Far Cry 4 Game Performance Review

Image Quality, Settings, And Test Setup

To the casual eye, Far Cry 4 looks like Far Cry 3. That's not a big surprise, since both are powered by the Dunia Engine 2. The landscape is noticeably different, though. Kyrat is no tropical jungle. It's a large, mountainous land with canyons, lakes and waterfalls. I can't think of a more expansive, beautiful game setting. It can be breathtaking. 

The animation above shows what you can expect from the various quality presets. Even low detail offers an attractive environment you can lose yourself in. With the ultra setting enabled, though, it's a sublime experience.

For our custom benchmark, we found a spot that's not only repeatable, but also demanding enough to really press the hardware. Vegetation and environmental effects seem to be the primary factor in dictating frame rate. A few animals or a small shootout won't have any meaningful impact. A 30-second run through the forest from inside an outpost is where we chose to test.

Test System And Hardware

As always, we strive to represent results across a wide range of graphics hardware. We tested every modern card we could get our hands on, from the Radeon HD 6450 to the new GeForce GTX 980 and dual-GPU Radeon R9 295X2. We even tried testing GeForce GTX 970 and 980 cards in SLI. However, our samples aren't identical models, and SLI wasn't cooperative when we used different third-party solutions in tandem.

We tested the game at low, medium and ultra detail levels across a number of resolutions from 1280x720 (720p) to 3840x2160 (Ultra HD). The 4K resolution is equivalent to four 1080p monitors. Despite that massive number of pixels, 4K monitors are becoming more popular every day thanks to sub-$600 options like Asus' PB298Q:

Unlike older models that require splitting a single video stream into two HDMI inputs, this 28" display is capable of 3840x2160 video at 60Hz over a single DisplayPort 1.2 cable. You can read more about the screen in Asus PB287Q 28-Inch 4K Monitor Review: Ultra HD For $650.


Test System
CPU
Intel Core i5-4690K (Haswell): 3.5GHz (3.9GHz max. Turbo Boost), 6MB Shared L3 Cache
Motherboard
ASRock Fatal1ty Z97 Killer, LGA 1150, Intel Z97 Express
Networking
On-Board Gigabit LAN controller
Memory
8GB G.Skill Trident (2 x 4GB) F3-2400C10D-8GTD, 1600MT/s, CAS 11-11-11-28 1T
Graphics
GeForce GT 730 512MB GDDR5
GeForce GTX 650 2GB GDDR5
GeForce GTX 750 Ti 2GB GDDR5
GeForce GTX 660 2GB GDDR5
GeForce GTX 760 2GB GDDR5
GeForce GTX 970 4GB GDDR5
GeForce GTX 980 4GB GDDR5

Radeon HD 6450 512MB GDDR5
Radeon R7 240 1GB DDR3
Radeon R7 250X 1GB GDDR5
Radeon R7 260X 1GB GDDR5
Radeon R9 270 2GB GDDR5
Radeon R9 285 3GB GDDR5
Radeon R9 290X 4GB GDDR5
Radeon R9 290X 8GB GDDR5
Radeon R9 295X2 8GB GDDR5
SSD
Adata Premier Pro SP920 128GB SATA 6Gb/s SSD
Power
In Win GreenMe 650W 80 PLUS Bronze PSU
Software and Drivers
Operating System
Microsoft Windows 8 Pro x64
DirectX
DirectX 11
Graphics Drivers
AMD Catalyst 14.9 Omega, Nvidia GeForce 344.75 WHQL
Benchmarks
Far Cry 4
Custom Tom's Hardware Benchmark, 30-second Fraps run

This thread is closed for comments
76 comments
    Your comment
  • gamebrigada
    Which drivers were you using on the AMD graphics side?
  • damric
    How much AA was used and what kind? None?
  • Ellis_D
    410076 said:
    How much AA was used and what kind? None?


    I'm assuming none since with my GTX 970 and i7 4790k, I was regularly bottoming out into the low-40s/upper-30s with SMAA enabled.
  • johnnyb105
    Kinda wondering why are they using a fx4170 and a 6300 WHEN THERE IS A FX 4350 AND 6350 CPU AND WHERE THE HECK IS THE 8350 AT???
  • stoned_ritual
    I have a gtx780 reference and an i5 4670k, I get BETTER framrates with SMAA enabled at 1080p than with 2xMSAA or the game-suggested level of 2xTXAA. I play this game on ultra. I do get fps drops in highly vegetated areas. The biggest performance gain is seen when I disable god rays.
  • Onus
    Hmmm, I got this game free with a 500GB Samsung 840 EVO. It will be the first shooter I've tried in years. I'm thinking my i5-3570K and HD7970 ought to manage "very high" reasonably well.
  • magic couch
    The AMD drivers used were the 14.9 Omega drivers, but the omega drivers are 14.12, not 14.9. Is it supposed to say 14.12 or were the 14.9 drivers used?
  • airborn824
    I am so BIASED. sigh what is this world coming to when we can trust nothing and no one for good info. I wanna see FX8350 with 290x with updated drivers and i5 4690 with GTX 9802 newest drivers so all of us can compare somewhat lol
  • Cryio
    @ Johnny: FX 4170 for old gen high clocked quad.

    6300, 6350, same thing mostly.

    You can OC an 8350 to that level of performance, so you can approximate.
  • airborn824
    THis game was so badly made. Why would there be such a difference in the CPU FPS? Sigh and very suprised such low FPS on the 295x2 which is the best money can buy these days. Very badly made game, glad i got it for FREE i would never buy it at its point now.
  • Nuckles_56
    Can you please put some laptops in there from time to time, as not everyone plays just on a desktop
  • ingtar33
    i noticed not a word about the glitches.
  • Ben Archer
    PB298Q 2560 x 1080 not 4k
  • Ben Archer
    4k chart is mislabeled as 1440p
  • clonazepam
    Frame-Rate Over Time chart for Ultra Details @ 1440p is mislabeled as "Middle-earth: Shadow of Mordor"

    Thanks Tom's for putting these test results up. I realize you picked a very specific 30 second run somewhere within the game, and that my subjective opinion on my experiences in-game may vary because I didn't approach my testing in the same manner. I realize it had to have been tedious to test this game on so many different hardware configurations. I'm glad you had what hardware you did have available for the testing. Cherry picking very specific hardware components or waiting for updated drivers would have postponed the article even longer. You must have been at this awhile ago, and it must have taken you quite a bit of time to complete.

    Thanks again, Happy New Year! (addressing various comments at once lol)
  • alidan
    47340 said:
    Hmmm, I got this game free with a 500GB Samsung 840 EVO. It will be the first shooter I've tried in years. I'm thinking my i5-3570K and HD7970 ought to manage "very high" reasonably well.


    i run a 280X that is factory clocked to a ghz edition level.
    i run the game with everything maxed but the fur sim and the AO, for ao, set it to the lowest one available, they fixed the really crappy implementation from far cry 3 that gave everything a black outline, and the difference between the lowest setting and the highest is visually negligible while you will notice a LARGE jump in framerate...

    though i cant figure out whats up with the random LARGE fps dips, im assumeing it has to do withloading areas apposed to it being just a graphics problem...

    now that i see the very high and ultra comparison. im going to find out what makes that blue haze and kill it, i think very high overall looks better than ultra because of that.

    -- update --

    turn off alpha to coverage in the game xml file, it costs a stupid amount of processing power for something you have to magnify to see

    i also turned off tree relief because it takes a stupid amount of processing power for not much visual benefit and i just knocked down the geometry level to low because in all honesty i play on the ground more often than i go in the air, and it culls trees so far out i don't notice it at all unless i specifically look for it.

    also personal preference, i turned shadows to either high or very high... i dont like soft shadows much, and the performance hit for them isnt worth it, i had originally though that anything lower than ultra lowered the shadow map and made it look blocky but i was wrong. with the power i saved, i turned hbao+ on because i saw that this game handles some areas stupidly where the only way to see detail with with the ssao... i wish games would stop doing this and just bake ao on anything that doesn't move and save me the processing power so i can turn the crap low of off completely.

    you also may want to turn water effects to low instead of ultra... the difference is negligible at best, but same with performance gains.

    sadly no way for me to get fps works so yay...
  • tomc100
    Can you run the game on Nvidia settings if you have an AMD gpu?
  • iam2thecrowe
    why not reduce the resolution to 720p with the cpu tests so we can see if there would be a difference between the i5 and i7? What another site also found was the game does not load at all with a dual core (non ht) processor. Doesn't even run. Apparently there is a fix for it, but still....
  • damric
    WHERE"S THE BEEF?
  • alidan
    132630 said:
    Can you run the game on Nvidia settings if you have an AMD gpu?


    yes you can, though im not sure if there is a performance hit when you use amd or not... i just did tweaking and got over all better performance with nvidia crap turned on than with it off... funny how that worked.
  • CauselessMango
    So does the game not have high and just goes from medium to ultra or did they not benchmark it?
    That was the benchmark I was looking for...
  • maestro0428
    Benchtastic article Don! Love how thorough this article is. As far as the game, I liked FC3 just fine and FC4 is more of the same. Breathtaking visuals in multi-monitor for sure!
  • spp85
    I am getting a solid 60+ average fps with HD7950 Vapor X overclocked at 1080p Ultra and SMAA anti aliasing and never dips below 40fps.
  • chenw
    On the 1440p performance page, one of the graphs is labeled Shadow of Mordor.

    Also, the graphs between frame time variance and frame time variance over time do not match. The first graph shows that 980 and 970 have similar values, and 290x having the worst, but in the over time graph, 980 is the worst (it seems like the colors for 290x and 980 have been mixed up).