(gamegpu) Battlefield 1 Benchmarks

Gregster · 13 Oct 2016 at 22:54

AMD sponsored game?

Brazo · 13 Oct 2016 at 22:55

Only game (ever) where the 970 beats the 1060!

TheRealDeal · 13 Oct 2016 at 22:57

Gregster said:
AMD sponsored game?

I don't think it is tbh. They are giving away some free bf1 stuff with rx480's but not the game. Amd do have a decent relationship with dice but so have Nvidia as far as i know. Most dice games run great on Nvidia whether Amd were involved or not.

Brazo said:
Only game (ever) where the 970 beats the 1060!

The 1060's performance is slightly weird in that sense. Possibly a miss print as it's beating it comfortably at 1440p.

southernorth · 13 Oct 2016 at 23:37

TheRealDeal said:
Your video was an empty server with nothing happening while the ops benchmark was done on the single player i think. The ops test has plenty happening on screen so should be closer to real world performance although Multiplayer usually differs from SP. From previous tests i have seen dx12 gave Amd a nice boost when on a server with people playing and a slight drop when in an empty server.

If it's a like-for-like comparison, with the same 'load' in either case why would results change so much, that makes no sense. It seems like quite an objective comparison. Due to the random nature of multiplayer games, getting a (repeatable) balanced benchmark would be exceedingly difficult between two cards.

Lokken86 · 13 Oct 2016 at 23:42

Gregster said:
AMD sponsored game?

Let's hope so

Poneros · 13 Oct 2016 at 23:44

Benchmarks for first 1-2 months are gonna be all over the place. The real question is how it will look in January.

shankly1985 · 13 Oct 2016 at 23:52

The battlefield beta site am sure had amd on it, I just looked and it's not there so I don't think this one is sponsored by anyone.

TheRealDeal · 14 Oct 2016 at 00:10

southernorth said:
If it's a like-for-like comparison, with the same 'load' in either case why would results change so much, that makes no sense. It seems like quite an objective comparison. Due to the random nature of multiplayer games, getting a (repeatable) balanced benchmark would be exceedingly difficult between two cards.

Well there's not much load happening on an empty server. Dx12 is at it's best when there is lots happening. An empty server is about as far removed as you can get from actual game play.

Here is the same guy doing both a controlled test and actual 64 player testing.

hyperseven · 14 Oct 2016 at 00:18

Do they let Boost 3.0 run rampant when conducting these/any benchmarks or are clocks locked to a set frequency?

JediFragger · 14 Oct 2016 at 00:21

shankly1985 said:
The battlefield beta site am sure had amd on it, I just looked and it's not there so I don't think this one is sponsored by anyone.

Probably last minute deal still being done tbh. Dice have always leaned towards AMD but know they can earn a shedload more moolah by allowing nVidia to sponsor it

TheRealDeal · 14 Oct 2016 at 00:29

hyperseven said:
Do they let Boost 3.0 run rampant when conducting these/any benchmarks or are clocks locked to a set frequency?

The cards are free to boost as they would out of the box or at least that's how most reviewers do it.

siriq111 · 14 Oct 2016 at 00:37

Do one more comparison when all or most of the features are enabled in the Frostbite engine.

GinG · 14 Oct 2016 at 08:33

Just ordered a 4gb 480 Nitro...will performance be similar to the 8gb @ 1080 assuming it does not utilize all the memory?

Also pleased to see the performance, I assume the 8350 clocked to 4.7ghz will be hitting similar to the 9590??

Maybe my investment in AMD 3 years ago is finally paying off

Supernaut91 · 14 Oct 2016 at 15:04

DX12 is a mess for nvidia.

tommybhoy · 14 Oct 2016 at 16:04

Supernaut91 said:
DX12 is a mess for nvidia.

Nvidia DX12-the new 'backwards compatible'!

GinG · 14 Oct 2016 at 16:30

Just been looking at a few benchmarks with 4gb vs 8gb 480's and pretty much nothing in it. Should see some good performance with my setup.

Kaapstad · 14 Oct 2016 at 16:39

Another game without mGPU support in DX12.

DX12 is shaping up to be the most useless API ever.

Bring back Mantle pleased......

Kaapstad · 14 Oct 2016 at 16:43

Almost forgot

The figures in the OP look very dodgy.

When you try and run a 2 way setup in DX12 where there is no support, it is not uncommon to get a drop in performance compared to a single card solution.

The graphs in the OP don't reflect this, are their mGPU figures just made up ?

Seanspeed · 14 Oct 2016 at 18:28

SiDeards73 said:
That 480 is spanking the 1060 in DX11 let alone DX12...

nashathedog said:
It's good to see the Fury giving the 1070 a run for it's money, actually it's giving it a sound thrashing.

Today I'm learning that something of a 5% performance gap or so is 'absolutely destroying' in comparison.

I'd call that a laughable exaggeration, but ok.

hyperseven said:
Do they let Boost 3.0 run rampant when conducting these/any benchmarks or are clocks locked to a set frequency?

Why would that be relevant?

Lokken86 · 14 Oct 2016 at 18:53

Seanspeed said:
Today I'm learning that something of a 5% performance gap or so is 'absolutely destroying' in comparison.

I'd call that a laughable exaggeration, but ok.

The Fury isn't beating the 1070 by anything eye raising but the 480 is a good bit ahead of the 1060 tbh.