F5F Stay Refreshed Hardware Desktop Issues with GPU speed on RTX 2080 Ti device

Issues with GPU speed on RTX 2080 Ti device

Issues with GPU speed on RTX 2080 Ti device

B
Back2Blaze
Member
204
09-17-2019, 02:20 PM
#1
I've been trying to diagnose poor gaming performance on my PC for a while now and have had no luck. These are my specs: RTX 2080ti Ryzen 9 3900x 2x16 3200MHz Memory Corasir RM 850x PSU My 3DMark Timespy score is only 8500, where the average is reported to be over 14 000! My GPU load is 100% the entire test, temperatures are only 68 degrees on the GPU, 48 degrees on the CPU. I do notice a hotspot of 108 degrees in GPU-Z, but nothing is indicating that there is thermal throttling. Core clock speeds are around 1700 MHz, and memory speed is 7250 MHz. I checked GPU-Z and it's reporting PWR as the limiting factor during the test. I noticed the power draw being reported during the test was only around 180W, though the card is rated at 250W TDP. I checked NVIDIA Control Panel and made sure to set Max Performance in the power consumption. So I suspect it's possibly not enough power is being delivered to the card, but I would expect some sort of instability for that to be the case? I never experience any crashing, artifacts, or anything other than low FPS to indicate the GPU is performing poorly. FWIW, I also ran performance tests in Linux and noticed similar low scores (in the Superposition benchmark, not 3DMark). I'm going to try tearing apart my computer and reapplying thermal paste, but figured I'd reach out to see if there were any other ideas. I am running a custom water cooling solution, which I applied thermal paste manually a couple years ago and had excellent performance, it's only recently that I've really noticed it not holding up. Thanks!
B
Back2Blaze
09-17-2019, 02:20 PM #1

I've been trying to diagnose poor gaming performance on my PC for a while now and have had no luck. These are my specs: RTX 2080ti Ryzen 9 3900x 2x16 3200MHz Memory Corasir RM 850x PSU My 3DMark Timespy score is only 8500, where the average is reported to be over 14 000! My GPU load is 100% the entire test, temperatures are only 68 degrees on the GPU, 48 degrees on the CPU. I do notice a hotspot of 108 degrees in GPU-Z, but nothing is indicating that there is thermal throttling. Core clock speeds are around 1700 MHz, and memory speed is 7250 MHz. I checked GPU-Z and it's reporting PWR as the limiting factor during the test. I noticed the power draw being reported during the test was only around 180W, though the card is rated at 250W TDP. I checked NVIDIA Control Panel and made sure to set Max Performance in the power consumption. So I suspect it's possibly not enough power is being delivered to the card, but I would expect some sort of instability for that to be the case? I never experience any crashing, artifacts, or anything other than low FPS to indicate the GPU is performing poorly. FWIW, I also ran performance tests in Linux and noticed similar low scores (in the Superposition benchmark, not 3DMark). I'm going to try tearing apart my computer and reapplying thermal paste, but figured I'd reach out to see if there were any other ideas. I am running a custom water cooling solution, which I applied thermal paste manually a couple years ago and had excellent performance, it's only recently that I've really noticed it not holding up. Thanks!

A
AimeeL9876
Junior Member
42
09-19-2019, 12:30 PM
#2
Did you connect two different PCIe power connectors to the card? Otherwise, it might only deliver around 220W or less.
A
AimeeL9876
09-19-2019, 12:30 PM #2

Did you connect two different PCIe power connectors to the card? Otherwise, it might only deliver around 220W or less.

P
Pauxyneu
Member
138
09-19-2019, 01:13 PM
#3
Absolutely, I've got individual connectors for each power input in my setup.
P
Pauxyneu
09-19-2019, 01:13 PM #3

Absolutely, I've got individual connectors for each power input in my setup.