ompi
libuv
Our great sponsors
ompi | libuv | |
---|---|---|
10 | 75 | |
2,016 | 23,241 | |
3.3% | 1.4% | |
9.7 | 9.0 | |
1 day ago | 3 days ago | |
C | C | |
GNU General Public License v3.0 or later | 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.
ompi
-
Ask HN: Does anyone care about OpenPOWER?
The commercial Linux world (see https://github.com/open-mpi/ompi/issues/4349) and other open source OSes (eg FreeBSD) seem to have lined up behind little-endian PowerPC. IBM still has a big-endian problem with AIX, IBM i, and Linux on Z.
-
Announcing Chapel 1.32
Roughly, the sets of computational problems that people used (use?) MPI for. Things like numerical solvers for sparse matrices that are so big that you need to split them across your entire cluster. These still require a lot of node-to-node communication, and on top of it, the pattern is dependent on each problem (so easy solutions like map-reduce are effectively out). See eg https://www.open-mpi.org/, and https://courses.csail.mit.edu/18.337/2005/book/Lecture_08-Do... for the prototypical use case.
-
How much are you meant to comment on a code?
One of the guys at the local LUG is one of the lead maintainers of Open MPI. He told us about a comment that ran into the hundreds of lines, all for a one-line change in the code.
-
Which license to choose when you want credit
But it would be very inconvenient to have to keep crediting everyone who's ever worked on it. If you look at old projects, their licenses can have like 10-20 of those lines (here's one I was recently looking into).
-
First True Exascale Supercomputer
I have a bit of experience programming for a highly-parallel supercomputer, specifically in my case an IBM BlueGene/Q. In that case, the answer is a lot of message passing (we used Open MPI [0]). Since the nodes are discrete and don't have any shared memory, you end up with something kinda reminiscent of the actor model as popularized by Erlang and co -- but in C for number-crunching performance.
That said, each of the nodes is itself composed of multiple cores with shared memory. So in cases where you really want to grind out performance, you actually end up using message passing to divvy up chunks of work, and then use classic pthreads to parallelize things further, with lower latency.
Debugging is a bit of a nightmare, though, since some bugs inevitably only come up once you have a large number of nodes running the algorithm in parallel. But you'll probably be in a mainframe-style time-sharing setup, so you may have to wait hours or more to rerun things.
This applies less to some of the newer supercomputers, which are more or less clusters of GPUs instead of clusters of CPUs. I imagine there's some commonality, but I haven't worked with any of them so I can't really say.
[0] https://www.open-mpi.org/
- Managing parallelism by process vs by machine
-
MPI + CUDA Program for thermal conductivity problem
I would suggest using OpenMPI because it's pretty easy to get started with. You can build OpenMPI with CUDA support, then you can pass device pointers directly to MPI_Send and MPI_Recv. Then you don't have to deal with transfers and synchronization issues.
-
Distributed Training Made Easy with PyTorch-Ignite
backends from native torch distributed configuration: nccl, gloo, mpi.
-
FEA computer simulation question
I use a linux ubuntu machine with MPI (https://www.open-mpi.org/). I had a question on making my computer simulations faster. Would be better to get an older AMD 9590 machine clocked at 4.7 ghz or continue using my Ryzen 7 1700 machine clocked at something like 3.5ghz?
-
C Deep
OpenMPI - Message passing interface implementation. BSD-3-Clause
libuv
- Epoll: The API that powers the modern internet (2022)
-
APIs in Go with Huma 2.0
I wound up on a different team with pre-existing Python code so temporarily shelved my use of Go for a bit, and we used Sanic (an async Python framework built on top of the excellent uvloop & libuv that also powers Node.js) to build some APIs for live channel management & operations. We hand-wrote our OpenAPI and used it to generate documentation and a CLI, which was an improvement over what was there (or not) before. Other teams used the OpenAPI document to generate SDKs to interact with our service.
- Python Is Easy. Go Is Simple. Simple = Easy
-
Notes: Advanced Node.js Concepts by Stephen Grider
In the source code of the Node.js opensource project, lib folder contains JavaScript code, mostly wrappers over C++ and function definitions. On the contrary, src folder contains C++ implementations of the functions, which pulls dependencies from the V8 project, the libuv project, the zlib project, the llhttp project, and many more - which are all placed at the deps folder.
- A Magia do Event Loop
-
A complete guide to the Node.js event loop
Libuv, the C library that gives Node.js its asynchronous, non-blocking I/O capability is responsible for managing the thread pool. Node.js gives you the capability of using additional threads for computationally expensive and long-lasting operations to avoid blocking the event loop.
-
What is Node.js?: A Complete Guide
Node.js is written in C, C++, and JavaScript. The core components of Node.js - the V8 engine and the libuv library - are written in C++ and C, respectively, since these languages provide low-level access to system resources, making them well-suited for building high-performance and efficient applications. JavaScript is mainly used to write the application logic.
-
Node v20.3.0 (Current) upgrade to libuv 1.45.0, including SIGNIFICANT performance improvements to file system operations on Linux
x8 apparently https://github.com/libuv/libuv/pull/3952
-
Node.js – v20.3.0
Notably upgrades to libuv 1.45 which has io_uring support. Faster file system access! Awhh yeah, it's on.
Remarkable what a mild & unintrusive PR adding io_uring was. https://github.com/libuv/libuv/pull/3952
-
Using Parallel Processing in Node.js and its Limitations
Well, the single-threaded nature ultimately leads to its biggest downfall. Node.js utilizes a synchronous event loop engineered using Libuv that takes in code from the call stack and executes it.
What are some alternatives?
gloo - Collective communications library with various primitives for multi-machine training.
libevent - Event notification library
Redis - Redis is an in-memory database that persists on disk. The data model is key-value, but many different kind of values are supported: Strings, Lists, Sets, Sorted Sets, Hashes, Streams, HyperLogLogs, Bitmaps.
Boost.Asio - Asio C++ Library
NCCL - Optimized primitives for collective multi-GPU communication
libev - Full-featured high-performance event loop loosely modelled after libevent
FlatBuffers - FlatBuffers: Memory Efficient Serialization Library
tokio-uring - An io_uring backed runtime for Rust
libvips - A fast image processing library with low memory needs.
uvw - Header-only, event based, tiny and easy to use libuv wrapper in modern C++ - now available as also shared/static library!
SWIFT - Modern astrophysics and cosmology particle-based code. Mirror of gitlab developments at https://gitlab.cosma.dur.ac.uk/swift/swiftsim
C++ Actor Framework - An Open Source Implementation of the Actor Model in C++