The furmark burnin isn't increasing GPU clock speeds.
The furmark burnin isn't increasing GPU clock speeds.
I chose to modify my Asus Strix R9 285OC to enhance my performance in Firestrike, and I updated Furmark to check its stability.
I increased the GPU frequency from 954Mhz (standard) to 1.1Ghz without changing the voltage. After about 20 minutes of testing, it ran smoothly, reaching a stable temperature around 78°C. Then I tested Firestrike, which failed during the second graphics test. I adjusted the voltage slightly and tried again. It crashed in Firestrike once more. I decided to look into the issue.
It turned out Furmark was handling my GPU at around 800-900Mhz instead of the expected 1100Mhz. After running a short Firestrike session, I confirmed the peak clock matched the expected value. I didn’t have the chance to monitor closely during the Firestrike run, but it seems Furmark was pushing the GPU too hard, which caused the crash.
I managed to get a Firestrike score of 1050Mhz, but stress testing remains challenging since high load doesn’t help if the clock is only at 85% capacity.
Any thoughts on what could be causing this?
By the way, I just purchased 3dmark’s full version on Steam, which now includes a stress test. It would be great if Furmark also worked well. I enjoy testing both CPU and GPU under full gaming load—this last one even turned my CoolerMaster HAF XB into a space heater.
I’ll update after running the 3dmark stress test to confirm it functions properly.
The throttling comes from the BIOS of the graphics card. The specific workload plays a role. Furmark and OCCT are built to extract as much power as the power supply can provide, leading to overheating. This behavior is called a "power virus." Previously, it was prevented at the driver level, but you could rename it to bypass the limit and damage the card. Now, with proper protections in place, any power virus will be properly throttled.
There are methods to circumvent it if you want to cause serious harm. You can modify the BIOS to disable these safeguards. The issue isn't about the voltages used for overclocking; it's the VRMs that control power distribution to the card, and the type of load Furmark applies causes excessive power draw.
SinxarKnights :
Furmark is being limited to avoid harm.
There might be attempts to rename the file to bypass it, though results are uncertain.
The goal seems to check stability, but throttling itself makes it tricky.
The main goal isn't just stability, but to trigger overheating. The system is throttled due to built-in overcurrent protection. If not controlled, the VRMs might ignite, leading to a short circuit when the graphics card overheats, possibly damaging your PC and even your home (which is why this protection exists).
SinxarKnights :
The goal is indeed to check stability, but the real aim is to trigger an overheating scenario. The throttling happens due to the built-in overcurrent protection. It's quite likely the VRMs would catch fire without it, which could lead to a short circuit when the graphics card overheats, risking damage to your PC or even your home.
Furmark is probably responsible for the throttling, but why would it do so unless something else is involved?
If voltage isn't adjusted, should I still be protected from VRM issues?
The throttling comes from the BIOS of the graphics card. The specific type of workload determines how much power is drawn. Furmark and OCCT are created to extract as much power as the power supply can provide, which can cause overheating. This behavior is referred to as a "power virus." Previously, it was possible to disable these protections at the driver level, allowing users to bypass the limits and damage the card. However, modern cards include built-in safeguards that ensure any power virus will be properly throttled.
There are methods to circumvent this if you wish to cause more severe damage. You can modify the BIOS settings to disable these protections. The issue isn't about the voltages used for overclocking; it's about the VRMs controlling power distribution. The type of load Furmark consumes leads to an excessive current draw. These VRMs maintain a stable voltage while adapting to current needs, though this process generates significant heat (the regulation circuitry is quite complex).
Some Nivida cards exhibit similar behavior, but I’m not certain about my 970 model. My R9 280 experienced the same problem as with your 285. This has been observed for several AMD cards over time, though it seems to vary by architecture. It appears to be an inherent feature in many GCN AMD cards, built directly into the BIOS.
I really value the responses you've received. It's logical that my card would limit itself to safeguard itself, though I'm still puzzled about why it only appears to act this way for Furmark while 3dmark remains unaffected. During the Furmark burn-in process, it didn't even attempt to reach maximum speeds; in fact, they rarely approached stock speeds.
By the way, I completed the 3dmark stress test and achieved 99.9% stability at 1050Mhz.
Try Valley using these configurations in the first image:
http://www.bit-tech.net/hardware/graphic...feat-xfx/8
Here is the link to the Valley benchmark:https://unigine.com/products/benchmarks/valley/
Here are the results I obtained with my R9 280:
Sadly I removed my previous 3DMark scores and kept only the best ones for my current card:
http://www.3dmark.com/fs/8093686
(This is for a 970, so it's not very useful).
What are your PC components?
I'm confident everything is functioning correctly and meeting expectations in terms of hardware performance. My current score ranks third among all AMD X4 860K systems with R9 285 single cards. I aim to thoroughly stress-test my card at overclocked speeds to determine if further improvements are possible without compromising safety.
I plan to use the Unigine benchmark to assess whether it will push the clocks even higher than the furmark.