snmalloc
verona
snmalloc | verona | |
---|---|---|
8 | 20 | |
1,491 | 3,553 | |
1.3% | 0.4% | |
6.3 | 6.6 | |
about 1 month ago | 20 days ago | |
C++ | C++ | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
snmalloc
-
Snmalloc: A Message Passing Allocator
https://github.com/microsoft/snmalloc#snmalloc mentions two biggest motivations as:
> Allocations on one thread are freed by a different thread
I can imagine one use-case for this: a task that is scheduled from and executed by a work-stealing thread-pool can allocate memory in one thread but by design there's no guarantee that the memory will be necessarily freed from that exact thread. Would that be a good use-case for snmalloc?
> Deallocations occur in large batches
This sounds much like a bump allocator use-case but which can do this exact thing by calling a single munmap(addr, len) and unmap multiple allocations all at once.
-
Is the JVM a upside or downside to Scala?
Yes, it's very efficient and that's not where the main problem lies. However, small allocations with modern C heap allocators like mimalloc or snmalloc has gotten extremely efficient as well. Would be interesting to see a benchmark comparison with Java's G1 and ZGC.
- Snmalloc 0.6 released, major redesign with security hardening
- Snmalloc: High-performance message passing based allocator
-
Maintenance status (jemallocator)
Did you ever benchmark against https://github.com/microsoft/snmalloc ?
verona
-
Snmalloc: A Message Passing Allocator
According to this FAQ, snmalloc was designed for the Verona language:
https://microsoft.github.io/verona/faq.html
Unfortunately, I cannot find any significant code samples for Verona on the website or in the GitHub repo. There are a few types defined in a pretty low-level way:
https://github.com/microsoft/verona/tree/master/std/builtin
- Microsoft Project Verona, a research programming language
-
Making C++ Safe Without Borrow Checking, Reference Counting, or Tracing GC
I think the future lies in figuring out how to get the benefits of that secret sauce, while mitigating or avoiding the downsides.
Like Boats said, the borrow checker works really well with data, but not so well with resources. I'd also add that it works well with data transformation, but struggles with abstraction, both the good and bad kind. It works well with tree-shaped data, but struggles with programs where the data has more intra-relationships.
So if we can design some paradigms that can harness Rust's borrow checker's benefits without its drawbacks, that could be pretty stellar. Some promising directions off the top of my head:
* Vale-style "region borrowing" [0] layered on top of a more flexible mutably-aliasing model, either involving single-threaded RC (like in Nim) generational references (like in Vale).
* Forty2 [1] or Verona [2] isolation, which let us choose between arenas and GC for isolated subgraphs. Combining that with some annotations could be a real home run. I think Cone [3] was going in this direction for a while.
* Val's simplified borrowing (mutable value semantics) combined with some form of mutable aliasing (this might sound familiar).
[0] https://verdagon.dev/blog/zero-cost-borrowing-regions-part-1... (am author)
[1] http://forty2.is/
[2] https://github.com/microsoft/verona
[3] https://cone.jondgoodwin.com/
-
A Flexible Type System for Fearless Concurrency
Their approach lines up pretty well with how we do regions in Vale. [0]
Specifically, we consider the "spine" of a linked list to be in a separate "region" than the elements. This lets us freeze the spine, while keeping the elements mutable.
This mechanism is particularly promising because it likely means one can iterate over a collection with zero run-time overhead, without the normal restrictions of a more traditional Rust/Cyclone-like borrow checker. We'll know for sure when we finish part 3 (one-way isolation [1]); part 1 landed in the experimental branch only a few weeks ago.
The main difference between Vale and the paper's approach is that Vale doesn't assume that all elements are self-isolated fields, Vale allows references between elements and even references to the outside world. However, this does mean that Vale sometimes needs "region annotations", whereas the paper's system doesn't need any annotations at all, and that's a real strength of their method.
Other languages are experimenting with regions too, such as Forty2 [2] and Verona [3] though they're leaning more towards a garbage-collection-based approach.
Pretty exciting time for languages!
[0] https://verdagon.dev/blog/zero-cost-borrowing-regions-overvi...
[1] https://verdagon.dev/blog/zero-cost-borrowing-regions-part-3...
[2] http://forty2.is/
[3] https://github.com/microsoft/verona
- Microsoft is rewriting core Windows libraries in Rust
-
Microsoft is to enable Rust use for Windows 11 kernel
Does this count? https://microsoft.github.io/verona/
-
Microsoft rewriting core Windows libraries in Rust
What about new Rust that "Microsoft Research" trying to "explore" https://github.com/microsoft/verona/blob/master/docs/explore.md ?
- Concurrent ownership in Verona
- Concurrent Ownership in Verona
-
Pony Programming Language
Fun fact: the person who created Pony, Sylvan Clebsch, has been working on a Microsoft Research project called Verona. From it's README [0]:
> Project Verona is a research programming language to explore the concept of concurrent ownership. We are providing a new concurrency model that seamlessly integrates ownership.
https://github.com/microsoft/verona/tree/master
What are some alternatives?
mimalloc - mimalloc is a compact general purpose allocator with excellent performance.
tour_of_rust - A tour of rust's language features
jemallocator - Rust allocator using jemalloc as a backend
PurefunctionPipelineDataflow - My Blog: The Math-based Grand Unified Programming Theory: The Pure Function Pipeline Data Flow with principle-based Warehouse/Workshop Model
Mesh - A memory allocator that automatically reduces the memory footprint of C/C++ applications.
dolt - Dolt – Git for Data
mimalloc_rust - A Rust wrapper over Microsoft's MiMalloc memory allocator
ante - A safe, easy systems language
jemallocator - Rust allocator using jemalloc as a backend
cone - Cone Programming Language
rpmalloc - Public domain cross platform lock free thread caching 16-byte aligned memory allocator implemented in C
felix - The Felix Programming Language