Test: AMD Radeon RX 5700 XT vs Nvidia RTX 2080 (2020)

| | , ,

Winner
ASUS GeForce RTX 2080 8G Turbo Edition GDDR6 HDMI DP 1.4 Graphics Card (TURBO-RTX2080-8G-EVO)
Best Price-Performance Ratio
PowerColor Red Devil Radeon RX 5700 XT 8GB GDDR6 Graphics Card
Model
ASUS GeForce RTX 2080
PowerColor Red Devil Radeon RX 5700 XT
Test Result
Test Result 9.8/10 Excellent May 2020
Test Result 9.6/10 Very Good May 2020
Manufacturer
ASUS
PowerColor
Performance
Video Memory
8 GB GDDR6
8GB GDDR6
Clock rate
1740 MHz
1400 MHz
Memory bandwidth
448GB/s
448GB/s
Displayport
HDMI
Value for money
Pros
  • Capable of 4K gaming at 60fps
  • DLSS is another exciting feature with a ton of potential
  • Sets a new bar for single-GPU performance
  • Great connectivity
  • Great mid-range 1440p graphics performance
  • Large performance improvement, beats RTX 2070 and is nearly as fast as the Radeon VII
  • Good design
  • Best price-performance ratio
Cons
  • Price
  • No real-time ray tracing
Recommended by us?
Winner
ASUS GeForce RTX 2080 8G Turbo Edition GDDR6 HDMI DP 1.4 Graphics Card (TURBO-RTX2080-8G-EVO)
Model
ASUS GeForce RTX 2080
Test Result
Test Result 9.8/10 Excellent May 2020
Manufacturer
ASUS
Performance
Video Memory
8 GB GDDR6
Clock rate
1740 MHz
Memory bandwidth
448GB/s
Displayport
HDMI
Value for money
Pros
  • Capable of 4K gaming at 60fps
  • DLSS is another exciting feature with a ton of potential
  • Sets a new bar for single-GPU performance
  • Great connectivity
Cons
  • Price
Recommended by us?
Best Price-Performance Ratio
PowerColor Red Devil Radeon RX 5700 XT 8GB GDDR6 Graphics Card
Model
PowerColor Red Devil Radeon RX 5700 XT
Test Result
Test Result 9.6/10 Very Good May 2020
Manufacturer
PowerColor
Performance
Video Memory
8GB GDDR6
Clock rate
1400 MHz
Memory bandwidth
448GB/s
Displayport
HDMI
Value for money
Pros
  • Great mid-range 1440p graphics performance
  • Large performance improvement, beats RTX 2070 and is nearly as fast as the Radeon VII
  • Good design
  • Best price-performance ratio
Cons
  • No real-time ray tracing
Recommended by us?

AMD Radeon RX 5700 XT vs Nvidia Geforce RTX 2080

Being a gamer is tough. Every 2 years or so a new graphics card gets introduced and you, again, have to ask yourself wether it is worth to upgrade or wait.

Which GPU gives you the most performance for the money and which is most future-proof, so you don’t have to think about upgrading your GPU every year.

General wisdom says: AMD offers better price-performance ratio, while Nvidia offers better high-end performance. If that is the case, we will find out in this review.

In this test we reviewed two of the most popular graphics cards amongst gamers: The AMD Radeon RX 5700 XT vs the Nvidia RTX 2080.

Both offer a lot of performance, not only for gaming, but also for video editing and workstations.

They are also priced differently. But do you really get your money’s worth getting the more expensive option?

We are taking a closer look at both GPUs and give you our opinion on which of both is best for you in the final Verdict below.

Test Results: AMD Radeon RX 5700 XT vs Nvidia Geforce RTX 2080

Winner
ASUS GeForce RTX 2080 8G Turbo Edition GDDR6 HDMI DP 1.4 Graphics Card (TURBO-RTX2080-8G-EVO)
Best Price-Performance Ratio
PowerColor Red Devil Radeon RX 5700 XT 8GB GDDR6 Graphics Card
Model
ASUS GeForce RTX 2080
PowerColor Red Devil Radeon RX 5700 XT
Test Result
Test Result 9.8/10 Excellent May 2020
Test Result 9.6/10 Very Good May 2020
Manufacturer
ASUS
PowerColor
Performance
Video Memory
8 GB GDDR6
8GB GDDR6
Clock rate
1740 MHz
1400 MHz
Memory bandwidth
448GB/s
448GB/s
Displayport
HDMI
Value for money
Pros
  • Capable of 4K gaming at 60fps
  • DLSS is another exciting feature with a ton of potential
  • Sets a new bar for single-GPU performance
  • Great connectivity
  • Great mid-range 1440p graphics performance
  • Large performance improvement, beats RTX 2070 and is nearly as fast as the Radeon VII
  • Good design
  • Best price-performance ratio
Cons
  • Price
  • No real-time ray tracing
Recommended by us?
Winner
ASUS GeForce RTX 2080 8G Turbo Edition GDDR6 HDMI DP 1.4 Graphics Card (TURBO-RTX2080-8G-EVO)
Model
ASUS GeForce RTX 2080
Test Result
Test Result 9.8/10 Excellent May 2020
Manufacturer
ASUS
Performance
Video Memory
8 GB GDDR6
Clock rate
1740 MHz
Memory bandwidth
448GB/s
Displayport
HDMI
Value for money
Pros
  • Capable of 4K gaming at 60fps
  • DLSS is another exciting feature with a ton of potential
  • Sets a new bar for single-GPU performance
  • Great connectivity
Cons
  • Price
Recommended by us?
Best Price-Performance Ratio
PowerColor Red Devil Radeon RX 5700 XT 8GB GDDR6 Graphics Card
Model
PowerColor Red Devil Radeon RX 5700 XT
Test Result
Test Result 9.6/10 Very Good May 2020
Manufacturer
PowerColor
Performance
Video Memory
8GB GDDR6
Clock rate
1400 MHz
Memory bandwidth
448GB/s
Displayport
HDMI
Value for money
Pros
  • Great mid-range 1440p graphics performance
  • Large performance improvement, beats RTX 2070 and is nearly as fast as the Radeon VII
  • Good design
  • Best price-performance ratio
Cons
  • No real-time ray tracing
Recommended by us?

Ranking Second: AMD Radeon RX 5700 XT

AMD Radeon RX 5700 XT

Pros

  • Large performance improvement, beats RTX 2070 and is nearly as fast as the Radeon VII
  • Easily faster than RTX 2060 Super
  • Good design
  • Best price-performance ratio

Cons

  • No real-time ray tracing

Great price-performance ratio

Since AMD released the Radeon RX 480 three years ago, only refreshes like the Radeon RX 580/590 have been available to buyers of mid-range graphics cards.

The Radeon RX Vega from 2017, on the other hand, were hardly available and expensive in the beginning, and they also require a lot of power.

Now AMD finally brings new models: The Radeon RX 5700 XT and the Radeon RX 5700 are the actual successors of the Radeon RX 480/470, but cost significantly more than their predecessors – because Nvidia has shown how it’s done.

With starting prices of 420 Dollar for the XT and 370 Dollar for non-XT, AMD positions both graphic cards above and below the Geforce RTX 2060 Super.

The competition has rated the pixel accelerators, which are internally called navigation devices, so strongly that new Super versions of the Geforce RTX 2060 and the Geforce RTX 2060 have been released with higher speeds to steal the show from AMD.

Whether this has worked depends on the angle from which the navigation maps are viewed.

For AMD, navi, the umbrella term for several graphics chips based on the new RDNA technology, is an extremely important design: It is not only in the Navi-10-GPU for the Radeon RX 5700 (XT), but also forms the basis for the systems-on-a-chip in the next Playstation and in the Xbox Anaconda alias Project Scarlett.

For the desktop versions, however, the manufacturer has decided not to support hardware-accelerated raytracing, but for the time being only for console implementation.

Later models with the Navi-20 chip could master fast real-time ray tracing, but we don’t expect any such consumer graphics cards before end of 2020.

AMD Radeon RX 5700 XT

But the Radeon RX 5700 (XT) are technically exciting and fast cards even without ray tracing: Both use the GPU called Navi 10, which is the first mainstream graphics chip after the Vega 20 of the Radeon VII to be produced by TSMC using the N7P process.

With 251 mm² at 10.3 billion transistors Navi 10 turns out quite compact, the Polaris 10/20 of a Radeon RX 480/580/590 is comparably large with 232 mm² at 5.7 billion transistors.

For AMD this means many GPUs per wafer and thus low prices, whereas Nvidia’s TU106 of the Geforce RTX 2060 Super is probably more expensive despite the cheap 12 nm production with 445 mm² at 10.8 billion transistors.

Like Nvidia, AMD is now using GDDR6 video memory on a 256-bit interface to achieve a high data transfer rate, which costs less than the HBM2 memory of a Vega 64.

The two navi-cards also remind us of the Geforce RTX Super in terms of the number of shader units and their frequencies: The Radeon RX 5700 XT is the full version with 2,560 ALUs with a typical AMD gaming clock of 1,755 MHz, whereas the Radeon RX 5700 has 2,304 computing cores at usually 1,625 MHz in games.

However, the GPU can also clock higher for applications that put less strain on the graphics card.

The Geforce RTX 2060 Super uses 2,176 cores at 1,650 MHz boost frequency for comparison.

Previous Polaris and Vega chips required much more shader units than Nvidia’s models and could only really utilize these in resolutions like 1440p or 4K-UHD. AMD has therefore developed a new design for navigation systems that differs significantly from the previous Graphics Core Next (GCN) as the technical basis: It is called RDNA, short for Radeon Deoxyribonucleic Acid, and is thus quasi the gene.

RDNA makes navigation strong

AMD Radeon RX 5700 XT

The Navi-10 chip contains the first version of RNDA technology, and as with GCN, AMD will expand and improve it over the next few years.

The basic structure of RDNA is well known: The chip integrates a display unit, a multimedia block, a root controller, various caches, two instead of four shader engines (like Vega 10) and in it the actual compute units.

New is the support for 16 PCIe Gen4 lanes, which doubles the speed compared to PCIe Gen3.

In order to use this advantage, a suitable processor is necessary – more precisely a Ryzen 3000 with X570 chip, which AMD also released today.

PCIe Gen4 is not relevant for games for the time being, but some applications like Davinci Resolve should benefit from the drastically higher data transfer rate during 8K playback in the timeline.

Due to lack of time we haven’t been able to reproduce this scenario so far, but special internal tests are planned.

Basically, RDNA is based on the same ISA (Instruction Set Architecture) as GCN, but the implementation in hardware is completely different:

The goal was to achieve significantly higher performance per clock cycle.

For this purpose AMD rebuilt the Compute Units and for the first time in many years radically changed the SIMD units and the cache system.

The manufacturer is talking about a 50 percent increase in speed for the same energy requirement, with a good half of this again being attributable to the architecture and the rest to 7 nm and frequency optimizations.

AMD Radeon RX 5700 XT

Each compute unit consists of 64 ALUs, which receive their instructions in the form of multiple threads through so-called wavefronts.

Up to now AMD used a Wave64 implementation, which is relatively coarse-grained.

Nvidia has been using a more fine-grained Warp32 implementation for years, which has advantages in branching in the code and thus improves the utilization of the shaders.

AMD is now also using a Wave32 approach for RDNA, and the ALUs themselves have also been changed to get the incoming work done in less time.

Instead of SIMD16 units at GCN, RDNA uses SIMD32 units, which means two Wave32 fronts are completed in one cycle instead of one Wave64 front in four cycles.

In the best case, this quadruples the workload and the SIMD units have to wait for data less often.

Optionally, the shader compiler can still deliver a Wave64 front to the SIMD32 slots, which then needs two cycles.

Besides the regular vector units there are twice as many scalar and scheduler units as with GCN. Furthermore, the asynchronous compute engines should provide the CUs with better data supply.

AMD speaks of Dual Compute Units, because always 2x 32 ALUs are combined in one.

They share the L0 cache with 16 KByte and doubled bandwidth, because AMD previously spoke of an L1 cache in this context.

This in turn is now shared by five Dual Compute Units; it has a capacity of 128 KByte. The L3 cache is even larger, but slower at 4 MByte.

This is twice the size of Polaris 10 (RX 480) and corresponds to Vega 10, but Navi 10 has only 40 CUs instead of 64 CUs and thus more buffer space per compute unit.

In addition, AMD claims to have reduced the cache latencies for RDNA and improved the lossless delta color compression. The latter reduces the effective demand for bandwidth, but not the necessary amount of space in the buffers or in GDDR6 video memory.

This is connected to four 64-bit controllers and is interleaved with 64 raster output stages, twice as many as Polaris 10.

RX 5700 (XT) in detail

AMD Radeon RX 5700 XT

An innovation introduced with Vega but never used – the Primitive Shader – has now been implemented in navigation systems in such a way that it is active, according to AMD: The Primitive Shader rejects hidden triangles earlier (culling), which increases performance under Direct3D and Vulcan.

Previously used implementations such as an FP16 speed, which is twice as high as FP32 commands, have also been integrated by Navi. Current games like Resident Evil 2 use this feature, upcoming titles like Borderlands 3 or Wolfenstein Youngblood will also use it.

The multimedia block alias VCN (Video Core Next) has undergone some changes, the same applies to the display controller: Navi 10 can decode H.264 content in 8K at 30 fps and encode in 4K at 90 fps, for H.265 AMD has now integrated 4K60 as well as 8K24 decode and 4K90 encode.

The widely used VP9 codec is decoded in hardware with 4K60, but our power consumption measurements show that the driver is not yet mature. The AV1 codec is still handled by the processor.

Navi now supports Displayport 1.4 including DSC (Display Stream Compression) for higher resolution monitors, but it remains at HDMI 2.0b.

The slot bezel of the Radeon RX 5700 XT and Radeon RX 5700 therefore again has three display ports and one HDMI, a virtual link via USB-C as Nvidia is missing.

However, AMD said that support is generally available in the Navi-10 chip.

Both Radeon graphics cards use a blower cooler, also called DHE (Direct Heat Exhaust) in the reference design. In addition to an evaporation chamber instead of heat pipes, AMD uses a phase-change graphite paste that is optically reminiscent of a thermal pad, as in the Radeon VII.

The cards are very quiet in idle mode, but quite loud under load.

The Radeon RX 5700 XT can be recognized by the dented radiator and backplate. The Radeon RX 5700 lacks the dent and the plate on the back, but this is only for visual purposes anyway and doesn’t cool.

Both graphic cards use an 8-pin and a 6-pin power connector, because the boards are identical. The XT simply has one more GPU voltage converter phase built in.

With the so-called Radeon Image Sharpening, navigation systems support a contrast-based post-processing filter, which is reminiscent of Nvidia’s Freestyle or the manufacturer-independent Reshader: The rendered image is sharpened afterwards, as various approaches of temporal edge smoothing slightly soften it – so fine texture details are shown to better advantage.

Radeon Image Sharpening works at any resolution under D3D9, D3D12 as well as Vulcan and is activated in the display options of the driver.

It is sufficient for the 2060S

For our benchmarks we use a Core i9-9900K with a standard clock rate and the PL1 target of 95 watts provided by Intel; it is combined with 16 GByte DDR4-2667 on a Z370 board.

All applications and games, as well as Windows 10 x64 v1930 are on a WD Black SN750 with 1 TByte, one of the fastest NVMe SSDs currently available. We use the Geforce 431.16 from Nvidia and Radeon Software 19.7.1 Beta5 from AMD as drivers.

AMD Radeon RX 5700 XT

According to our measurements, the Radeon RX 5700 XT is on average almost 20 percent faster than a Radeon RX Vega 64. GTA 5, Jurassic World Evolution and No Man’s Sky are particularly well suited to the navigation card, where it is a quarter to a third faster.

The lead over the non-XT is 10 percent, which means that the Radeon RX 5700 also overtakes a currently similarly expensive Radeon RX Vega 64 on average. AMD’s previous top model, the Radeon VII, can hardly distinguish itself from the Radeon RX 5700 XT, despite double the memory bandwidth and 4,096 instead of 2,560 shader units; the lead is only 5 percent.

Applied to Nvidia’s portfolio, this means that a Radeon RX 5700 XT overtakes a Geforce RTX 2060 Super and is also slightly ahead of a regular Geforce RTX 2070, which is currently on sale for 450 DOollar, i.e. 50 Dollar more than the new Radeon RX 5700 XT.

The Radeon RX 5700 without XT has almost the same speed as a clearly more expensive Geforce RTX 2060 Super, it is therefore also faster than an older Geforce GTX 1080. Compared to the most widespread AMD card according to Steam, the Radeon RX 580, the navigation cards provide roughly twice the performance.

As so often in a duel between AMD and Nvidia, it depends on which game is currently running with which resolution: In Ark Survival Evolved for example – still in the Top10 on Steam – the Radeons are drastically behind in 1080p.

Only from 1440p it gets better, but with epic details the title stutters on every graphics card. But AMD has gained a lot in GTA 5, also a very popular game until today, which is still demanding with ultra settings. In general, the Radeon RX 5700 (XT) is well suited for 1440p gaming, even maximum settings are usually rendered at over 60 fps.

Away from games, AMD’s cards convince with excellent performance in Davinci Resolve, the XT is faster than a Geforce RTX 2080!

Regarding power consumption, the navigation pixel accelerators achieve a drastically better energy efficiency than the previous models.

The Radeon RX 5700 XT, which is almost as fast as the Radeon VII, only needs 220 watts under load instead of 290 watts. And the Radeon RX 5700 manages to overtake the Radeon RX Vega 64, but is content with just under 180 watts instead of 285 watts.

Compared to a Radeon RX 580, this means an almost doubled Fps/Watt ratio, but the Polaris chip is operated far above its sweet spot.

AMD Radeon RX 5700 XT

Compared to Nvidia’s Geforce RTX, the efficiency drops slightly: A Geforce RTX 2070 only needs 175 watts instead of 220 watts for the speed of the Radeon RX 5700 XT and is faster than a Radeon RX 5700 with almost identical power consumption.

This shows that AMD achieves a comparable performance per compute unit in navigation as Nvidia’s Turing, but at least at similar frequencies has the less efficient design despite the 7 nm process.

Availability and conclusion

AMD sells the Radeon RX 5700 XT for 420 Dollar, the Radeon RX 5700 is offered for 370 Dollar. For the time being there will only be reference models, the partners’ custom designs followed in the next few weeks after launch.

AMD includes a three-month access to the Xbox Game Pass for PC with the navigation cards.

Verdict: Great price-performance ratio

RX 5700 xt

If the competition quickly brings in a few super cards and the regular Geforce RTXs drop in price, this speaks for AMD’s navigation models: Both graphics cards deliver high performance, because the Radeon RX 5700 XT beats the Geforce RTX 2070 by a narrow margin and the Radeon RX 5700 is on par with the new Geforce RTX 2060 Super.

The navigation pixel accelerators cost clearly less, whereby cheaper custom designs from partners are to follow.

With the Navi architecture, AMD has succeeded in drastically improving the computing power per shader unit: It is now roughly comparable to what Nvidia implemented with the Turing technology and GDDR6 video memory is used by both graphics card manufacturers anyway.

What the AMD models lack, however, is hardware-accelerated raytracing for games like Cyberpunk 2077, Doom Eternal, Metro Exodus, Vampire Bloodlines 2 and Wolfenstein Youngblood.

The Radeon RX 5700 (XT) also has the disadvantage in terms of energy efficiency, despite 7 nm chips, although the difference is by far not as big as it is with a Radeon RX 580 or a Radeon RX Vega 64.

The XT variant apparently clocks above the sweet spot, whereas the non-XT is on par with the Geforce RTX 2060 Super in terms of performance per watt.

In practice, however, the differences are small, neither the electricity bill nor the noise level of the cooler in custom designs should be greatly affected by this. The small chip shouldn’t be a technical problem either.

Ranking First: Nvidia Geforce RTX 2080

Nvidia Geforce RTX 2080

Pros

  • Great for 1440P /1080P gaming
  • DLSS is an exciting feature with a ton of potential
  • Ray tracing and deep learning
  • New and improved design

Cons

  • 650W power supply requirement

Amazing performance

Under the codename Turing, Nivida worked for a long time on the next GPU generation, which is now officially available for a longer time already – apart from the small delay with the RTX 2080Ti.

In the run-up Nvidia only published information about how strong the graphics cards really are. On September 19th 2018, the time had come: Nvidia lifted the ban on publishing test reports, which finally brought independent performance values to light.

First of all, it is important to mention that Nvidia used much more technology from the workstation/server-oriented Volta architecture for Turing than its direct predecessor Pascal.

A central feature, which also gives it its name: Raytracing. If you followed the live presentation, this word might have been burned into your memory by now.

ModelRTX 2080TiRTX 2080RTX 2070
Architecture12nm TU10212nm TU10412nm TU104
CUDA cores435229442304
Base clock135015151410
Boost clock151517101620
Storage11GB GDDR68GB GDDR68GB GDDR6
Memory bus352 bits256 bits256 bits
Bandwidth616GB / s448GB / s448GB / s
TDP250W215W175W
Price1259 Dollar849 Dollar639 Dollar

But if there was one thing that left a bland aftertaste, it was the price tag of Nvidia.

The RTX 2080Ti for instance costs 1260 Dollar in the Founders Edition and thus almost twice as much as the predecessor GTX 1080Ti for the launch.

Does the company shamelessly exploit its market power in the graphic card segment here or are there good reasons why the price increase is so immense? We’ll find that out in the following.

The RTX 2080Ti and RTX 2080 use the Turing TU102 respectively TU104 chip. Even in the 12nm process, the technical performance on paper is impressive: 18.6 billion transistors including the additional raytracing and tensor hardware find room here.

The CUDA cores also work more efficiently and promise a considerable increase in performance compared to the previous generation.

As already mentioned, we find the TU104 chip in the RTX 2080 again.

Based on the technical data, the GPU should be significantly slower than the RTX 2080Ti, but still faster than the GTX 1080Ti.

Because in the previous generation the 1070 model corresponded to the GTX 980Ti in terms of performance.

Nvidia Geforce RTX 2080

The GDDR6 memory with 14Gb/s via 256 bit bus should certainly run well.

But the pure memory size is the same for both RTX cards with 8GB – the GTX 1080Ti on the other hand still had 11GB video memory.

How this affects the real test is shown in the following paragraphs.

Before we move on to the benchmarks, we’ll take a closer look at the RTX 2080Ti, which doesn’t differ optically from the RTX 2080, if you disregard the power plug (8+8 pin instead of 8+6 pin).

In terms of colour, the many metal components in light grey and black dominate.

With the new RTX models, Nvidia has completely turned the reference design upside down and implemented a new cooling mechanism for a long time. Instead of a radial fan that blows the exhaust air away to the rear, there is now a cooler with two fans with 13 blades each.

These run in idle with 1500 revolutions per minute, which is almost inaudible in practice. We’ll find out later how loud the fans turn up under load.

From a purely visual point of view, it’s of course a question of taste whether the look of the new cards is pleasing.

But Nvidia does a good job in the workmanship, as the RTX 2080Ti FE feels very high-end.

Nevertheless, everything remains in the 2-slot form factor, so Nvidia doesn’t experiment in this respect. With a length of just under 27cm, the card should also fit into most PCs without any problems.

Nvidia Geforce RTX 2080

The manufacturer does without RGB lights, instead there is a “Geforce RTX” lettering on the side, which stands out from the rest of the design.

The VirtualLink connector (USB type C), on the other hand, is a novelty and is installed at the rear right.

It supplies four lanes via DisplayPort and a maximum of 23 watts, which are intended for VR headsets, among other things. In addition, there is one HDMI and three DisplayPorts.

3DMark is the current version of the very popular synthetic benchmark.

Although this benchmark doesn’t provide any exact conclusions about the actual gaming performance, one can still interpret a rough direction regarding the GPU performance from the values.

In this category, both the RTX 2080 and the RTX 2080Ti can show off with good performance values.

In the comparison the 2080Ti is more than a third faster than the direct predecessor model and also the RTX 2080 is ahead by at least twelve percentage points.

In the strategy game, Civilization VI, it already becomes clear: The increased performance of the graphics card only becomes noticeable at higher resolutions.

The three cards just mentioned are almost equally fast under FullHD. Only at 4K resolution does the RTX 2080Ti perform around 40 percent and thus considerably more FPS than the GTX 1080Ti.

The RTX 2080 is meanwhile in between, but with a slight tendency to the GTX 1080Ti in the FPS rates.

Nvidia Geforce RTX 2080

In Shadow of the Tomb Raider we are dealing with a dynamic and action-packed game.

In this case the FPS rates are already clearly distinguishable under FullHD, whereby the RTX 2080 is only minimally ahead of the GTX 1080Ti. The RTX 2080Ti is also about 40 percent faster than GTX 1080Ti under 4K.

The widely announced raytracing feature is still a long way off, though.

The first games, which are supposed to support the feature, are still in development or will be patched later accordingly.

For the time being it remains to be seen whether raytracing is really worthwhile considering the performance required for it.

Although the RTX 2080Ti’s power consumption has slightly increased in comparison to its predecessor, this is only marginally noticeable in idle.

It looks a bit clearer under heavy gaming use, where the new RTX models consume 324 or 282 watts.

As expected, the new cooling design shows its strengths in temperature measurement, as both new cards are cooler than all other comparable models.

At the same time, the cards are also very quiet with only 28 decibels in idle and around 37 decibels in gaming.

Such results would probably have been impossible with the old cooler with radial fan.

Verdict: Amazing performance

Nvidia Geforce RTX 2080

Let’s start with the obvious in conclusion: The Nvidia Geforce RTX 2080Ti is currently the fastest graphics card aimed at end customers.

In comparison to its predecessor, it works significantly faster, especially in higher resolutions, and sometimes reaches more than a third more FPS.

Meanwhile, we don’t even need to talk about the competitor AMD: Their graphics cards are in a different league and are simply not competitive.

The new cooling design is a great help here, which also contributes to the improved performance.

Nvidia has a bit more justification need in the RTX 2080, as it often proved to be only marginally faster than the GTX 1080Ti in the test.

A rule of thumb could still be deduced from the past years that the xx70 model is always as strong as the xx80Ti model of the previous generation. In view of the results, it seems unlikely that this formula will work again with Turing, because the RTX 2070 would then have to calculate almost as fast as the RTX 2080.

Which compromises gamers have to make when raytracing also remains an open question.

We have already questioned the high prices of Nvidia in advance and can now confirm:

The called price recommendations are excessive. This already starts with the RTX 2080Ti, which costs almost twice as much as its predecessor for the launch at 1260 Dollar (prices went down since launch), but often doesn’t bring more than a third more performance.

In short: Who does not want to do without the achievement of Turing – and for which the price plays a subordinated role – can access with the new RTX models.

But if you focus on a fair price-performance ratio or simply want to set an example against Nvidia’s cheeky price policy, you should probably look at the AMD Radeon RX 5700 XT.

Verdict: AMD Radeon RX 5700 XT vs Nvidia Geforce RTX 2080

All in all the verdict in this review is pretty simple. If you value performance over all, go with the Nvidia RTX 2080. Yes, it costs a bit of money for the performance you get, but it will als be a very futur-proof GPU and good to go for at least 3 years.

If you are budget conscious and don’t need to play every game on ultra graphics settings while maintaining 180 FPS you should go with the AMD Radeon RX 5700 XT.


Are you interestet in more hot tech deals? Subscribe to our Email-list and we let you know when the best deals come up.

Leave a Comment