Skip to content
Catalyst Note CN markCNCatalystNote

Noting the changes that set others in motion.

AI hardware

Open-source UEFI tool unlocks full tensor performance on RTX 20, 30, 40-series GPUs

A bootable USB utility removes Nvidia’s AI math cap on Windows systems without flashing firmware or patching drivers, according to the published report

Open-source UEFI tool unlocks full tensor performance on RTX 20, 30, 40-series GPUs
Depth
65%Tom's Hardware reported a 65% increase in FP16 math performance on a developer’s RTX 4060 laptop · reported
1 outlet · newest source 357h ago

Key points

  • nv-uefi-unlocker lifts the FP16 tensor rate limit on RTX 20, 30, and 40-series desktops and laptops.
  • The tool runs from a FAT32 USB drive before the OS loads, avoiding VBIOS flashes or driver modifications.
  • Tom's Hardware reported a 65% increase in FP16 math performance on a developer’s RTX 4060 laptop.

Tom's Hardware reported that an open-source UEFI utility called nv-uefi-unlocker removes Nvidia’s tensor performance cap on RTX 20, 30, and 40-series GPUs under Windows. The tool operates from a bootable USB drive, allowing users to bypass the half-speed AI throttle without flashing video BIOS or patching drivers.A new USB tool lets Windows users remove Nvidia's speed limit on AI math for older and newer RTX graphics cards.The nv-uefi-unlocker UEFI payload bypasses Nvidia's tensor rate cap on RTX 20-40 series by manipulating speed-selector registers before driver initialization. This development simplifies a previously complex modification process for GeForce owners.

Evidence: reportedOpen-source USB tool unlocks Nvidia's half-speed GPU AI limiter on Windows — lifts GPU's FP16 math performance up to 65%, UEFI unlocker covers RTX 20, 30, and 40-series desktops and laptops

aifleet-tomshardware

The claim

The nv-uefi-unlocker project uses the community V67 payload with read-back verification to trick Nvidia’s signed Booter firmware into unlocking speed-selector registers. According to Tom's Hardware, this allows tensor units to run at full rate. The release includes scripts that download and verify Nvidia’s public 610.43.03 driver packages to rebuild firmware files locally, ensuring no proprietary firmware is distributed directly.

Evidence: reportedOpen-source USB tool unlocks Nvidia's half-speed GPU AI limiter on Windows — lifts GPU's FP16 math performance up to 65%, UEFI unlocker covers RTX 20, 30, and 40-series desktops and laptops

aifleet-tomshardware

Evidence: reportedOpen-source USB tool unlocks Nvidia's half-speed GPU AI limiter on Windows — lifts GPU's FP16 math performance up to 65%, UEFI unlocker covers RTX 20, 30, and 40-series desktops and laptops

aifleet-tomshardware

Evidence: reportedOpen-source USB tool unlocks Nvidia's half-speed GPU AI limiter on Windows — lifts GPU's FP16 math performance up to 65%, UEFI unlocker covers RTX 20, 30, and 40-series desktops and laptops

aifleet-tomshardware

The response

The tool requires systems to boot in UEFI mode with Secure Boot disabled or set to trust a self-signed EFI. Tom's Hardware noted that the unlock does not persist through power loss, meaning users must run the utility from a boot menu at every startup. The project credits Zagitta’s gpu-unlock as its porting source and builds on earlier community efforts like gpu-unlock-ada.

Why this matters now

Previously, unlocking these limits relied on manually installed Linux modules, with Windows support remaining experimental. The new UEFI approach makes the modification safer through read-back verification and more accessible to Windows users. Tom's Hardware stated that having one tool for many GeForce owners on Windows should significantly open up the hack to a wider audience.

What buyers should weigh

The tool modifies firmware behavior before the OS loads, requiring Secure Boot to be off or set to trust a self-signed EFI. The unlock does not persist through power loss and must be run again on every boot. No official Nvidia fix or statement was included in the report.

Why it matters

  • The tool enables higher FP16 math performance on consumer-grade hardware, potentially affecting cost-benefit analyses for edge AI workloads.
  • Bypassing vendor-imposed performance limits raises questions about hardware lifecycle management and support eligibility.
  • The ease of use on Windows expands the attack surface for unauthorized hardware modifications in enterprise environments.
  • Read-back verification adds a layer of safety, but the process still requires manual intervention at every boot.

Topics: Nvidia, GPU, AI performance, UEFI, open-source

Updated

The strongest doubt

Counterpoint from the sources

The unlock requires Secure Boot to be disabled or configured to trust a self-signed EFI, which may reduce system security.

Ask the article

Local keyword search of this article; no generated answers.

Did this article give you the evidence you needed?

No vote saved on this device.
Your choice stays on this device. No public tally.

Questions readers ask

Does this tool require flashing the GPU's BIOS?

No, the tool runs from a USB drive before the OS loads and does not require flashing the video BIOS or modifying drivers.

Which Nvidia GPU series are compatible?

The tool supports RTX 20, 30, and 40-series desktop and laptop GPUs, according to the published report

Is the performance boost permanent?

No, the unlock does not persist through power loss and must be run again on every boot.

Connected stories

This storyUS Commerce Dept delayed Entity List update, omitting Megaspeed despite agency pushUS Commerce Dept delayed Entity Nvidia Board Meets to Discuss CEO Jensen Huang’s Future Amid Spy ScandalNvidia Board Meets to Discuss CE600,000 Nvidia H200 chips pile up in Hong Kong awaiting Beijing's China import clearance600,000 Nvidia H200 chips pile uNvidia board moves to oust Huang after second affair report surfacesNvidia board moves to oust Huang

Reading receipt

  • nv-uefi-unlocker lifts the FP16 tensor rate limit on RTX 20, 30, and 40-series desktops and laptops.
  • The tool runs from a FAT32 USB drive before the OS loads, avoiding VBIOS flashes or driver modifications.
  • Tom's Hardware reported a 65% increase in FP16 math performance on a developer’s RTX 4060 laptop.
Time spent: 0:00 · Depth: Full

Sources

  1. Tom's Hardware

Related stories

Background reading