Adjusting performance beyond limits leads to visible red dots and distortion effects.
Adjusting performance beyond limits leads to visible red dots and distortion effects.
I understand that these signs suggest your GPU might be overheating its VRAM/VRMs. Based on my experience, ignoring such issues could damage your graphics card, so proceed with caution if you plan to overclock. AIDA 64 offers an on-screen display feature allowing you to monitor multiple temperature diodes, including those on your GPU, which helps identify overheating VRAM. I’ve configured it to show temperatures on my second monitor while gaming for better oversight.
Even though your card claimed a slight overclock, it’s overheating more than expected. Most cards handle this issue well, but the AMD R9 290 series are particularly known for that. Although the overall temperature stays within limits, some internal parts could still be affected by too much heat.
Adjust the fan speeds to raise RPM quickly at higher temperatures, and adapt the card’s behavior based on temperature changes, airflow, and overclocking settings.
Improving case ventilation is crucial—strong 120mm fans with adjustable speeds should be used, especially with drive cages directing airflow horizontally across the card. This creates consistent airflow both above and below the device.
Make gradual overclock tests step by step, as results vary due to differences in cards and environments.
Running a card at higher clock speeds doesn’t always boost performance or efficiency; it can even cause instability, overheating, or reduced functionality without any visible signs of issues.
Rogue Leader:
lasdant :
Rogue Leader :
lasdant :
Back that memory clock down should help lower your temperatures. That’s what we’re seeing, the system is overheating.
I adjusted the memory clock back to 1260MHz and it worked again. I played for a short time then dots appeared on the screen.
Curious, it seems like the issue is still related to heat.
In my opinion, I should start fresh with everything at default settings and observe the results. Let’s determine the acceptable temperature range.
Sorry, I didn’t catch it in time—I just reset everything back to stock except the core clock, which stayed at 1020MHz. The red dots reappeared later, but after I left and checked the temps, the card was at 76°C when it started acting up. I’ll try again with all settings back to default.
20 MHz shouldn’t cause problems. Please let me know what happens next.
Done, no artifacts or strange behavior on screen. The card was running between 70-75°C. It looks like the card is automatically downclocking to manage heat—when it hit 75°C, the core clock dropped to around 500MHz, and once temperatures fell to 65°C, the core clock returned to the base 1000MHz.
https://imgur.com/a/Ws2meF3
Although your card accused a mild overclock, it's overheating more than usual. Most cards do this, but the AMD R9 290 series are known for it. Even though they stay within specs overall, some internal parts may be affected by too much heat.
- Adjust the fan speeds to increase RPM, making them spin faster at the same temperature. This helps the card adapt better to heat and overclocking.
- Focus on enhancing case airflow and CPU cooling. These cards benefit from strong 120mm fans at the bottom and top of the card, with horizontal airflow through the drive cage, creating steady currents above and below.
- Make small incremental changes in overclocking or testing the card. Results can vary due to differences in cards and environments.
- Running a card at higher clocks doesn’t always improve performance; it can even cause overheating and instability, especially without visible issues.
Currently, I have three case fans—two intake and one exhaust. This might be causing poor airflow. I’ll add more fans for better ventilation.
Should I prioritize adding more intake or exhaust fans? It’s a bit unclear.
Rogue Leader:
Lasdant:
Rogue Leader:
Lasdant:
Rogue Leader:
Back that memory clock down to reduce heat. That’s what’s happening—your memory is getting too hot.
I lowered the memory clock back to 1260MHz and it repeated, then dots appeared on the screen after a short play.
Still likely a thermal issue.
I’d restart everything at stock settings and monitor the temps. Let’s find out what range we’re dealing with.
Sorry, I missed it in time. I reset everything except the core clock, which stayed at 1020MHz. The red dots came back after some time, but now when I left and checked the temperature, the card was at 76°C and acting up. I’ll try again fully at stock settings.
20 MHz shouldn’t cause artifacts. Please let me know what happens.
Done, no artifacts or strange behavior on screen. The card was running between 70-75°C. It seems like the GPU is trying to cool itself by downclocking when it hit 75°C; once temps dropped to 65, the core clock returned to the base 1000MHz.
https://imgur.com/a/Ws2meF3
Temperatures in the 70s are okay. It definitely sounds like you have a GPU from the lower end that resists overclocking. I’ll address the memory overheat issue. If adding just 20 MHz to the core clock causes artifacts, then your GPU isn’t performing well at all.
My prediction is that your GPU's VRMs are getting too hot. On my 280X, the core often reached around 90°C under full load. The 390X tends to hit 94°C during heavy use... These cards are built to handle high temperatures, so 75°C is actually normal. You shouldn't be seeing issues at that level.
When I first noticed artifacts on my 280X before it failed, the VRMs were consistently reaching nearly 100°C. I recommend downloading a trial of AIDA 64 and enabling GPU VRM temperature controls via the on-screen menu. You might find they reach about 100°C when artifacts appear.
Another suggestion... if your card is overheating, consider using the FPS cap in the Adrenalin control panel. I've seen my 390X VRMs stay below 82°C after I started limiting frame rate in games. If I turn off the capping feature, they can spike to 100°C when pushed.
Rogue Leader:
lasdant :
Rogue Leader :
lasdant :
Rogue Leader :
Back that memory clock down should help lower your temps. That’s exactly what’s happening—your memory is getting too hot.
I adjusted the memory clock back to 1260MHz and it worked again. I played for a while then dots started showing up on the screen.
Still a bit of a heat issue, though. I’d recommend starting fresh with everything at stock settings and checking the temperatures. Let’s see what range we’re dealing with here.
Sorry, I didn’t catch it in time. I just reset everything back to stock except the core clock, which I kept at 1020MHz. The red dots came back after a while, but now when I left and checked the temps, the card was sitting at 76°C when it started acting up. I’ll try it again completely at stock settings.
20 MHz shouldn’t cause any artifacts. Please let me know what happens.
Done, no artifacts or weird behavior on screen. The card was running between 70-75°C. It seems like the GPU is trying to cool itself by downclocking when it hits 75°C. Once the temps dropped to 65°C, the core clock went back up to the base 1000MHz.
https://imgur.com/a/Ws2meF3
Temps in the 70s are okay. It definitely sounds like you have a GPU from the lower end of the range and it doesn’t handle overclocking well. I’ll take the memory overheat issue off my plate. If adding just 20 MHz to the core clock causes artifacts, then your GPU isn’t up to the task at all.
I’ve had this card for almost four years now—maybe that’s the reason. But I guess I’ll have to accept running it at stock speeds. It doesn’t seem to enjoy being overclocked.
I've had this card for almost four years now, maybe it's connected to that. But I guess I just have to keep it running at normal speeds. It doesn't seem to enjoy being pushed too hard. If it's been overclocked a lot, the VRMs might be damaged. To be honest, these GPUs aren't usually known for lasting well under heavy overclocking.
Rogue Leader:
If it's been heavily overclocked all this time, the VRMs might be damaged. My overclocking isn't extreme, but these GPUs aren't usually built for long-term overclocking. I've only begun recently, mostly sticking to stock performance. Now that it's nearing the end of its lifespan and I don't mind what happens, I'm enjoying testing its limits.