Benchmarking von selbst gehosteten LLMs (clt26) episode artwork

EPISODE · Mar 29, 2026 · 54 MIN

Benchmarking von selbst gehosteten LLMs (clt26)

from Chaos Computer Club - recent audio-only feed · host David Georg Reichelt

Large Language Models (LLMs) revolutionieren derzeit Arbeitsprozesse in der Dokumentenverarbeitung, der Softwareentwicklung und dem Wissensmanagement. Doch die Bequemlichkeit der Nutzung großer Cloud-Anbieter hat einen hohen Preis: Den Verlust der digitalen Souveränität. Wer weder Datenschutz noch volle Kontrolle über die zugrunde liegenden Prozesse garantieren kann, begibt sich in eine riskante Abhängigkeit. Die Lösung ist der Betrieb eigener Modelle auf kontrollierter Infrastruktur. Da High-End-Enterprise-GPUs (wie die H100) oft das Budget sprengen, haben wir am Rechenzentrum der Universität Leipzig untersucht, wie sich LLMs auf „erschwinglicher“ Hardware im Rechenzentrums-Maßstab realisieren lassen. In diesem Vortrag präsentieren wir die Ergebnisse unserer Benchmarking-Studie im Kontext des Kiara-Clusters. Wir analysieren die Performance aktueller Open-Source-Modelle – verschiedene Varianten von Llama 3, Llama 4 sowie DeepSeek – auf NVIDIA L40s-Karten. Dabei berichten wir von unseren Erfahrungen bei der Einrichtung und den Fallstricken im Betrieb. Licensed to the public under http://creativecommons.org/licenses/by/4.0 about this event: https://chemnitzer.linux-tage.de/2026/de/programm/beitrag/279

Large Language Models (LLMs) revolutionieren derzeit Arbeitsprozesse in der Dokumentenverarbeitung, der Softwareentwicklung und dem Wissensmanagement. Doch die Bequemlichkeit der Nutzung großer Cloud-Anbieter hat einen hohen Preis: Den Verlust der digitalen Souveränität. Wer weder Datenschutz noch volle Kontrolle über die zugrunde liegenden Prozesse garantieren kann, begibt sich in eine riskante Abhängigkeit. Die Lösung ist der Betrieb eigener Modelle auf kontrollierter Infrastruktur. Da High-End-Enterprise-GPUs (wie die H100) oft das Budget sprengen, haben wir am Rechenzentrum der Universität Leipzig untersucht, wie sich LLMs auf „erschwinglicher“ Hardware im Rechenzentrums-Maßstab realisieren lassen. In diesem Vortrag präsentieren wir die Ergebnisse unserer Benchmarking-Studie im Kontext des Kiara-Clusters. Wir analysieren die Performance aktueller Open-Source-Modelle – verschiedene Varianten von Llama 3, Llama 4 sowie DeepSeek – auf NVIDIA L40s-Karten. Dabei berichten wir von unseren Erfahrungen bei der Einrichtung und den Fallstricken im Betrieb. Licensed to the public under http://creativecommons.org/licenses/by/4.0 about this event: https://chemnitzer.linux-tage.de/2026/de/programm/beitrag/279

NOW PLAYING

Benchmarking von selbst gehosteten LLMs (clt26)

0:00 54:27

No transcript for this episode yet

We transcribe on demand. Request one and we'll notify you when it's ready — usually under 10 minutes.

Breaking News Show | eTurboNews Juergen Thomas Steinmetz News is relevant to the global travel and tourism industry, human rights and global issues.Breaking news when it happens and only from the source. That Hoarder: Overcome Compulsive Hoarding That Hoarder Hoarding disorder is stigmatised and people who hoard feel vast amounts of shame. This podcast began life as an audio diary, an anonymous outlet for somebody with this weird condition. That Hoarder speaks about her experiences living with compulsive hoarding, she interviews therapists, academics, researchers, children of hoarders, professional organisers and influencers, and she shares insight and tips for others with the problem. Listened to by people who hoard as well as those who love them and those who work with them, Overcome Compulsive Hoarding with That Hoarder aims to shatter the stigma, share the truth and speak openly and honestly to improve lives. HOMELAND HOMELAND The Church is a body not a building. It's the bride of Jesus Christ! Jesus is coming back for a mature bride. That means it's time for the church of Jesus Christ to move from milk to meat. This is the hour of maturity!HOMELAND is an announcement that the church is being set free. Only the church has the ability to transform the world. The kingdom's of this world will become the kingdoms of our Lord and Savior!All of creation has been waiting for this moment! Sons and daughters of God are rising up and taking their seat! LIGHTS, CAMERA, SMILE! Creatives Club Media Lights, Camera, Smile, is a podcast for anyone with a dream to share something with the world, out of the overflow of themselves - be it their mind, their heart, their personalities, and much more. Each of us are alive in this moment in time, with an innate ability to have ideas and create various things to benefit both ourselves and the people around us for a reason, and here, you will find the encouragement, the inspiration, and the motivation to do just that. Hosted by Cicily, founder of Creatives Club, she dives into various topics surrounding creativity and business. Exploring entrepreneurship for creatives in a corporate reality, sharing tips and tricks in a media centered company, answering questions regarding what a creative actually is are just a few of the things discussed on this podcast. Be encouraged to create for yourself as Cicily gets vulnerable by pivoting the camera to herself for the first time.To submit questions for Cicily to answer, or have her address certain t

Frequently Asked Questions

How long is this episode of Chaos Computer Club - recent audio-only feed?

This episode is 54 minutes long.

When was this Chaos Computer Club - recent audio-only feed episode published?

This episode was published on March 29, 2026.

What is this episode about?

Large Language Models (LLMs) revolutionieren derzeit Arbeitsprozesse in der Dokumentenverarbeitung, der Softwareentwicklung und dem Wissensmanagement. Doch die Bequemlichkeit der Nutzung großer Cloud-Anbieter hat einen hohen Preis: Den Verlust der...

Can I download this Chaos Computer Club - recent audio-only feed episode?

Yes, you can download this episode by clicking the download button on the episode player, or subscribe to the podcast in your preferred podcast app for automatic downloads.
URL copied to clipboard!