Veganism Social

Howard Chu @ SymasArcticDB is a high performance, serverless DataFrame database built for the <a href="https://mastodon.social/tags/Python" class="mention hashtag" rel="nofollow noopener" target="_blank">#Python</a> Data Science ecosystem. Uses <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> for local storage, or Amazon S3 for cloud. <a href="https://anaconda.org/conda-forge/arcticdb" rel="nofollow noopener" translate="no" target="_blank">https://anaconda.org/conda-forge/arcticdb</a>

Howard Chu @ SymasSotohp: photo management using <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> <a href="https://github.com/dacr/sotohp" rel="nofollow noopener" translate="no" target="_blank">https://github.com/dacr/sotohp</a>

Howard Chu @ SymasUsing <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> from <a href="https://mastodon.social/tags/python" class="mention hashtag" rel="nofollow noopener" target="_blank">#python</a> - article in Japanese <a href="https://zenn.dev/amenaruya/articles/83ce42f1e0b1cf" rel="nofollow noopener" translate="no" target="_blank">https://zenn.dev/amenaruya/articles/83ce42f1e0b1cf</a>

Howard Chu @ SymasPro-tip: answers I wrote about <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> on monero.stackexchange.com are specific to LMDB in the <a href="https://mastodon.social/tags/monero" class="mention hashtag" rel="nofollow noopener" target="_blank">#monero</a> codebase. They are not applicable anywhere else.<a href="https://www.threads.net/@shbekti/post/DInw_GguYja" rel="nofollow noopener" translate="no" target="_blank">https://www.threads.net/@shbekti/post/DInw_GguYja</a>In particular, while 32- and 64-bit builds of Monero use a fully compatible DB format, that isn't true by default for LMDB.LMDB was not written to be a generic system library, and people using it as such (cough redhat cough) are morons. It's intended to be built into specific applications, with build flags set as needed.

Howard Chu @ SymasFor cloudsmith, <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> is the only key/value store <a href="https://cloudsmith.com/navigator/search?q=keywords%3A+key-value+store&registry=all" rel="nofollow noopener" translate="no" target="_blank">https://cloudsmith.com/navigator/search?q=keywords%3A+key-value+store&registry=all</a>

Howard Chu @ SymasSpam filtering for evolution, with <a href="https://mastodon.social/tags/lmdb" class="mention hashtag" rel="nofollow noopener" target="_blank">#lmdb</a> <a href="https://archlinux.org/packages/extra/x86_64/evolution-bogofilter/" rel="nofollow noopener" translate="no" target="_blank">https://archlinux.org/packages/extra/x86_64/evolution-bogofilter/</a>

Howard Chu @ SymasMainsail EVM - switched from their homegrown <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> integration layer to using LMDB via heed (from meilisearch). <a href="https://arkscic.com/blog/mainsail-evm-development-highlights-whats-coming" rel="nofollow noopener" translate="no" target="_blank">https://arkscic.com/blog/mainsail-evm-development-highlights-whats-coming</a>

Howard Chu @ SymasTorn Write Detection and Protection with references to <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> <a href="https://transactional.blog/blog/2025-torn-writes" rel="nofollow noopener" translate="no" target="_blank">https://transactional.blog/blog/2025-torn-writes</a>LMDB is immune by design; the essential write that marks a txn as committed is only 120 bytes long. Guaranteed to be atomic even on ancient HDDs with 512 byte sectors.

Howard Chu @ SymasLow-Level Memory Management in Scalable Distributed Architectures: Approaches to Improving Reliability and Performance of Digital Services Terletska Khrystyna Lviv Polytechnic National University, 79013, Ukraine, LvivThe study includes practical examples from <a href="https://mastodon.social/tags/RocksDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#RocksDB</a> and <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> <a href="https://ijrpr.com/uploads/V6ISSUE4/IJRPR42112.pdf" rel="nofollow noopener" translate="no" target="_blank">https://ijrpr.com/uploads/V6ISSUE4/IJRPR42112.pdf</a>

Howard Chu @ SymasLooks like a Chinese blog post about reDB. Their benchmark shows nothing else touches <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a>'s read performance. <a href="https://learnblockchain.cn/article/14194" rel="nofollow noopener" translate="no" target="_blank">https://learnblockchain.cn/article/14194</a>

Howard Chu @ SymasMeta Open Materials 2024 (OMat24) Dataset"The OMat24 dataset contains a mix of single point calculations of non-equilibrium structures and structural relaxations. The dataset contains structures labeled with total energy (eV), forces (eV/A) and stress (eV/A^3). The dataset is provided in ASE DB compatible <a href="https://mastodon.social/tags/lmdb" class="mention hashtag" rel="nofollow noopener" target="_blank">#lmdb</a> files."<a href="https://huggingface.co/datasets/facebook/OMAT24" rel="nofollow noopener" translate="no" target="_blank">https://huggingface.co/datasets/facebook/OMAT24</a>Note - distributing reference data in LMDB files isn't really recommended, since LMDB files are byte order- and wordsize-dependent.

Howard Chu @ SymasHelixDB is a graph-oriented DBMS designed for nearest-neighbor searches over vectors, written in <a href="https://mastodon.social/tags/rust" class="mention hashtag" rel="nofollow noopener" target="_blank">#rust</a>. It uses <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> as its underlying storage engine.<a href="https://dbdb.io/db/helixdb" rel="nofollow noopener" translate="no" target="_blank">https://dbdb.io/db/helixdb</a>

Howard Chu @ SymasSiaqodb Abandoned Project Failed Company NoSQLSiaqodb is a NoSQL embedded document-oriented DBMS designed for .NET environments, built on <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a>HistoryThe project was originally closed-source and proprietary. In 2021, the source code was released to GitHub and the company backing its development ceased operations.<a href="https://dbdb.io/db/siaqodb" rel="nofollow noopener" translate="no" target="_blank">https://dbdb.io/db/siaqodb</a>

Howard Chu @ Symas"Popular" github repos for AI projects using <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> <a href="https://www.aibase.com/repos/topic/lmdb" rel="nofollow noopener" translate="no" target="_blank">https://www.aibase.com/repos/topic/lmdb</a>

Howard Chu @ SymasdpBento, a unified benchmark suite for measuring data processing tasks centered on DPUs, built with <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> <a href="https://arxiv.org/html/2504.05536v1" rel="nofollow noopener" translate="no" target="_blank">https://arxiv.org/html/2504.05536v1</a>DPUs are programmable network interface cards (NICs) or SmartNICs that allow for data-path customizations and optimizations based on System-on-a-Chip (SoC) architectures. The target data-path operations include packet processing and disaggregated storage protocols in data center networks.

Howard Chu @ SymasBest way to store ~30 lists each with 2 values?"<a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> smokes everything out there" - this guy gets it. <a href="https://www.reddit.com/r/csharp/comments/1jrnogt/comment/mlho3v6/" rel="nofollow noopener" translate="no" target="_blank">https://www.reddit.com/r/csharp/comments/1jrnogt/comment/mlho3v6/</a>

Howard Chu @ SymasYes, I recommend all our competitors just use <a href="https://mastodon.social/tags/postgres" class="mention hashtag" rel="nofollow noopener" target="_blank">#postgres</a>; makes it trivial to blow them away in competitive benchmarking... <a href="https://mastodon.social/tags/lmdb" class="mention hashtag" rel="nofollow noopener" target="_blank">#lmdb</a> <a href="https://mastodon.social/tags/openldap" class="mention hashtag" rel="nofollow noopener" target="_blank">#openldap</a>

Howard Chu @ Symas<a href="https://mastodon.social/tags/BoltDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#BoltDB</a> is a Go rewrite of <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> (mostly; it has a lot more limitations). It's no longer maintained. Since 2021 some forks had a backdoor giving remote command access to machines.<a href="https://snyk.io/blog/go-malicious-package-alert/" rel="nofollow noopener" translate="no" target="_blank">https://snyk.io/blog/go-malicious-package-alert/</a>All of this is inconceivable for LMDB, since it has no other dependencies. Also, the thought of an embedded DB engine having access to any networking APIs at all is just mindboggling.The Go build system, and its automatic pulling of dependencies from github, is ludicrous.<a href="https://mastodon.social/tags/golang" class="mention hashtag" rel="nofollow noopener" target="_blank">#golang</a>

Howard Chu @ Symas<a href="https://udongein.xyz/users/lispi314" class="u-url mention" rel="nofollow noopener" target="_blank">@lispi314</a> by the way, fyi, a Twitter engineer wrote an <a href="https://mastodon.social/tags/LMDB" class="mention hashtag" rel="nofollow noopener" target="_blank">#LMDB</a> backend for <a href="https://mastodon.social/tags/git" class="mention hashtag" rel="nofollow noopener" target="_blank">#git</a> back in 2015. <a href="https://lore.kernel.org/git/1435020656.28466.8.camel@twopensource.com/" rel="nofollow noopener" translate="no" target="_blank">https://lore.kernel.org/git/1435020656.28466.8.camel@twopensource.com/</a>"The [LMDB] backend runs git for-each-ref about 30% faster than the files backend with fully-packed refs on a repo with ~120k refs. It's also about 4x faster than using fully-unpacked refs. "The git devs then spent the next 5 years nitpicking the patch, literally to death. In 2020 the patchset died, the original author left twitter for a new job.

Recent searches

Search options

Administered by:

Server stats:

#lmdb