Conclusion
The NVIDIA GeForce RTX 5090 Founders Edition is NVIDIA’s new flagship $1,999 video card, equipped with 32GB of GDDR7, DLSS 4 feature suite, and a whole host of other goodies. It doesn’t come cheap, and it also doesn’t come cheap on the power demands. In our launch review, we put the GeForce RTX 5090 FE through its paces in native resolution performance, upscaling performance, raster performance, ray tracing performance, and other metrics as well such as latency.
The NVIDIA GeForce RTX 5090 Founders Edition is indeed the fastest consumer gaming graphics card on the planet right now. It offers between 25-35% performance uplift over the previous generation GeForce RTX 4090 at its default performance. In our overclocking review today, we pushed the NVIDIA GeForce RTX 5090 Founders Edition to its limits to find the highest stable overclock and see how much gas is left in the tank.
Overclocking Summary
In our overclocking, we were able to increase the Power Limit by just 4%, but this makes perfect sense as this is at the limit of the 12VHPWR connector and cable itself. The TDP of the GeForce RTX 5090 FE is 575W, and a 4% Power Limit increase takes it to 600W, which is max for the cable. Of course, you will get power excursions that exceed this, and the cable and power supply needs to be able to deal with this. When we push the Power Limit up 4%, we are at the limits on power. Though we had the ability to raise the Voltage, ultimately this would degrade GPU clock speed potential as it exceeds the TDP by quite a lot.
When we raised the GPU clock frequency, we managed to set it to +270 in MSI Afterburner, which did show positive gains in the GPU clock speed while gaming. We briefly were able to hit +300, and we think that with added Voltage, the GPU could do it. However, if you increase Vthe oltage, the power demand of this video card is going to skyrocket. In addition, we managed to push the memory frequency from 28Gbps up to 31Gbps without it throttling the GPU clock speed. We were actually able to hit 32Gbps, but temperatures rose a lot on the memory (without fan speed adjustments), and ultimately it could end up throttling GPU frequency with the current TDP. Still, at 31Gbps we are almost hitting 2TB/s of memory bandwidth, which is just incredible.
At +270 to the GPU frequency, we saw the GPU clock speed rise from an average of 2715MHz, at default, up to an average of 2968MHz when overclocked. We peaked a few times at 3GHz, but the sustained clock speed was generally just under 3GHz, in that 2950MHz range. This overall was a 9% GPU clock speed overclock, and 11% memory overclock that we achieved on the Founders Edition of the GeForce RTX 5090.
In terms of cooling, the Founders Edition cooling is brilliant and very capable of allowing overclocking on the GeForce RTX 5090. However, the caveat here is that you will probably want to set a custom fan curve to allow for higher fan speeds when overclocked. The memory temperature is already very hot at default, at 92c, and overclocking it further just increases that even more. We highly recommend raising the fan speeds if you plan to overclock. We set 80% fan speeds, just to be sure, but you could get away with a bit slower fan speeds, perhaps around 60%. The fan noise will be noticeable, however, anything above 60% can be considered annoying fan noise.
Final Points
The performance of overclocking the NVIDIA GeForce RTX 5090 FE can be summed as, well, underwhelming. Overall, we saw a 5-6% performance improvement in games from overclocking the NVIDIA GeForce RTX 5090 FE. There were some cases where it was 7% or even 8%, but those were more rare. The overclock didn’t improve the gameplay experience to any noticeable degree. It helped most to smooth out some very high game settings, or when DLSS was enabled, but it wasn’t Earth-shattering.
Overall, it does not seem worth the power demand or resulting temperature increases and fan speed increases to overclock the NVIDIA GeForce RTX 5090 Founders Edition. The actual gameplay performance advantage you get from overclocking is minimal, and the advantages do not outweigh the negatives.
Though the Founder Edition is built well and can take the heat, it seems that the GeForce RTX 5090 GPU is pretty tapped out or near the limits of its potential. On the one hand, that is a good thing; it means you are getting all the performance you can get and leaving nothing on the table, but on the other hand, it means the GeForce RTX 5090 may not be the best-suited GPU for enthusiast tweaking. While the GeForce RTX 5090 may not be best suited for overclocking, perhaps the GeForce RTX 5080 will be, we will have to find out how that one does, so stay tuned.
Discussion (7 replies)
Join Discussion →Thanks for the great write-up! As always, these end up being my reference templates for OCing whatever I end up with. Very torn on whether to invest the extra $500 for an AIO card like the Suprim SOC (read a couple of reviews and very, very similar performance-wise, just can be cooled more), or anything else for that matter. These FE cards seem amazing for their engineering and cooling capabilities. The next trick will be to try and snag one.
I'm waiting unitl March before I even think about it.
My question is, are there larger gains from CPU increase or memory increase? Can you decrease one to allow more power limit to increase the other? For example, if the GPU is the bottleneck, can mem clocks be reduced to increase power limit available to the GPU to eek out some extra MHZ?
Good question. @David_Schroth taught me years ago about lowering mem freq to free up some power for higher GPU clockrates. I'd imagine so but the silicon lottery also comes into play. I tend to lose out with the silicon side of things so using this strategy often helps when dialing things in.
Yes, GDDR7 like GDDR6X takes a lot of power, reducing memory allows the GPU to sustain a higher clock, HOWEVER, there is a cap based on the bin, so you hit a point that the GPU just can't do that clock speed, thus Voltage is needed to overcome that, and that increases power again by a lot on these GPUs in this node.
Good to know. Thanks @Brent_Justice!
That makes things more complex if the GPU clocks cap out at a certain place, especially if that is before you've reached the full power target for the card. If the GPU is only going to go to 3ghz and that's it regardless of any extra power limit on the card, might as well increase the mem clocks too.