Modders have developed a software solution, CMP Unlocker, to restore disabled VRAM on Nvidia's CMP 170HX cryptocurrency mining GPUs. This software increases the memory capacity of the 8GB variant to 64GB and the 10GB variant to 40GB, addressing limitations imposed by Nvidia's firmware.
The CMP 170HX uses the GA100 (Ampere) silicon, which is also found in the A100 accelerator. The GA100 is a monolithic chip with on-package memory, meaning the total memory is physically present. Nvidia disabled certain memory stacks through firmware on the CMP 170HX, which is typically made from binned GA100 silicon that did not meet A100 quality standards. The unlocker reactivates these disabled memory stacks.
The exploit is entirely software-based and does not require any physical modifications to the CMP 170HX. It functions by bypassing Nvidia's Falcon security microprocessor, a method detailed in Jon Pry's research paper, "A Canary in the Crypto Mine: Defeating Stack Protection in a GPU Secure Coprocessor."
This software modification allows users to access the full memory potential of the CMP 170HX, which could be beneficial for applications requiring large amounts of VRAM, such as AI workloads. The ability to unlock significant memory on these older mining cards offers an alternative in a market with high GPU prices and memory shortages.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's briefNew every morning, and the back catalogue is archived by date.
A software-based solution called CMP Unlocker has been developed to restore previously disabled VRAM on Nvidia's CMP 170HX cryptocurrency mining GPUs, increasing their memory capacity from 8GB to 64GB or 10GB to 40GB. This development allows users to access the full physical memory present on the GA100 silicon, which was previously limited by firmware, potentially offering a way to address memory shortages for AI and other demanding applications.