BF4 CPU usage

[JAG87]

It's evident that most people in this thread do not know what a thread is. And I'm so sorry for the pun.

 

[tweakboy]

This is still BF4 beta so I expect engine performance improved by launch time. More optimized and smoother game play on release. gl

 

[sm625]

Who the heck is going to seige Shanghai? Anyway, it is clear that 4 core + HT is pretty much no gain vs 4 core without HT. It might even be worse since the i5 has only 6M of cache and yet somehow manages to keep up with the 8M i7.

 

[Fjodor2001]

New consoles are 6 threads for games. 2 cores are reserved for other tasks.

Any source?

I don't see why they would use such a design approach. When gaming, why not let the game get basically all the CPU power?

Sure there could be some background tasks that need to be running concurrently, but you don't need to allocate a separate CPU core for each task! Just create a thread/process and let the OS scheduler deal with it instead. I.e. that background task may only get 10% of the processing cycles on one of the cores, and the rest of the 90% on that core can be used for the game.

 

[ShintaiDK]

Any source?

Its been covered several times in the console threads. Just like games wont have more than 4½-5½GB memory. And thats including VRAM.

 

[tweakboy]

Im waiting for Hawaii Core in 2017. giggles

 

[Fjodor2001]

Its been covered several times in the console threads. Just like games wont have more than 4½-5½GB memory. And thats including VRAM.

What does "covered" mean? Is it just guessing, or do you have any official source?

 

[soccerballtux]

What does "covered" mean? Is it just guessing, or do you have any official source?

yeah threads are not a source

it probably started with some confused journalist making it up and then wikipedia sourced it and then everyone believed it.

on the other hand real time recording and multitasking could be why.

 

[escrow4]

1680x1050 seems wonky - why is the 6300 only getting 33/41 when the six core usage is so high? And where are some 1080p/1200p CPU graphs for that 6300? Don't see it in PC labs.

 

[Shivansps]

How in hell could the BF4 max out a FX6300 or about 70% of a FX8350 and still run in a console? The octa core Jaguar cant be faster than a FX6300, ill surprised if can cath up with the FX4300.

We are talking about either a some serious diference on CPU usage with a low level API, or there is something to inflate the numbers so they can make mantle look good when presented.

 

[ShintaiDK]

How in hell could the BF4 max out a FX6300 or about 70% of a FX8350 and still run in a console? The octa core Jaguar cant be faster than a FX6300, ill surprised if can cath up with the FX4300.

We are talking about either a some serious diference on CPU usage with a low level API, or there is something to inflate the numbers so they can make mantle look good when presented.

You just remove features etc.

 

[ShintaiDK]

What does "covered" mean? Is it just guessing, or do you have any official source?

http://kotaku.com/the-five-possible-states-of-xbox-one-games-are-strangel-509597078

http://www.eurogamer.net/articles/digitalfoundry-ps3-system-software-memory

Forgot to add, around 10% of the GPU power on the Xbox One is also reserved for non gaming usage. Most likely kinect.

 

[sm625]

Can we get some benchmarks taken while recording (ie streaming)?

 

[Enigmoid]

The Pclab is rather interesting. A 4770K scores ~20% more than a 2600K. And is on pair or better than a 4.5Ghz 2600K.

Very interesting to see that results are much better with a 770. Look at the gain you get by switching to a 770 from a 7970ghz with the 8350. Looks like nvidia's drivers have significantly less overhead.

(Tremendously ironic that you need a nvidia gpu to get the best result from an AMD CPU).

 

[BrightCandle]

I just had a play with this and while I had some FPS issues (wasn't using NVidias drivers) I did however find CPU usage hovering around 30%. That is around 3-4 cores. I was expecting to see most of my cores maxed out with a lot of CPU usage in these charts but it didn't happen. I seemed to be completely limited to around 80 fps most of the time despite testing on low graphics.

Weird I will have to look into it, these review sites are seeing a very different thing to me.

 

[soccerballtux]

http://kotaku.com/the-five-possible-states-of-xbox-one-games-are-strangel-509597078

http://www.eurogamer.net/articles/digitalfoundry-ps3-system-software-memory

Forgot to add, around 10% of the GPU power on the Xbox One is also reserved for non gaming usage. Most likely kinect.

wow, thats crazy

 

[Maximilian]

It's evident that most people in this thread do not know what a thread is. And I'm so sorry for the pun.

Would it be more accurate to say this game supports X amount of cores and the amount of threads it has dosent matter?

 

[BrightCandle]

Would it be more accurate to say this game supports X amount of cores and the amount of threads it has dosent matter?

I don't know. Technically the abstraction the software uses is a thread but because its a real time program not all threads are 100% on every core all of the time. I don't really know what to call the intermix of the two, I never have. Core utilisation would seem to be a better term than multithreaded in this case.

 

[ShintaiDK]

wow, thats crazy

At least they have plans to possible release the 10% of the GPU...in 3-4 years time.

http://www.neowin.net/news/microsof...rs-access-xbox-one-gpu-reserves-in-the-future

But still 6 cores for both consoles.

 

[Shakabutt]

The Pclab is rather interesting. A 4770K scores ~20% more than a 2600K. And is on pair or better than a 4.5Ghz 2600K.

Be fair and mention to people :

All tests were performed at a resolution of 1920x1080, changing the level of detail using the built-in profiles: Ultra High. In addition to the difference in the level of detail of the image, lower settings have no anti-aliasing (MSAA x4). Is it worth it at all interested in the details Ultra, you can check on the sixth page of this article. Impact of graphic detail the smooth operation of the program checked for example, the Radeon HD 7850, and the results from a comparison of Ultra, High, Medium and Low can be found on page seven. Tests performed on the map Siege of Shanghai in Domination mode, below a snapshot of the settings:

Domination Mode has 24 players only (i think), infantry only.

 

[AtenRa]

Be fair and mention to people :




Domination Mode has 24 players only (i think), infantry only.

Domination is up to 32 player (infantry only, small maps) in the PC(at least in BETA).

 

[Makaveli]

It's evident that most people in this thread do not know what a thread is. And I'm so sorry for the pun.

lmao :biggrin:

 

[tweakboy]

Again, we cant conclude this cpu usage as its beta. Much more improvement in engine speed still to come.

 

[AtenRa]

Just to put things in perspective,
Spawn point number C in Domination map is the Spawn Point number D in Conquest map.

Conquest Map

Domination Map

 

[Makaveli]

I haven't downloaded the beta yet but for you guys that did how big is it?

I just upgraded BF3 to premium and that takes up like 34GB's.......