GTA V, Rough benches with 14 AMD / Nvidia GPU's (From 650ti - TitanX and 260x - 295x2

[Dzzope]

So as the title says, this is really a rough and ready guide as to what you can expect over a range of cards and settings.

It's not an in-depth analysis by any means (and I would have loved if he'd included minimum frames as well as average)

 

So here you go:

http://www.forbes.com/sites/jasonevangelho/2015/04/15/grand-theft-auto-v-pc-version-benchmarked-across-14-nvidia-and-amd-graphics-cards/

 

BTW for those saying "you NEED more than 2GB vram".. Just no. 2GB cards run fine, for their power, more ram would have minimal effect on performance.

The game seems pretty well optimized for mid range or older cards, No you won't run all the settings max'd but then when could you ever run a Trip A title max'd with a mid range or aging GPU.

 

Anyway, I hope this is useful as it covers a wide range of cards and performance ranges / budgets and lets you know roughly what you can expect to get fps wise.

 

For those who prefere a video, NCIX released this which will also give you a decent idea of the settings / hardware to get your desired results. Thanks to SnowComet for posting:

 

[Mitrailleuse]

forbes?

have you no shame?

[Dzzope]

forbes?

have you no shame?

Quick and dirty guide for a wide range of performances, Fro what I've seen, the higher performing cards are about right so it's more meant for those with lower tier GPU's..

As I said, it's rough guide as to what to expect and not an in-depth performance analysis.

[William5849]

I would also check out DudeRandom84 on youtube. https://www.youtube.com/user/DudeRandom84 He has some great video's showing the higher end GPU performance on the game. At high settings, the game does struggle in places to run smoothly.

[Mew]

Holy cow their website is awful. I can't imagine how difficult it would be to read their article on an ultra wide. Wait a minute.... why are they running with tesselation off?

[William5849]

Yeah, the testing seems odd. I've seen the 970 perform like crap when it went over 3.5gb of vram but he fails to mention this. I'd assume playing in 4k would be fine when playing at sub par settings, but if your playing in 4K maxed out, he hardly wants to touch that subject because they are perform like crap at settings that high.

[Dzzope]

Yeah, the testing seems odd. I've seen the 970 perform like crap when it went over 3.5gb of vram but he fails to mention this. I'd assume playing in 4k would be fine when playing at sub par settings, but if your playing in 4K maxed out, he hardly wants to touch that subject because they are perform like crap at settings that high.

4k was tested with everything max except AA. Whats your issue?

 

BTW, I've no issue with the site (though granted I've only scrolled through the one article).. But with screen width of 1600px or 3200 it works fine for me..

[1800-GPU-MILK]

Do people hate the 780 or something?

Or are they just a**holes?

[Dzzope]

Do people hate the 780 or something?

Or are they just a**holes?

Why? What?

 

If your wondering why it wasn't tested, thats because there are cards that are there with roughly similar performance...

[1800-GPU-MILK]

Why? What?

 

If your wondering why it wasn't tested, thats because there are cards that are there with roughly similar performance...

I've been having a very poor performance with my cards.

I guess this is the nail in the coffin for me to sell the two :L

[Tagiau]

The benchmarking thing in gta5 is shit. 

[Manny Calavera]

Nvidia cards have lower score on 1440p, interesting.

[Sakkura]

I would put much more faith in the Guru3D article. It also includes FCAT measurements, so the frame pacing can be quantified. There is some microstuttering in Crossfire, but not quite as bad as Forbes makes it seem.

[cesrai]

Yeah, the testing seems odd. I've seen the 970 perform like crap when it went over 3.5gb of vram but he fails to mention this. I'd assume playing in 4k would be fine when playing at sub par settings, but if your playing in 4K maxed out, he hardly wants to touch that subject because they are perform like crap at settings that high.

Where did you see that ??? members here said they didn't suffer from stuttering from the 970.

[HippY]

I like the fact that he shows other option rather than " maxed out" and "low end", so it does show me that I should get a 40 FPS with decent High settings 1080p on my HD 7770!

[William5849]

https://www.youtube.com/watch?v=T-I9LX4GNrI

[Sakkura]

https://www.youtube.com/watch?v=T-I9LX4GNrI

 

That's literally just some random guy on youtube.

[Dzzope]

I like the fact that he shows other option rather than " maxed out" and "low end", so it does show me that I should get a 40 FPS with decent High settings 1080p on my HD 7770!

This is exactly why I posted this, Hopefully those with mid or even lower end cards will find it useful.

[Dzzope]

I would put much more faith in the Guru3D article. It also includes FCAT measurements, so the frame pacing can be quantified. There is some microstuttering in Crossfire, but not quite as bad as Forbes makes it seem.

Likewise, This is a rough and ready guide to those with non top end systems.

[snowComet]

NCIX uploaded a video on their youtube channel, simply classified video cards for GTA 5 based on the resolution....

 

I think it is the simplest classification I have yet seen on youtube.....

 

[mr moose]

Yeah, the testing seems odd. I've seen the 970 perform like crap when it went over 3.5gb of vram but he fails to mention this. I'd assume playing in 4k would be fine when playing at sub par settings, but if your playing in 4K maxed out, he hardly wants to touch that subject because they are perform like crap at settings that high.

 

The most probable answer is he fails to mention it because it didn't happen.

[Dzzope]

NCIX uploaded a video on their youtube channel, simply classified video cards for GTA 5 based on the resolution....

 

I think it is the simplest classification I have yet seen on youtube.....

 

Great vid.. adding to OP

[Sauron]

According to TB's port report the included benchmark utility doesn't work very well and gives inaccurate results.

[Inkz]

Do people hate the 780 or something?

Or are they just a**holes?

I was wondering about that too, as a 780 owner I would like to know exactly. I know I can guess-timate inbetween 770 and 780ti but is just not the same.

[MEC-777]

The in-game benchmarking tool isn't very consistent, which makes it not completely ideal for testing. But, it is good enough to give you a rough idea, a ball-park estimate as to what performance to expect on your own machine with similar components.

 

Only thing I used the benchmark for was to optimize my graphics settings. It at least contains a number of scenes with varying loads on the GPU and I found it quite useful to fine-tune the settings and get them "just right".