast majority of dynamic allocation is going to be referenced from a single scope at one time.
The reality is that it takes gross incompetence to have a speed impact from reference counting.
> In C++ it's rare that you would even use reference counting, since that essentially means you don't know the lifetime of your value
Sure, because it is a manual memory managed language with RC being an escape hatch only. But there are plenty of problems/programs where you simply can’t know the lifetime of your objects, e.g. Chrome uses a proper GC for C++ as well.
I don't know what you mean by escape hatch, but it generally just isn't necessary and memory is managed automatically by scope. The bigger point here is that it just isn't a significant part of execution time.
But there are plenty of problems/programs where you simply can’t know the lifetime of your objects
Like what? I can only think of one, which is passing memory allocated in one thread to another thread.
Chrome uses a proper GC for C++ as well
This is an anecdote, it doesn't prove or disprove anything in the bigger picture.
That’s just an implementation detail. The point is that the object’s lifetime is only known at runtime and will be reclaimed when a counter reaches zero, this is reference counting. Whether you have to manually inc/dec that counter, or the language does it for you through some abstraction is besides the point, it is automatic memory management either way, as it.. manages memory automatically.
> Like what? I can only think of one, which is passing memory allocated in one thread to another thread
Any programming language, both parsing into an AST, AST manipulations, interpretation (and that is a very wide category, not only for things you would think of as proper languages). But even some games may want to use GC for some in-game objects, as the lifetime of those is fundamentally dependent on user action.
Would the litany of managed languages and their widespread usage be less anecdotal?
both parsing into an AST, AST manipulations, interpretation even some games may want to use GC for some in-game objects, as the lifetime of those is fundamentally dependent on user action
Here you are conflating the lifetime of resources inside the various scopes of a program with dynamic resources in a game. These are not the same thing. Language level reference counting will not save you or help you to know when to unload a level or a texture. Just because there is control over resources doesn't mean reference counting. Likewise even in something like java you need to set links to heap allocated objects to null so that they can be garbage collected. The language doesn't magically know when you need to unload a level.
> only necessary when giving memory allocated in one thread to another thread
RC count can be larger than one even when only a single thread using it. But I’m not familiar with this usage and it is not really RC for memory management anymore, more like a lock-less data structure.
I wasn’t talking about texture/level loading/unloading because it is more complex, but things like using scripting languages for part of the game logic.
I think you're just repeating yourself, but I'm not sure what question you're answering.
RC count can be larger than one even when only a single thread using it. But I’m not familiar with this usage and it is not really RC for memory management anymore, more like a lock-less data structure.
I don't understand what you are saying here.
things like using scripting languages for part of the game logic.
Scripting languages are slow for a lot of reasons, like pointer chasing and excessive memory allocations. Reference counting is a very small piece of that puzzle.
I don't think anyone is debating that.
not suitable for lock-free algorithms and may overflow the stack.
This you will have to explain. I see people make vague assertions like this but I never see a good explanation.
You can do that in multiple ways.
First you can use the extra bits of a pointer for a counter to fit it all into 64 bits.
Second, you can use a 128 bit compare and swap which has been supported by CPUs for about 20 years now.
Third, you can not use pointers and use indices of whatever bit resolution you want, using the extra bits for a counter.
Finally, how does a garbage collector change this ?
If make a long list using shared_ptr will overflow the stack when the head destructor executes.
If we set aside for a second the insanity in making a linked list where every pointer destruction calls the next pointer in the list's destructor, how is this unique to a shared_ptr ?
Stack overflow is not unique to a shared_ptr, but GC pointers don't have this problem
No one should ever have this problem. It is a ridiculous way to make a linked list in the first place.
You told me something was impossible to do without garbage collection and I explained three different ways that I've already done it, then you just keep trying to talk about something that was never up for discussion in the first place. You hallucinated shared_ptr into the conversation from nowhere.
Without talking about smart pointers, what am I missing from the list above? Why do some lock free algorithms need garbage collection?
Why don't you answer my questions above? They confront the ABA problem directly since I explained three ways to keep counts paired with pointers or indices. What can't be done without a garbage collector? Why do you keep giving vague recommendations to read about general topics? Give me a specific deeply technical answer if you can.
When you want to allocate an index you check the current index and version, and replace it with the index points to if the version is the same. Freeing is the reverse since you have an index to give the list.
These indices are used to coordinate to a second array where you can store whatever data you want.
Here are some other techniques.
https://people.csail.mit.edu/shanir/publications/Lock_Free.p... https://www.boost.org/doc/libs/1_55_0/boost/lockfree/stack.h... https://lumian2015.github.io/lockFreeProgramming/lock-free-s...
Still, I'm not sure what garbage collection changes about these techniques. Lock free lists have been studied for a long time, they have nothing to do with memory allocation.
Described algorithms ignore the ABA problem.
I literally wrote a method for doing that, an index with a version that can be checked to make sure nothing changed.
Also if you're going to say that heavily tested implementations ignore the ABA problem you need to explain why you think that or why you think they won't work and again, why garbage collection changes anything.
stores pointers on 48 bits which is not enough on new architectures.
48 bits is the size of the memory controller on modern CPUs and exceeding that would need over 281 terabytes of memory.
Again, the original question is what lock free algorithm can be done with garbage collection that can't be done without it?
New processors use 56 bits of virtual address. It doesn't matter if you have that much memory, because addresses are virtual. Also, newer versions of Android do not allow the use of unused address bits.
This is again, your assertion, it isn't evidence or an explanation of any kind, you just keep saying the same thing. The link you have is people discussing a bunch of surrounding issues.
Fundamentally, allocation of arbitrary memory just doesn't have to be ingrained in the lock free data structure. As soon as you can deal with 64 bits at a time, you can store pointers. There are lock free heap allocators and lock free block allocators that can be combined with whatever you are using to deal lock free with integers/pointers.
Freeing memory is going to be a matter of ownership. If you pop a pointer, that thread should own it. Not only that, but a pointer combined with a reference count can always be used if necessary and again, 128 bit compare and swap has been around for 20 years.
I didn't see what you're talking about in the 15 year old message board discussion and I think if there was something specific and clear you would have copied and pasted it.
I also think that if you had any understanding of what you're saying, you would have given an explanation yourself.
So go ahead and actually put something here that you can back up. It is a common scenario where someone has no real evidence to link something adjacent and then tell someone to 'go find it in this link'.
int lfstack_pop(_Atomic lfstack_t \*lfstack)
{
lfstack_t next;
lfstack_t orig = atomic_load(lfstack);
do
{
if (orig.head == NULL) // undefined behavior !!!
{
return -1;
}
next.head = orig.head->next;
} while (!atomic_compare_exchange_weak(lfstack,&orig,next));
free(orig.head);
return 0;
}
Can you focus and nail down one claim with specifics before moving on to something else?
I've already done all the things you claim are impossible. That's how I know what you're saying is nonsense. Avoiding a double free on a pointer can always be done with atomic reference counting so that only one thread claims ownership. That's the whole story. You haven't given a shred of evidence to explain what garbage collection enables something that was previously impossible.
It is obvious that if you had something you could say that directly applies to what you claimed before (some algorithms can't be done without garbage collection), that you would have already said it a long time ago.
How would you implement a graph that requires heap allocation for each node and supports adding/removing nodes and links with purely scope-based memory management (no shared_ptr or equivalent, no tracing GC), while still reclaiming memory as soon as possible (so no arena-based solutions)?
There are also single-threaded concurrent scenarios where object ownership can be ambiguous, but those are similar to the multi-threaded scenarios you discuss.
What’s manual rc like?
E.g., the Arc type in Rust. Rust, for example, has a separate Rc type. Rc should outperform Arc, but cannot be shared across threads.