Mantle vS 780Ti in BF4

[Ian Evey]

You might have posted another graph ?

 

[chief barker]

Naw one of my posts was deleted and im pretty sure i didnt say anything bad. THat said i can't quite remember what i said now. That reminds me, where am i and what's my name?

Same here

 

[Silent_Scone]

Naw one of my posts was deleted and im pretty sure i didnt say anything bad. THat said i can't quite remember what i said now. That reminds me, where am i and what's my name?

Set yourself up with that one

 

[chief barker]

Set him up with that one

I know how it works tin foil hat on all the time (well when the wife and kids are out).

 

[LtMatt]

Tin foil double intruder for me. #dontjudgeme

 

[chief barker]

Tin foil double intruder for me. #dontjudgeme

Never


Edit.

Ban incoming

 

[Gregster]

My post also got deleted and not sure why?

I said that nVidia drivers (latest 334.67) gave bigger gains than a whole new API and from the Swedish site that Matt linked, this lists the gains.

Mantle on the 290X gave 4.41% according to their testing
334.67 gave the 780 11.54% according to their testing.

Both Mantle and nVidia owners can be happy

Ohhh and just to point out. When the 770 was released, in most tests it showed the 770 was faster than the 7970. Now we all know that the 7970 was clocked fairly low and was capable of clocking faster than the 770. This is now the case with the 290X against the 780 Ti. The Ti is clocked conservatively but when you give it some, it seriously flies.

 

[gpuerrilla]

Really? Synthetic bench graphs now lol. One of the last few forcewares did give quite a substantial performance boost in BF4. For what it's worth anywho

Is this your way of sticking up for poor old greg?

If anyone takes stick it's generally as they have holes in their arguments or hang their neck out too far over something trivial.

The only annoying thing about these threads is there is pages of posts and only a handful of genuinely good information (oh and graphs)

 

[LtMatt]

Is this your way of sticking up for poor old greg?

If anyone takes stick it's generally as they have holes in their arguments or hang their neck out too far over something trivial.

The only annoying thing about these threads is there is pages of posts and only a handful of genuinely good information (oh and graphs)

I guess they (almost) had to catch up eventually. I bet Greg's hoping Mantle/AMD drivers don't improve, in single player at least. In multiplayer, well the gains are in the 20-50%+ range already.

 

[chief barker]

I guess they (almost) had to catch up eventually. I bet Greg's hoping Mantle/AMD drivers don't improve, in single player at least. In multiplayer, well the gains are in the 20-50%+ range already.

I've see it on my 'unsupported' crossfired 7970s and I can't wait.

 

[gpuerrilla]

I guess they (almost) had to catch up eventually. I bet Greg's hoping Mantle/AMD drivers don't improve, in single player at least. In multiplayer, well the gains are in the 20-50%+ range already.

What you guys play MP on BF4???

/sarcasm

 

[PaulyD]

So, did a singapore SP on ultra 4msaa 16af etc, counted down on my g500s stopwatch from 90 secs, went all the way to the river where you get on the boat.

On a 3570k 4.2 and 670 @ 1160 core and 6400 memory. 1920x1200

driver 332.21.

Frames, Time (ms), Min, Max, Avg
124222, 2127016, 36, 108, 58.402

Was going to do a graph but thought i would get shot , means nothing at all except most graphs look way off these days.

How do you get an output like that?

 

[humbug]

Single player benchmark, highest in game settings. Worst possible case for Mantle % improvements. None of that TechReport FUD. Latest drivers.










Source
http://www.nordichardware.se/Grafik...standa-i-bf4-med-nya-geforce-drivrutiner.html



Greg knows best, even though he's never used Mantle or a 290.

And thats still with the game mostly GPU bound on an overclocked 12 thread Intel CPU.

Take away the performance of the CPU and the Nvidia cards fall in performance while the AMD performance remains high, building a bigger and bigger performance delta over Nvidia.

 

[Kaapstad]

And thats still with the game mostly GPU bound on an overclocked 12 thread Intel CPU.

Take away the performance of the CPU and the Nvidia cards fall in performance while the AMD performance remains high, building a bigger and bigger performance delta over Nvidia.

Actually that is the default clock for a 3960X in a RIVE.

They could not run the CPU any slower in the RIVE without altering the bios settings lol.

 

[Silent_Scone]

Is this your way of sticking up for poor old greg?

If anyone takes stick it's generally as they have holes in their arguments or hang their neck out too far over something trivial.

The only annoying thing about these threads is there is pages of posts and only a handful of genuinely good information (oh and graphs)

No? You know what synthetic benchmarks are, right? Mainly, not relevant at all?

He was talking about a 15% performance improvement through a driver branch. What's that got to do with synthetic benchmarks?


Greg has SLI Titans, not sure why he needs sticking up for. Still faster than most of your junk .

Oh, sorry did I speak my mind again.

Come on guys, there isn't an argument here. I'm sure Mantle will reap more of a appeal in time, but right now you're comparing gnats **** lengths for the sake of it frankly. I'll just get back to playing TR with 4X SSAO down sampled to 4K. Give that a go on your 290s if you can get down
sampling to work on 14.1

*runs away*

x

 

[humbug]

Actually that is the default clock for a 3960X in a RIVE.

They could not run the CPU any slower in the RIVE without altering the bios settings lol.

It says here that £750 CPU is 3.6Ghz http://www.overclockers.co.uk/showproduct.php?prodid=CP-500-IN&groupid=701&catid=6&subcat=1672

Nice if you can afford it, if you can't and have a 290X you don't need it for that kind of performance

 

[Ian Evey]

How do you get an output like that?

Run fraps, tick boxes for desired output, under fps tab, min/max/ave, then press F11 while in any game to benchmark that level or part of the game you want to benchmark, It stores it under c:/Fraps/Benchmark and open it with wordpad or whatever you like.

 

[Ian Evey]

That first 90 seconds of singapore is probably a great way to bench bf4 in single player. very consistent.

 

[gpuerrilla]

No? You know what synthetic benchmarks are, right? Mainly, not relevant at all?

He was talking about a 15% performance improvement through a driver branch. What's that got to do with synthetic benchmarks?

Eh? He mentions tons of performance gains when it's quite obvious only very recently have they made such improvements. Before it has only been marginal improvements.

Greg has SLI Titans, not sure why he needs sticking up for. Still faster than most of your junk .

Oh, sorry did I speak my mind again.

So having expensive hardware (willy waving) means you command respect? Get a grip.

I seen he had to RMA one to scan or was that another Gregster?

edit- added after:

Come on guys, there isn't an argument here. I'm sure Mantle will reap more of a appeal in time, but right now you're comparing gnats **** lengths for the sake of it frankly. I'll just get back to playing TR at 1440P with 4X SSAO down sampled to 4K. Give that a go on your 290s if you can get down
sampling to work on 14.1

*runs away*

 

[Kaapstad]

It says here that £750 CPU is 3.6Ghz http://www.overclockers.co.uk/showproduct.php?prodid=CP-500-IN&groupid=701&catid=6&subcat=1672

Nice if you can afford it, if you can't and have a 290X you don't need it for that kind of performance

It is an i7 3960X they are using, in most X79 mobo's they run @3.5ghz. If you put one in the RIVE it will run @3.9ghz by default as the turbo on all cores is set to max by default.