Welcome to Bankruptcy

This is an expensive piece of hardware. Yes, That’s the line I’m going to start this review off with. It’s the elephant in the room and I won’t bury the lead. This card is fast but it’s also one of the cheaper 2080 Ti’s that exist. Coming in at $1,149.99 straight from EVGA at the time of me writing this it is double the price of the previous generation 1080 Ti. That makes upgrading from the previous gaming flagship (ignoring Titan cards) a tricky proposition. Before that was the 980 Ti which came in at a MSRP of $649. We’ve gone from $649 to $699 to $1200 and that is hard to ignore.

But is it ‘High-End’?

I previously wrote an article calling Nvidia out on their neglect of the high-end market with the announcement and launch of Turing. I mostly don’t feel that way anymore since I wrote that article. Now that I’ve had time to finally play with the raytracing technology in games and in the Unreal Engine 4, I am thoroughly impressed. Beyond those new visual features, I am also impressed by the performance gain from the 1080 Ti. This is very much a high-end product SKU but is it worth the price?

Testing

Specs

The test bench for today will be my personal machine.

  • Intel Core i7 8700k @ 5GHz Core 4.4GHz cache
  • Gigabyte Aorus 370z Gaming 7
  • 32GB Corsair DDR4-2800
  • 480GB SanDisk Ultra SSD Boot Drive
  • 960GB PNY CS1311 SSD Data Drive
  • SeaSonic Prime Titanium 850W

The GPUs used for comparison will be

  • Gigabyte 980 Ti Xtreme Gaming
  • EVGA 1080 Ti FTW3
  • EVGA 2080 Ti XC Black

For some of the tests the CPU will be different due to historical reasons.

I do not have complete data for every benchmark for each of these cards, but I’ll do my best to give a fair comparison.

Synthetic Benchmarks

I am a firm believer in using synthetic benchmarks for assessing performance of computer hardware. Gaming benchmarks, while being more relevant to gamers in general who would be using the hardware, are subject to updates that you cannot roll back and other strangeness that are specific to how you play the game in question. It is valuable to learn how to extrapolate game performance from synthetic benchmarks as a reader of reviews. There isn’t some secret understanding going on in the hardware testing world and the SKUs used today are very logical and once you understand where each model stands on a tiered list of performance you should be able to essentially do away with needing to even read reviews and be able to make your own mind up on if a product is fast enough for what you need.

3DMark FireStrike

AIDA64 GPGPU

Memory

FLOPS

IOPS

AES-256

SHA-1 Hash

Single-Precision Julia

Double-Precision Mandel

Subjective Gaming Performance

I’ve taken the time out of my hectic existence to put some hours into playing RTX enabled games to give you my opinion of how they felt. I played BF:V multiplayer with RTX on and off in 1080p using the driver version 418.91 and 417.75. In BF:V framerate sits around 100-140fps with RTX enabled and ‘ultra’ settings. You could easily lock 120fps by setting ‘RTX Reflection Quality’ to ‘Medium’. I never felt like I was given an advantage by having the RTX Reflections enabled nor did I feel at a disadvantage from any performance hit.

The Concept of Headroom

Having more compute power in your graphics hardware then you need affords you the ability to do some interesting things. While you could run every game with unlimited frame rate and stress every component constantly you could also take another approach and use vsync or an adaptive sync technology to limit things down. Doing this in more laid-back games has very minimal effect on responsiveness but it also lets your system breathe a little. It should be obvious that your headroom potential decreases as you move to lower end parts. This is something to think about when the time comes to pick out a graphics card.

Bugs & Issues

Being an early adopter of a technology, let alone a new GPU architecture, comes with some caveats. There are still some teething issues going on with the drivers from Nvidia for the Turing series. Some very popular games may crash more frequently, or you may experience a regression in general stability overall. As of me writing this article there is a bug with the 418 drivers where iRay rendering crashes and falls back to using the CPU for rendering. The solution is to go back to the 417.75 driver or earlier. There are also issues related to the Nvidia services that are installed with the drivers where you may experience abnormal background CPU usage. There are modified drivers available on Guru3D forums which strip out GeForce Experience and other things which seems to fix the CPU usage bug at least.

Summary

Overall, I think this is a suitable update to the 1080 Ti, in fact it’s the only update that is worth doing if you own the 1080 Ti. I’ve supplied a handy table to assist you in your decision process. Note my use of the word upgrade in the table which is to describe the process of obtaining a graphics card that will provide you with better performance then you currently have.

Current Graphics Card Potential Upgrade Card
GTX 1080 Ti RTX 2080 Ti
GTX 1080 RTX 2080
GTX 1070 Ti RTX 2070 / 2080
GTX 1070 RTX 2070
GTX 1060 RTX 2060

If this were a sidegrade table, it would look more like this. A sidegrade is where you purchase a part which has equal performance but is newer or better in other ways.

Current Graphics Card Potential Sidegrade Card
GTX 1080 Ti RTX 2080
GTX 1080 RTX 2070
GTX 1070 Ti RTX 2060 / 2070
GTX 1070 RTX 2060
GTX 1060 N/A

I’m going to end my review here for now. I like to live with a product for a while so I can develop a deeper opinion of it. I’ll be taking a look at streaming performance, more games, video encoding and 3D rendering in other future articles. Thank you for taking the time out of your day to read this and I hope you have a wonderful rest of your day!