er, writes SS0=0x88888888 / SS1=0x8, does a PCIe FLR, then chains to the Windows boot manager.No driver patch. No VBIOS flash.

Source, build steps, capture procedure and benchmark script (GPL-2.0): github.com/ZeroHackz/gpu-unl…

Measured on 2× RTX 4060 Ti 16GB: 46.6 → 77.5 TFLOPS single 77.5 / 77.9 TFLOPS dual (+66%) INT8/FP8 and LLM decode are unaffected — this is the FP32-accumulate path only, so real gains land in prefill, diffusion and LoRA.

I ported the unlock to Windows. A small UEFI app runs before the NVIDIA driver, writes SS0=0x88888888 / SS1=0x8, does a PCIe FLR, then chains to the Windows boot manager. No driver patch. No VBIOS flash.

RTX 40-series: consumer cards ship with FP16/BF16→FP32 tensor throughput halved vs the professional parts on the same silicon. Same die, same tensor cores — just gated by firmware speed selectors.

RTX 40-series: consumer cards ship with FP16/BF16→FP32 tensor throughput halved. I ported the unlock to Windows. A tiny UEFI app flips the speed selectors before the driver loads, resets the card, then boots Windows. 4060 Ti: 46.6 → 77.5 TFLOPS (+66%) github.com/ZeroHackz/gpu-unl…

so it sucks.

@bundebunden @eurofootcom @ElementaryFPL

@wingedwork funny. Your logical is also illogical. My premise starts at that they studied all the data on the web and considered them a book to study. But when they "finished" their own book, they turn around and say don't read our book to study from.

thanks so much

I have never been more dissapointed in someone that I dont know

Best take, I've heard all week.

yeah this is insane.

no.

Bro, fix the formations glitch. I can't swap formations in-game without it being messed up positions-wise.

any updates on this question?

Very cool flex.

Geo politics and the long game. The Veterans game.

Yeah, that's impressive. Are you open to share workflow?

yeah im gonna pass. chinese models are way more value for money.