This year’s CES had major announcements from all of the big names in the industry with Intel and AMD both announcing new motherboard chipsets and AMD also announcing their upcoming GPUs. Nvidia had a lot to announce as well. The biggest of those was their new 50 series of GPUs based on their Blackwell architecture. That included four GPUs which they planned to launch in January and February. They have the RTX 5090 which is their flagship, the RTX 5080, RTX 5070 Ti, and the RTX 5070 all announced. They also announced Multi Frame Generation which along with the new Transformer based DLSS ray Reconstruction and Super Resolution models make DLSS 4 and Nvidia Reflex 2 as well. RTX Neural Shaders and Neural Faces were both announced as well. Today the performance benchmarks for the upcoming RTX 5090 lift and while I take a look at the new RTX 5090 Founders Edition we can also check out DLSS 4 alongside of seeing how the new flagship GPU performs. It’s a lot to cover, so let’s get to it!

Product Name: Nvidia RTX 5090 Founders Edition

Review Sample Provided by: Nvidia

Written by: Wes Compton

Amazon Affiliate Link: HERE

What is new?

Nvidia announced a lot with this one so let’s try to touch on as much as we can. They of course had a few hardware announcements for the 50-Series. They have announced four cards, the RTX 5090, the RTX 5080, the RTX 5070 Ti, and the RTX 5070. That is the new 50-series RTX family. The slides below include their focus on them. The 5090 for example is targeted at 4k 240 Hz and the other three are targeted at 2k or 1440p performance. They all have GDDR7 VRAM with the flagship RTX 5090 having 32 GB and a 512-bit interface. The RTX 5080 cuts that in half to 16 GB, the RTX 5070 Ti also has 16GB, and then the RTX 5070 has 12 GB.

The new GPUs are based on the Blackwell architecture and they do have new updated RT and Tensor cores with that being the 4^th gen for the RT Cores and 5^th gen for Tensor cores. The new tensor cores can now also handle floating point (FP) 4 along with FP8 and FP16. They also have introduced an AMP processer which is the AI Management Processor to help schedule AI tasks alongside of graphics rendering. They have increased the number of NV encoders and decoders, it now depends on the card model and isn’t a flat number of them across the entire generation of cards. They have also moved to PCIe Gen 5 and DisplayPort 2.1b including UHBR20. For pricing the RTX 5090 comes in at $1999, the RTX 5090 is half that (just like its memory) at $999. The RTX 5070 Ti is $749 and the RTX 5070 is $549.

Some of the main goals with Blackwell were to optimize the neural workloads and reduce the memory footprint. No big surprises there. Memory is one of the more expensive parts of the cards and anyone who has been paying attention knows that neural workloads and AI have been where Nvidia has been seeing the biggest improvements. Combining those things they have brought AI into shaders with neural shaders. The example of this they show is a hair being rendered with ray tracing and how using spheres rather than triangles helps use less data which means less VRAM and higher frame rates.

Nvidia announced Transformer which replaces CNN as the model they use with Super Resolution. They have improved the design significantly to get better detail when scaling this up. On top of that, they announced DLSS 4 which has improved on the frame generation that they introduced with DLSS 3. It is now Multi Frame Generation. Where before they were able to generate every other frame, they can now do x3 or x4. This gets interesting once they add in DLSS Super Resolution as well because that is already rendering ¾ of the image. With both, they are rendering 15 of 16 pixels using AI. Of course how well this works then depends a lot on how good the renders are. But it gives huge improvements in performance and with frame generation, we know that those improvements still happen even if you are PCU limited for example. The example they show has DLSS off at 27 FPS, turning on Super Resolution it goes to 71 FPS. DLSS 3.5 gets you to 140 FPS and DLSS 4 is 248 FPS. DLSS is already supported by a LOT of games and Nvidia is saying that DLSS 4 will have 75 games and apps supporting DLSS 4 at Day 0. Some of those will be by using the Nvidia App, which can override the DLSS settings on some games. It is also important to note that some of these features will go back and work with legacy cards as well. Specifically, DLAA is going to work back to the 20 series of cards, same with the improved DLSS Super Resolution. The new multi-frame generation however only works with 50 series cards.

They also introduced a new Nvidia Reflex. Reflex helps improve responsiveness to get lower latency through the entire pipeline. Were Reflex offered 50% faster responsiveness the new Reflex 2 gets you 75% by using frame warp. They say that is coming first to 50 series and will be available in games like Valorant soon.

Below I have the specifications for the RTX 5090 as well as the last two xx90 cards. We can see that the number of GPCs hasn’t changed from the 4090 to the 5090 but they did increase the SMs from 128 to 170. The CUDA core SM count is still the same but there are now 21760 CUDA cores to the previous 16384, a 32% increase. We still have 4 Tensor SMs but the new 5^th gen design has 680 cores now. For clock speed, the RTX 5090 is set lower than the RTX 4090 with a clock speed of 2407 MHz. There are 32%more RT cores which is a 32$ increase but the RT FLOPS has jumped up 66% with the new RT design. The memory has increased from 24 GB up to 32GB and it now uses GDDR7. They have also increased the memory interface up to 512-bit from 384-bit which the 3090 and 4090 both had. The memory data rate has gone from 21 Gbps up to 28 Gbps and the bandwidth increased with that and the larger pipeline from 1008 GB/sec up to 1792 GB/sec a 77% increase. The cache has increased in side with the L1 cache going from 16384 KB up to 21760 KB and the L2 increasing as well from 73728 KB up to 98304 KB. The TGP also had a big jump going from 450 watts on the RTX 4090 to 575 watts here on the RTX 5090. This is an even bigger gap when you include the RTX 3090 which was 350 watts. The manufacturing process is still the TSMC 4nm 4N process so no changes there I mentioned it earlier but all of the 50 series cards moved to PCIe gen 5 and the RTX 5090 is no different. For pricing, it has an MSRP of $1999 which is $400 more than the RTX 4090 but is similar to the RTX 3090 Ti which also launched at $1999 back in 2022.

Before getting into testing I did also run GPUz to double-check that our clock speeds matched up with the specifications. The RTX 5090 Founders Editon had its boost clock at 2407 MHz which matches the specifications. For testing all testing was done using the Nvidia 571.86 driver which was provided to press ahead of the launch.

Packaging

The PC hardware industry as a whole isn’t exactly known for being extremely sustainable, especially with how hardware is always on a crash course with being out of date. But we have at least seen improvements on the packaging side of things. Nvidia has been experimenting with a few designs including the previous generation of Founders Edition cards which used a corrugated material but the overall design did take up a lot of space. For this generation, the RTX 5090 Founders Edition’s box has a lot going on. The most surprising is that the outer shipping box is officially part of the packaging now where before they did have a shipping box that perfectly fit each card but the box inside was still what you might expect to see on a store shelf. This time around I’m told this brown box is the outer packaging and what retailers who carry the Founders Edition will have. This came to me with the shipping labels on it and also on the outside it has a large white sticker, similar to a laptop box, with the part number, BIOS version, SKU, and everything else printed on it. There is a second white sticker on the back edge as well with your serial number and all of the certification logos in one place. The box has a seal that you rip off which reveals the quote “Inspired by Gamers. Enhanced by AI. Built by Nvidia”

When you open the brown box up all of the Nvidia branding is hidden away on the inside, which makes sense. If this is the box it will ship out in, you don’t want it to screen I have an expensive video card in here steal me. The front edge has stripes and the RTX 5090 model name and the stripes and Nvidia branding are on the underside of the lid. The box then has two arrows on the sides, you can pull these out and the side open to make it easier to get to the second box inside. This second box isn’t black like Nvidia has always done. This is also plastic-free where the previous design used a lot of plastic in it. The inner box is inkless and has flaps on the top and bottom edges that once you pull out allow the two halves of the box to come apart. They have die-cut the corrugated cardboard to fit the RTX 5090 Founders Edition inside so that it can't move around. In the past, the documentation and accessories were hidden up under the card and that is still partially true here. But now they are hidden up under the cool box inside the brown box. There is a smaller box and inside you have the quick start guide and the new power adapter.

The 12VHPWR connection ended up being a point of contention with the last generation of Nvidia’s cards. There were issues with melted connections and a lot of misinformation and quick reactions when it started to happen. The main reason they found was that the connection was sometimes not plugged in tightly and was often in tight bends because of it being on top of the cards. Combine that with a lot of the cables being stiff and you have a recipe for problems. Along with other changes that I will touch on when I get to look at the card itself. Nvidia did change the power adapter cable that comes with the cards. This new adapter has the same four 8-pin PCIe plugs but they have individually sleeved the cables and made the cable longer. Those two things mean that it is significantly more flexible and won't put as much stress on the connection. The 12VHPWR end is a lot more robust as well with a thick body once you get past the cable needed to terminate all of those sleeved cables.

I do also have a comparison between the old adapter design and this really shows how much longer the new adapter is in comparison.

Card Layout and Photos

The new design for the RTX 5090 Founders Edition is an interesting one in a few ways. What stands out to me is Nvidia has changed the design in a few really big ways but it still somehow just fits with the styling they have been doing for a few generations now. The RTX 5090 Founders Edition is the same 304 length that we saw with the 4090 Founders Edition, same with the 137mm height. So it is still a very large card, you can see that when you look at how much the card sits above the top of the PCI bracket. What is new however is the move to having both fans on one side and this looks a lot more like a traditional card design. They call this a double flow-through design, which uses a blow-through design on both fans. A lot of cards, including the Founders Edition cards have been using one blow-through fan so this is the natural progression. But it does mean that the PCB is all packed into the middle of the card which is interesting. The other big change is in the card's thickness. They have a new SFF focus by making this a 40 mm dual-slot card whereas the flagship cards from the last two generations have all been triple-slot cards. I would still love to see these cards move back down to a size closer to what we used to see, but I think we might be beyond that at this point and a lot of those cases (both full-sized and SFF) that can’t support the longer and taller cards might just be toast. Including my own Caselabs case as well RIP.

The double flow-through design while keeping the length and height of the old designs means that Nvidia was able to pack in really large fans here. The fans are both 115mm edge to edge and they have 7 large axial blades that blow down into the card. To help with stability the outside of the fans have a ring around it. The fans as well as the centercaps are a very dark grey that goes well with the black shroud design. The shroud is all metal all around the card, giving it a lot of strength (and weight) and these shrouds are what make the Founders Edition designs feel so premium compared to all but the highest-end aftermarket cards. When we look through the fan we can see the blow-through design and if you look closely you might be wondering what on earth is going on at the display connection end of the card. We can see right through there, how are there any connections there at all? Nvidia has a flexible board with high-speed signaling connecting that I/O PCB to the main PCB in the center of the card. They also have a split PCB for the PCIe connection down at the bottom of the card as well.

With both fans being on the one side the back side of the RTX 5090 Founders Edition now looks like the back side, unlike in the past. This side has the same X shape as the accents that create the infinity loop shape. If you look closely that does have the Nvidia logo etched on it. The RTX 5090 is also printed here in the top section of the middle. The two sides are then just filled with vertical heatsink fins. They are all black and match the card. It is hard to see it but both sides also have a slight shape of the fins that are concave right where the fans are located.

Up on the top edge of the RTX 5090 Founders Edition Nvidia has put the same backlit GeForce RTX logo that a lot of their cards have had. I do wish that this was the model name itself, not the generic brand name. It’s always fun to be able to show off what card you have, especially when it is a flagship model like the RTX 5090. Also in this area is the power connection and there is a lot there that needs to be touched on. For starters, the connection has been recessed down into the card and set at an angle, similar to how it was on the RTX 3090. With the RTX 4090, it pointed directly up and was flush so all of the connections would end up needing a tight bend right there to fit in most cases. This gives that connection a lot of relief but that isn’t the only change. The other big change is the move from the 12VHPWR connection to what they now call the 12V-2x6. The same power supply cable works with it, the size hasn’t changed at all as well. They have just changed the lengths of the pins inside. The new connection has .25mm longer pins for the 6 main power connections and then the sense pins are 1.5mm shorter. This means that the connection engages better but if you don’t get it plugged in perfectly the sense pins will pick that up and it won’t work rather than have a loose connection and have it melt. This alone is most likely enough to avoid a majority of the issues that happened last generation, but the angled and recessed connection along with the new adapter as well will help avoid issues as well.

Moving around on the edges of the RTX 5090 Founders Edition doesn’t show too much really. The shroud design wraps around the top, end, and bottom. The two black center sections do have two slots in them for some ventilation in these areas, but that’s it. The shroud also wraps tightly up against the PCB which helps keep air flowing in the directions you want it to go, but it does mean we can’t peek at any of the inside from there. A big benefit with this means that finally, the card isn’t pushing hot air down onto your motherboard, right where a lot of the M.2 slots are hidden away.

For display connections, the RTX 5090 Founders Edition has four total connections. There are three DisplayPort 2.1 connections and one HDMI down at the bottom. With all of the board features, what I might be the happiest about is the move to a proper black PCI bracket. Not only does it better match the card, but once installed in a PC it should also match your case a lot better as well.

Getting the RTX 5090 Founders Edition sitting next to the previous RTX 4090 Founders Edition puts the RTX 5090 Founders Edition’s size into perspective. It has the same shape from the side profile with the same height, length, and even the bends on the ends. The end view however shows the big difference in width. I also have a shot with the power adapters plugged in and this does a great job of showing how the new adapter and the revised angle of the connection change how the power connection on top works. I don’t mind a top connection when it is angled like this to make it easier to route but the old design with the plug going directly up and also being flush to the top of the card makes it hard to fit these tall cards in a lot of cases without putting the connection under a lot of strain, the new adapter is more flexible, is long enough to extend past the end of the card so you don’t have a mess of PCIe power connections all on top of the card and the connection is now recessed and angled so there is almost no need to bend the cables at all.

I mentioned the lighting, but because most of it is hidden it's better to see when everything is lit up. So before getting into testing, I did get a few pictures of the double V accents on both sides as well as the backlit GeForce RTX branding on top. The white looks good, I do wish they were RGB because part of me thinks this would be cool in the old Nvidia green or being able to match it to my build. Being white though, it shouldn’t clash at least.

Test Rig and Procedures

     Test System

CPU: AMD Ryzen 7 9800X3D – Live Pricing

Motherboard: ASUS ROG Crosshair X870E Hero – Live Pricing

Cooling: Enermax LIQMAXFLO 360mm Liquid CPU Cooler - Live Pricing

Noctua NT-H2 Thermal Paste - Live Pricing

      Memory:       G.SKILL Trident Z5 Neo RGB Series (2x16GB) 6000MT/s CL28-36-36-96 – Live Pricing

Storage:       Viper VP4300 Lite 4TB – Live Pricing

Power Supply: be quiet! Dark Power Pro 13 1600W- Live Pricing

Case: Primochill Wetbench - Live Pricing

OS: Windows 11 Pro 64-bit - Live Pricing

Synthetic Benchmarks

As always I like to start my testing with a few synthetic benchmarks. 3DMark especially is one of my favorites because it is very optimized in both Nvidia and AMD drivers. It's nice to not have to worry about it being favored too much either way and the repeatability of the results makes it a nice chance to compare from card to card, especially when comparing with the same GPU. With this being the first GPU in Nvidia’s new 50 Series of GPUs when doing our testing today I am most interested to see how the RTX 5090 Founders Edition compares to Nvidia’s previous flagship with the RTX 4090 as well as AMDs RX 7900 XTX as well.

The first round of tests was done in the older Fire Strike benchmark which is a DX11 test. There are three detail levels, performance, extreme, and ultra. The RTX 5090 Founders Edition changes the scale of our charts here and is sitting WAY out in front of the RTX 4090 and everything else tested. It was a 33% improvement in Fire Strike, 44% in Fire Strike Extreme, and 30% in Fire Strike Ultra.

The next two were both based on the Time Spy benchmark. One is the standard test and then there is the extreme detail level. The RTX 5090 Founders Edition is once again way out in front here. Both tests show its performance up 33% over the RTX 4090 putting a big gap in between them and thinking ahead at other upcoming Nvidia GPUs it does make me wonder if the 5080 and maybe the 5070 Ti will fill in that gap or come in behind the 4090.

I did also test using the new 3DMark Speed Way which is their latest benchmark. Speed Way is DX12 as well but combines more future-focused tech like Ray Tracing which up until its release where only used in feature tests, not full benchmarks. The RTX 5090 Founders Edition scored a 14457 and was 42% ahead of the RTX 4090 in Speed Way. The Port Royal benchmark is similar as well, also including ray tracing and it was 42% in that test as well.

The last test is the newer 3Dmark Steel Nomad benchmark. Officially this is the replacement for the Time Spy benchmark. It is a DX12 benchmark and doesn’t include ray tracing but is updated to better take advantage of modern cards. The RTX 5090 Founders Edition scored a 14194 here and this is the biggest gap between it and the RTX 4090 from all of the 3Dmark tests with an improvement of almost 54%.

In-Game Benchmarks

Now we finally get into the in game performance and that is the main reason people pick up a new video card. To test things out I ran through our new benchmark suite that tests 8 games at three different resolutions (1080p, 1440p, and 4k). Most of the games tested have been run at the highest detail setting and a mid-range detail setting to get a look at how turning things up hurts performance and to give an idea of whether turning detail down from max will be beneficial for frame rates. Cyberpunk 2077 is also tested with Super Sampling (DLSS/FSR/XeSS). In total, each video card is tested 60 times and that makes for a huge mess of results when you put them all together. To help with that I like to start with these overall playability graphs that take all of the results and give an easier-to-read result. I have one for each of the three resolutions and each is broken up into four FPS ranges. Under 30 FPS is considered unplayable, over 30 is playable but not ideal, over 60 is the sweet spot, over 120 FPS is for high refresh rate monitors, and 240 helps show the performance ideal for the latest higher refresh displays.

So how did the Nvidia RTX 5090 Founders Edition perform? Well, this is a flagship card and while our tests have been updated with newer more demanding games it still steamrolled right through all of the 1080p tests with the 9800X3D still being the limit in a lot of those tests. All but one result was up over 240 FPS with that last result just being under. At 1440p we see a few more results in the 120-239 FPS range with 14 results up over 240 FPS and 4 over 120. Then at 4k the 5090 still handled everything with ease. Every result was over 120 FPS but only 7 were up over 240 FPS at this resolution.

To get a better look at some of the cards that are the closest competition to the RTX 5090 I put together all of the results from our tests and averaged them out by resolution. The RTX 5090 is out in front as expected but this better shows us the average result. At 4k which is the only resolution I would imagine anyone would be looking at this card for the 7900 XTX averaged 145.7, the RTX 4090 averaged 170.7, and the RTX 5090 was up over 200 at 226.3 FPS. That is almost a 59% increase over the 7900 XTX and 32% over the RTX 4090.

Of course, I have all of the actual in game results as well for anyone who wants to sort through the wall of graphs below. There weren’t any big surprises in any of the testing here. The RTX 5090 Founders Edition was at the top of the charts in all of the games tested. As you get down below 4k, even with the updated test suite and new faster CPU we still see games being CPU limited, especially Total War: Pharaoh Dynasties and Borderlands 3 at their lower detail settings. No one is going to be running the RTX 5090 at anything but the highest detail but those medium detail settings will be important as we test some of the medium to lower-end cards once they come out.

Another new addition to my testing was a few additional tests using Cyberpunk 2077. This is one of only a few games that support most of the tech from all three of the GPU companies. So I did tests at medium and ultra detail while having Super Sampling on for all of the cards. Using whatever the latest and greatest is supported. In this case, I tested with DLSS, specifically DLSS 3 with frame generation. Cyberpunk does support the newer DLSS 4 with a beta version of the game. Once that hits a widespread release I will most likely change to those results. Don’t worry though, I will check that out later in our DLSS section. This gives us a good look at the performance you can expect to see. Just a note here, the AMD cards only allowed FSR when running windowed mode whereas Nvidia only performed well in fullscreen mode. The 7900 XTX does jump up over the RTX 4090 in this comparison but the 5090 is still way out in front.

Compute and AI Benchmarks

Now some people don’t need a video card for gaming, they need the processing power for rendering or 2D/3D production, or in some cases, people who game also do work on the side. AI performance importance has increased quickly recently as well. So it is also important to check out the compute and AI performance on all of the video cards that come in. That includes doing a few different tests. My first test would normally be Passmark Performance Test 11’s compute test but the OpelCL issues caused problems there so our first was today is Geekbench AI. The RTX 5090 is sitting up at the top of the chart in all three of the results but the half-precision score is especially impressive sitting at 69957, a 45% improvement over the RTX 4090.

Blender is always my favorite compute benchmark because the open-source 3D rendering software is very popular and it isn’t a synthetic benchmark. With the latest version of Blender, they redid the benchmark so we now have a new test that runs three different renderings and gives each a score. I have all three stacked together so we can see the overall performance. The RTX 5090 Founders Edition is up top of the chart. It was 33% faster than the 4090 in two of the three results and 45% higher on the third.

For CUDA-based cards, I also check out V-Ray Benchmark 5 to check out CUDA and RTX performance in the 3D rendering and simulation software. The RTX 5090 outperformed the 4090’s RTX score with its CUDA score to put things into perspective. Overall it improved in the 4090’s performance by 29% in the CUDA test and 28% in the RTX tests.

RTX and DLSS

Alongside the introduction of the new Blackwell-based GPUs, Nvidia had multiple new announcements on the tech side of things with their newest version of DLSS 4 with its addition of Multi Frame Generation. They also have Deep Learning Anti-Aliasing (DLAA) and Transformer based models to continue to improve the detail and accuracy of the AI-generated frames. So I am excited to see how it all performed. First up though I do have our standard 3Dmark DLSS comparison which I ran on every Nvidia card from the last two generations. For this test, I tested them all at 4k using the performance setting. I ran the same test with DLSS off, and then with DLSS 1, 2, and 3 to compare the performance between them. No surprise that the RTX 5090 is up at the top. But more importantly, this is a chance to see the performance improvements that can be capable. The 5090 went from 83 FPS up to 146 with DLSS 1. It then went farther up to 186 with DLSS 2 then ended up at an impressive 259 with DLSS 3. The graph doesn’t include it because it was still in beta but I did get to test the new DLSS 4 performance with Multi Frame Generation on at x2, x3, and x4 levels. The jump from the DLSS 3 result to x2 shows there was a small difference between my x2 results. With x3 the improvement was noticeable, it scored a 358.65 FPS result. But with it at the highest setting the 5090 finished the test with an FSP of 441.81. That’s an improvement of 426% over the non-DLSS result and is just mind-blowing. This is a synthetic benchmark and it’s a lot easier for deep learning super sampling to handle things when there isn’t any variability to the result. But it is still impressive to see what it can be capable of in a best-case scenario.

I also wanted to take a look at DLSS 4 performance as well a little more than my initial look at it with 3DMark’s benchmark. For this, I put Nvidias Frameview to the test to run a few benchmarks using the games that currently support DLSS 4. For games with a built-in benchmark, I ran the benchmark but used frameview so we could get the FPS and 1% lows. All of the tests were done at 1080p, 1440p, and 4k with the highest detail settings including the highest RTX settings. A few of these games had to be run on beta copies, namely Cyberpunk 2077, Hogwarts Legacy, Alan Wake 2, and Star Wars Outlaws. Marvel Rivals and Dragon Age both were on the public version but just used NV App’s override.

The first game tested was Dragon Age: The Veilguard and for this one, I did a test with DLSS 3 and 4 as well as with it off on all three resolutions. The performance improvements I saw in 3Dmark were here as well but a little smaller. At 4k for example the 5090 went from 94 FPS up to 269, a 186% improvement. The DLSS 3 to 4 gap was smaller at 31% but still notable. You can also see the 1% lows improve but the gap between those and the FPS gets wider.

For Hogwarts Legacy, I kept things simple with DLSS 4 on or off but it was interesting that the performance between the 4k and 1440p results were almost exactly the same with DLSS 4 on. At 4k the RTX 5090 improved 365% when running DLSS 4.

Next up was Cyberpunk 2077 and I changed things up here. This game had the option to switch between the Transformer and CNN models so I wanted to see what the performance difference between them was. At 1080p transformer was a little slower but it was just one FPS behind at 1440p and 4k. More importantly, though, I was more impressed with the overall quality of them. Pictures don’t show the improvement as much as seeing the improvement in smoothness and clarity but it is a noticeable difference. At 4k with all of the ray tracing options completely cranked up the 5090 still struggled with DLSS off coming in at just 33 FPS but at 119 with DLSS 4 on it was silky smooth.

With Marvel Rivals I had issues with my 1080p and 1440p testing (no performance improvement, sometimes less performance) but 4k is the only thing important with the RTX 5090 so I have included those. The 5090 went from 241 up to 366 which was a smaller improvement than I saw on some of the other games. But still a welcomed improvement. I imagine as we see this get fully implemented the issues I saw will get worked out and this is one of the games that will benefit the most from it.

My last test was in Alen Wake 2 and I was some inconsistencies here as well, namely the DLSS 4 1440p result which acted like it had a frame limit. Both the 1% low and total frame rate were locked in at 119 FPS. The other resolutions didn’t have this issue and you see a big improvement on both. At 4k it went from 47 FPS up to 165 and the 1% lows improved as well from 43 up to 76. Then at 1080, it jumped from 134 up to a crazy 418 FPS.

Cooling Noise and Power

For my last few tests, rather than focusing on in game performance, I like to check out other aspects of video card performance. These are also the most important ways to differentiate the performance between cards that have the same GPU. To start things off I took a look at power usage.

For this, our test setup utilizes the Nvidia-designed PCat v2 along with cables to handle both traditional 6 or 8-pin connections as well as 12VHPWR. The PCat also utilizes a PCIe adapter to measure any power going to the card through the PCIe slot so we can measure the video card wattage exclusively, not the entire system as we have done in the past. I test with a mix of applications to get both in game, synthetic benchmarks, and other workloads like GeekbenchAI and AIDA64. Then everything is averaged together for our result. I also have the individual results for this specific card and I document the peak wattage result. I do have to note that during my testing of the RTX 5090, I ran into issues with a few tests, especially with OpenCL which is what AIDA64 uses to do their stress tests. I ended up having to replicate our test results using OCCT. To get as close as possible I used the 3D Adaptive stress test in OCCT and using the 4090 I stress tested it and played around with settings to get to a setting where OCCT replicated the same power usage and card temperatures. In the future, once that issue is fixed I will be redoing our 5090 tests with AIDA64, but this issue was confirmed to be reproducible by Nvidia the day before launch and I didn’t have time to retest the whole list of cards tested. That said the OCCT result wasn’t the peak power usage here, so any difference between the tests will only change the average result slightly (but that same testing affects our temperature tests and power efficiency graphs. The RTX 5090 Founders Edition peaked at 648 watts and averaged 613, this is without a doubt the highest power draw I’ve seen to date and is 136 watts more than the RTX 4090. At one point in my testing, I had issues with 3Dmark Fire Strike Ultra crashing and as it turned out, it was because our motherboard needed the supplemental power to the PCIe lanes plugged in. So going in, you are going to need more power to run a 5090 than cards in the past. The average across all of the tests was 613 watts and the results ranged from 550 watts in FarCry 6 up to 648 in Time Spy Extreme.

With having exact peak wattage numbers when running Time Spy Extreme I was also able to put together a graph showing the total score for each watt that a card draws which gives us an interesting look at overall power efficiency in the popular and demanding benchmark. Even with the 648 watts pulled, the overall performance of the RTX 5090 still put it right at the top of our efficiency chart and with a good margin between it and the RTX 4090 which was below it. 

My next round of tests were looking at noise levels. These are especially important to me because I can’t stand to listen to my PC whirling. Especially when I’m not in game and other applications are using the GPU. For my testing, though I first tested with the fan cranked up to 100% to get an idea of how loud it can get, then again at 50% to get an idea of its range. The RTX 5090 Founders Edition with its new cooler design was a little noisier than the other Founders Edition cards at 50% fan speed but interestingly at 10)% fan speed it dropped back down into the middle of the chart. The reason for it being lower in the 100% chart can be seen in the fan RPM chart, it is down in the bottom ¼, they have the fans running slow but not slower than the previous generation 4090 and 3090 Ti. To be fair though, this is a 2 slot card and is using a lot more power I’m surprised they weren’t running the fans even higher.

I also take a look at noise performance while under load. For that when running AIDA64’s stress test I wait until the temperature of the card has leveled off and then measure how loud things are when the card is at its worst-case scenario with the stock fan profile. Here the RTX 5090 Founders Edition was once again up near the top like it was in the 50% fan speed chart. That makes sense though given that when under load the fans were running at 46%. That tells us even before getting into the temperatures that the card isn’t working too hard to keep things cool. You can see for example the last generation of AMD reference cards that were also two-slot were pushing their fans a lot harder, the 5090 is tied with the RTX 4090 here.

To finish up my testing I of course had to check out the cooling performance. To do this I ran two different tests. I used AIDA64’s Stress Test run for a half-hour each to warm things up (on everything except the 5090 which was tested on a similarly matched OCCT workload). Then I documented what temperature the GPU leveled out at with the stock fan profile and then again with the fans cranked up to 100%. With the stock profile, the RTX 5090 Founders Edition came in at 72c. For comparison, the RTX 4090 FE was at 77c in the same test. For GPU memory it was running at 83c which is warm for sure. Then with the fans cranked up, the RTX 5090 Founders Edition ran at 58c, again noticeably cooler than the 4090 in the same test and below the 3090 Ti and 3080 as well. The memory cooled down to 68c. The delta between the two results was 14c which is on the high side showing that if you want to play with the fan profile and don’t mind a little noise there is a lot of cooling headroom to be had especially if you want to overclock.

While running the stock fan profile testing I also took the time to get a few thermal images so we could see what is going on. The new thinner cooler design is able to pack in large fans by keeping the large length and height of past flagship Founders Edition cards and you can see just how much cooler things are running on the front and back side of the card where the fans are running. In the center with the covered sections, things are warmer with the power connection and then the rear triangle on the back side of the card being the two hottest spots. Even then on that back side picture, the thermal camera picks up the hotspot as being just below the card where there is less airflow between it and the M.2 heatsink. I’m sure larger aftermarket designs are going to improve on this, but I am impressed that Nvidia was able to design a card that cools this well while being a dual-slot card. I still wish it was a normal card height, but it's good to know they are paying attention to the SFF guys as well.

Overall and Final Verdict

Nvidia’s design of the RTX 5090 Founders Edition somehow managed to be completely new while looking right at home if you were to put it next to a Founders Edition card from the last few generations. That means you still get the clean black and dark grey styling that is always one of or the best looking card available in my opinion. The lighting accents are subtle, backlighting some of the X design on both the front and back of the card as well as the GeForce RTX branding up on top. I do wish they had included RGB just as a way to customize the card a little and I will always push to see the actual model name used up on the top edge over the generic branding. Being able to show off that you have an RTX 5090 would be cool. I was really happy to see that they finally went with a fully blacked-out PCI bracket, I’ve been pushing for this for years, and not only does it look great but when you have the card installed in a case it will blend in better as well.

They cut the design down in size from a three-slot design to a two-slot design which should help with some SFF designs and help make sure there is room for airflow in builds that want to use multiple cards (compute and AI situations). They did keep the same height and length, so this still isn’t a small card by any means. Once again the build quality is a step above what you might see from some aftermarket designs with an all-metal shroud design that isn’t ever going to sag.

Performance for the new generation of cards in my testing had the RTX 5090 outperforming the RTX 4090 by around 32% which is right in line with the increase in CUDA cores for the card. There were some tests which saw an even bigger increase and the RTX 5090 was at the top of the chart across the board in every applicable test. What was even more impressive to me was the improvements with DLSS 4, the performance difference that it can make is sometimes shocking, but on top of that Nvidia has improved the smoothness and picture quality. At the end of the day, there wasn’t anything that I threw at the RTX 5090 that slowed it down, but if you do run into something that it can’t handle DLSS 4 is going to fix you right up. I did see some bugs in my DLSS testing, mostly when trying down resolutions, but I suspect some of those will be smoothed out once the updates are released. The biggest issue I ran into performance-wise was that a few of our benchmarks just wouldn’t run at all and they were all OpenCL. Nvidia is aware and is working to get support for those tests.

The big increase in performance without any change in manufacturing size does have the RTX 5090 having a significantly higher power consumption. I saw it pulling up to 648 watts at peak, combine that with today's highest-end CPUs and we are swinging back to needing high-wattage power supplies. Speaking of power, the power connection has been improved in a whole list of ways including moving from the original 12VHPWR connection to the changed design that is called 12V-2-6. It looks the same and all of the power supplies will still connect. But they have changed the pin heights to get a better connection and the sense pins are shorter and are more likely to catch when the plug isn’t connected all the way. On top of that Nvidia’s card design has recessed the connection down into the card and angled it to reduce any strain on the connection. They have also included a much nicer power adapter as well. All of that power does mean there is more heat but the double blow-through design handled it surprisingly well running similarly in temperatures to the RTX 4090 Founders Edition even with a thinner card design and a lot more wattage going through.

For pricing, the RTX 5090 Founders Edition has an MSRP of $1999 which isn’t cheap but isn’t entirely out of line with some of Nvidia’s flagship cards at launch. The previous generation RTX 4090 launched at $1599 but before that, the RTX 3090 Ti launched at $1999. That said, I did put together a chart which is up above that graphs out card pricing to 3Dmark Time Spy Extreme scores. I included both launch MSRP pricing as well as current low prices from PCPartPicker. This also gives us a unique look to see what cards maybe weren’t a value when they launched but are now. The RTX 5090 Founders Edition, being a flagship card isn’t going to be a value leader but it is in line with the RTX 4080 and RTX 3080 and is a lot better value than the RTX 4090 and RTX 3090. The $1999 price point is well beyond a price point that I could afford, but if you are looking for the best possible performance this is the only option.

Live Pricing: HERE

Author Bio

Author: garfi3ldWebsite: http://lanoc.org

Editor-in-chief

You might call him obsessed or just a hardcore geek. Wes's obsession with gaming hardware and gadgets isn't anything new, he could be found taking things apart even as a child. When not poking around in PC's he can be found playing League of Legends, Awesomenauts, or Civilization 5 or watching a wide variety of TV shows and Movies. A car guy at heart, the same things that draw him into tweaking cars apply when building good looking fast computers. If you are interested in writing for Wes here at LanOC you can reach out to him directly using our contact form.