Battlefield Bad Company 2 GPU Benchmarks Here

[SpinachEater]

A lot of people are asking "can I run it" questions so here is the only badco2 benchmarks I could scrounge up. If anyone finds more please post.

http://www.techspot.com/article/255-battlefield-bad-company2-performance/page4.html

 

[FATAL STR1K3]

spinach ftw.

 

[JDV28]

Nice man! thanks!

 

[SpinachEater]

Found a new BC2 benchmark from xbit.

http://www.xbitlabs.com/articles/video/display/battlefield-bad-company-2.html

 

[jay2tall]

The graphs were produced on a system running an i7 920 @ 3.70GHz. I can agree with the 5870 test as I get around that or slightly under @ 4GHz and all setting maxed, 8xAA, 16xAF.

The problem is BC2 is CPU demanding and unless you have a newer quad core it's going to really hold your graphics back. I wonder if there are any benches out there that have older test beds with the same card. Maybe a 5770,GTX260, GTX285,5870 would be good on maybe 3-4 systems. It would draw a good picture.

 

[SpinachEater]

It is weird because for dual cores, it seems to put a heavy load on them. I was trying to mimic the techspot CPU tests where they disabled two of their cores. With two cores, I was getting 90%+ usage in both of them while running BC2. With the four cores, I was only getting 40-50% usage. Below, techspot was getting something to that order as well with 4 cores at 2.6Ghz.

It seems like BC2 makes good use out of dual cores but doesn't put a load on quad core CPUs. What is weird about that is they overclocked their i7 920 all the way up to 3.7Ghz and saw very significant performance gains yet the cores are probably still below 50% load. Seems like BC2 isn't quad core optimized.

I am curious Jay...have you looked at your CPU load being at 4Ghz?

 

[jay2tall]

I am curious Jay...have you looked at your CPU load being at 4Ghz?

I have not. I can check it out tonight. I also have this loaded on my parents laptop as I make my way over there for a few day visit at times. At night when i'm bored I load this up. They have a T6600 core2 and 4650mobile. IT runs fine on low settings and isn't that bad to look at, 40fps'ish In this case the CPU load is over 90% ALL the time.

 

[gamerk316]

It seems like BC2 makes good use out of dual cores but doesn't put a load on quad core CPUs. What is weird about that is they overclocked their i7 920 all the way up to 3.7Ghz and saw very significant performance gains yet the cores are probably still below 50% load. Seems like BC2 isn't quad core optimized.

I see the opposite: BC2 is very well optimized for 4 cores. The reason you don't get 100% is because the CPU is fast enough to handle all the game data without needing to use 100% of its resources.

100% CPU usage does not mean scaling, it means you need a faster CPU.

 

[jay2tall]

I see the opposite: BC2 is very well optimized for 4 cores. The reason you don't get 100% is because the CPU is fast enough to handle all the game data without needing to use 100% of its resources.

100% CPU usage does not mean scaling, it means you need a faster CPU.

I would agree. I have a buddy who has a dual core and it pegs his CPU as well. I can't recall his model, I believe its an E8400 or something.

 

[SpinachEater]

I see the opposite: BC2 is very well optimized for 4 cores. The reason you don't get 100% is because the CPU is fast enough to handle all the game data without needing to use 100% of its resources.

100% CPU usage does not mean scaling, it means you need a faster CPU.

I know that workload doesn't directly translate to well optimized. 100% load with junk coding could run slower that 80% load of optimized code. Here is where my understanding of processing breaks down. As an alternative to increasing the processor speed, can't you make use of the processing bandwidth (ie the open resources) to get a similar performance gain?

The techspot CPU tests were showing that the CPU became the bottleneck when it was under ~3Ghz. If the CPU is the bottleneck at 3Ghz and only using 33% of the resources...isn't there room for optimization there? BC2 seems to just take the workload of two threads and spread it evenly across 4 when available without any increase of workload...and here is were I might probably going wrong or misspeaking but doesn't that indicate that the coding is optimized for dual cores and not quads?

 

[Vengeance187]

The techspot CPU tests were showing that the CPU became the bottleneck when it was under ~3Ghz. If the CPU is the bottleneck at 3Ghz and only using 33% of the resources...isn't there room for optimization there? BC2 seems to just take the workload of two threads and spread it evenly across 4 when available without any increase of workload...and here is were I might probably going wrong or misspeaking but doesn't that indicate that the coding is optimized for dual cores and not quads?

You are completely correct, which is why there has been no response/rebuttal. You don't even have to know anything about processors...it's just common sense.
If you were getting a certain fps at below 50% cpu usage, then increase the cpu speed with a relative increase in fps with the same % of cpu usage, then it's clearly not optimized for quad core. If it were truely optimized and cpu usage wasn't 100% to begin with, then fps would remain the same and % cpu usage would start off higher and drop as clock speeds increased. Seems like simple common sense to me.