What's the optimal method to verify error correction during memory clock overclocking?
What's the optimal method to verify error correction during memory clock overclocking?
are there any apps that assist in testing this, or would I need to perform benchmarks and monitor the fps? gpu: gtx 1080ti
The fixes in GDDR5 work without being obvious, but the main issue is FPS stops rising and then begins to decline as time goes on, just as expected. I don’t know of any auto-overclocking tool that tracks FPS versus clock speed, so this needs to be adjusted individually for each card.
Are you referring to checking the stability of the GPU during memory clock overclocking? Yes, I typically perform a stress test followed by gaming to evaluate performance. Benchmarking or stress testing serves as troubleshooting; the practical results come from real gameplay.
The fixes in GDDR5 work without being obvious, but the main issue is FPS stops rising and then begins to decline as time progresses, which is normal. I don’t know of any auto-overclocking tool that tracks FPS against clock speed, so this needs to be adjusted individually for each card.
I should also highlight that what is referred to as the "error correction" for GDDR5 is actually more about error detection followed by retransmission. This makes it slower because the data is repeatedly sent until the checksum is accurate.
Consider it more like parity rather than ECC, but instead of stopping the system when a parity error occurs, the incorrect data is ignored and sent again—similar to how Ethernet or Wi-Fi works.