A day after DLSS 5's nvngx_dlssnr.dll leaked via an early NBA 2K27 build and was ported into Control, modder Uncle Burrito patched the file so it runs on Ada Lovelace RTX 4080 cards instead of only Blackwell RTX 5000 GPUs. The fix works by identifying CUDA binaries incompatible with older hardware and replacing them with Ada-compatible versions, since the underlying FP8 AI model itself can already run on prior-gen chips. Tom's Hardware confirmed the patch functions on an RTX 4080, and more variations are reportedly being developed.
tomshardware.com
· 2026-08-28
OpenAI revealed its first custom AI chip, called Jalapeño, claiming it delivers industry-leading speed and efficiency for inference workloads. The chip, co-developed with Broadcom, is set to be deployed in OpenAI's own infrastructure by year-end, with two more generations already in development.
cnbc.com
· 2026-08-26
Nvidia is working to extend CUDA compatibility beyond x86-64 and aarch64 to RISC-V CPUs, according to a Hot Chips 2026 presentation. The company laid out strict hardware requirements, including RVA23 compliance, adherence to RISC-V server SoC and platform specs, ACPI support, and PCIe coherency, to ensure CUDA runs efficiently rather than falling back on lowest-common-denominator code.
chipsandcheese.com
· 2026-08-24
GoKawiil