Prey Performance Review

After releasing the excellent (once its early bugs were fixed) Dishonored 2, Lyon, France-based Arkane Studios hits it out of the park again with Prey, a first-person shooter published by Bethesda Softworks as a reboot from 11 years ago.

Unveiled at E3 2016, Prey uses the CryEngine graphics engine and audio from Audiokinetic's Wave Works Interactive Sound Engine (Wwise). Of course, our interest in both technologies is their cross-platform support. Prey is in fact available for consoles (Xbox One and PlayStation 4) and on the PC (Windows). However, the game remains exclusively DirectX 11-based.

Minimum & Recommended System Requirements

Steam's page for the game lists minimum and recommended system configurations, which are suggested by Bethesda to facilitate acceptable performance. Prey does not appear to be particularly hungry for modern CPUs or GPUs, so hardware from five years ago should yield playable frame rates.

ConfigurationMinimumRecommended
ProcessorIntel Core i5-2400 or AMD FX-8320Intel Core i7-2600K or AMD FX-8350
Memory8GB16GB
Graphics CardGeForce GTX 660 or Radeon HD 7850GeForce GTX 970 or Radeon R9 290
Operating SystemWindows 7, 8.1, 10 (64-bit only)Windows 7, 8.1, 10 (64-bit only)
Disk Space20GB20GB
AudioDirectSound-compatibleDirectSound-compatible

Radeon vs GeForce

As usual, let's start with a graphics quality comparison between GeForce and Radeon cards. This being a multi-platform game destined to run on AMD-based consoles and a diverse range of PC components, it'd be disappointing if the rendering wasn't identical on cards from both discrete GPU vendors. Take note as well that AMD and Bethesda are partners, suggesting that Prey could be better-optimized for the Graphics Core Next architecture.

We're pleased to report that there isn't much difference between the output of a GeForce or Radeon, after all.

Graphics Options

The adjustable graphics settings are few, but the most important options are available for customization. Originally, it was only possible to adjust FoV manually through a configuration file. But as of patch v.1.03, a slider is available in the Advanced options menu.

Four quality presets are available: Low, Medium, High, and Very High. It is also possible to manually change the settings.

The High preset roughly correlates to what you'd see on a console. Very High offers a slight improvement to texture quality and lighting/shadows, particularly at long range.

The Medium preset imposes a slight reduction in texture quality, the distance at which shadows appear, and certain visual effects (like smoke). Even so, the game still looks great if your hardware necessitates these less demanding settings.

Low quality, on the other hand, sacrifices too much of what makes Prey look good. Textures lose a lot of their fidelity, there are fewer shadows, lighting quality is reduced, and anti-aliasing goes down the drain. The Low preset is clearly reserved for weak hardware configurations; avoid it if at all possible.

MORE: Mass Effect Andromeda Performance Review

MORE: Ghost Recon Wildlands Performance Review

MORE: For Honor Performance Review

This thread is closed for comments
24 comments
    Your comment
  • Bryan_B103
    I love the "Frame Time" chart on page 3. This provides a really nice graphical representation of what your experience would be in-game. I'm excited to see charts like that one used in future reviews.
  • barryv88
    Why was yesterday's RX480 used and not the newer 580?
  • c4s2k3
    373149 said:
    Why was yesterday's RX480 used and not the newer 580?


    I suspect they are going for cards in the mainstream. I'm sure there are plenty of 480s in use out there. Not too many 580s yet since it is new.
  • ykki
    No Ryzen benches? Are you planning to redo all the gaming benches for Ryzen once R3 hits and AGESA 1006 is in full effect?
  • coolitic
    Dishonored 2 was bad, Toms, even ignoring the bugs.
  • coolitic
    Why no high-end tests?
  • elbert
    Only a half review without Ryzen.
  • 10tacle
    746565 said:
    Why no high-end tests?


    45049 said:
    Only a half review without Ryzen.


    Well there's a lot of hardware missing I'd have liked to have seen as well like a GTX 1080. However in all fairness, they said they chose their hardware on what people *currently* have (according to Steam user hardware surveys), not what they may get in the future. Not many have Ryzen builds and GTX 1080s yet.

    In any event, as a 1440p GTX 970 SLI owner, I'd have liked to have seen the VRAM consumption at that resolution since there are benchmarks for it here. For me it's a little concerning that even at 1080p it uses 3.1GB. I'm trying to squeeze another year out of my GPUs holding out for Volta. But Pascal is due for a refresh later this summer/early fall (2xxx series).
  • elbert
    202972 said:
    746565 said:
    Why no high-end tests?
    45049 said:
    Only a half review without Ryzen.
    Well there's a lot of hardware missing I'd have liked to have seen as well like a GTX 1080. However in all fairness, they said they chose their hardware on what people *currently* have (according to Steam user hardware surveys), not what they may get in the future. Not many have Ryzen builds and GTX 1080s yet. In any event, as a 1440p GTX 970 SLI owner, I'd have liked to have seen the VRAM consumption at that resolution since there are benchmarks for it here. For me it's a little concerning that even at 1080p it uses 3.1GB. I'm trying to squeeze another year out of my GPUs holding out for Volta (expected to be nearly twice as powerful as Pascal).

    I see the argument but how big of a percentage will not buy the game thinking the Ryzen was left out because of very poor optimization? Its a bad move for a game developer to risk limiting its potental customers. I was wanting the game but I'll wait till I see Ryzen benchmarks.
  • 10tacle
    45049 said:
    I was wanting the game but I'll wait till I see Ryzen benchmarks.


    You can get an idea of that right now with other games and what to expect: http://www.pcgamer.com/gaming-performance-of-ryzen-7-vs-core-i7-with-geforce-gtx-1080-ti/

    My guess is based on that, it will be fine for above 1080p resolutions as the GPU is used more no matter how powerful the CPU is.
  • elbert
    202972 said:
    45049 said:
    I was wanting the game but I'll wait till I see Ryzen benchmarks.
    You can get an idea of that right now with other games and what to expect: http://www.pcgamer.com/gaming-performance-of-ryzen-7-vs-core-i7-with-geforce-gtx-1080-ti/ My guess is based on that, it will be fine for above 1080p resolutions as the GPU is used more no matter how powerful the CPU is.

    Your links are to far cry primal not prey. Here is the only one I have found but its on a questionable site.
  • 10tacle
    45049 said:
    Your links are to far cry primal not prey.


    I meant those links for another topic, lol. Look again at the PC Gamer link (it has a Bethesda title Fallout 4 in benchmarking).
  • Dugimodo
    looks tempting, I remember playing the original.
    Regarding the charts, am I the only one who finds the CPU utilisation chart confusing. Even knowing the AMD card uses a little less I can't figure out how the coloured bars tell me that? Wouldn't a min, max, and average bar be simpler?
  • envy14tpe
    I like that this review uses a setup that is most commonly used and in-budget for gamers. If people would look at (data and conculsion) how easily this game plays at 1080p with a midrange GPU n CPU I think you'd realize there is no need to test every possible GPU n CPU on the planet. Personally I wish more people tested games using normal gamers setup.
  • bigpinkdragon286
    45049 said:
    Your links are to far cry primal not prey. Here is the only one I have found but its on a questionable site.

    If your benchmark graph is to be believed, the AMD system appears to perform better than the Intel system, which doesn't exactly help the narrative that Ryzen should be skipped for pure gaming systems.

    The review is incomplete.

    While it's true to say that Ryzen isn't in a lot of systems yet, as it's too new, Ryzen is something a lot of people are looking at purchasing. Reviews shouldn't just be for what people currently have, but also should include what people are shopping for. People look at reviews to help with product choices. Leaving Ryzen out of the review here seems to promote a level of disservice to consumers.
  • elbert
    938891 said:
    45049 said:
    Your links are to far cry primal not prey. Here is the only one I have found but its on a questionable site.
    If your benchmark graph is to be believed, the AMD system appears to perform better than the Intel system, which doesn't exactly help the narrative that Ryzen should be skipped for pure gaming systems. The review is incomplete. While it's true to say that Ryzen isn't in a lot of systems yet, as it's too new, Ryzen is something a lot of people are looking at purchasing. Reviews shouldn't just be for what people currently have, but also should include what people are shopping for. People look at reviews to help with product choices. Leaving Ryzen out of the review here seems to promote a level of disservice to consumers.

    While I would agree its the slowest 6 core broadwell-e vs the fastest 8 core Ryzen. Actually looks like the core support isn't upto par. Could be the game only supports 4 cores.
  • cinergy
    Interesting that this is tested with the old RX480 and not the new RX580 although there isn't that huge difference. Should still be enough to pass 1060.
  • FritzEiv
    247398 said:
    Interesting that this is tested with the old RX480 and not the new RX580 although there isn't that huge difference. Should still be enough to pass 1060.
  • photonboy
    Cinergy,
    I believe the RX-580 is close to 10% faster than the RX-480. However, the GTX1060 may have slightly more OC left so it's probably at least 5%.

    Careful with PRICING though because the RX-480 (4GB and 8GB) and GTX1060 6GB prices vary between $200 and $400USD.

    It's hard to justify spending more than $260USD for the Asus Strix GTX1060 6GB though because if you go much higher you're getting close to a GTX1070 which starts at $335.
  • 10tacle
    67821 said:
    It's hard to justify spending more than $260USD for the Asus Strix GTX1060 6GB though because if you go much higher you're getting close to a GTX1070 which starts at $335.


    Also keep in mind that Nvidia is going to come out with a Pascal refresh later this summer or early fall with a 2000 series to extend a little more life before releasing consumer grade Voltas in 2018. The last time they did this was with Kepler going from the 6-series to the 7-series to extend Kepler's life until Maxwell (which gained about 10% in performance, like the 680->770).
  • mapesdhs
    On Win7, FF 53.0.3, most of the images on page 1 are not visible. They show up in IE, though I notice some of the slider quality comparisons don't appear to show the same scene (or at least that's what it looks like).
  • takeshi7
    I'm interested by the CPU Utilization because I ran this game on a 4 core i7 and I was consistently getting more than 60% utilization, indicating the game was using at least 5 threads. On a 4 thread i5 I figured it would be pegged at 100%
  • mapesdhs
    takeishi7, try using Process Explorer, etc. to more closely examine what's happening.
  • photonboy
    TAKESHI7,
    I don't like how the article states "there is still more headroom" because a CPU can rarely use 100% in a game. A game can easily be bottlenecked due to the performance of a SINGLE CORE (running the main thread of game code).

    Heck, I've got games that are bottlenecked when my system reports 30% CPU usage (i.e. using only two threads in eight-thread i7 CPU).