NVIDIA GeForce RTX 2060 and RTX 2070 Review (w/ BF V Ray Tracing Results)

Add Your Comments

Better late than never is a popular phrase coined by Geoffery Chaucer from way back in the day that is perfectly applicable now, particularly for this review. As most know by now, both of NVIDIA’s mid-range cards have been released in the RTX 2060 FE and RTX 2070 FE. The latter being released in 10/2018 while the RTX 2060 was released (well arrived on my doorstep) January 7th – our first day of CES.

Both cards aim to bring ray tracing (RT) capabilities and its Tensor Cores to a lower price point where most graphics cards are sold. To that end, the MSRP for the RTX 2070 FE will be $599 with reference versions at $499. The RTX 2060 FE on the other hand, is priced at $349. Though the pricing brackets have seemingly moved up a notch with Turing’s release, both cards strive to offer a better value than the higher-tier cards while still allowing for a good gaming experience including the use of Ray Tracing in Battlefield V (now tested!!!). Read on to see how both cards stacked up against the full RTX lineup which was retested using the latest drivers.

Specifications

At a high level, the 2060 and 2070 are not too much different from its bigger brothers. Both offer the RT and Tensor cores but in this flavor, use the TU106 core. The 2070 uses the full version of this while the 2060 is cut back slightly.

The 2070 includes 2944 CUDA core, 288 Tensor Cores, and 36 RT cores. The 2060 meanwhile sports 1920 CUDA cores, 242 Tensor Cores, and 30 RT cores. The memory bandwidth is 256-bit on the 2070 and 8GB GDDR6, while the 2060 is 192-bit and sports 6 GB GDDR6. Both cards memory is clocked at 1750 MHz/14 Gbps. Power use for the RTX 2070 FE is rated at 185 W with the 2060 coming in a bit lower at 160 W.

If one needs a features refresh for Turing, head on over to our RTX 2080 and RTX 2080 Ti review where we covered that material in detail.

Specifications
RTX 2060 RTX 2070
CUDA Cores 1920 2304
Tensor Cores 242 288
RT Cores / RTX-OPS 30 / 37T 36 / 45T
GPU Base Clock 1365 MHz 1410 MHz
GPU Boost Clock 1680 MHz 1710 MHz
Frame Buffer 6 GB GDDR6 8 GB GDDR6
Memory Interface 192-bit 256-bit
Memory Clock (Data Rate) 14 Gbps (1750 MHz)
Memory Bandwidth 448 GT/s
ROPs 48 64
Texture Units 80 144
Texture Fillrate (Gigatexels /sec) 201.6  246.2
GigaRays /sec 5 GR/s 6 GR/s
TDP (Watts) 160 W 185 W
Price $349(FE) $599(FE) / $499

GPUz

Below is our always gratuitous GPUz picture confirming the specifications above.

GPUz – RTX 2060 FE

GPUz – RTX 2070

Retail Packaging and Accessories

The retail packaging for the RTX 2060 and RTX 2070 FE cards come in the same style packaging we saw the RTX 2080 and 2080 Ti show up in, but these are not as wide (the cards are not as long). Slide the box open to expose the card and we see it sitting securely in foam protecting it from transport damage.

This slideshow requires JavaScript.

Meet the NVIDIA RTX 2060/2070

NVIDIA GeForce RTX 2060 and RTX 2070

Like the 2080 and 2080 Ti, both cards use NVIDIA’s dual fan, dual slot cooling solution. We see two large fans with a grey aluminum outline and shroud, while between the fans is the name of the card. Since both cards look exactly the same and have the same underpinnings/power delivery, we’ll use the RTX 2060 in all of these pictures

When reading out our RTX 2080/RTX 2080 Ti review, one will likely notice the little FE cards look the same as the big ones except they are a couple of inches shorter. They use a different PCB and power bits here versus the more powerful cards, and they retain the same size fans, but have a different heatsink hidden below.

A Closer Look

Power needs on both the 2060 and 2070 come in the form of a single 8-pin PCIe plug located on the end of the card opposite the I/O. This combination yields in spec power delivery of up to 225 W. As far as outputs are concerned, both cards have 2x DisplayPort, 1x HDMI, 1x USB Type-C, and DVI. This is a bit different than the 2080/2080 Ti which doesn’t include DVI, but adds another DisplayPort.

Taking this thing apart was, for me, quite an exercise in patience. I think I lost count of how many screws, and tiny screws at that, were removed to get these images. One thing of note here is the 8-pin PCIe power connector is soldered to the board with wires as opposed to being directly soldered to the board. Outside of that, we are able to see the Micron memory on this 2070 (our 2060 has Samsung, note) surrounding the TU106 core.

Zooming in a bit on the power delivery, the GPU VRM uses six-phases and controller by a uP9512R controller. The memory VRM is 2-phase and is managed by a uP1666Q controller.

Power Delivery

 

Test System and Benchmark Methods

Our test system is based on the latest mainstream Intel platform, Z370, and uses the i7-8700K 6c/12t CPU. The CPU is overclocked to 4.7 GHz on all cores/threads with cache set to 4.3 GHz. The clock speeds used provides a good base to minimize any limitations the CPU may have on our titles, particularly when using the lower resolutions, and should be attainable with a good air cooler or better. The DRAM is in a 2×8 GB configuration at 3200 MHz with CL15-15-15-35-2T timings which is a middle of the road option that balances performance and cost.

Test System Components
Motherboard ASRock X370 Taichi
CPU Intel i7 8700K @ 4.7 GHz / 4.3 GHz Cache
CPU Cooler EVGA CLC 240
Memory 2×8 GB G.Skill Trident Z 3200 MHz CL15-15-15-35
SSD Toshiba OCZ TR200 480 GB (OS + Applications)
Power Supply EVGA 750W G3
Video Card RTX 2060/RTX 2070 (417.71 drivers)

Thanks go out to EVGA for providing the CLC 240 CPU Cooler and 750 W G3 Power Supply to cool and power the system, G.Skill for the Trident Z DRAM, and Toshiba OCZ for the 480 GB TR200 SSDs storage running the OS, benchmarks, and games. With our partners helping out, we are able to build matching test systems to mitigate any differences found between using different hardware. This allows for multiple reviewers in different locations to use the same test system and compare results without additional variables.

   

Below are the tests we run with a brief description of the settings. We have made some significant changes since the last update adding a few new titles and dropping some of the older games. More details can be found in the GPU Testing Procedure article which we have updated with our latest benchmarks.

  • UL 3DMark Time Spy – Default settings
  • UL 3DMark Fire Strike (Extreme) – Default settings
  • Shadow of the Tomb Raider – DX12, “Highest” preset (will add RTX when it has been patched)
  • The Division – DX12, Ultra preset, VSync Off
  • Ashes of the Singularity: Escalation – DX12, Crazy preset, GPU focused
  • Far Cry 5 – Ultra defaults
  • F1 2018 – Very High defaults, TAA, and x16 AF, Australia track, show FPS counter
  • World of Tanks: Encore Benchmark – Ultra defaults
  • Final Fantasy XV Benchmark – High defaults

Synthetic Benchmarks

Our first set of benchmarks hail from Underwriters Laboratories who acquired Futuremark back in 2014. Earlier in 2018, a rebrand occurred and since that time, Futuremark is now UL. The benchmarks have not changed, just the name. We chose to stick with 3DMark Fire Strike (Extreme) and 3DMark Time Spy as these tests give users a good idea of performance on modern titles.

3DMark Fire Strike (Extreme) is a DX11-based test which UL says the graphics are rendered with detail and complexity far beyond other DX11 benchmarks and games. This benchmark runs at 1920×1080. 3DMark Time Spy is a DX12 benchmark designed for Windows 10 PCs. It supports new API features such as asynchronous compute, explicit multi-adapter, multi-threading, and runs at 2560×1440 resolution.

3DMark Fire Strike (Extreme) and 3DMark Time Spy

Our first glimpse of performance results shows the RTX 2060 (blue) running just a bit behind the 1070 Ti here and 15% slower than a Strix Vega 64. The RTX 2070 is over 21% faster than the 2060 here and easily beats out Vega 64.

Our Timespy results show the RTX 2060 gaining some momentum and easily beating out the 1070 Ti and matching Vega 64 here. The 2070 is nearly 19% faster than the 2060 and Vega 64 here.

Gaming Benchmarks

Moving on to the gaming benchmarks, we have updated our testing suite to bring more modern titles into the mix. Gone are GTA V, Crysis 3, and Rise of the Tomb Raider, which were replaced with Shadow of the Tomb Raider, World of Tanks: enCore benchmark, F1 2018, Final Fantasy XV benchmark, and Far Cry 5. We kept The Division and Ashes of the Singularity (though we updated to AOTS: Escalation). The games should provide a good view of the overall performance of the card. Many of these are DX12 games.

We will not test DLSS features in FF XV as the benchmark itself is a bit flawed. Ray Tracing will also not be tested here even though BF V and Windows have now been updated to support it. In the future, SoTR will have it along with many other titles so we will circle back when appropriate.

World of Tanks: enCore and F1 2018

For WOT:e, the 2060 yields 164 FPS while the 2070 reached 194 FPS. 2060 performance here is faster than the 1070 Ti and the Vega 64 as well. In F1 2018, the 1070 Ti/RTX 2060/Vega 64 all perform around the same at 113 FPS. Our 2070 hits 137 FPS a good 23 FPS above Vega 64.

Far Cry 5 and The Division

Far Cry results show the RTX 2060 reaching 107 FPS with the 1070 Ti and Vega 64 beating it out (108 and 117 FPS respectively). The 2070 steps up and hits 126 FPS here.

In The Division, both NVIDIA GPUs get smoked here with 2060/2070 hitting 91/107 FPS and the 1070 Ti and Vega 64 reaching 100/115 FPS. Is this an AMD title?

Shadow of the Tomb Raider and Final Fantasy XV

Moving on to Shadow of the Tomb Raider, here again, the 1070 Ti and Vega 64 have a good showing versus the 2060 (88 FPS) and the 2070 (104 FPS) where the two non-RTX cards perform on par with the 2060.

In our FF XV testing (which will be dropped moving forward and replaced with BF V) the RTX cards come back out on top with the 2060 reaching 81 FPS and 2070 94 FPS. The 1070 Ti matched the 2060 here while Vega 64 was well behind.

Ashes of the Singularity: Escalations

For AOTSe, the 2060 hit 54 FPS and the 2070 62 FPS. The 1070 Ti makes a strong showing here at 64 FPS as does the Vega 64 at 56 FPS. This is a fairly CPU heavy game so on the low side we are not seeing much difference at 1080p.

BF V Results

This is the first review in which we have added Battlefield V (BF V) and Ray Tracing results. For this review, we tested all capable RTX cards at three different resolutions (FHD, WQHD, UHD) with the RT settings set to Ultra. We used the single player campaign name Tirailleur. We start FPS recording as soon as we see the soldiers in the truck as it is on rails and has some ray tracing in that initial scene. Once out of the truck we manually make our way up the middle/left side taking out enemies until we reach the top of the hill and move on to the next checkpoint. There is plenty of ray tracing in puddles and flames throughout as well as a lot of gunplay.

BF V – 1080p

BF V – 2560×1440

BF V – 4K UHD

Starting with 1080p, we can see that the full range of RTX cards is able to play the game at 1920×1080 resolution with graphics set to ultra as well as ray tracing set on ultra. The RTX 2060 dips to an average of 52 FPS here but is still plenty playable. A couple of quick image quality adjustments and users can reach that magic 60 FPS number. One thing we can walk away from with all of this testing is RT really cuts the FPS down by around half, give or take.

Moving up to 2560×1440, the FPS drop considerably here with the only cards ‘playable’ using ultra RT (close to 60 FPS) being the RTX 2080 and 2080 Ti at 53 and 64 FPS respectively. The 2070 and 2060 simply do not have enough RT horsepower to push adequate FPS coming in at 46 FPS and 35 FPS.

Last, moving up to 4K UHD, both the 2080 and 2080 Ti can push this title past 60 FPS without RT enabled with the 2070 hitting 55 FPS and 2060 an admirable 40 FPS. After enabling RT, the 2080 Ti manages about 40 FPS and at these settings are not playable with the game experience being quite choppy and slow. Changing the RT settings to low, the FPS bounce back up close to 60 FPS and is playable once again.

Overall, we like what RT adds to BF V though certainly at times noticing the effects will be difficult due to fast action. But between battles, it is a noticeable difference for the better. We should see a lot more titles coming out this year that support both DLSS and RT capabilities. The sooner these come out and better they are implemented will play a role in how this technology gets applied moving forward. Here is to hoping AMD gets involved in some form as well to push this forward.

2560×1440 and 4K UHD Results

Below are the higher resolution results starting with 2560×1440 and the gaining in popularity 3840×2160 (4K UHD). These resolutions are more fitting for the cards we are testing as the 1080p results with these cards can have a ceiling on them from the CPU (even at 4.7 GHz).

2560 x 1440 Results

4K UHD Results

Looking at our 2560×1440 results, we can see that in most titles, both the 2060 and 2070 can reach that magical 60 FPS value people strive for. Its only in AOTSe where it is considerably below 60 FPS. That said, that title works well with lower FPS and isn’t a worry.

Moving on up to 4K UHD resolution, the 2070 holds its own in some titles notably WoTe and F1 2018 but still falls below 60 FPS at our settings. Lowering IQ will be a must for 4K UHD gaming with the 2070. The 2060 is not a 4K UHD card at these settings.

Overall the RTX 2070 is a very capable 2560×1440 card while the RTX 2060 will play many titles here, it is more of a 1080p card if you want Ultra settings and 60 FPS.

Overclocking

When overclocking these cards, we followed the same method of using the OC scanner to find a ‘stable’ core speed and then push up from there. In the case of the RTX 2060, we were able to push the core up to +141 which yielded core clocks around 2025 MHz. The Samsung memory (not Micron!) on the 2060 hit 2000 MHz from its 1750 which is about the average. The RTX 2070 was able to push its clocks to a +166 which yielded a 2010 MHz on the core. The Micron memory on this card also hit 2000 MHz before it started to artifact and act wonky.

Overall we see a few % increases in games and benchmarks when using the OC scanner values and 2000 MHz on the GDDR6.

RTX 2060 FE – PTL

 

RTX 2070 FE – PTL

Temperatures and Power Use

We test power consumption by running through the game benchmarks of Shadow of the Tomb Raider and F1 2018 at both stock speeds, and while overclocked. We monitor temperatures throughout this testing with the peak temperature what is listed in the data below. In order to more accurately simulate real gaming conditions, the benchmarks are extended (time) to allow the card to settle/saturate.

RTX 2060 FE – Temperatures

RTX 2070 FE – Temperatures

Temperatures on both cards landed in the acceptable range with the RTX 2060 peaking at 74 °C overclocked while the 2070 managed to reach 76 °C. The fans do get audible when running at those temperatures, however, it is not an intrusive sound.

RTX 2060 FE – Power

RTX 2070 FE – Power

Power use on these cards was fairly minimal for the performance. The RTX 2060 FE with a 160 W TDP sipped on power to the tune of 323 W peak when overclocked and 304 W peak when at stock speeds. Average use for these titles is around 30 W+ less. The RTX 2070 FE and its 180 W TDP peaked at 339 W.

Conclusion

The NVIDIA RTX 2070 and RTX 2060, for now, fills out their lineup from top to bottom and hopes to offer users an ‘affordable’ entry path into the latest Turing architecture. With pricing set at $599 and $349 respectively for the FE cards, they should be able to get into more hands, especially the RTX 2060 which, according to statistics from TPU, is the only Turing based GPU that has improved on its price to performance ratio.

Performance on both cards was solid performing as good or better than their last generation counterparts with ease and also beating out the next step up (i.e.. RTX 2070 beats GTX 1080). So the performance is certainly there, but the price increase in like classed cards from generation to generation can really be a kick in the pants. We are used to seeing improvements and similar pricing or small price increases, but Turing took it to a different level… except for the RTX 2060. If only we had some competition to drive the pricing down! The RTX 2060’s performance is in the ballpark of Vega 64 with around half the power use and similar price points.

So where do these land? The price, especially on the higher-end parts really wasn’t palatable for many users who were spoiled with generational gains and minimal price increases. However, this is new technology which performs better and adds new features that, on paper, are pretty exciting with ray tracing and DLSS. However, we currently only have one title that uses RT (BF V) and DLSS is slowly creeping its way into more titles.

The RTX 2060 gets hands down approved status since it was the only card able to improve upon its price to the performance metric while still offering improved performance as well as RT and DLSS capabilities (at 1080p). The RTX 2070 also gets approved, but it is a bit more tempered due to pricing.

Click to find out what this means

Joe Shields (Earthdog)

Related Articles:

Leave a Reply

Your email address will not be published. Required fields are marked *

Discussion
  1. What do you think is going on with the memory - both Samsung and Micron only going to ~2000, hardware cap/already pushed to the limit/architecture ?
    Microns go up to ~2000, Samsung up to ~2100. My RTX2060 with Microns can make exactly 2000MHz stable ... up to 2100+ in some benchmarks.
    In most reviews are not presented real max clocks.