The American cards now reach the Europe! The EVGA GeForce RTX 3080 FTW3 Ultra compares well with the top models of other manufacturers in terms of performance. In addition, the large cooling block keeps the RTX 3080-GPU cooler and quieter than the Founders Edition. Whether the FTW3 Ultra also makes it difficult for the top models of, for example, ASUS and MSI in that area, we do not dare to say without direct comparison, but it is clear that EVGA would compete with this model in the top. Based on the good performance and cooler, we give this video card an Ultimate award.
- Excellent performance
- Extensive vrm and XOC bios with 450W power limit
- Efficient, quiet cooler
- Fan and argb headers
- As poorly available as any other RTX 3080
EVGA RTX 3080 XC3 and FTW3 Test
EVGA is the largest brand of Nvidia video cards in the United States, but you usually notice little of that in the Benelux. It was also often difficult for us in the past to obtain samples from this manufacturer. However, as of this year, EVGA has a renewed focus on our region, allowing us to get started with two RTX 3080s from the brand that was previously missing from our round-ups of Nvidia’s next-generation graphics cards. In this review, we test the RTX 3080 XC3 Ultra and the RTX 3080 FTW3 Ultra.
Partly direct sales to consumers
Normally you would probably have noticed EVGA’s greater presence in our region mainly because the products became more available. Of course, this is hardly the case in the current situation. Unlike almost all other manufacturers, EVGA also sells video cards directly to consumers, in addition to regular sales through hardware stores. To this end, it uses a queuing system where you receive an email as soon as there is stock and it is your turn. You then have eight hours to place an order before the purchase option rolls over to the next in the queue.
Step-Up, Advanced RMA and Extend Warranty
Before we get into the tested cards, let’s also talk about the additional service options that EVGA offers over other manufacturers. For example, there is Step-Up, with which you can upgrade to a faster or newer video card within ninety days of purchase for an additional price difference. If your video card is broken, you can use an Advanced RMA, where you immediately receive a new card against collateral and you have thirty days to return your broken card. In addition, EVGA offers the option of extending the manufacturer’s warranty against payment to five or even ten years.
On the next page, we’ll take a closer look at the two RTX 3080s we’ll cover in this review. Do not skip pages; there you will find pictures of the inside of the cards.
XC3 Ultra versus FTW3 Ultra
EVGA has four series of RTX 3080s in its range, the main ones being the XC3 and FTW3. The other two series, Hybrid and Hydro Copper, use water cooling and are therefore more niche products. Within each series there are another two or three variants. The Ultra versions tested by us have the highest clock speed, the Gaming versions have somewhat lower clocks and the XC3 also has a Black version, which is supplied without a backplate.
EVGA RTX 3080 XC3 Ultra
The RTX 3080 XC3 Ultra is a relatively compact video card, which is about as long and wide as Nvidia’s Founders Edition. The cooler protrudes marginally above the thickness of two expansion slots; the manufacturer calls it a 2.2-slot video card. We will go into more detail about the heat sink equipped with three fans later in the review. EVGA has set a power limit of 340W for the RTX 3080 XC3 Ultra, which is 20W more than the FE. The maximum adjustable power limit is just a bit lower at 366W.
EVGA RTX 3080 FTW3 Ultra
Compared to the XC3 Ultra, the FTW3 Ultra is a giant, although the dimensions are still relatively modest if you compare it with the top models of some other manufacturers. The VRM is a lot heavier – more details about that on the next page – and that’s allowed, because the maximum power limit is 400W. For die-hard overclockers, EVGA has even released an XOC bios, which allows the card to consume a maximum of 450W. That is 40 percent more than the standard tdp of an RTX 3080!
Both the XC3 and the FTW3 are equipped with RGB lighting, but the FTW3 is a lot more exuberant; the large, light gray area on the front lights up completely on that model.
|RTX 3080||Nvidia Founders Editon||EVGA XC3 Ultra||EVGA FTW3 Ultra|
|Target price||€ 719, –||€ 895, –||€ 934, –|
|Boost clock frequency||1710MHz||1755MHz||1800MHz|
|Number of fans||2||3||3|
|Thickness||Dual slot||‘2.2 slot’||Triple lock|
|Image connections||1x HDMI, 3x DP||1x HDMI, 3x DP||1x HDMI, 3x DP|
|Power supply||1x 12-pin
|2x 8-pin||3x 8-pin|
|Vrm||15 + 3||13 + 3||19 + 3|
|Power limit (standard / max.)||320 / 370W||340 / 366W||380W / 400W *
* 450W with XOC bios
On this page we take a closer look at the PCBs of the RTX 3080 XC3 Ultra and RTX 3080 FTW3 Ultra in more detail. The RTX 3080 XC3 Ultra is a relatively compact card and therefore has a PCB that does not exceed the standard height of a PCIe slot. In the world of RTX 3080s, this is quite exceptional. Nevertheless, EVGA still had enough space to make various recesses on the back of the PCB, which correspond with holes in the rear backplate. The airflow of the outer fan can thus partly be blown through the video card. Nvidia applies the same concept to its Founders Edition, but Nvidia went a lot further by keeping the length of the pcb to a minimum.
The XC3 Ultra has a power supply consisting of sixteen phases, of which thirteen for the GPU and three for the memory. The power stages come from Alpha & Omega Semiconductor and have a rating of 50 amps. It is striking that EVGA uses various types of mosfets interchangeably. The copies with AL00 imprint correspond to the AOZ5332QI and BLN0 is code language for the AOZ5311NQI . The specifications hardly differ, so we may be dealing with a creative solution to the shortages of parts that plague the hardware world.
The larger pcb of the FTW3 Ultra is used, among other things, to house a huge battery of 55A power phases from OnSemi. We count 22, of which 19 for the GPU and the remaining 3 for the memory. In addition, we do not count the empty spot in the middle left, because in theory 23 could have been placed. A large part of the extra PCB space is used for air passages or is simply empty.
To the left of the three PEG connectors is a bios switch with which you can switch to the OC bios; effectively you apply a more aggressive fan profile, whereby the fans no longer stand still when the video card is idle. There are fan and argb headers at the end of the circuit board, with which you can synchronize a case fan or LED strip with your video card.
|RTX 3080 XC3 Ultra||RTX 3080 FTW3 Ultra|
|Controller||UPI uP9511R + uS5650Q||OnSemi NCP81610 + NCP45491|
|Power stages||16x 50A
Alpha & Omega AOZ5332QI / AOZ5311NQI
Our starting point for video card testing is that we want to minimize the bottleneck for the rest of the system. Currently, AMD unquestionably holds the gaming performance crown with its Ryzen 5000 processors. That is why we have chosen to test all video cards with the Ryzen 9 5950X as the processor. In addition to the two EVGA cards, we also (re) tested Nvidia RTX 3080 Founders Edition in the same conditions.
We have tested previously tested RTX 3080 variants on a Ryzen 3000 system and, moreover, with much older drivers. In cases where the test method is practically unchanged and where drivers have little influence, such as noise production, the results will still be reasonably comparable. Where applicable, we refer to this in the review. The performance is often not one-on-one to compare, if only because many games have now had updates. If you want to compare the EVGA cards with that caveat in mind with the RTX 3080s tested last year, you can make a comparison with previously tested models in the Pricewatch.
The exact specifications of the test system can be found in the table below.
|Test system GPUs|
|Processor||AMD Ryzen 9 5950X @ 4.65GHz all-core|
|Motherboard||ASUS RoG Crosshair VIII Hero (bios 2502, Agesa 18.104.22.168 Patch C)|
|Random access memory||G.Skill Trident Z Royal 32GB (2x 16GB) DDR4-3600 CL16-16-16-36|
|Ssd||Samsung 970 EVO 250GB (OS) + Kioxia Exceria + 2TB (games)|
|Nutrition||Be quiet Dark Power Pro 12 1500W|
|Cooling||Alphacool Eisblock XPX, Alphacool XT45 480mm radiator, Alphacool D5 water pump, be quiet Pure Wings 2 fans|
|Operating system||Windows 10 Pro 64bit (20H2)|
In this review we focus on the differences between the RTX 3080 versions. Broadly speaking, the performance of all cards with the same GPU will be comparable. To be able to discover small differences, such as differences arising from the firmware, choices in the PCB design and of course the quality of the cooler, we test fewer games than in a complete GPU review. We do test a lot more extensively per game.
We have selected three games in which it is quite possible to perform a longer test of ten minutes. This consists of the following per game:
- Metro Exodus – A loop of the built-in benchmark.
- F1 2020 – We let the AI drive around the circuit in Zandvoort for ten minutes, with the cockpit view and the weather on ‘very wet’.
- Shadow of the Tomb Raider – After warming up the video card for seven minutes in this game’s live rendered menu, we’ll run all the scenes from the built-in benchmark, which lasts three minutes in total.
During each test we log just about everything that can be logged: performance with PresentMon of course, but also temperature, clock frequency, fan speed and so on. Those familiar with the endurance test concept could see this as three individual endurance tests. Ten minutes is more than enough for this; with the majority of cards, the stable clock speed and temperature are reached within minutes. We test all games at 4k resolution with Ultra settings.
We tested the EVGA RTX 3080s and the Nvidia Founders Edition with the same driver: GeForce 461.72. Because we use other, longer tests, the figures are not one-on-one comparable to those from our GPU reviews. We also selected the games mainly for suitability for a longer-term benchmark and, for example, not for a ‘fair’ ratio between games with an AMD or Nvidia bias. This review is therefore explicitly not intended to make comparisons between video cards with different GPUs. For this we would like to refer you to our extensive GPU reviews. The most recent to include an RTX 3080 is the RX 6900 XT review.
Noise production and power consumption
In addition to performance, we also measure the noise production and power consumption of video cards. We measure the sound production in a soundproof box at a distance of 30 cm, with the decibel meter aimed diagonally at the video card. The threshold for our measurements is 18dB (A); you can interpret that as completely silent. During the idle test, the video card is obviously not loaded. The load test is based on an extended game test of ten minutes, so that the video card has time to warm up and reach its maximum noise production.
If a video card has a double bios, for example a quiet or OC mode, we run the synthetic benchmarks (3DMark and Unigine Superposition) and one game double in addition to the sound and temperature tests.
We carry out the current measurement with a so-called riser card from the manufacturer Adex, which we place between the PCIe slot and the video card. That way we can measure not only the current flowing through the loose power cables, but also the power that the video card draws directly from the PCIe slot. For the measurement itself we rely on different copies of the Tinkerforge Voltage / Current Bricklet 2.0. The riser card is equipped with such a bricklet for the final measurement. To measure the current of the PEG cable, the bricklet, together with the necessary connectors, was placed on a printed circuit board specially designed for our test. Depending on the number of power cables that a video card requires, we naturally use various meters.
We start with a number of synthetic benchmarks. 3DMark may not be a game, but as a benchmark it does give a first impression of the mutual relationships. Fire Strike is 3DMark’s DirectX 11 test, Timespy runs in DirectX 12 and Port Royal adds ray tracing. We run the Unigine Superposition benchmark in DirectX 11 mode.
Except for the Fire Strike Ultra test, all synthetic tests indicate the same order. The RTX 3080 XC3 Ultra is slightly faster than the Nvidia Founders Edition, but the FTW3 Ultra clearly distances itself and manages to outperform the FE by up to four to five percent. If we take a look at the numbers through our eyelashes next to those of all previously tested RTX 3080s, the FTW3 Ultra would definitely compete in the top.
In Metro Exodus, we see that the XC3 Ultra barely distinguishes itself from the Founders Edition. The FTW3 Ultra, on the other hand, clocks up to over 1920MHz, around seven percent higher.
In F1 2020, the RTX 3080 XC3 Ultra clocked just a bit lower than the Founders Edition, but the difference cannot be named. The FTW3 Ultra is clearly faster, which is of course due to the higher clock speeds.
In Shadow of the Tomb Raider, the XC3 Ultra manages to outsmart the Founders Edition if we look at the clock speed, but the performance is slightly lower.
The cooling compared
The coolers of the RTX 3080 XC3 and FTW3 are quite different in size. With some imagination you can call the XC3 a dual slot card, the FTW3 is a lot thicker. Yet the structure is comparable; both coolers use a large copper baseplate from which six or seven heat pipes distribute the heat throughout the cooling block. The baseplate is large enough to also cool the memory chips, while separate heatsinks have been placed for the VRMs. The typical shape of some heat pipes, such as the long heat pipe that makes a 180° bend in the XC3, serves to maximize the contact area with the cooling fins, according to EVGA.
Both video cards have three fans with a hydrodynamic bearing, which is said to have a longer life than traditional plain bearings. The blades of the fans are printed with small E’s in EVGA’s brand font. According to the manufacturer, they reduce the turbulence, but from a physical point of view that should not have to be EVGA logos.
The RTX 3080 FTW3 is also equipped with what EVGA calls ‘iCX3 technology’. Where most video cards only have a temperature sensor in the GPU, no fewer than nine are placed on the circuit board of this model. You can read all sensors via EVGA’s own Precision X1 software. In addition, they are used to individually determine the rotation speeds of the fans, in order to provide additional cooling of the parts of the video card that get the hottest.
There are two sides to the performance of a cooler: the temperature and the noise production. For this review we used the endurance test in Metro Exodus, which loads the video cards a bit more than the game we used before. So you can’t compare the values one-to-one with those in our previous RTX 3080 reviews, but because we retested the FE, we still have a point of reference.
The XC3 Ultra stays two degrees cooler during gaming than the Founders Edition; the FTW3 Ultra saves another three degrees from that. We do not actually measure any difference in this area between the standard and OC bios; the only noticeable difference is that the fans run continuously in idle. Fortunately, this does happen at a very low noise level.
If we look at the noise production, we see that the XC3 Ultra is slightly louder than the Founders Edition. That difference of half a decibel is not really noticeable. The FTW3 Ultra is clearly quieter.
If you want to set the fan profile yourself or simply want to buy the video card with the best possible cooler, the efficiency of the cooler is of great importance. After all, you want as much cooling as possible with as little noise production as possible.
To test this, we measure at which fan profile the card is as close to 40dB (A) as possible, measured at a distance of 30cm. Then we set the clock speed and voltage to fixed values of 1935MHz and 950mV. In this way we align the power consumption of all cards as much as possible after all, the power consumption is the ‘load’ of the cooler. Finally, we run the Metro Exodus test again to see how high the maximum temperature of the card is. As a result, we basically capture the capacities of the cooler in one number.
With a fixed noise production, the FTW3 Ultra naturally emerges as the winner, with a temperature that remains eleven degrees lower than with the FE. The XC3 Ultra also cools a bit better, but the difference with the FE is not huge. On our first encounter with the Founders Edition, we already referred to it as Nvidia’s best reference card ever, and apparently, the limitations of the dual slot form factor don’t allow much more.
Idle the RTX 3080 XC3 Ultra consumes practically the same as the Nvidia Founders Edition. Under load consumption rises to 329W, 20W more than the FE. This difference can be fully explained by the higher power limit with which the RTX 3080 XC3 comes. The FTW3 Ultra lets idle record a slightly higher consumption, but that is not much in an absolute sense. The energy consumption under load is clearly higher at 368W.
Normally, we would undoubtedly have warmly welcomed EVGA’s renewed focus on the Benelux market. In the current situation you don’t notice much of it as a consumer, but if the result is that a little more cards are now coming to the low countries than would otherwise have been the case, that is of course already a profit. Just with regard to the extras and warranty conditions offered, EVGA can put other video card manufacturers on edge.
It is difficult to draw firm conclusions about the tested RTX 3080 XC3 Ultra and FTW3 Ultra, because we can no longer compare the test results with those of all the models that we tested last year and, of course, have returned long ago. Still, we can say something, with the retested Founders Edition and less update-sensitive benchmarks as reference points.
The XC3 Ultra performs fairly similar to the FE, while the FTW3 Ultra compares well with the top models of other manufacturers, such as the MSI Suprim X and the Gigabyte Aorus Xtreme, in terms of performance. The XC3’s cooler is marginally better than the FE’s, but EVGA takes more out of the physical limitations of a dual-slot design than Nvidia itself. The copy on the FTW3 Ultra knows how to cover much more distance and keeps the RTX 3080 both significantly cooler and quieter than the FE. Whether the FTW3 Ultra also makes it difficult for the top models of, for example, ASUS and MSI in that area, we do not dare to say without direct comparison, but it is clear that EVGA would compete with this model in the top.
In a normal situation, the XC3 Ultra should also have a bit of its price and there is now very little meaningful to say about that. This applies less to the FTW3 Ultra, which we give an Ultimate award on the basis of its good performance and cooler.
The practice is of course more unruly at the moment. The actual buying of an RTX 3080 now mainly comes down to buying what you can buy, where for some time you also have to pull your wallet as soon as there is a card in stock somewhere. EVGA alone cannot do much about this, but by popular request we will soon publish an extensive article about the price increases of video cards. So stay tuned.