Bad Company 2 Beta Benchmark

[schneiderguy]

How is it possible to use 50+% of quad cores when dual core is using only 70% load in BC2? Does 54% CPU usage in BF2 mean it's dual core optimized even though it's only optimized for 1 thread?

You're GPU limited? Try running in DX9 mode.

How do you know BF2 is single threaded? Did you develop it?

Was your screen shot in actual game play?

Can you post screen shots of your processor usage in your desktop with bc2beta in graphs not a single point in time where you could have easily taken the shot when you connect and disconnect off the server as that eats cycles.

http://pics.bbzzdd.com/users/schneiderguy/bc2cpu2.jpg

(task manager was on my second monitor)

multi- a combining form meaning “many,” “much,” “multiple,” “many times,” “more than one,” “more than two,” “composed of many like parts,” “in many respects,” used in the formation of compound words: multiply; multivitamin.

http://dictionary.reference.com/browse/multi

:hmm:

I don't care what dictionary.com says "multi" means. Multithreaded means "having more than one thread". You're the first person that I have ever seen use it to mean "having more than 2 threads".

edit: Also, when you say optimized code always uses 100% of the processor - I guess that means you wouldn't mind if Windows used 100% of the processor all of the time? Because it would be optimized, right

An optimized code uses all of the available CPU

 

[AzN]

Alright, you win - you're right. I obviously have no business being employed in the software development industry. Guess I should go talk to my manager now as I obviously don't know and haven't known for the past 15 years what I'm talking about.

Here's a suggestion - finish high school and go to college studying under a compsci degree. Pay particular attention in your systems architecture and operating systems classes. Then you might have an idea of what's going on inside of a computer and application.

You sound like an idiot. "My credentials is better than yours, I'm right your wrong. Go get education. bla bla" Did anyone ask you? I presume software companies let idiots work in software development industry for 15 years.

Regarding Dice's comments - DUH. But then again you seem to have never developed software, so you probably wouldn't have any clue how much overhead something like simple debug code can affect an application. Their comment still has no bearing on your "100% CPU usage" fallacy.

I have programmed stuff back in high school but I'm no software developer. I have made hacks and cracks and such. Nothing mind boggling as debugging. :hmm:

BC2 is and unoptimized engine that doesn't use 100% of the CPU. Same as your buggy software you've been debugging. :hmm:

Ok, that's enough. It seems you CANNOT hold a conversation without reducing it to grade school level. Take a few days and try to think of a new approach for your posting style.
Whether you are right or wrong regarding the subject matter is irrelevant.
Anandtech Moderator - Keysplayr

 

[SunnyD]

I have programmed stuff back in high school but I'm no software developer. I have made hacks and cracks and such. Nothing mind boggling as debugging. :hmm:

BC2 is and unoptimized engine that doesn't use 100% of the CPU. Same as your buggy software you've been debugging. :hmm:

So you admit you have absolutely no reasonable background to make your incorrect assumptions, and yet you're sticking by them. That takes some balls. I applaud you. Of course you're going to know better than a professional software game development company like Dice, forgive me for insulting you. So when is your perfectly written game engine coming out again? Make sure it uses 100% of whatever the latest modern CPU is so it can run like crapso kickassly perfect while you're at it.

 

[AzN]

You're GPU limited? Try running in DX9 mode.

How do you know BF2 is single threaded? Did you develop it?

LOL. So now BF2 is multithreaded?

http://pics.bbzzdd.com/users/schneiderguy/bc2cpu2.jpg

(task manager was on my second monitor)

You still haven't answered this question.

How is it possible to use 50+% of quad cores when dual core is using only 70% load in BC2?

I don't care what dictionary.com says "multi" means. Multithreaded means "having more than one thread". You're the first person that I have ever seen use it to mean "having more than 2 threads".

So now dictionary is wrong and you are right?

I don't care if I'm the first person to use it. The world multi means many or it could also mean more than 2. You assumed wrong without reading the context why the other person used it in a sentence. :hmm:

edit: Also, when you say optimized code always uses 100% of the processor - I guess that means you wouldn't mind if Windows used 100% of the processor all of the time? Because it would be optimized, right

That sounds so idiotic. When windows is doing nothing and idling it shouldn't use 100% processor. That would be unoptimized.

 

[thilanliyan]

In the definition you gave, "multi" also means "more than one" and I'm sure in the computer industry, multi-threaded means more than one, not more than two.

 

[AzN]

So you admit you have absolutely no reasonable background to make your incorrect assumptions, and yet you're sticking by them. That takes some balls. I applaud you. Of course you're going to know better than a professional software game development company like Dice, forgive me for insulting you. So when is your perfectly written game engine coming out again? Make sure it uses 100% of whatever the latest modern CPU is so it can run like crapso kickassly perfect while you're at it.

I don't need a software development background to know BC2 is a buggy engine. All engines are but some are more refined than others.,

I've been gaming long enough to know optimized game use 100% CPU usage relative to the cores that was designed for. There is parallel execution problem when a dual core doesn't use 100% CPU usage when quad core is using same amount of CPU usage as dual core. In short Frostbite is unoptimized engine that was designed for quad it looks like from Phil's benches leaving dual core users with less than desirable results.

You are the software developer you should be designing software that corresponds with dual and quad core in order. Not all over the place.

 

[AzN]

In the definition you gave, "multi" also means "more than one" and I'm sure in the computer industry, multi-threaded means more than one, not more than two.

That's the thing. English words are never precise. That is why you need to read the context before replying as I've told schneiderguy.

Personally multi means many to me. I use it describe more than 2 or 3 or 5 and even sometimes more than 1.

 

[thilanliyan]

Personally multi means many to me. I use it describe more than 2 or 3 or 5 and even sometimes more than 1.

I know multi can mean more than one/two/three/etc but like I said...in the computer industry I'm fairly certain multi-threaded is taken to mean more than one. Like it was pointed out, you're the first person I've seen to take multi-threaded to mean more than two.

 

[v8envy]

That's the thing. English words are never precise. That is why you need to read the context before replying as I've told schneiderguy.

Personally multi means many to me. I use it describe more than 2 or 3 or 5 and even sometimes more than 1.

Your personal meanings are nonsense. Multi as a prefix has a very precise meaning, and it means "more than one." It does not mean two, the prefix for that is "duo-". It does not mean three, the prefix for that is "tri-."

A dual core processor is a multi-core processor. Reversing that statement doesn't make sense -- a multi-core processor can be any processor with more than one core (including quads), not just a dualie.

 

[v8envy]

I've been gaming long enough to know optimized game use 100% CPU usage relative to the cores that was designed for.

This is bogosity of the highest order. To start with, code can be optimized for performance, memory usage, responsiveness, maintainability, extendability and lots of other parameters familiar to even a first year software engineering student.

In no way does "optimized" imply 0% CPU spoilage. That would be impossible in a general case -- targeting a 4 ghz quad would mean your code would NOT perform well on a 1.8 ghz quad.

 

[AzN]

Your personal meanings are nonsense. Multi as a prefix has a very precise meaning, and it means "more than one." It does not mean two, the prefix for that is "duo-". It does not mean three, the prefix for that is "tri-."

A dual core processor is a multi-core processor. Reversing that statement doesn't make sense -- a multi-core processor can be any processor with more than one core (including quads), not just a dualie.

Point taken. I shouldn't use the word multi to describe more than 2? In the dictionary it says that I can which I used the word to describe quad optimization.

 

[AzN]

This is bogosity of the highest order. To start with, code can be optimized for performance, memory usage, responsiveness, maintainability, extendability and lots of other parameters familiar to even a first year software engineering student.

In no way does "optimized" imply 0% CPU spoilage. That would be impossible in a general case -- targeting a 4 ghz quad would mean your code would NOT perform well on a 1.8 ghz quad.

I'm talking about optimized for the highest performance. You are just taking it out of context now.

 

[v8envy]

Point taken. I shouldn't use the word multi to describe more than 2? In the dictionary it says that I can which I used the word to describe quad optimization.

If the word was used to mean "more than one", then yes it was used correctly. You were also arguing for "a minimum of two" with your statements, which this prefix does NOT imply.

 

[AzN]

If the word was used to mean "more than one", then yes it was used correctly. You were also arguing for "a minimum of two" with your statements, which this prefix does NOT imply.

good to know I will use quad optimization for now on.

 

[v8envy]

I'm talking about optimized for the highest performance. You are just taking it out of context now.

No, I'm not. Highest performance does not imply 100% CPU usage nor does 100% CPU usage imply efficient, optimized code.

I could busywait in 8 threads and peg an i7 on the task manager, but app performance would be significantly worse. Latency as well as throughput are always an issue. If you have an inherently serial problem then distributing it and waiting on mutexes will gain you LESS performance.

For example, if the application fired up 4 threads and 3 of them were waiting for I/O it could still be multi-threaded and possibly optimized. And it could use less than 25% of the lowest end AMD quad.

 

[RussianSensation]

I've been gaming long enough to know optimized game use 100% CPU usage relative to the cores that was designed for. There is parallel execution problem when a dual core doesn't use 100% CPU usage when quad core is using same amount of CPU usage as dual core.

I am not sure I understand your logic here. A game can have 100% CPU usage simply because the CPU is overloaded to the max (i.e., it is too slow). For instance, a 3.0ghz Core 2 Duo may be 50% loaded, and in the perfect world a 1.5ghz Core 2 Duo will be 100% loaded. If the CPU load is not 100%, you cannot possibly conclude whether or not the game is properly optimized. Therefore, with your logic, you would conclude that the game is not optimized if you were running it on a 3.0ghz Core 2 Duo....because you are not pegged at 100%, but you'd conclude that it was perfectly optimized if you had a 1.5ghz C2D processor. That doesn't make sense.

Go to Post #138:
"4. I know PC gamers are constantly looking for that edge over the console crowd, so I'll say this about performance, since this is that type of thread. If you have a faster CPU, our audio engine will be able to do more "stuff" in the same amount of time, and as such provide an even more awesome experience for you guys. This is however not something we would be able to do had we used hardware acceleration, since those resources are the same for everyone (kind of like on consoles, ironically enough). So just know that as you buy better computers, our engine will scale fairly well along with it. Even more so in future revisions of the engine." - http://forums.electronicarts.co.uk/...y-2-pc/870182-multithreading-quad-core-6.html

What's a faster CPU: Post #155
"I talked about having a "faster CPU", which might be a bit vague, so let's elaborate a little on that. The way the revision of the audio engine that's in BC1/BC2 works is that it'll use a single HW thread to do all audio logic and mixing, ie. it'll help more with a CPU with a higher frequency. This actually also goes for the consoles as well, so on PS3 all the mixing will be done on a single SPU whereas the audio logic still runs on the PPUs. So this touches on the "future revisions" bit as well. As a game becomes "feature complete", any major changes will be done on some other branch of the engine so as to not endanger the game's stability etc. So I can't say much about it, but the current revision of the audio engine will utilize multiple HW threads in parallel for both audio logic and mixing. This obviously goes for consoles as well, since we basically don't have any platform-specific code (in the audio engine, apart from the lowest level)." http://forums.electronicarts.co.uk/...y-2-pc/870182-multithreading-quad-core-7.html

So it appears that the game will use the CPU for its use of software mixing. The way I understood that is the game use will at least 2 threads (since 1 is for the Audio source).

 

[AzN]

No, I'm not. Highest performance does not imply 100% CPU usage nor does 100% CPU usage imply efficient, optimized code.

So you would let CPU cycles go to waste for max performance in a game? Why would an optimized code not use all of CPU cycles to execute a function?

I could busywait in 8 threads and peg an i7 on the task manager, but app performance would be significantly worse. Latency as well as throughput are always an issue. If you have an inherently serial problem then distributing it and waiting on mutexes will gain you LESS performance.

For example, if the application fired up 4 threads and 3 of them were waiting for I/O it could still be multi-threaded and possibly optimized. And it could use less than 25% of the lowest end AMD quad.

Why not put these threads in order or combination for peak performance instead of waiting for I/O. I just don't see that these threads are being optimized at all if it doesn't use all of it's CPU cycles for max performance.

 

[v8envy]

Why not put these threads in order or combination for peak performance instead of waiting for I/O. I just don't see that these threads are being optimized at all if it doesn't use all of it's CPU cycles for max performance.

Because sometimes there's no work to do. It's that simple. If you're waiting for the user to shoot the bad guy in the head there are no computations to make. Once the user does pull the trigger only THEN do you need to compute impact, physics of blood splatter and body recoil, sound, etc.

Seeing where this is going? You can't simply store up computations and use them later. (People bringing up CPU out-of-order execution just to be difficult will be ignored.)

Old games running on single cores did have busywait event loops. They'd peg any single core 100%. That's not efficiency, that's poor design.

 

[AzN]

So it appears that the game will use the CPU for its use of software mixing. The way I understood that is the game use will at least 2 threads (since 1 is for the Audio source).

Don't you think that's a waste? 1 core just for audio?

 

[Keysplayr]

Azn has received some time off for personal insults. I recommend that nobody posts any retaliations to any of his posts at this point. Unless it has to do with the subject matter.
Thanks in advance.

Anandtech Moderator - Keysplayr

 

[SunnyD]

I don't need a software development background to know BC2 is a buggy engine. All engines are but some are more refined than others.,

I've been gaming long enough to know optimized game use 100% CPU usage relative to the cores that was designed for. There is parallel execution problem when a dual core doesn't use 100% CPU usage when quad core is using same amount of CPU usage as dual core. In short Frostbite is unoptimized engine that was designed for quad it looks like from Phil's benches leaving dual core users with less than desirable results.

You are the software developer you should be designing software that corresponds with dual and quad core in order. Not all over the place.

That's the thing. You don't design for a specific core count, you design around a paradigm. If you go multithreaded, then inherently having additional cores will speed things up - if your code is designed properly. But you also have to take the least-common-denominator in terms of hardware into play. A single core can handle multiple threads; it does so as I've described - serially. Having multiple cores on the other hand can be just as bad, for instance, say you have one thread waiting on another. You've just stalled a core waiting for input. There are many reasons for an application to simply release it's slice of CPU time back to the operating system, which in turn means that application isn't utilizing 100% of a core. That doesn't mean it's not efficient. Quite the contrary - it means the code is doing what it is supposed to do in the minimal amount of time necessary. That is what is called efficiency. Running a CPU up to 100% is simply brute forcing your way through, but it doesn't mean it's doing so efficiently.

 

[yacoub]

Anyone running Win7 with the CPU Usage widget installed can confirm it's definitely multi-threaded.

Also, why is the 5770 so far behind the GTX-275 and 260-216? Is that typical? I thought they were much closer in performance.

Thanks for posting those charts, though. Definitely nice to see where the cards are in relation to one another. It will be interesting to see how much it changes for the final product which should have some significant optimization passes on the maps as well as code improvements.

 

[schneiderguy]

Also, why is the 5770 so far behind the GTX-275 and 260-216? Is that typical? I thought they were much closer in performance.

From the translation it looks like they thought all of the cards were running in DX10 mode. However the 5xxx cards should default to DX11 mode which is higher quality afaik (I know the devs said the shadow quality is better, not sure what else). So that could account for the lower than expected performance vs. the 2xx cards which would all run the DX10 path.

 

[yacoub]

Oh interesting, thanks. IIRC, Dice said they didn't include the highest quality level of textures in the beta, so it will be interesting to see how the final product performs.

 

[MrK6]

Oh interesting, thanks. IIRC, Dice said they didn't include the highest quality level of textures in the beta, so it will be interesting to see how the final product performs.

I think they only included the highest quality performance drop .