Mastouille

5 messages3 participants0 message aujourd’hui

Howard Chu @ SymasI've dedicated most of my career to writing the most efficient code.<a href="https://ngmx.com/@mejofi/114889780968597360" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://ngmx.com/@mejofi/114889780968597360</a>But the mention of storing way more data than you're ever going to use also triggered me here.It's a common symptom of poorly designed apps looking for write-optimized databases like RocksDB. If your workload is more write heavy than read, then by definition you're writing more than you'll use.It's another reason why <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> is read optimized, because we only care about data we'll actually use.

Howard Chu @ Symaslmdb-queueA lightning fast disk based queue, meta data is stored in <a href="https://mastodon.social/tags/lmdb" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#lmdb</a> <a href="https://lib.rs/crates/lmdb-queue" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://lib.rs/crates/lmdb-queue</a>

Howard Chu @ Symas"This article introduces <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> (an efficient key-value storage database) and the command line tool `lmcmd` developed based on Python. Since LMDB uses binary files for storage, which is inconvenient to manage and debug, `lmcmd` was developed to provide a command line operation interface similar to Redis, supporting database operations, data import and export, and search functions." <a href="https://developer.aliyun.com/article/1672386" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://developer.aliyun.com/article/1672386</a>

Howard Chu @ Symas"If I abandon Vercel and switch to VPS, I can abandon Firestore and build my own query layer with <a href="https://mastodon.social/tags/lmdb" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#lmdb</a> ... just thinking about it makes me happy"<a href="https://xcancel.com/apolkingg8/status/1943623985666679152#m" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://xcancel.com/apolkingg8/status/1943623985666679152#m</a>

Howard Chu @ Symas"xs-lib is a command-line tool (xs) and a Rust library (xs_lib) for recording and retrieving sequential streams of payloads. It uses Lightning Memory-Mapped Database (<a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a>) for local data storage."<a href="https://lib.rs/crates/xs-lib" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://lib.rs/crates/xs-lib</a>I don't understand why you'd need to store a sequentially accessed stream into a database. This looks like an over-engineered `cat` to me.

Howard Chu @ SymasIf you can't beat 'em, join 'em. Oracle now uses <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> <a href="https://docs.oracle.com/en/middleware/fusion-middleware/fmwlc/licensing-information.pdf" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://docs.oracle.com/en/middleware/fusion-middleware/fmwlc/licensing-information.pdf</a>

Howard Chu @ Symas"I’m experiencing significant slow-downs when converting data from <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> to HDF5 format."Solution: don't do that. <a href="https://mastodon.social/tags/HDF5" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#HDF5</a> is a pig.<a href="https://stackoverflow.com/questions/79706522/hdf5-write-performance-degrades-over-time-when-converting-from-lmdb-3-7m-entri" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://stackoverflow.com/questions/79706522/hdf5-write-performance-degrades-over-time-when-converting-from-lmdb-3-7m-entri</a>

Howard Chu @ SymasARTIQ (Advanced Real-Time Infrastructure for Quantum physics) is a leading-edge control and data acquisition system for quantum information experiments. It is maintained and developed by M-Labs and the initial development was for and in partnership with the Ion Storage Group at NIST. ... Persistent datasets are now stored in a <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> database for improved performance. <a href="https://m-labs.hk/artiq/manual-beta.pdf" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://m-labs.hk/artiq/manual-beta.pdf</a>

Howard Chu @ SymasSquawk - a compact <a href="https://mastodon.social/tags/XMPP" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#XMPP</a> desktop messenger<a href="https://git.macaw.me/vae/squawk" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://git.macaw.me/vae/squawk</a>Uses <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a>

Howard Chu @ Symas<a href="https://mastodon.social/tags/Flutter" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#Flutter</a> Local DBA high-performance local database for Flutter that leverages Rust's <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> embedded database through FFI (Foreign Function Interface). This library provides a robust, efficient, and safe way to store local data in your Flutter applications across multiple platforms. <a href="https://pub.dev/packages/flutter_local_db" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://pub.dev/packages/flutter_local_db</a>

Howard Chu @ SymasOSM Express: fast data format for OpenStreetMap data, built on <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> <a href="https://github.com/bdon/OSMExpress" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://github.com/bdon/OSMExpress</a>

Howard Chu @ Symas"Top 8 Embedded SQL Databases in 2025"<a href="https://www.explo.co/blog/embedded-sql-databases" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.explo.co/blog/embedded-sql-databases</a>Lists a few non-SQL DBs like <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a>, but ok whatever.Also includes a couple written in Java, which I'd personally exclude because they're only usable in Java apps, which are categorically too heavyweight to be considered "embedded" IMO.

Howard Chu @ SymasVideoMind: Giving LLMs a Real Memory of the Visual World<a href="https://www.linkedin.com/pulse/videomind-giving-llms-real-memory-visual-world-atharva-date-htb1f" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://www.linkedin.com/pulse/videomind-giving-llms-real-memory-visual-world-atharva-date-htb1f</a>"What if we could give LLMs structured visual memory? Not just a transcript of captions—but episodic, grounded, object-aware memory that mirrors human understanding? ... Captioning BLIPDecoder generates and caches a caption per frame (<a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a>-backed). "

Howard Chu @ SymasSimplifying Synchronization with Lightning Stream for Enterprise<a href="https://blog.powerdns.com/simplifying-synchronization-with-lightning-stream-for-enterprise" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://blog.powerdns.com/simplifying-synchronization-with-lightning-stream-for-enterprise</a>"By leveraging <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> backends, Lightning Stream delivers high-speed, scalable replication with one-way and two-way synchronization between authoritative instances across different data centers."

Howard Chu @ SymasCytoplasm: A general-purpose C library for creating high-level, networked, and multi-threaded C applications.<a href="https://git.telodendria.io/Telodendria/Cytoplasm" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://git.telodendria.io/Telodendria/Cytoplasm</a><a href="https://git.telodendria.io/Telodendria/Cytoplasm/pulls/43" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://git.telodendria.io/Telodendria/Cytoplasm/pulls/43</a> Supports <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> in addition to some kludgy corruption-prone JSON file writer. But sure, it's important to have the latter?

Howard Chu @ SymasComprehensive RPC framework with support for C++, CSharp, Java, JavaScript, Python and more.<a href="https://vcpkg.link/ports/zeroc-ice" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://vcpkg.link/ports/zeroc-ice</a>Uses <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a>

Howard Chu @ SymasMCCD: A Multi-Attribute Chinese Calligraphy Character Dataset Annotated with Script Styles, Dynasties, and Calligraphers<a href="https://arxiv.org/html/2507.06948v1" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://arxiv.org/html/2507.06948v1</a>"the total MCCD dataset of 7765 Chinese character annotations and its three subsets, MCCD-Style for 10 calligraphic styles, MCCD-Dynasty for 15 major historical periods, and MCCD-Calligrapher for 142 famous calligraphers. ... To enhance data retrieval efficiency, all training and testing sets are also converted into the <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> format (binary form)."

Howard Chu @ SymasInvestigate <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> as HLR storage <a href="https://projects.osmocom.org/issues/3649" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://projects.osmocom.org/issues/3649</a>This Open Source Mobile Comm project appears to be an open source impl of a 3G network's backend software. But they're not using the standard 3GPP protocols, they're using their own home-grown stuff instead.Looks like a toy, using SQLite3 as its HLR storage. Real 3G/4G networks use LDAP, lots of OpenLDAP in particular because nothing else scales enough, efficiently enough.

Howard Chu @ SymasA reminder that social media platforms are not <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> support channels. And just RTFM, instead of getting misinfo from twitter <a href="https://x.com/cachecrab/status/1942295240628404691" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://x.com/cachecrab/status/1942295240628404691</a>Obviously yes, the docs say LMDB supports multiprocess concurrency. That has always been true.

Howard Chu @ SymasSpatia: Multimodal Model for Prediction and Generation of Spatial Cell Phenotypes <a href="https://arxiv.org/html/2507.04704v1" rel="nofollow noopener noreferrer" translate="no" target="_blank">https://arxiv.org/html/2507.04704v1</a>"Each cell-gene example consists of this uint8 image patch and the corresponding per-cell transcript vector for a single gene, serialized into <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener noreferrer" target="_blank">#LMDB</a> for efficient training (MIST-C)."

Recherches récentes

Options de recherche

Administré par :

Statistiques du serveur :

#lmdb