Your RTX 3090 Gets 25–67% More BF16 GEMM With GPU-unlock
GPU-unlock, a Linux kernel module, doubles FP/BF16 issue rate on RTX 30-series cards: 3090 Ti BF16 GEMM 87.5 to 157.4 TFLOPS. Decode speed does not change.
10 posts
GPU-unlock, a Linux kernel module, doubles FP/BF16 issue rate on RTX 30-series cards: 3090 Ti BF16 GEMM 87.5 to 157.4 TFLOPS. Decode speed does not change.
Nvidia says the RTX 5090 is not discontinued but supply is 'somewhat constrained'. Cheapest new unit in September: $4,999.99 vs a $1,999 MSRP; listings hit $8,500.
The Register benchmarks HP's N1X-powered OmniBook Ultra 16: 67.6 tok/s on Qwen3.6-35B, 128 GB unified memory, 300 GB/s bandwidth, and a 50:50 CPU/GPU memory split. Full llama.cpp numbers and tradeoffs.
NVIDIA's CUDA Rust gives Rust developers two tracks — SIMT via cuda-oxide and Tile via cutile-rs — to write GPU kernels natively in Rust, compiled to PTX, with compile-time memory safety.
AMD Ryzen AI Max PRO 390 with 64GB unified memory runs 32B models at 8.6 t/s — parity with a 16GB RTX 5080 laptop, no discrete GPU needed.
AMD's Threadripper Halo Station packs a 96-core CPU with dual MI350P accelerators and up to 576GB of HBM3E — a $150K desktop that runs trillion-parameter models locally.
NVIDIA's NVHBM moves memory control to the HBM stack itself, not the GPU. This architectural shift reveals memory bandwidth, not compute cores, is the real scaling bottleneck for AI inference.
vLLM 0.27.0 (Aug 10) ships JIT warmup and runner-owned Triton warmup that eliminate first-request compile stalls, plus Kimi K3 support — but the torch 2.13.0 upgrade is a breaking change that needs migration testing before you bump.
NVIDIA cut RTX 50 supply by 20% and sidelined the 5070 Ti. Street prices are up 19%. Here's how it changes your local AI build.
AMD has notified its supply chain of a ~10% GPU price increase driven by a DRAM shortage. DRAM contract prices rose 90–95% QoQ in Q1 2026. Here's what to buy, hold, or skip.