here's a random paperpost, from 29 October 2010:
Source Magazine: Music of the Avant Garde, Pt 2
https://www.paperposts.me/source-magazine-music-of-the-avant-garde-pt-2
#magazine #art #fluxus #music #avantgarde #print #60sand70s #performance #paperpost
RT @rumgewieselt: Gerade 59 Token/Sekunde auf 3× GTX 1080 Ti (Pascal, 2017) erreicht
mehr auf Arint.info
<p>RT @rumgewieselt: Gerade 59 Token/Sekunde auf 3× GTX 1080 Ti (Pascal, 2017) erreicht</p> <p><a href="https://arint.info/@Arint/116545453166339435">mehr</a> auf <a href="https://arint.info/">Arint.info</a></p> <p>#GPU #llamacpp #LLM #MoE #Performance #Qwen #arint_info</p> <p><a href="https://x.com/rumgewieselt/status/2052457151625642356#m">https://x.com/rumgewieselt/status/2052457151625642356#m</a></p>
Removing fsync from our local storage engine
https://fractalbits.com/blog/remove-fsync/
#HackerNews #removingfsync #localstorage #engine #performance #optimization #database

Recently, The Document Foundation published an open letter to European citizens. We asked Euro-Office – the new coalition forming around a European alternative for productivity – whether ODF (the Open Document Format) would be its native document format. Unfortunately, we have not yet received a reply, and this confirms – at least in part – the suspicion that Euro-Office will join Microsoft’s allies in a strategy to lock in European citizens, who will see their content snatched away by a company that – in words only – presents itself as a defender of digital sovereignty. With the open letter, we have raised an issue that the general debate is not yet grasping: digital sovereignty is not determined solely by the terms of the licence and the location of the server, but by the format in which documents are created, stored and exchanged. We were able to pose our question publicly, with confidence, because we represent something extremely solid – support for the single open and standard format: ODF – which has been built up over twenty years by many people, whose names rarely appear in press releases. The foundations underpinning the political moment Germany has established by law that ODF
Part 3 of the StyloBot Release Series is up.
https://www.mostlylucid.net/blog/stylobot-release-reliability
This one is less about bots and more about the reality of long-running .NET based systems: everything that learns from traffic eventually accumulates.
Came from one of my periodic reliability reviews where StyloBot’s vector layer had drifted to 13GB on the .NET Large Object Heap due to the wrong abstraction (in-process HNSW behaving like an unbounded cache).
The interesting part wasn’t the fix. It was recognising that the architecture itself was wrong for the runtime pattern.
Covers:
how I periodically review long-running services
using dotnet-counters, dotMemory and dotTrace to find growth
why “just add a cap” is often the wrong answer
replacing unbounded ANN structures with bounded hot caches + compacted persistence
taking the vector layer from 13GB LOH to <6MB
The broader point applies to any system that “remembers”:
bot detection, fraud scoring, recommendations, anomaly detection, RAG pipelines, adaptive systems.
Fix the shape, not the symptom.
#dotnet #aspnetcore #performance #architecture #ai #rag #observability
Making Julia as Fast as C++
https://flow.byu.edu/posts/julia-c++
#HackerNews #Julia #C++ #speed #optimization #programming #languages #performance