BF4 Retail CPU scaling measured

[CAT-THE-FIFTH]

More details here:

http://gamegpu.ru/action-/-fps-/-tps/battlefield-4-test-gpu.html

 

[CAT-THE-FIFTH]

I more shocked at how well the £75 to £85 FX6300 is doing against much more expensive CPUs. Add the new game bundles into the mix(hopefully we will get them in the UK),and the value for money is very decent IMHO,especially for a game like BF4.

 

[CAT-THE-FIFTH]

Pcgameshardware should also have results at some point too.

 

[CAT-THE-FIFTH]

Snobbery has nothing to do with it. These bundles are great value, IF THE ONLY THING YOU'RE LOOKING TO DO is play BF4/Crysis 3.

That's a pretty huge 'if'.

Plus they are great for plenty of other games. I helped spec dozens of builds on various forums(and in real life too),and the FX6300 has decent performance when compared to Intel CPUs in many games.

The people bitching and moaning about how slow an FX6300 is,are those who have little or no experience of using sub £110 CPUs in any meaningful way. I have used the whole gamut of them from recent Core i3,FM2 Athlon II X4 and the FX63** series CPUs. Heck,I have a Xeon E3 myself,but I can acknowledge the makings of a good budget CPU.

The more games they can do well in the better for all of us. We should be having low end £110 to £120 quad cores from Intel now,not a reheated Core i3 which cannot be overclocked.

Heck,look in general hardware,plenty of builds with the FX6300 are specced by loads of members(people such as Stulid and the like),if the budget does not allow for a £150 CPU.

 

[CAT-THE-FIFTH]

I was looking at this article earlier:

http://www.eurogamer.net/articles/digitalfoundry-battlefield-4-next-gen-vs-pc-face-off-preview

They allowed journalists to look at the PS4,XBox One and PC versions of the games.

It seems the PC DICE was using at the event was an FX8350 with dual HD7970 cards. Interesting as it was not some AMD event.

 

[CAT-THE-FIFTH]

Hi, ok can anyone answer this. Currently I have a 2500K and xfire 7950. Would I get less bottlenecking with an FX6300 or FX8320 in Crysis3, FarCry3 or BF4?

TBH,with a Core i5 2500K,especially if you overclock it,it will still be a decent CPU IMHO. If there is a situation where you need more performance,you can always get a Core i7 3770K at a later date and you also get PCI-E 3.0 too.

 

[CAT-THE-FIFTH]

TressFX works quite well on my GTX660,so it seems to be not artificially locked out like PhysX.

 

[CAT-THE-FIFTH]

I would be careful about pclab as they seem to get weirdly low results with AMD CPUs at times,which can be contradicted by other reviews(CPU throttling?). In their article where they tested the BF4 Beta,their results were using a small domination map which could only support upto 32 players and no vehicles,and not the significantly larger conquest maps with upto 64 players and vehicles:

http://forums.anandtech.com/showpost.php?p=35563228&postcount=45

Both GameGPU and pcgameshardware,will test multiplayer maps at some point so I would wait for their articles to get a better view of performance in this case. If they roughly agree with what pclab then the results can be taken at face value.

 

[CAT-THE-FIFTH]

Sweclockers has tested the R9 290X in BF4 MP using the same map:

http://translate.google.com/transla...7810-prestandaanalys-battlefield-4/4#pagehead

Interestingly at Ultra settings at 2560X1440 and 1920X1080,the FX8350 is within 10% of the Core i5 4670K. However,if you drop the settings down to medium,it appears the CPU loading profile changes,and appears less multi-threaded it seems.

However,the results seem to contradict the pclab review,as at Ultra settings with an R9 290X the FX8350 and Core i5 4670K are reasonably close.

 

[CAT-THE-FIFTH]

I wonder how well the Intel Core i7 980X would do in the list with BF4?

I think you will be safe!

 

[CAT-THE-FIFTH]

AMD core temperatures aren't accurate, (That should be evident by the fact it shows all at the same temperature, but we'll ignore that)

It's an offset, so it could be X temp +20c.


Which is why AMD's keep it under 62c seems so conservative, but in reality it's going to be much hotter.

Unless someone can answer AMD's thermodynamic law breaking with their higher TDP and "lower" temperatures. (Compare it to Sandy, as Ivy/Haswell are gimped as they can't get rid of the heat due to the paste rather than solder, but the heat to get rid of is less.)

Temperature is also dependent on the surface area of the die. The FX CPUs have a much larger die than a socket 1155 SB CPU. If anything they are more comparable in surface area to the socket 2011 32NM CPUs.

 

[CAT-THE-FIFTH]

Actually looking on another forum,it seems the Sweclockers result might be on a 64 player conquest map.

 

[CAT-THE-FIFTH]

Good point, there's a decent amount of factors.

The 3930K was a pretty big chip IIRC, 8 core with 2 disabled?

More the Core i7 3820,although the FX8350 is nearly 10% larger.

IIRC,AMD has less dense caches than Intel,and the CPU has to be clocked higher too which does not help.

 

[CAT-THE-FIFTH]

If we look at the server roadmap Piledriver is still staying for high end 12 and 16 core servers(probably an MCM),but in an improved version called Warsaw but with better performance per watt.

So,I would suspect,we will find higher clocked versions of the current FX CPUs,with improved power consumption,ie,a new stepping for next year.

TBH,Intel is doing the same next year for desktop with Haswell still being the desktop performance part.

I expect AMD will bypass Steamroller for their 8 thread cores,and move onto Excavator in 2015,probably with a new socket and probably with DDR4 compatability.

 

[CAT-THE-FIFTH]

What PSU do you have?? What temperature does your GPU have?? What temperature is your CPU?? What clockspeeds are your CPU actually running at?? Is it throttling??

How is your RAM arranged?? Is it a 4GB stick and a 2GB stick?? If so,the RAM is not running in dual channel.

The BF4 Beta also had an issue with load balancing with six thread CPUs,which is not present in the retail game. I would make sure all the latest patches are installed.

 

[CAT-THE-FIFTH]

It seems Windows 8.1 improves performance a decent amount over Windows 7 in the game:

http://www.pcgameshardware.de/Battl...ts/Battlefield-4-Test-Benchmarks-CPU-1095298/

 

[CAT-THE-FIFTH]

It seems the Sweclockers results were for 64 player maps and used an average of three 180 second runthroughs. There were at least 50 players during each runthrough.

 

[CAT-THE-FIFTH]

What? You just said the i5s mostly beat the i7s (true, from the charts you linked for ultra in Windows 7), now you're saying there's no difference?

It seems like a crap review site. The Core i7 and FX8350 are outperformed by cheaper lower clocked CPUs with less cores in their respective ranges. I would discard that review as it seems to be inconsistent and ignores what people are seeing with the game(and what other reviewers are seeing too). They don't indicate what maps they used or how many players were on each map.

One of the Sweclockers reviewers posted this Anandtech:

I am one of the main contributors to the SweClockers article (I hate Siege of Shanghai now..), and without trying to hijack the thread I just want to clarify some things. There is always a lot of information lost in translation, especially as Google Translate is less than stellar sometimes..


* All our tests took place on the 64p version of "Siege of Shanghai" during real multiplayer on real servers. I can't swear we had 60+ players all the time, but I can almost guarantee we played on servers with 50+ almost every benchmark run.

* We outline our benchmark methodology and our settings as detailed as we can (with screenshots!). Of course we couldn't do identical multiplayer playthroughs, but we tried to compensate by collecting data during pretty long timeframes (at least 3x 180 sec). If our three runs didn't give a plausible value, we did a few more until satisfied.
http://www.sweclockers.com/artikel/17810-prestandaanalys-battlefield-4/2

* If you read the comments below the graphs (I know, translations..), we actually doesn't call "a winner" when the FPS difference is to narrow. The nature of multiplayer is to random, so a few FPS must be considered within margin of error. We also try to point out the CPU and GPU bottlenecks in the different scenarios.

Again, I am sorry for "hijacking" the thread a bit, I just wanted to clear some things up. Happy hunting in BF4!

They made multiple runthroughs and made sure maps had at least 50 people. Moreover their use of 180 second runthroughs is good practice as many review sites tend to use 30 seconds of gameplay only.