Nvidia GeForce RTX 2060 Review & Mega Benchmark

[pcnthuziast]

Nvidia's business practices are nearly criminal. That is all.

 

[Capaill]

Well done Nvidia. You priced yourself out of the market by including Ray Tracing. So you had to introduce this card at a massively reduced price (compared to the 20xx series) to get back in the game.
Nvidia won't truly return to common sense land until they release a series with no RT.

 

[Nightfire]

Techpowerup recently retested a lot of GPU's with the latest Nvidia and AMD driver. View the full review here:
https://www.techpowerup.com/reviews/Palit/GeForce_RTX_2070_GameRock_Premium/

Techspot claims to have done the same in their latest RTX 2060 review. But, the statement "We used AMD's Adrenalin 2019 Edition 19.1.1 drivers for the Radeon GPUs and Game Ready 417.35 WHQL for the GeForce GPUs." is simply not true.

This is how I noticed that the latest benchmarks are fake.
Battlefield V performance review:
https://www.techspot.com/review/1746-battlefield-5-gpu-performance/
Battlefield V performance in latest RTX 2060 review:
https://www.techspot.com/review/1781-geforce-rtx-2060-mega-benchmark/

In the TPU reviews, ( Original Battlefield V performance review vs RTX 2070 Gamerock review ) you clearly see the performance improvements done by AMD in the latest drivers in Battlefield V while on techpost.com all AMD benchmarks show the exact same results as on driver 18.11.1 from two months ago. Not only that, Nvidia has also improved performance going form driver 416.81 to driver to driver 417.35 , clearly visible when you compare the original Battlefield V performance review on TPU with the latest RTX 2070 review, but on techspot.com the Nvidia performance on 1440p stays exactly the same on all cards, both the minimum and avarage fps. Simply not possible with the driver improvements form the last couple of months.

And in case you might think that the charts are dynamic and updated with new results, have a look at the YouTube video with these charts:
https://www.youtube.com/watch?v=FUhCf6t3RBA

Clear evidence that they have not retested AMD cards on 19.1.1 and they have not retested Nvidia cards on 417.35.

True, but I didn't see a difference at TPU, at least on GTAV.

 

[Adi6293]

" Are today’s GPUs more complex than they were a few years ago? Have manufacturing costs increased? We know the cost to build a modern fab is astronomical, and yields for these big complex GPUs aren’t great. Also, is more investment being made into research and development? The answer to that one is unquestionably yes."

If that is the explanation, why is nVidia having record profits in gaming?

Mining....

I have to disagree hete. nVidias revenue has been going up every year in past few years. They have 63% margins. Yes it might cost them more to make those cards but they chose to go ahead with RT so they should take a hit not gamers.

 

[Adi6293]

Excellent review! Its good to see that my 1080 is still a good card at 1440p and hopefully I can hold off until another generation to do a complete rebuild. Still rocking my 3770k from 2013!

the 3770 is def holding you back. I moved from it to Ryzen1600 (Massive boost in -.1% lows) im running 1070Ti

But do you play at 1440p? I really dont think that overclocked i7 3770K would hold back a GTX1080 at 2560x1440 unless he kept it stock then maybe a little

 

[Steve]

Techpowerup recently retested a lot of GPU's with the latest Nvidia and AMD driver. View the full review here:
https://www.techpowerup.com/reviews/Palit/GeForce_RTX_2070_GameRock_Premium/

Techspot claims to have done the same in their latest RTX 2060 review. But, the statement "We used AMD's Adrenalin 2019 Edition 19.1.1 drivers for the Radeon GPUs and Game Ready 417.35 WHQL for the GeForce GPUs." is simply not true.

This is how I noticed that the latest benchmarks are fake.
Battlefield V performance review:
https://www.techspot.com/review/1746-battlefield-5-gpu-performance/
Battlefield V performance in latest RTX 2060 review:
https://www.techspot.com/review/1781-geforce-rtx-2060-mega-benchmark/

In the TPU reviews, ( Original Battlefield V performance review vs RTX 2070 Gamerock review ) you clearly see the performance improvements done by AMD in the latest drivers in Battlefield V while on techpost.com all AMD benchmarks show the exact same results as on driver 18.11.1 from two months ago. Not only that, Nvidia has also improved performance going form driver 416.81 to driver to driver 417.35 , clearly visible when you compare the original Battlefield V performance review on TPU with the latest RTX 2070 review, but on techspot.com the Nvidia performance on 1440p stays exactly the same on all cards, both the minimum and avarage fps. Simply not possible with the driver improvements form the last couple of months.

And in case you might think that the charts are dynamic and updated with new results, have a look at the YouTube video with these charts:

Clear evidence that they have not retested AMD cards on 19.1.1 and they have not retested Nvidia cards on 417.35.

Going over my notes the only game where I saw note worthy gains was Fortnite. The BFV results look strong for AMD and I just went a re-checked, no more than 1-2 fps (margin of error) difference on the retest for the RX 580.

That said I'll happily check other maps and areas of the game when I get a chance.

 

[Peter Farkas]

I just picked up an ex-miner RX 580 for $150 which is a bloody good deal where I live. I see no reason upgrading at the moment with these prices...
Maybe in a year.
Great review from the Benchmark King! Thanks Steve!

 

[noel24]

Conclusively faster than Vega 56 on immature drivers, 100 watts less power draw under load.

That's an astronomical gulf in performance per watt. Navi better be something special.

Not exactly. It's 7 percent faster on mature drivers (this is 2070/80/80Ti driver after all, It won't get much better), while simple Vega shrink to 7nm would bring either 25% performance boost for the same TDP, or 50% reduction in TDP at the same performance, or could simply be somewhere in between. Add that Vega 56 got 8GB of VRAM, so is more futureproof. And when It's available, It can cost under $350 on sale. The same goes for RX580: die-shrink it, add 25% more of those Compute Units, and maybe support for GDDR6 and You get a $250-300 card that will equal 2060/1660 with 6GB of VRAM. Navi must just be priced right and it will take a big chunk of nVidias under-$300 market. Have a look what is the most popular card on Steam survey. the problem is, AMD is underdog and They don't price their cards right, They price them against nVidia to squizzz out as much as possible.

 

[Nightfire]

Conclusively faster than Vega 56 on immature drivers, 100 watts less power draw under load.

That's an astronomical gulf in performance per watt. Navi better be something special.

Not exactly. It's 7 percent faster on mature drivers (this is 2070/80/80Ti driver after all, It won't get much better), while simple Vega shrink to 7nm would bring either 25% performance boost for the same TDP, or 50% reduction in TDP at the same performance, or could simply be somewhere in between. Add that Vega 56 got 8GB of VRAM, so is more futureproof. And when It's available, It can cost under $350 on sale. The same goes for RX580: die-shrink it, add 25% more of those Compute Units, and maybe support for GDDR6 and You get a $250-300 card that will equal 2060/1660 with 6GB of VRAM. Navi must just be priced right and it will take a big chunk of nVidias under-$300 market. Have a look what is the most popular card on Steam survey. the problem is, AMD is underdog and They don't price their cards right, They price them against nVidia to squizzz out as much as possible.

Agreed, but I think they would be better off with 8 GB of GDDR5X in 256 bit instead of 6 GB of pricier GDDR6 in 192 bit. Bandwidth and cost should be close, and you get a more future-proof card.

 

[OcelotRex]

Techpowerup recently retested a lot of GPU's with the latest Nvidia and AMD driver. View the full review here:
https://www.techpowerup.com/reviews/Palit/GeForce_RTX_2070_GameRock_Premium/

Techspot claims to have done the same in their latest RTX 2060 review. But, the statement "We used AMD's Adrenalin 2019 Edition 19.1.1 drivers for the Radeon GPUs and Game Ready 417.35 WHQL for the GeForce GPUs." is simply not true.

This is how I noticed that the latest benchmarks are fake.
Battlefield V performance review:
https://www.techspot.com/review/1746-battlefield-5-gpu-performance/
Battlefield V performance in latest RTX 2060 review:
https://www.techspot.com/review/1781-geforce-rtx-2060-mega-benchmark/

In the TPU reviews, ( Original Battlefield V performance review vs RTX 2070 Gamerock review ) you clearly see the performance improvements done by AMD in the latest drivers in Battlefield V while on techpost.com all AMD benchmarks show the exact same results as on driver 18.11.1 from two months ago. Not only that, Nvidia has also improved performance going form driver 416.81 to driver to driver 417.35 , clearly visible when you compare the original Battlefield V performance review on TPU with the latest RTX 2070 review, but on techspot.com the Nvidia performance on 1440p stays exactly the same on all cards, both the minimum and avarage fps. Simply not possible with the driver improvements form the last couple of months.

And in case you might think that the charts are dynamic and updated with new results, have a look at the YouTube video with these charts:

Clear evidence that they have not retested AMD cards on 19.1.1 and they have not retested Nvidia cards on 417.35.

Going over my notes the only game where I saw note worthy gains was Fortnite. The BFV results look strong for AMD and I just went a re-checked, no more than 1-2 fps (margin of error) difference on the retest for the RX 580.

That said I'll happily check other maps and areas of the game when I get a chance.

That's a level of civility lacking in the OP's comment. Always best to give the benefit of the doubt in these situations, Belgium...

 

[amghwk]

I'm not sure which is more depressing:

#1 The feeling that this generation of cards is woefully inept to replace the power of the last generation

or

#2 The games that are being used to test these cards are so boring.

It's more like the former. Though the second point is also true.

No thanks to current gen gamers, companies are more than happy to churn out more of the same BS and POS like pubg or battle royale or whatever it's called.

 

[darkzelda]

Great review as always, this is the best RTX gpu, and it's not even compelling.

 

[BigBoomBoom]

" Are today’s GPUs more complex than they were a few years ago? Have manufacturing costs increased? We know the cost to build a modern fab is astronomical, and yields for these big complex GPUs aren’t great. Also, is more investment being made into research and development? The answer to that one is unquestionably yes."

If that is the explanation, why is nVidia having record profits in gaming?

They aren't using top of the line, monopolistic 7nm fabs, to compete with AMD usage of 7nm fabs. 12nm is just a refined 16nm, not an optical shrink. They also keep pushing prices because not only AMD isn't competing at top end, AMD is also not interested in competing in term of price. RX Vega VII same price same performance as GTX 1080 Ti /RTX 2080, AMD is clearly isn't interested in lowering price of the GPU market.

 

[Vulcanproject]

They aren't using top of the line, monopolistic 7nm fabs, to compete with AMD usage of 7nm fabs. 12nm is just a refined 16nm, not an optical shrink. They also keep pushing prices because not only AMD isn't competing at top end, AMD is also not interested in competing in term of price. RX Vega VII same price same performance as GTX 1080 Ti /RTX 2080, AMD is clearly isn't interested in lowering price of the GPU market.

It seems many people here are failing to realise this. Not least those who can't see the ridiculous architecture advantage Nvidia have. Denying the enormous performance per watt disparity!

Yes, AMD will shrink Vega to 7nm, but Nvidia can do exactly the same thing with Turing. The difference is they don't actually need to unless AMD deliver with Navi.

As soon as AMD turn up with something worthwhile Nvidia can then invoke the 7nm node to maintain the gap. It's a desperate situation at the moment for the graphics industry.

Perhaps only Intel can save it and make the biggest difference in the next couple of years. Bring some real competition.

 

[NightAntilli]

They aren't using top of the line, monopolistic 7nm fabs, to compete with AMD usage of 7nm fabs. 12nm is just a refined 16nm, not an optical shrink. They also keep pushing prices because not only AMD isn't competing at top end, AMD is also not interested in competing in term of price. RX Vega VII same price same performance as GTX 1080 Ti /RTX 2080, AMD is clearly isn't interested in lowering price of the GPU market.

Yeah I also wouldn't be interested in offering competing products, only for everyone to go out and buy my competition when they drop prices.

 

[Aksae]

My 1060 runs Just Cause 4 and Battlefield V at 4k 30fps

 

[Colonel Blimp]

I don't think 6gb of ram will age well and the RTX portion is a complete waste of time and extra money, see https://www.hardocp.com/article/2019/01/20/battlefield_v_nvidia_ray_tracing_rtx_2060_performance/

for why.

If I saw a Vega56 on offer I'd rather have that, or wait for Navi to replace my 280x.

 

[Colonel Blimp]

My 1060 runs Just Cause 4 and Battlefield V at 4k 30fps

So you bought BFV to not play multiplayer then?

 

[cldmstrsn]

Excellent review! Its good to see that my 1080 is still a good card at 1440p and hopefully I can hold off until another generation to do a complete rebuild. Still rocking my 3770k from 2013!

the 3770 is def holding you back. I moved from it to Ryzen1600 (Massive boost in -.1% lows) im running 1070Ti

meh its really not. I have it OC to 4.2 and it still runs everything I want on Ultra at 60 frames so until that doesnt happen I will wait another year or so till I do a full upgrade.

 

[JRMBelgium]

That's a level of civility lacking in the OP's comment. Always best to give the benefit of the doubt in these situations, Belgium...

He already posted on reddit that he didn't run all the cards with the latest drivers. But I still don't see an edited article. So what doubt is there really?

 

[OcelotRex]

He already posted on reddit that he didn't run all the cards with the latest drivers. But I still don't see an edited article. So what doubt is there really?

Steve's reply: "That said I'll happily check other maps and areas of the game when I get a chance." Emphasis mine.

It has been a less than a day since an editor here has offered to retest benchmarks. Having come to TechSpot for years I am willing to give the benefit of the doubt that if discrepancies arise from the retest the article will be updated and a response will be placed in the thread. I think most long-time readers would extend that as well. That said it will take some time to complete the benchmarks (as it took to complete the review since Nvidia chose not to provide them with a sample to test.)

Flat-out calling someone's hard work "fake" is not very civil nor is it productive.

 

[poohbear]

Well, my question is if 2060, replaces 1070 (from My point of view - mostly due to currency exchange fluctuations here in Central Europe), and two years ago wasn't enough of an upgrade over GTX 970, and Navi is nowhere to be seen, would 2060 make sense to be a viable upgrade for at least next 3 years? My main concern is only 6 GB of VRAM, which today seems OK, but what if next gen consoles introduce more demanding textures. Did someone test the above selection of games with some modded HD texture packs?
Generally, I think I'll wait for Navi mid-range, better yet for the next gen consoles to see what they pack.

I agree you should wait. When AMD releases their product, Nvidia will slash prices to compete. Thank you competition!!!

Personally, I'll just stick with my GTX 1070. Love my 8gb of RAM and I bought it cheap used. Plays everything with maxed settiongs @ 1440p (I dont use any AA though).

 

[Nightfire]

I don't think 6gb of ram will age well and the RTX portion is a complete waste of time and extra money, see https://www.hardocp.com/article/2019/01/20/battlefield_v_nvidia_ray_tracing_rtx_2060_performance/

for why.

If I saw a Vega56 on offer I'd rather have that, or wait for Navi to replace my 280x.

A version of this card with no RTX or tensor cores would be perfect if it was $100 cheaper. 6 GB is enough if you are not running BFV on ultra dx12 like your link has.

 

[Colonel Blimp]

I don't think 6gb of ram will age well and the RTX portion is a complete waste of time and extra money, see https://www.hardocp.com/article/2019/01/20/battlefield_v_nvidia_ray_tracing_rtx_2060_performance/

for why.

If I saw a Vega56 on offer I'd rather have that, or wait for Navi to replace my 280x.

A version of this card with no RTX or tensor cores would be perfect if it was $100 cheaper. 6 GB is enough if you are not running BFV on ultra dx12 like your link has.

Agreed, RTX on this card is pointless therefore you're paying that much extra for something that is unusable on this card.
Maybe other games will release with better implementation but by the time they do, Navi will be here and we probably won't be far off RTX Version 2.0 from nVidia.

 

[Sausagemeat]

I think this is a really good card for the money. In the U.K. stock Vega 56 coats around the same as an aftermarket 2060 and the 2060 is clearly faster, uses less power, has tensor cores etc. And there doesn’t seem to be anything else comparable. Good to see at least one company pushing new tech in the midrange at this point.