Mantle vS 780Ti in BF4

LtMatt · 6 Feb 2014 at 21:14

Sugarhell said:
So this article is so BS. First no graphs from cpu,gpu usage no video from the run.If this particular spot is cpu limited the graph will also show it.

I already test the first mission of bf4 with my 3770k at 3.5 and 7970 at 1050/1425 at high,motion blur 50% fov 70 like TR and i never dropped below 80 fps

You can check it here:http://www.overclock.net/t/1464668/...performance-in-battlefield-4/60#post_21726293

Meanwhile on rts games like total war 290x and titan are tied. Which is huge cpu limited game.http://www.anandtech.com/show/7457/the-radeon-r9-290x-review/14

Do you actually believe that nvidia has over 30% cpu improvements with dx and they dont put it on a label?

Good work Sugar. I have to admit, the results did not sit entirely right with me either.

TechReport did publish that Nvidia sponsored PR thing not long back. Something to do with 290 cards. Can't remember what, but Nvidia went to them cap in hand, told them what to do and they published an article. Can't remember what it was about now though.

kitch9 · 6 Feb 2014 at 22:43

Ian Evey said:
He is not an Nvidia fan if he is wearing a red tie.

If your going to bait, At least post something that could be believable

Uhm, he's a Nvidia fan who's been fixed.

That's the joke.

Mantle has generated a lot of very touchy forum dwellers.... It's amusing.

l2ez4m · 6 Feb 2014 at 22:45

It's really hard to come by unbiased press in these dayz.
But in a strange way - [TR] prolly done more to casual AMD user than all other tech sites combined.
I mean in the end of 2012 they were the first ones who 'discovered', pushed and made AMD publicly acknowledge and eventually fix their frame latency/pacing, memory management problems (first single card and much much later on cf configurations).
It even wouldn't be a stretch of imagination to speculate that this debacle 'forced' them to rediscover their old forgotten Sideport feature - implemented on R200 series in a form of Xdma.
Constructive criticism is always better than unqualified praise.

Buckster · 6 Feb 2014 at 22:47

R290x vs 780 Ti must be quite an easy decision

290 vs 780 Ti more difficult - but to me the 290 doesn't seem quite value enough to not consider the 780 Ti

Mantle is only a very few select games -even in near future

if the customer cooled 290s were 300 it would be a no brainer - at 380 odd - well thats more difficult IMO

Buckster · 6 Feb 2014 at 22:49

rolypolyman said:
Amd, forcing down the prices of the 780Ti.

well the prices of the 290 customs aren't exactly dropping are they - yes supply and demand agreed

andybird123 · 6 Feb 2014 at 23:09

Sugarhell said:
So this article is so BS. First no graphs from cpu,gpu usage no video from the run.If this particular spot is cpu limited the graph will also show it.

I already test the first mission of bf4 with my 3770k at 3.5 and 7970 at 1050/1425 at high,motion blur 50% fov 70 like TR and i never dropped below 80 fps

You can check it here:http://www.overclock.net/t/1464668/...performance-in-battlefield-4/60#post_21726293

Meanwhile on rts games like total war 290x and titan are tied. Which is huge cpu limited game.http://www.anandtech.com/show/7457/the-radeon-r9-290x-review/14

Do you actually believe that nvidia has over 30% cpu improvements with dx and they dont put it on a label?

you say it never dipped below 80, but your link with the graph does show dips below 80, and an average of 96fps, where as the article shows a 290X with an average of 118fps

looking at some other 290X reviews that does seem to be the right differential, so your results match up with theirs :confused:

Sugarhell · 6 Feb 2014 at 23:18

andybird123 said:
you say it never dipped below 80, but your link with the graph does show dips below 80, and an average of 96fps, where as the article shows a 290X with an average of 118fps

looking at some other 290X reviews that does seem to be the right differential, so your results match up with theirs

First i am using a 7970.Second i failed once with the car when you have to shot the heli so i had to redone this part so on the loading screen i had 30 fps. Also all the autosave gets the fps down to 30-40.So no

brawl3r · 6 Feb 2014 at 23:19

4737Joker said:
Greg and Matt should setup a youtube channel, you could discuss (argue) with eachother about gfx cards and benchmarks and then fight each other with large foam graphs or something

It would save me some time reading your threads which seem to conclude the same way

You really don't know how much you made me chuckle with this comment.

Silent_Scone · 6 Feb 2014 at 23:27

Sugarhell said:
So this article is so BS. First no graphs from cpu,gpu usage no video from the run.If this particular spot is cpu limited the graph will also show it.

I already test the first mission of bf4 with my 3770k at 3.5 and 7970 at 1050/1425 at high,motion blur 50% fov 70 like TR and i never dropped below 80 fps

You can check it here:http://www.overclock.net/t/1464668/...performance-in-battlefield-4/60#post_21726293

Meanwhile on rts games like total war 290x and titan are tied. Which is huge cpu limited game.http://www.anandtech.com/show/7457/the-radeon-r9-290x-review/14

Do you actually believe that nvidia has over 30% cpu improvements with dx and they dont put it on a label?

Maybe, just maybe, The 290x was having issues on their test bed? The latest drivers aren't great for DX from what I've heard.

So much conspiracy. Given the settings in TR tests, averaging as high as 140FPS seems quite plausible to me. In fact, very much so.

Sugarhell · 6 Feb 2014 at 23:38

MjFrosty said:
Maybe, just maybe, The 290x was having issues on their test bed? The latest drivers aren't great for DX from what I've heard.

So much conspiracy. Given the settings in TR tests, averaging as high as 140FPS seems quite plausible to me. In fact, very much so.

Maybe. By the way i am using 14.1 and my dx performance is the same.I expect 140 fps from 780ti and 290x. Nvidia cards tends to produce a lot more Max fps than amd cards so on low res or low settings they will gonna have a bit higher average. But this dont explain the claims from TR. Nothing on the first mission was cpu bound. Also DC on nvidia drivers gives an increase but not that huge like 30% as TR claims.

More about DC/commandlist

https://developer.nvidia.com/sites/...dev/docs/GDC_2013_DUDASH_DeferredContexts.pdf

humbug · 6 Feb 2014 at 23:40

14.1 Drivers in DX are exactly the same performance as 13.12.

gpuerrilla · 6 Feb 2014 at 23:42

Buckster said:
well the prices of the 290 customs aren't exactly dropping are they - yes supply and demand agreed

Buckster said:
for extra £100-£120 - I think the Ti is worth it

I'm just making the same decision now - 3rd party cooled 290 (380 odd) - vs Stock cooler 780 Ti (499 or so depending on current ocuk offers)

I think £300-350 is enough to be fair. Spending £550 (the extra £100-120) is not great mileage considering it doesn't beat the 290x by such a margin to warrant that spend. If it was £20 or even £50 it would be a different story, but to me that money goes a long way.

Silent_Scone · 7 Feb 2014 at 00:24

humbug said:
14.1 Drivers in DX are exactly the same performance as 13.12.

Well, the BF4 boards would tend to disagree...There are a few people having issues in the Mantle thread as well ** Reign it in a bit with the personal remarks please **

Sugarhell · 7 Feb 2014 at 00:29

MjFrosty said:
Well, the BF4 boards would tend to disagree...There are a few people having issues in the Mantle thread as well

No matter what when you change drivers some people agree some people disagree. You see like for example 100 people to complain.The others? That applies both on amd and nvidia

Silent_Scone · 7 Feb 2014 at 00:33

I'm talking about people rolling back due to severe frame drops from one day to the next. It's in the Mantle thread for all to see.

So going back to my original point, 14.1 may have been having problems on the TR test bed. 145FPS seems a pretty accurate number for medium settings. I'm sure if I disable SLi and run the same settings on my Ti I'd see similar averages. Of course that doesn't prove anything indefinitely but the number seems very feasible is my point. Basically, I don't think it's fair to say the test is fixed. It's probably not a completely ideal representation given other users are seeing higher gains.

K1LaW8tt5 · 7 Feb 2014 at 00:39

MjFrosty said:
I'm talking about people rolling back due to severe frame drops from one day to the next. It's in the Mantle thread for all to see.

So going back to my original point, 14.1 may have been having problems on the TR test bed. 145FPS seems a pretty accurate number for medium settings. I'm sure if I disable SLi and run the same settings on my Ti I'd see similar averages. Of course that doesn't prove anything indefinitely but the number seems very feasible is my point. Basically, I don't think it's fair to say the test is fixed. It's probably not a completely ideal representation given other users are seeing higher gains.

I wouldn't rule it out. definitely 14.1 is causing problems for some

TAZOO7 · 7 Feb 2014 at 00:51

Am i the only one that dont give a dam about numbers?

And there is more than one game out there mind

Give me anyone of them cards, id be just as happy with either

THE END

humbug · 7 Feb 2014 at 00:52

MjFrosty said:
Well, the BF4 boards would tend to disagree...There are a few people having issues in the Mantle thread as well ***

I'm also having issues with Mantle in BF4, i'm not having issues with Mantle in Star Swarm or 14.1 at all in D3D. that does not constitute "rose tinted welded in AMD eye ball" unless your hyperbolic and upset reading what is not there.

Silent_Scone · 7 Feb 2014 at 07:41

humbug said:
I'm also having issues with Mantle in BF4, i'm not having issues with Mantle in Star Swarm or 14.1 at all in D3D. that does not constitute "rose tinted welded in AMD eye ball" unless your hyperbolic and upset reading what is not there.

I'm reading Tech Report. Not sure what you're reading Humbug.

You don't have a 290x either AFAIK? So we aren't really talking about AMD performance as a whole on 14.1. I do however have a 780ti reference, so back at you at seeing what isn't there. Gettin' real old now.

The graphs are what they are, I've tried to rationalise the results but if in your opinion there is yet another instance of NVs'seemlessly bottomless bribe money pit then well, that's swell.

Either way, the experience on both cards will be near identical, everyones happy.

andybird123 · 7 Feb 2014 at 08:34

Sugarhell said:
First i am using a 7970.Second i failed once with the car when you have to shot the heli so i had to redone this part so on the loading screen i had 30 fps. Also all the autosave gets the fps down to 30-40.So no

Yes, I accounted for that, I looked at other reviews that have 7970 and 290x results and a 20-25% differential in bf4 seems correct, so your average agrees with TR's results

You cant really post up results and then say "but im going to arbitrarily add X to these because X" if you did something wrong then redo the run surely, or cut the bit where you know the mistake was, which is bad journalism, posting results and letting them stand, or trying to massage your results to fit an idea?