Category desk

Research

Technical deep dives, reproducible tests, and tool evaluations.

Infographic showing DeepSeek V4.1 Flash served across three DGX Spark nodes with key throughput figures

Multi-Node DGX Spark: Tensor-Parallel Sizing for 2x and 4x Clusters

A 135 GB model against a 128 GB box forces a cluster. Here is the decision order, the parallelism maths, and what the interconnect really costs per token.

Conceptual 3D illustration of sparse Mixture-of-Experts architecture routing in Laguna S 2.1.

Vllm Vs Sglang Research

SGLang vs vLLM vs TensorRT-LLM: Choosing a Production Inference Engine

Paged KV, radix-tree prefix sharing and ahead-of-time compilation all pull in different directions. Work out which one your own workload should pay for.

Open-weight AI model fit guide for a 24GB Apple Silicon Mac in 2026

Kv Cache Vram Research

KV Cache Explained: Why Long Context Costs More VRAM Than Your Weights

A model that fits can still die on one long prompt. Here is the arithmetic behind the second allocation that grows with context, and how to shrink it.

Protocol schematic poster for a 128GB Apple Silicon local AI stack led by Qwen3.6

Gpu Vram Requirements Research

Will It Fit? VRAM Math for Open-Weight Models Across Every Quantization Format

Weights are only one of four memory buckets, and past roughly 115k tokens the KV cache costs more than the model does. Here is the arithmetic in full.

Infographic mapping one to four NVIDIA DGX Sparks to the recommended local model at each node count: Qwen3.8 Flash Next, DeepSeek V4 Flash, GLM-5.3 Flash, DeepSeek V4.1 Flash.

Inference Engineering Research

Inference Engineering: A Practical Primer for Serving Your Own Models

Prefill is compute-bound, decode is bandwidth-bound, and KV cache is what really caps concurrency. The metrics, memory maths and tradeoffs behind serving.

Illustration of a DDR5 RAM module depicted as a soaring price peak towering over small PCs and a server rack, symbolizing surging memory prices.

Nvfp4 Quantization Research

NVFP4 vs FP8 vs MXFP4 vs EXL3: The Quantization Format Guide for Local LLMs

The same 320B model is 328 GB in FP8 and near 135 GB at 4 bits. Bytes-per-parameter math, per-format tradeoffs, and how to size a box that actually fits.

FIPS 140-2 sunset infographic detailing the September 21, 2026 transition to Historical status, procurement blocks, and FIPS 140-3 requirements.

Fips 140 2 Sunset Research

FIPS 140-2 Sunset September 2026: The Migration Guide to FIPS 140-3

Prepare for the September 2026 FIPS 140-2 sunset. Learn the procurement impacts, technical breaking changes, and exact steps to migrate stacks to FIPS 140-3.

Architecture flow diagram of EU Cyber Resilience Act Article 14 showing PSIRT notifications routing through the ENISA Single Reporting Platform to national CSIRTs within 24-hour and 72-hour statutory deadlines.

Cyber Resilience Act Research

EU Cyber Resilience Act Article 14: Navigating Mandatory Vulnerability Reporting Requirements for Enterprise PSIRTs

Explore the EU Cyber Resilience Act Article 14 mandatory 24-hour vulnerability reporting rules, ENISA platform architecture, and PSIRT compliance steps.