Erotic RTX2060 graphics card

[IanJ]

Went fishing for a new graphics card for my Dev PC...…..found the new RTX2060 at Overclockers.co.uk......Looks like a real erotic card!

What's wrong with this spec (copy n paste direct from their site):-

Specification:
- GPU: RTX 206
- Core name: TU106
- Process: 12nm
- Cuda Cores: 1920
- Memory capacity: 6GB
- Memory interface: 192-bit
- Memory type: GDDR6
- Base clock: 1365MHz
- Boost clock: 1830MHz
- Memory speed: 14,000MHz
- Memory bandwidth: 336Gb/sex
- RTX-Ops: 39
- TDP: 190W
- Dimensions: 235 x 112 x 40 mm
- Bracket: 2-slot
- Cooler: 2-slot
- Warranty: 2yr

Ian.

[Ampera]

RTX 2070 is probably going to be my next upgrade, to replace my GTX 970, but the RTX cards are pretty impressive, even beyond the iffy raycasting people berate them for. Personally I think raycasting is the future, it's a definitively superior 3D rendering method, but someone has to put the first step somewhere.

[wraper]

RTX 2070 is probably going to be my next upgrade, to replace my GTX 970, but the RTX cards are pretty impressive, even beyond the iffy raycasting people berate them for. Personally I think raycasting is the future, it's a definitively superior 3D rendering method, but someone has to put the first step somewhere.

Compared with previous generation not that impressive, especially considering their price. Raytracing is basically the only thing they have to brag about.
For example GTX 1080Ti is basically on par with non Ti RTX 2080.

[Ampera]

My numbers showed 2070 was on par with the 1080Ti, but that was passmark, so grain of salt I guess.

It still beats the shit out of my 970, and anything that AMD makes. I note your chart doesn't have a 2070 as well, which could very well be an important budget contender. There are a lot of factors to graphics cards' worth. Prices are never static, and while Passmark might be good for ballpark, it's not a real world example. Even within real world tests, the 1080Ti has more memory than the 2070/2080, but the memory on the 2070/2080 is faster and I believe also has considerably higher bandwidth. Tasks at which the cards will perform better or worse will definitely vary.

There's also the question of how much can you get one new, RTX or GTX, because even though I'd be alright in getting some hardware used, used GPUs seems like a game of russian roulette. It may be 50 chamber russian roulette, but still a gamble regardless, with how many people there are on ebay (and especially craigslist) who are out to steal your money.

All of these numbers can literally change day by day, task by task, person by person. I like the 2070 because it has what I would consider to be good performance, significantly above what I have, a good value (1080Tis are still clocking in over 1000 USD on Newegg), and has the neat hardware raytracing, which while I believe to be the future, has gimmick-class implementation. Would be neat to mess around with, but I'm not sure I'd use it all the time (on any games that even support it).

[wraper]

2070 is on par with non Ti 1080 and they cost about the same when new. But you can save a lot buying used. In the nutshell max performance of new series have gone a half step up in lineup over previous series. Only RTX 2080 Ti beats top end of previous generation.

[Ampera]

2070 is on par with non Ti 1080 and they cost about the same when new. But you can save a lot buying used. In the nutshell max performance of new series have gone a half step up in lineup over previous series. Only RTX 2080 Ti beats top end of previous generation.

I question your numbers, it's not that cut and dry, because the technology and memory architecture is very different to the 10 series. Memory is much faster with more bandwidth than the 1080 (as I said). Power draw and limits could be different, more room for overclocking and the effectiveness of boost clocks could be different. Different tasks may work significantly better on a 2070 than a 1080. There's also risk and lack of warranty on used cards, which can be a significant issue, since parts /do/ die, depending on how they were treated.

[wraper]

I question your numbers

Game benchmarks, 2070 is  around 5% faster than 1080 in average. Synthetic benchmarks do not represent real life performance.

[Ampera]

I question your numbers

Game benchmarks, 2070 is  around 5% faster than 1080 in average. Synthetic benchmarks do not represent real life performance.

If you actually read what I wrote, you would know that's not what I meant. I also said that passmark was a ballpark measure. What I mean is that I question how cut and dry you make it out. There are so many thing you can use a graphics card for, and with how different the two cards, even though it's just a new generation, there will be gaps closer than others. That's why I question your numbers.