We benchmarked an RTX 3090 over USB4 and profiled every kernel. GPUs use 1.2-1.6% of their memory bandwidth. The bottleneck is the compiler, not the cable.
egpu for mac - tinygrad docs
TinyGPU app lets you use AMD and NVIDIA GPUs on macOS over USB4/Thunderbolt with tinygrad.
the tiny corp on Twitter / X
If you have a Thunderbolt or USB4 eGPU and a Mac, today is the day you've been waiting for! Apple finally approved our driver for both AMD and NVIDIA. It's so easy to install now a Qwen could do it, then it can run that Qwen... pic.twitter.com/daUsyBHh1W— the tiny corp (@__tinygrad__) April 1, 2026
Combining NVIDIA DGX Spark + Apple Mac Studio for 4x Faster LLM Inference with EXO 1.0
Disaggregating Prefill and Decode: Faster First Tokens, Faster Streams
Combining NVIDIA DGX Spark + Apple Mac Studio for 4x Faster LLM Inference with EXO 1.0
Disaggregating Prefill and Decode: Faster First Tokens, Faster Streams
vik on Twitter / X
Photon, our inference engine, isn't fast just because of GPU kernels. A lot of the speedup comes from engine-level work: request scheduling, prefix caching, image processing, all tuned to keep the GPU saturated. https://t.co/3M7eFcFKo5— vik (@vikhyatk) May 2, 2026
0xSero on Twitter / X
I told y’all this is the move. Heterogenous hardware is the way forward. Large cheap pools of mixed memory + specialized accelerators (Nvidia GPUs, DGX Spark, Cerebras wafers) The next year will be dominated by solutions that split the stack. - 3000$ for a used Mac Studio… https://t.co/zMlScSnJ0X— 0xSero (@0xSero) April 1, 2026
XiongjieDai/GPU-Benchmarks-on-LLM-Inference
Multiple NVIDIA GPUs or Apple Silicon for Large Language Model Inference?
Nvidia invests $5 billion into Intel to jointly develop PC and data center chips
Intel will help build x86 chips with Nvidia RTX GPU chiplets
InferenceMAX™: Open Source Inference Benchmarking
NVIDIA GB200 NVL72, AMD MI355X, Throughput Token per GPU, Latency Tok/s/user, Perf per Dollar, Tokens per Provisioned Megawatt, DeepSeek R1 670B, GPTOSS 120B, Llama3 70B
eGPUs on NixOS
Getting an Nvidia eGPU working on NixOS: Thunderbolt authorization, kernel/module setup, and fixing Gamescope glitches with the latest beta driver.
the tiny corp on Twitter / X
Qwen 3.5 27B getting 18.5 tok/s on Mac Mini with external 7900XTX. It should be able to be 3x faster than this with work, SSM stuff is still in PR. Hopefully Mac eGPU support brings in devs. pic.twitter.com/2aMkUpXY1S— the tiny corp (@__tinygrad__) April 1, 2026
Open Source AI Inference Benchmark | InferenceX
Compare AI inference performance across GPUs and frameworks. Real benchmarks on NVIDIA GB200, B200, AMD MI355X, and more. Free, open-source, continuously updated.
NVIDIA Shatters MoE AI Performance Records With a Massive 10x Leap on GB200 'Blackwell' NVL72 Servers, Fueled by Co-Design Breakthroughs
Scaling performance on MoE AI models is one of the industry constraints, but it appears that NVIDIA has managed to make a breakthrough.
ASUS Ascent GX10
Desktop AI supercomputer delivering up to 1 petaFLOP performance, powered by NVIDIA GB10 Grace Blackwell Superchip, supporting OpenClaw and Hermes Agent.
Building Linux kernel on macOS natively