florianleibert
Re:5.6x throughput on Kimi K2.6 by speculating less
on Tuesday April 21, 2026 @06:24AM (#47846892)
hi!
open sourced a serving config for Kimi K2.6 from 90 tok/s 508 tok/s on 8xMI300X. Same weights / 0 quality loss.
Scaling is linear @15.8 tok/s per slot latency is constant. REpo has command launcher, Dockerfile, benchmark tool. Known limitations: BF16 KV only (FP8 crashes due to an AITER 384-expert constraint)
Nvidia's moat is not what it used to be
For years, the lock-in was dead simple: CUDA.Want top-tier performance? You wrote CUDA. Do that once, and you were all-in on NVIDIA. The ecosystem compounded—libraries, tooling, docs, talent—everything reinforcing the same gravity well.That world is starting to crack.We’re entering a phase where low-level code isn’t a rare skill anymore. Models are capable at generating kernels, bindings, and...
Re:Wilbur: A backtracking and self-training agent for performing tasks...
on Friday April 12, 2024 @12:19PM (#40014606)
This is pretty cool.
First Interaction with Kubernetes on DC/OS
3 Definitions of Cloud and Why They Matter
Uber Drives Distributed Deep Learning Forward
Training Your Custom Model with the DC/OS TensorFlow Package
Distributed TensorFlow on Any Infrastructure with Mesosphere DC/OS
Mesosphere DC/OS Brings Large-Scale Real-Time Processing to Geospatial Data
Distributed TensorFlow on Any Infrastructure with Mesosphere DC/OS
Challenges in Cycle Counting
Containers and Distributed Systems: Where They Came from and Where They’re Going
Re:Continuous Delivery for DC/OS with Spinnaker
on Saturday September 09, 2017 @10:46PM (#15210298)
There is also a tutorial on the Spinnaker website.
https://www.spinnaker.io/guides/tutorials/codelabs/dcos-sour...
Continuous Delivery for DC/OS with Spinnaker
DC/OS 1.10 – includes Kubernetes
Re:Using Tensorflow to Recognize Saimaa Ringed Seals from Webcam
on Thursday June 22, 2017 @05:07PM (#14615376)
Here is a tutorial on the easiest way to run Tensorflow: https://dcos.io/blog/2017/tutorial-deep-learning-with-tensor...
Deep Learning with TensorFlow
Re:Using Apache Spark with Intel BigDL on Mesosphere DC/OS
on Thursday June 22, 2017 @04:51PM (#14615255)
Here is another good tutorial - about Tensorflow on DC/OS:
https://mesosphere.com/blog/2017/05/11/deep-learning-tensorf...
Using Apache Spark with Intel BigDL on Mesosphere DC/OS
The SMACK Stack Is the New LAMP Stack – Mesosphere
Accelerating Machine Learning with GPUs and DC/OS 1.9 – Mesosphere
Networking for Docker Containers (a Primer) – Mesosphere
Serverless Computing on DC/OS with Galactic Fog
Container orchestration on DC/OS with ESRI
Re:Concord – High Performance Stream Processing with C++ and Mesos
on Friday May 13, 2016 @02:33PM (#11692291)
Super impressive performance!
Re:Azure Container Service is now generally available
on Tuesday April 19, 2016 @02:05PM (#11528867)
Here is more information about the underlying technology, which also powers e.g. Twitter (Apache Mesos and DC/OS).
https://azure.microsoft.com/en-us/blog/microsoft-joins-the-n...
Re:DC/OS – Datacenter Operating System
on Tuesday April 19, 2016 @04:58AM (#11525721)
Here is a reference to the original mesos (core of DC/OS) paper by Ben Hindman, Matei Zaharia and Andy Konwinsky: https://www.cs.berkeley.edu/~alig/papers/mesos.pdf