Tracked topic
… It's impressive, but read between the lines: we're several years into the LLM craze, and if that kind of run can still be newsworthy, it tells you how CUDA-first the training world remains. The broader ecosystem also moves CUDA-first. New models? CUDA first. AI software packages? CUDA first. …
… And yet, most of the conversation revolves strictly around Nvidia, and all of it has to do with CUDA's head start against ROCm . …
… AMD is going about it differently Unified platform with quantization AMD has been testing its latest Ryzen AI Max+ chipset using ROCm, Ollama, and Qwen 3.5 models. …
… Sign in to your XDA account Nvidia and Microsoft recently unveiled the much-awaited N1X and N1 SoCs, now called RTX Spark . These Arm-based chips are meant to reinvent Windows PCs for the AI era, with Nvidia claiming that laptops powered by RTX Spark are purpose-built for agentic AI. …
… That said, given similar development devices like AMD's Ryzen AI Halo PC and Nvidia's DGX Spark Mini PC cost roughly $4,000, you can expect the Surface RTX Spark Dev Box to be pricey. …
… It supports almost every local AI tool most people actually use, and ROCm has come a long way. AMD's Gorgon Halo will further bolster its local AI capabilities in the high-end market. What remains is the gap between ROCm and CUDA, and closing it might take AMD a bit longer than it hopes.
… It's a $3,800 AI workstation running Arm Linux with 6,144 CUDA cores the same count as an RTX 5070 and 128GB of unified memory. The machine was designed to serve language models, not render frames, so it kind of went against the entire idea behind that chip. …
… My 4080 Founders Edition features 9,728 CUDA cores. The 5080 offers only a minor, roughly 10% increase in shading units, making the raw generation-over-generation rasterization gains an absolute joke for a card that launched years later. …
… The 128GB LPDDR5X memory that makes these machines interesting for AI workloads is the same memory that's been constrained enough to trigger massive price hikes, which likely explains why Nvidia is also planning RTX Spark configurations that start at 16GB of memory. …
… It's not a CUDA equivalent in maturity or scope, and plenty of local LLM tooling still uses Metal directly, but it shows Apple is aware that unified memory has become one of its strongest AI advantages. …