shikokuchuo

shikokuchuo / mori

Public

Shared Memory for R Objects

shikokuchuo.netmori r shared-memory
19
1
100% credibility
Found Apr 22, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
C
AI Summary

mori is an R package for sharing data objects across processes without copying them, saving memory and speeding up parallel computations.

How It Works

1
🔍 Discover mori

While crunching big datasets in R and wanting faster parallel analysis, you find the mori package that lets data share smartly across workers.

2
📦 Install easily

In your R console, you add mori with a quick one-line command, and it's ready to use.

3
📊 Share your data

You tell mori to share your large table or list of numbers, creating a magic view that multiple parts of your program can peek at without copying.

4
⚡ Run parallel tasks

Launch several worker helpers on the same data; they all access it instantly together, feeling super speedy and light on memory.

5
📈 See the speedup

Your computations finish much quicker, using far less computer memory than before, with results popping up smoothly.

🎉 Big analysis wins

Your heavy data project now runs blazing fast and efficient, letting you focus on insights instead of waiting.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 19 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is mori?

mori is an R package that shares vectors, lists, and data frames via OS shared memory on POSIX and Windows. Use `share()` to write data once and get ALTREP-backed views for zero-copy access across processes; `map_shared()` opens by name. It tackles parallel R's memory bloat by sending tiny references (~100 bytes serialized) instead of full copies, with automatic GC cleanup.

Why is it gaining traction?

Zero-copy integration with mirai daemons cuts serialization, transfer, and RAM use—benchmarks show 30% faster bootstraps on 200MB data frames across 8 workers. Lazy materialization pays only for touched columns, copy-on-write keeps shares safe, and no external deps make it drop-in for github shared storage usage in R pipelines.

Who should use this?

R data scientists running mirai parallel jobs on big matrices or data frames. Analysts doing simulations or bootstraps where workers duplicate datasets, eating RAM on multi-core machines.

Verdict

Grab it for mirai speedups if you're in R parallel land, but at 19 stars and 1.0% credibility it's experimental—docs shine, tests cover well, so prototype safely before prod. (187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.