The Crew Benchmarked, Performance Review

[Julio Franco]

Set in a huge open world that allows players to race their way across the United States, Ubisoft's new action-driving MMO has 7,000 miles of roads, 15 cities and 15 million individual objects. Depending on your vehicle and skills behind the wheel, it's said to take about an hour and a half to drive coast to coast in the game and four hours to circumnavigate the entire map.

The Crew is primarily an online multiplayer game, but it does have a single player component that offers roughly 20 hours of gameplay. The game has been designed for the latest generation consoles and PCs, though the former are capped at 30fps frames and a resolution of 1080p while our preferred platform ought to look and feel better with a 60fps cap and a higher res.

The Crew has been built using the Babel game engine along with developer Ivory Tower's proprietary tools. Considering the fact that we haven't tested a title powered by this combo before, we don't know how well optimized the game will be. For what it's worth, Ubisoft has had a poor track record in this area lately so we're interested to see how The Crew performs.

Read the complete review.

 

[Jad Chaar]

Poorly optimized game if it performing that badly at 1080P... Dang I am losing respect for Ubisoft by the day xD. Just curious, why did you guys throw in the 1366x768 resolution suddenly?

 

[Steve]

Poorly optimized game if it performing that badly at 1080P... Dang I am losing respect for Ubisoft by the day xD. Just curious, why did you guys throw in the 1366x768 resolution suddenly?

It is a demanding game on GPUs so we went with the lower res.

I am not sure that it is fair to say this game is poorly optimized. Considering the massive open world and impressive graphics the performance seems about right.

 

[Guest]

Poorly optimized game if it performing that badly at 1080P... Dang I am losing respect for Ubisoft by the day xD.

Yeah, guise! There is no such thing as demanding software, everything that performs slow is because it's poorly optimized. The correct way if for my mid-range GPU to always perform extraordinarily well, regardless of how many visual effects the game has or how complex its geometry is.

Just look at those dumb universities and research centers that build supercomputers to perform very poorly optimized tasks. I'm pretty sure protein folding research and climate simulations could be done just as quickly on my Core i5 laptop if they optimized it properly.

 

[Guest]

Poorly optimized?!? You clearly haven't played the game. This is to date the best launched game for pc that I have ever played. Absolutely buttery smooth in every aspect. 1080p 60 fps on a fx6300 and two 760s. I can't maintain 60 in far cry 4 with the same settings.

 

[MrBlkfx1]

If "insert my card" isn't performing well. The game is poorly optimized. You just can't please some people. Anyways, good review as always guys.

 

[EEatGDL]

I think if you base poorly optimized by using 4xMSAA, then you have no idea. It seems fine to me in both image quality and performance. 1080p with 4xMSAA is like rendering 4K internally, so take those FPS as if "running at 4K without AA".

 

[GhostRyder]

Cool review @Steve, I am actually shocked by some of the results as I was personally not expecting it to be that demanding on the GPU side especially at 1440p (Albeit, it is a giant open world with nice graphics so I am not sure what I was expecting lol). The CPU side is the most shocking, it looks to me like its more optimized for multi-cores to a point where its happy with a very wide range of cpus. I am glad to see that as it opens everyones CPU up to this game without complaining.

I actually may want to play this MMO, I have been kicking it around as I personally am not a big MMO fan but this one sounds pretty sweet.

 

[Guest]

I think if you base poorly optimized by using 4xMSAA, then you have no idea. It seems fine to me in both image quality and performance. 1080p with 4xMSAA is like rendering 4K internally, so take those FPS as if "running at 4K without AA".

Actually, it's not like rendering in 4K.
What you're thinking of is SSAA (super-sampling). MSAA is an optimization of SSAA where multiple subpixel samples are taken only on the pixels at the edges of triangles, not on every pixel on the screen. Also, it doesn't sample texture and some shader effects multiple times, since they are not prone to aliasing like geometry is (textures are, but in a different way, which is handled by AF). That means lower bandwidth and fillrate requirements, which makes 1080p with 4x MSAA a lot faster than 1080p with 4x SSAA or 4K.
Also, just for the record, 1080p with 4x SSAA and 4K aren't the same either, because anti-aliasing uses rotated grids or random locations for taking samples inside each pixel. 1080p with 4x SSAA doesn't result in a 3840x2160 grid.

 

[Cryio]

Considering the massive open world and impressive graphics the performance seems about right.

Doesn't the game look just ... ok-ish? It's certainly not on the level of GRID Autosport, NFS Most Wanted 2012 or Rivals. Not even talking about Forza 5 or Horizon 2.

 

[Cryio]

Two questions:

1) Why did you use SSAO? HDAO+ isn't tied to Nvidia hardware. I didn't heard of anything regarding Crew's HDAO+ being exclusive to Nvidia hardware ... so ?

2) How much of a performance impact is 4xMSAA? If would much prefer playing with HDA0+ or a better resolution and FXAA/MLAA than 4xMSAA. I'm more than willing to trade a few jaggies for an overall better picture on in case of resolution, crispness.

 

[Steve]

Two questions:

1) Why did you use SSAO? HDAO+ isn't tied to Nvidia hardware. I didn't heard of anything regarding Crew's HDAO+ being exclusive to Nvidia hardware ... so ?

2) How much of a performance impact is 4xMSAA? If would much prefer playing with HDA0+ or a better resolution and FXAA/MLAA than 4xMSAA. I'm more than willing to trade a few jaggies for an overall better picture on in case of resolution, crispness.

1) You cannot select HDAO on AMD hardware in the game, not sure why but you can't. That left us with SSAO if we wanted to test the same level of ambient occlusion on both AMD and Nvidia cards.

2) 4xMSAA is a big hit but it looks very nice. Those with lower end hardware are better of with FXAA for sure. We generally test with the highest quality settings or near enough.

 

[MasterElwood]

Two questions:

1) Why did you use SSAO? HDAO+ isn't tied to Nvidia hardware. I didn't heard of anything regarding Crew's HDAO+ being exclusive to Nvidia hardware ... so ?

2) How much of a performance impact is 4xMSAA? If would much prefer playing with HDA0+ or a better resolution and FXAA/MLAA than 4xMSAA. I'm more than willing to trade a few jaggies for an overall better picture on in case of resolution, crispness.

1) You cannot select HDAO on AMD hardware in the game, not sure why but you can't. That left us with SSAO if we wanted to test the same level of ambient occlusion on both AMD and Nvidia cards.

2) 4xMSAA is a big hit but it looks very nice. Those with lower end hardware are better of with FXAA for sure. We generally test with the highest quality settings or near enough.

Sorry Steve, but you did the benchmarks completly wrong - and here is why:

I have a GTX970. With 4xMSAA I can drive the whole day and run constantly in the 60fps limit. BUT only in the city or on the open road.

When I drive in the woods or a canyon, the fps drop down to 30-40 - what in the crew means almost unplayable.


In the crew there is a BIG fps difference between different races - but its clear you didnt bench one of the demanding courses...

 

[Steve]

Sorry Steve, but you did the benchmarks completly wrong - and here is why:

I have a GTX970. With 4xMSAA I can drive the whole day and run constantly in the 60fps limit. BUT only in the city or on the open road.

When I drive in the woods or a canyon, the fps drop down to 30-40 - what in the crew means almost unplayable.

In the crew there is a BIG fps difference between different races - but its clear you didnt bench one of the demanding courses...

I have finished the game now and I didn't notice the slow down you are talking about in any of the races that took place in either the canyons or woods.

That being said you do realize I didn't have 20 hours plus to finish the game at the time and analyze each section of the game?

The test can't be completely wrong if all cards were tested under the same conditions, we clearly displayed normal/expected game-play performance.

Our test was also ideal because it allowed us to replicate the results for accurate testing.