Derek Lewis
@dlewis.io
CTO & Data Scientist at Silex Data Solutions & CODEHR.ai. Opinions expressed are my own.
Wrote a prime sieve in Chip8 assembly to calculate primes from 0 to 255 and display them as blocks on the screen:
November 11, 2025 at 11:05 PM
Wrote a prime sieve in Chip8 assembly to calculate primes from 0 to 255 and display them as blocks on the screen:
You probably wouldn't know it from this top output, but I have a FSDP training run going on the DGX Spark cluster. No wasted CPU time spent processing interrupts or copying between buffers. RDMA networking is a wonderful thing.
October 31, 2025 at 10:45 PM
You probably wouldn't know it from this top output, but I have a FSDP training run going on the DGX Spark cluster. No wasted CPU time spent processing interrupts or copying between buffers. RDMA networking is a wonderful thing.
2x performance by adding the 2nd DGX Spark w/ the 200GbE interconnect to a distributed training run with Karpathy's nanochat. Brings base training down from 10 days to 5 days. Token throughput is 4x compared to single node run, but only because grad accumulated steps changed from 8 to 4.
October 31, 2025 at 8:48 PM
2x performance by adding the 2nd DGX Spark w/ the 200GbE interconnect to a distributed training run with Karpathy's nanochat. Brings base training down from 10 days to 5 days. Token throughput is 4x compared to single node run, but only because grad accumulated steps changed from 8 to 4.
200GbE network is up and running between the DGX Sparks. Having a high throughput cluster on a desk that consumes less than 400W of power under full load is awesome. NCCL benchmarks show near line-speed for AllGather.
October 31, 2025 at 4:32 PM
200GbE network is up and running between the DGX Sparks. Having a high throughput cluster on a desk that consumes less than 400W of power under full load is awesome. NCCL benchmarks show near line-speed for AllGather.
Waiting for a 200GbE interconnect cable to come in to connect my NVIDIA DGX Sparks. Did some NCCL connectivity and validation testing with the 10GbE ports in the meantime:
October 30, 2025 at 10:37 PM
Waiting for a 200GbE interconnect cable to come in to connect my NVIDIA DGX Sparks. Did some NCCL connectivity and validation testing with the 10GbE ports in the meantime:
NVIDIA DGX Spark #2 is up and running.
October 30, 2025 at 7:20 PM
NVIDIA DGX Spark #2 is up and running.
Womp womp - looks like NVIDIA NIM images aren't updated to CUDA 13.1, yet. That means no NIM on the DGX Spark for the time being except for a few custom images they have done. Unfortunate, because I really wanted to see mxfp4 & trt-llm w/ gpt-oss-120b.
October 29, 2025 at 5:40 PM
Womp womp - looks like NVIDIA NIM images aren't updated to CUDA 13.1, yet. That means no NIM on the DGX Spark for the time being except for a few custom images they have done. Unfortunate, because I really wanted to see mxfp4 & trt-llm w/ gpt-oss-120b.
Long context llama.cpp testing with the NVIDIA DGX Spark & gpt-oss-120b.
October 28, 2025 at 7:03 PM
Long context llama.cpp testing with the NVIDIA DGX Spark & gpt-oss-120b.
NVIDIA DGX Spark is up and running. Setup process was seamless. Now for some fine-tuning and CUDA development.
October 27, 2025 at 2:11 AM
NVIDIA DGX Spark is up and running. Setup process was seamless. Now for some fine-tuning and CUDA development.
Made the plunge and ordered a DGX Spark. Less interested in the inferencing performance and more interested in having the full Nvidia DGX stack on my desk for development.
October 17, 2025 at 9:10 PM
Made the plunge and ordered a DGX Spark. Less interested in the inferencing performance and more interested in having the full Nvidia DGX stack on my desk for development.
With gpt2-xl you can drop the positional encodings entirely and get decent output. The smaller the model the more dependent it is on the positional encodings to generate non-garbage output.
April 14, 2025 at 2:05 AM
With gpt2-xl you can drop the positional encodings entirely and get decent output. The smaller the model the more dependent it is on the positional encodings to generate non-garbage output.
We’re excited to announce the open sourcing of our AI Foundry Starter Template at Silex Data! This production-ready starter kit empowers you to build and deploy AI apps with LangChainAI/LangGraph, featuring streaming chat, robust Keycloak authentication, Kong's multi-model gateway, and OpenShift.
March 12, 2025 at 7:36 PM
We’re excited to announce the open sourcing of our AI Foundry Starter Template at Silex Data! This production-ready starter kit empowers you to build and deploy AI apps with LangChainAI/LangGraph, featuring streaming chat, robust Keycloak authentication, Kong's multi-model gateway, and OpenShift.
Doing some retro programming: Cocoa-based LLM streaming chat application on OS X 10.5 circa 2009 making calls to a FastAPI back-end, which streams events from a LangChain/LangGraph agent capable of making tool calls. 10.5 did not have a NSJSONSerialization, yet.
January 25, 2025 at 5:05 PM
Doing some retro programming: Cocoa-based LLM streaming chat application on OS X 10.5 circa 2009 making calls to a FastAPI back-end, which streams events from a LangChain/LangGraph agent capable of making tool calls. 10.5 did not have a NSJSONSerialization, yet.
Getting Siri to use ChatGPT isn’t 100% and the TeX formatting isn’t available in the Siri output when ChatGPT is used. Markdown formatting seems to there at a glance.
December 12, 2024 at 3:57 AM
Getting Siri to use ChatGPT isn’t 100% and the TeX formatting isn’t available in the Siri output when ChatGPT is used. Markdown formatting seems to there at a glance.
And logged into ChatGPT now on iOS 18.2.
December 12, 2024 at 3:56 AM
And logged into ChatGPT now on iOS 18.2.
And that global rate limit got hit:
December 11, 2024 at 11:59 PM
And that global rate limit got hit:
Looks like OpenAI is getting overloaded after the iOS 18.2 rollout. Trying to sign into my account on my phone after the upgrade to test out the new integration.
December 11, 2024 at 11:44 PM
Looks like OpenAI is getting overloaded after the iOS 18.2 rollout. Trying to sign into my account on my phone after the upgrade to test out the new integration.
At least the tracing works...
December 5, 2024 at 2:08 AM
At least the tracing works...
SNO (Single Node OpenShift) is up and running on the new MacBook Pro M4 Max w/ VMware Fusion. Being able to run a full OpenShift environment locally is very beneficial for local development. Have the ArgoCD, Logging (Loki), Observability, Serverless (Knative), and OpenTelemetry operators installed.
December 4, 2024 at 5:21 PM
SNO (Single Node OpenShift) is up and running on the new MacBook Pro M4 Max w/ VMware Fusion. Being able to run a full OpenShift environment locally is very beneficial for local development. Have the ArgoCD, Logging (Loki), Observability, Serverless (Knative), and OpenTelemetry operators installed.
New M4 Max w/ 128GB of memory and 4TB of storage is finally here! Planning to use it for MLX training/inference and OpenShift SNO in VMware Fusion. It will be replacing my M2 Max which “only” has 32GB of memory.
December 1, 2024 at 8:13 PM
New M4 Max w/ 128GB of memory and 4TB of storage is finally here! Planning to use it for MLX training/inference and OpenShift SNO in VMware Fusion. It will be replacing my M2 Max which “only” has 32GB of memory.