Bad Company 2 Beta Benchmark

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

scooterlibby

Senior member
Feb 28, 2009
752
0
0
Same. My dualie is performing much better after the patch. Seriously can't believe how many people I've seen lately writing that a highly clocked dual just won't cut it. Utter BS. I'm not saying some games don't benefit more from quads, but no benchmarks have given me any reason to go to a Core 2 Quad or Core Ix. Maybe Sandybridge or Bulldozer will justify a new build for a gamer like myself.

Haha me of one day ago, such a fool! Anyways, with this Fermi non-announcement and the lack of available 5970's I said 'fuck it' and used some of the money I'd saved for a Q9550 for $194 (including tax at micro center). I currently have it @ 3.2 and have definitely noticed a substantial increase in minimum frame rates with this game.
 
Last edited:

SunnyD

Belgian Waffler
Jan 2, 2001
32,674
146
106
www.neftastic.com
Please do explain how E8400@ 3ghz is just as fast Q6600 @ 2.4ghz in GTA 4.

http://www.pcgameshardware.com/aid,...ark-review-with-13-processors/Reviews/?page=2
As has been pointed out already by others (Sorry, I stopped reading this thread regularly) in addition to my previous explanation, it should be abundantly obvious. I'm not sure why you keep picking examples that explicitly prove what I already explained in an attempt to disprove it.

I can only assume that you're confused as to why the dual core CPU that is only 25% faster in clock speed can match a slower CPU with 100% more cores. The explanation is simple.

As I pointed out in my example, the numbers I used implied a pure 100% scaling when comparing many factors:

Obviously then, in order to compensate a two core processor would need twice the clock frequency to keep up (simplifying out a lot of things, in reality clock speeds aren't the end-all be-all of an architecture). Ergo the reverse is true, a quad core processor could run at half the clock speed of a dual core processor to achieve the same thread throughput. Again, it's not QUITE that trivial, but the math is fundamental.

I left out a lot, but that was just for brevity's sake.

Simply put, 100% scaling is almost never possible with so many variables involved. In the case of software, you have to take into account the total workload on the the computer in addition to the actual work the application itself is doing, which may or may not scale the same on different core counts. Also, as was pointed out even while fundamentally the same for the most part, there are some architectural differences between an E8400 CPU and a Q6600 CPU that go deeper than just core counts. But the fundamentals are the same - it takes a faster dual core CPU to match a slower quad core CPU's performance (in this example).
 

toyota

Lifer
Apr 15, 2001
12,957
1
0
STALKER Call of Prypiat uses one of my 4 cores at full 100% all the time, that's the usage I found monitoring the CPU usage with Rivatuner, so Windows 7 not all the time spreads the thread across all the cores. Bioshock uses two cores with an average usage of 27% to 54% while the other two remaining never exceeds 24% and can go as low as 7%, it would means that it's optimized for Dual Core only.

I find odd that in every thread you keep hailing the i7, you should buy one. The i7 architecture isn't any wider than the Core 2 Quad architecture, so the IPC isn't that much different, both C2Q and C2D shares the same IPC, it isn't any different at all.

The main advantage of the i7 is their server approach tweaks that helped to boost performance like an integrated memory controller, more SSE instructions, cache tweaks, Turbo Boost for single threading and hyper threading which helps to maximize the front end utilization which went underutilized a great deal of time with the Core 2 architecture, plus it might be not a big deal but the i7 is a true quad core, the C2Q communication between the first set of two cores with the second set of two cores was abysmally slow. The Core i7 has a higher latency cache compared to the C2Q, but the C2Q besides of the underutilized front end issue, it was very limited by the Front Side Bus and its non native quad core design.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3382&p=3

"Conroe was the first Intel processor to introduce this 4-issue front end. The processor could decode, rename and retire up to four micro-ops at the same time. Conroe’s width actually went under utilized a great deal of the time, something that Nehalem did address, but fundamentally there was no reason to go wider. "

GTA4 will never be the best representation of future games, the game has crappy graphics that can't even tax enough a HD 3870 in terms of raw power, it is so CPU bound, plus has a lots of uncompressed textures that will simply fill a 2GB card easily, making it a VRAM/Texture limited game, it is simply a crappy game with no optimizations. If the game was properly done, it would run great on a Dual Core and a 9600GT with no problems, so much hardware for so little in return, Crysis did it much better as it offered much better graphics, and running it on a Quad Core won't give you lots of benefits.

http://www.behardware.com/articles/778-10/giant-roundup-131-intel-and-amd-processors.html

Here the benchmarks shows that Nehalem is barely faster than the C2Q and Phenom II in terms of gaming performance, in everything else, the biggest difference is only 50%, which shows that talking out of context, Nehalem is simply a rehashed C2Q with a True Quad design, IMC, minor tweaks for better threading performance, server oriented performance tweaks and a redesigned cache. At the execution engine level, they're not too different. But there's no reason to dump the design since it is a very good one which dates from the P6 design and its first incarnation, the Pentium M.
I am most certainly aware that the i7 doesnt offer very much over the Core 2 quads in most cases. people were pointing out GTA 4 performance on different cpus so all I did was link to the i7 benchmark for the game.
 

happy medium

Lifer
Jun 8, 2003
14,387
480
126
Haha me of one day ago, such a fool! Anyways, with this Fermi non-announcement and the lack of available 5970's I said 'fuck it' and used some of the money I'd saved for a Q9550 for $194 (including tax at micro center). I currently yave it @ 3.2 and have definitely noticed a substantial increase in minimum frame rates with this game.

Good choice your gtx 260's are just fine till Christmas. When they launch Crysis 2.
 

AzN

Banned
Nov 26, 2001
4,112
2
0
Because GTA4 only uses 3 threads (as far as I can tell, it never goes over 75% usage on my quad) and the e8400 is faster per clock than the q6600.

If we have 3 cores at 2.4ghz (the q6600 running gta4), that is around 7.2ghz worth of power (assuming linear scaling, no cache sharing issues, etc). If we have 2 cores at 3.0ghz, then we have 6.0ghz worth of power, but if it's 10-15% faster per clock, that 6ghz from the e8400 is more like 6.8ghz in terms of Q6600 power, which would put it right up with the q6600 in this application.

So now if it uses 75% core it's only optimized for 3 threads? Didn't you say in your previous posts that it could still be 4 thread even if it doesn't use 100% CPU usage?
 
Last edited:

AzN

Banned
Nov 26, 2001
4,112
2
0
Different cpus. 45nm vs. 65. Cache size is different. FSB is different. Wolfdale has massive cache very good for games. Newer architecture.

Compare E6600 with Q6600. Tells the real story. Same 65nm, same cache, same generation, same clock speed, same FSB.

64% improvement for min. frames by going from 2 to 4 cores. Very good core scaling!

If you want to compare wolfdale core, compare E8400 with QX6850 (same 45nm generation, same cache, same FSB, same clock). Also great scaling.

You are quite wrong in your assessment. Q6600 has 2x 4meg cache total of 8meg cache between all cores or even 2 cores. A E8400 has 6meg cache between only 2 cores. Tell me which has higher cache? Clock for clock Q6600 is superior when compared to dual cores with less cache.
 

AzN

Banned
Nov 26, 2001
4,112
2
0
Ex falso quodlibet?

Win7 does a lot of thing but it tries to not utilize all cores if there's no reason for that.. core parking and stuff

I was being sarcastic.
 
Last edited:

AzN

Banned
Nov 26, 2001
4,112
2
0
As has been pointed out already by others (Sorry, I stopped reading this thread regularly) in addition to my previous explanation, it should be abundantly obvious. I'm not sure why you keep picking examples that explicitly prove what I already explained in an attempt to disprove it.

I can only assume that you're confused as to why the dual core CPU that is only 25% faster in clock speed can match a slower CPU with 100% more cores. The explanation is simple.

As I pointed out in my example, the numbers I used implied a pure 100% scaling when comparing many factors:



I left out a lot, but that was just for brevity's sake.

Simply put, 100% scaling is almost never possible with so many variables involved. In the case of software, you have to take into account the total workload on the the computer in addition to the actual work the application itself is doing, which may or may not scale the same on different core counts. Also, as was pointed out even while fundamentally the same for the most part, there are some architectural differences between an E8400 CPU and a Q6600 CPU that go deeper than just core counts. But the fundamentals are the same - it takes a faster dual core CPU to match a slower quad core CPU's performance (in this example).

That's exactly the point. The fact BC2 beta had awful fast dual core results compared to slow quad is somehow optimized?
 

MrK6

Diamond Member
Aug 9, 2004
4,458
4
81
You are quite wrong in your assessment. Q6600 has 2x 4meg cache total of 8meg cache between all cores or even 2 cores. A E8400 has 6meg cache between only 2 cores. Tell me which has higher cache? Clock for clock Q6600 is superior when compared to dual cores with less cache.
Ironically no, you're incorrect. The Q6600 is simply two E6600 dies on the same package. They each have 4MB of cache, and they can only share it between the two cores, not all four. Wolfdale-based chips have 50% more cache per core.
 

SunnyD

Belgian Waffler
Jan 2, 2001
32,674
146
106
www.neftastic.com
That's exactly the point. The fact BC2 beta had awful fast dual core results compared to slow quad is somehow optimized?

I'm sorry, I must have missed that graph or presentation somewhere. Be that as it may, as I said it depends on a lot of other factors too. As was pointed out, simply comparing clock speeds isn't enough to determine what is going on.
 

evolucion8

Platinum Member
Jun 17, 2005
2,867
3
81
When I upgraded from the QX6850 @ 3.6GHz to this CPU, besides of some very small improvements in some games, WinRaR, the system felt slighly smoother when doing heavy jobs like media encoding. Heck, my QX6850 when was taxed with Linpack 64, it slowed down my system, with this CPU, I can even continue to use my PC and not feeling any slowing down at all.

But when I'm doing some stuff like copying multiple files to different drives and stuff that can fill the Windows cache up, QX6850 responded slighly better, considering that my current CPU has more tweaks at the RAM level like better timings and some optimizations at the chipset level. Probably the QX6850 couldn't tax the FSB much like the Q9650, or may be some issues with cache alignment. Odds of architectures.

Cache performance is very important with Intel CPU's for some reason. AMD's cache dependency isn't that high.
 

vj8usa

Senior member
Dec 19, 2005
975
0
0
You are quite wrong in your assessment. Q6600 has 2x 4meg cache total of 8meg cache between all cores or even 2 cores. A E8400 has 6meg cache between only 2 cores. Tell me which has higher cache? Clock for clock Q6600 is superior when compared to dual cores with less cache.

Even ignoring L2 cache, the E8400 has an inherent advantage over the Q6600 clock for clock since it's a 45nm (Penryn) chip.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3069&p=3

So now if it uses 75% core it's only optimized for 3 threads? Didn't you say in your previous posts that it could still be 4 thread even if it doesn't use 100% CPU usage?

Use some common sense here - even if an application is using 76% of a quad core CPU, it has to have at least four threads. In a similar vein, if an application is using 51% of a dual core CPU, it has to be at least dual threaded.
 
Last edited:

Scionix

Senior member
Feb 25, 2009
248
0
0
Sorry if it's been mentioned, but the "performance" patch didn't do anything to solve my GTX 295 problems. I get stutter no matter what (even with all low settings), and nothing gets me a stable 60 fps. Turning shadows off low also wrecks my fps. Anyone else with a 295 having issues?

Bleh. I'm not expecting to never drop below 60fps with max settings or anything, but...

c'mon D:
 

Udgnim

Diamond Member
Apr 16, 2008
3,680
124
106
Sorry if it's been mentioned, but the "performance" patch didn't do anything to solve my GTX 295 problems. I get stutter no matter what (even with all low settings), and nothing gets me a stable 60 fps. Turning shadows off low also wrecks my fps. Anyone else with a 295 having issues?

Bleh. I'm not expecting to never drop below 60fps with max settings or anything, but...

c'mon D:

use 1 GPU

BC2 beta does not like SLI or crossfire setups atm
 

glorylin

Junior Member
Sep 20, 2008
13
0
0
Sorry if it's been mentioned, but the "performance" patch didn't do anything to solve my GTX 295 problems. I get stutter no matter what (even with all low settings), and nothing gets me a stable 60 fps. Turning shadows off low also wrecks my fps. Anyone else with a 295 having issues?

Bleh. I'm not expecting to never drop below 60fps with max settings or anything, but...

c'mon D:

I believe it's some kind of incompatibility with nvidia drivers + quad core. Go to the settings.ini and change direct x version to 9. Temporary fix for the beta.
 

Scionix

Senior member
Feb 25, 2009
248
0
0
I believe it's some kind of incompatibility with nvidia drivers + quad core. Go to the settings.ini and change direct x version to 9. Temporary fix for the beta.

Changed to DX9, game is buttery smooth now, everything at max (besides AA), and I'm not dropping under 60fps. Hurf.
 

AzN

Banned
Nov 26, 2001
4,112
2
0
Even ignoring L2 cache, the E8400 has an inherent advantage over the Q6600 clock for clock since it's a 45nm (Penryn) chip.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3069&p=3

Besides marketing sham of technical garbage how is the performance compared to 65nm at the same cache? The difference? it's mostly smoke and mirrors. There are plenty of reviews comparing E4700 vs E5200. At the same cache levels they perform relatively same.


Use some common sense here - even if an application is using 76% of a quad core CPU, it has to have at least four threads. In a similar vein, if an application is using 51% of a dual core CPU, it has to be at least dual threaded.

Wasn't this my argument from the start? BC2 could be quad optimized but in CPU usage on my dual core it is most likely not. Now you say GTA4 isn't quad optimized but triple optimized? I'm sure you could go find benches comparing x3 vs x4 cpu's....
 

AzN

Banned
Nov 26, 2001
4,112
2
0
Ironically no, you're incorrect. The Q6600 is simply two E6600 dies on the same package. They each have 4MB of cache, and they can only share it between the two cores, not all four. Wolfdale-based chips have 50% more cache per core.

You do not know what you are talking about. It does not matter if Q6600 is 2 E6600 die in package.

The way quad core works is that 1 core could still use full 4 meg cache the other could also use the other set of 4 meg cache in a dual core optimized game. So in a game both 2x 4 meg cache can be used!
 

schneiderguy

Lifer
Jun 26, 2006
10,801
91
91
You do not know what you are talking about. It does not matter if Q6600 is 2 E6600 die in package.

The way quad core works is that 1 core could still use full 4 meg cache the other could also use the other set of 4 meg cache in a dual core optimized game. So in a game both 2x 4 meg cache can be used!

That's assuming Windows is smart enough to put the threads on the right cores :p
 

MrK6

Diamond Member
Aug 9, 2004
4,458
4
81
You do not know what you are talking about. It does not matter if Q6600 is 2 E6600 die in package.

The way quad core works is that 1 core could still use full 4 meg cache the other could also use the other set of 4 meg cache in a dual core optimized game. So in a game both 2x 4 meg cache can be used!
lol, no. EDIT: not even worth it, I've been down this road