PDA

View Full Version : Tested Intel® Core™2 Duo with 480GTX for CPU bottleneck


gerardfraser
07-31-10, 01:02 AM
My test were not truly CPU scaling test ,due to the fact I used high resolution and maxed out the games,it was just some test to show a highly clocked Intel Core 2 Duo with a 480 plays games fine.


So take the benches with a grain of salt,seems some people take the benches to serious.your results may vary.These were just my results.


CPU: Intel® Core™2 Duo Processor E8500 @4.0GHZ
Motherboard:ASUS P5QL/EPU CORE 2 DUO INTEL P43
RAM: G.SKILL F2-8500CL5D-2GBPK
HardDrive: 750GB USB External Hard Drive
Video:EVGA SC 480GTX @843Core/1686Shader/Memory 1997
Monitor: 24Inch @ 19200x1200

Setting used
http://img688.imageshack.us/img688/973/settingsd.th.jpg (http://img688.imageshack.us/i/settingsd.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)


Well I put together the above computer to test to see how bad the bottleneck is on a 480 GTX card.
To my surprise it really is not that bad ,and I see maybe three games I tested were limited by the Intel® Core™2 Duo Processor E8500 4.0GHZ CPU speed.
Dirt2,FarCry2,REVIL5

All games were maxed out fully,all games played great no stuttering with one exception.
Metro 2033 which can not be maxed out fully and still play great.This applies to all setups on a single card.

Let me know if you do not agree with any benches here.

Battlefield Bad Company 2
DirectX 11 enabled
32x Anti aliasing
16 Anisotropic filtering All image quality settings enabled at maximum
Fraps at airfeild and jeep ride.Tons of action.No real bottleneck here.

MIN:46
AVG:54.6
MAX:60

http://img827.imageshack.us/img827/2005/bfbc2game20100730214353.th.jpg (http://img827.imageshack.us/i/bfbc2game20100730214353.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)


Dirt 2
DirectX 11 enabled
8x Multi-sample Anti aliasing
16 Anisotropic filtering (Forced NCPL)
All image quality settings enabled at maximum
London BatterSea Track.Bottleneck here.Compared to core I7 920 @4.0GHZ (APPROX 10-20FPS)

MIN:51
AVG:62.9
MAX:

http://img14.imageshack.us/img14/4981/dirt2game20100730210023.th.jpg (http://img14.imageshack.us/i/dirt2game20100730210023.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)



JustCause2 2
DirectX 11 enabled
32x CSAA
16 Anisotropic filtering
All image quality settings enabled at maximum
No real bottleneck here.
MIN:
AVG:52.98
MAX:

http://img188.imageshack.us/img188/9566/justcause22010073021183.th.jpg (http://img188.imageshack.us/i/justcause22010073021183.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)



BattleForge
DirectX 11 enabled
8x Anti aliasing
16 Anisotropic filtering (Forced NCPL)
All image quality settings enabled at maximum
No real bottleneck here.

MIN:38.5
AVG:75.6
MAX:185.2

http://img6.imageshack.us/img6/1926/battleforge201007302221.th.jpg (http://img6.imageshack.us/i/battleforge201007302221.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)



Metro 2033
DirectX 11 enabled
4x Multi-sample Anti aliasing
16 Anisotropic filtering
All image quality settings enabled at maximum
Fraps just before chapter one starts ,room with all the monsters.Lots of action.No real bottleneck here.
Settings will cause stuttering maxed out.Down too 15 FPS with sunrays.

MIN:22
AVG:29.7
MAX:36

http://img580.imageshack.us/img580/9738/metro203320100730222734.th.jpg (http://img580.imageshack.us/i/metro203320100730222734.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)


Aliens VS Predator 2010
DirectX 11 enabled Ran @1920x1080
4x Full Screen Anti aliasing
16 Anisotropic filtering
All image quality settings enabled at maximum
Japamd benchmark tool.No bottleneck here
MIN:
AVG:46
MAX:

http://img202.imageshack.us/img202/2069/avpmax.th.jpg (http://img202.imageshack.us/i/avpmax.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)


Stalker:COP
DirectX 11 enabled
4x Multi-sample Anti aliasing
16 Anisotropic filtering
All image quality settings enabled at maximum
No bottleneck here.

Rain
MIN:26.6
AVG:52.2
MAX:89.8

http://img196.imageshack.us/img196/509/copvq.th.jpg (http://img196.imageshack.us/i/copvq.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)


Crysis
DirectX 10 enabled
Different levels Anti aliasing
16 Anisotropic filtering (Forced NCPL but does not work with Crysis and nvidia)
All image quality settings enabled at maximum
Island Map.No bottleneck here.ATI 5850 OC is faster then 480GTX.

MIN:36
AVG:42.81
MAX:57

http://img12.imageshack.us/img12/6358/843480crysis.th.jpg (http://img12.imageshack.us/i/843480crysis.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)


FarCry 2
DirectX 10 enabled
4x Multi-sample Anti aliasing
16 Anisotropic filtering (Forced NCPL)All image quality settings enabled at maximum
Ranch Map.Bottleneck here.Compared to core I7 920 @4.0GHZ (APPROX 10-20FPS)

MIN:51.6
AVG:69.84
MAX:100.98

http://img709.imageshack.us/img709/9066/843480farcry.th.jpg (http://img709.imageshack.us/i/843480farcry.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)


Resident Evil 5
DirectX 10 enabled
8x Multi-sample Anti aliasing
16 Anisotropic filtering (Forced NCPL)All image quality settings enabled at maximum
Varible Benchmark.Small Bottleneck here.Compared to core I7 920 @4.0GHZ (APPROX 5-10FPS)

MIN:
AVG:101.2
MAX:

http://img413.imageshack.us/img413/4416/re5dx102010073023301013.th.jpg (http://img413.imageshack.us/i/re5dx102010073023301013.jpg/)

Uploaded with ImageShack.us (http://imageshack.us)

ragejg
07-31-10, 02:03 AM
Ok, thanks for the info. I see you've aggregated quite a bit of data, but what are your thoughts on this?

gerardfraser
07-31-10, 02:21 AM
Ok, thanks for the info. I see you've aggregated quite a bit of data, but what are your thoughts on this?

I really do not have any thoughts on it really.
Except a core duo can play games as good as a New quad core I7 machine.

I bought the computer parts just to test for bottlenecks that everyone always throws out there when people have problems in games.

People say yeah its your CPU,MEM,motherboard with the bottleneck,while actual fact theres no real bottleneck (except a few games but playable anyway)with highly clocked good CPU from intel.
Thats about it really,Im not a big benchmarker,I just like playing games and did the test for fun.

Madpistol
07-31-10, 09:13 AM
I really do not have any thoughts on it really.
Except a core duo can play games as good as a New quad core I7 machine.

I bought the computer parts just to test for bottlenecks that everyone always throws out there when people have problems in games.

People say yeah its your CPU,MEM,motherboard with the bottleneck,while actual fact theres no real bottleneck (except a few games but playable anyway)with highly clocked good CPU from intel.
Thats about it really,Im not a big benchmarker,I just like playing games and did the test for fun.

Hold up there buddy... For BC2, you really needed to do a 32 person match online and see what you got. You may not see a bottleneck in single-player, but as soon as you crank the player numbers in a multiplayer match, you'll get a different story.


I'll do a couple of 10-minute benchmarks in multiplayer (2 core and 4 core tests) and get back with you

Madpistol
07-31-10, 09:48 AM
This benchmark doesn't use a GTX 480, but if you'll look in my sig, you'll see there's more than enough power in those components.

The only difference is that my CPU is @ 3.8Ghz


Settings: 1920x1080
4xAA, 16xAF, All settings high, (HBAO and Vsync turned off)

First test: 32-man Hardcore server, 2 cores (set in task manager, cores 3 and 4)

Frames, Time (ms), Min, Max, Avg
25427, 539153,____27, 110, 47.161


Second test: Same server (next round), same map, 4 cores.

Frames, Time (ms), Min, Max, Avg
34266, 501699,____40, 104, 68.300




That's just going from 2 to 4 cores. No difference in settings at all. I didn't even leave the server. I just took 2 separate benchmarks in game (1 full round each) I also noticed that when I was only using 2 cores, there were "hitches" in the framerate about every second or 2. As soon as I set the game to use all 4 cores, those hitches went away and gameplay became very smooth.

In newer games, you will see a boost in FPS going from a high-powered dual core to a Quad Core. I can run benchmarks all day in different 32-man servers, but I can guarantee that the results will be similar in all instances.

Quad Cores are preferable for gaming now. Maybe not all games, but if you want liquid smooth framerates in multi-threaded games, a Quad core is the way to go.

spajdr
07-31-10, 09:52 AM
Same apply for Starcraft II, which is limited noticeably too with any dualcore cpu or quadcore AMD cpu.
http://static.techspot.com/articles-info/305/bench/CPU.png

Bah!
07-31-10, 10:17 AM
Hold up there buddy... For BC2, you really needed to do a 32 person match online and see what you got. You may not see a bottleneck in single-player, but as soon as you crank the player numbers in a multiplayer match, you'll get a different story.


I'll do a couple of 10-minute benchmarks in multiplayer (2 core and 4 core tests) and get back with you

QTF.

There is a huge difference between a C2D and an i7 in BFBC2.

Madpistol
07-31-10, 10:57 AM
Here's another test, this time using single player in BC2.

Test: First level (from point that it's first playable) for 10 minutes

Same settings as a few posts above.



First run: 2 cores (cores 3 and 4)

Frames, Time (ms), Min, Max, Avg
36491, 600000,____24, 94, 60.818



Second run: 4 cores

Frames, Time (ms), Min, Max, Avg
45334, 600000,____21, 116, 75.557



The first thing that you'll notice is that the minimum framerate on 4 cores is actually lower. Each run was slightly different, so that's what I got. However, it's impossible not to see the difference in Average FPS. 15 fps average gain going to 4 cores. In other words, there is more CPU to use than just 2 cores.

The same effect happened in this game with framerates. Although they look similar in terms of numbers, the gameplay difference was immense. On the 2 core run, I constantly felt the framerate hitching and stuttering inconsistently. However, on the 4 core run, it was very smooth and incredibly enjoyable. As I said, the numbers don't tell the whole story; the 4 core run was much more enjoyable simply because the framerate was much smoother.


Once again, Dual cores are starting to show their age. Quad Cores are the preferable CPU's for gaming. Hell, I've even tested BC2 on my fiance's computer, which has an e8400 @ 4.0Ghz. The framerate does the same thing as 2 cores on the PII 965; it constantly hitches and is less enjoyable. I just wish I had a Core 2 Quad to compare it against. I have a feeling that Core 2 Quads do much better in BC2 than Core 2 Duo's.

Yaboze
07-31-10, 12:31 PM
Most games do run well on a Dual Core Intel. I'm really surprised how much BFBC2 and SC2 really tax my machine (i7 + 480). I have SC2 set on Ultra and it ran really stress it when there's lots going on.

ViN86
07-31-10, 01:43 PM
So how did you determine there was no bottleneck? Turn your graphics down to minimum with a low resolution and run your tests again. If the FPS is about the same, then you definitely have a bottleneck with your CPU.

The term "bottleneck" is misleading, because it needs to be with respect to the other components. A bottleneck implies that there is a rate-limiting step, which is typically either the CPU or GPU. My E8400 at 3.6GHz bottlenecks in BC2. I can play at 1920x1200 with 4xAA 4xAF or at 1680x1050 with no AA/AF and I get the same framerates in the mid 50's. Meanwhile, people with i7's typically get over 100 FPS in BC2.

EDIT: Did you compare them to a similar review with an i7?

gerardfraser
07-31-10, 02:46 PM
@Madpistol
Thanks for the benches.I did not run multi player.Its fraps all runs will be different.So I guess I should buy dual CPU board and run 12 cores/24 thread and see how much better BC2 is.

@ViN86
Well I understand how to setup for a CPU bottleneck,damm theres always going to be a bottleneck somewhere.Its was just benches and it showing a older machine can run a 480GTX just fine.Thats all it is.

I agree some games are going to run better on more cores,there few games that do.The benches are for people wondering if a buy a better video how will the game play,most people can not afford new CPU/motherboard/ram for a few fps differeence in a couple couple games.

ViN86
07-31-10, 04:06 PM
@Madpistol
Thanks for the benches.I did not run multi player.Its fraps all runs will be different.So I guess I should buy dual CPU board and run 12 cores/24 thread and see how much better BC2 is.

@ViN86
Well I understand how to setup for a CPU bottleneck,damm theres always going to be a bottleneck somewhere.Its was just benches and it showing a older machine can run a 480GTX just fine.Thats all it is.

I agree some games are going to run better on more cores,there few games that do.The benches are for people wondering if a buy a better video how will the game play,most people can not afford new CPU/motherboard/ram for a few fps differeence in a couple couple games.
Well thanks for spending the time compiling everything. My point was just that you were comparing CPU's, so to determine bottleneck you need more than one setting for your graphics card. I was just wondering if you did that or used another source.

gerardfraser
07-31-10, 04:49 PM
Well thanks for spending the time compiling everything. My point was just that you were comparing CPU's, so to determine bottleneck you need more than one setting for your graphics card. I was just wondering if you did that or used another source.

Well the tittle of thread is misleading.Its not a true CPU scaling test.Basically bought rig to test some games and see if they play well on an older machine.

For the comparing part Iam just going from what I done with an I7 920 machine,but even that I always run trisli/trifire.Also compared from other peoples settings/reviews.

Maybe someone can try some benches same settings on I7 system and compare with the benches I done.

onmikesline
07-31-10, 07:19 PM
and he didnt even max res to 2560x1600 for the 30'' monitors. now your opening a new can of worms

ATOJAR
07-31-10, 07:45 PM
a core duo can play games as good as a New quad core I7 machine.

Lol. :headexplode:

Noriega
07-31-10, 08:14 PM
I would like to see some benchmarks from games that really need good CPU... ArmA 2 and GTA IV for example where 4 cores are welcome.

gerardfraser
07-31-10, 11:22 PM
Lol. :headexplode:

You trying to be an asshole ,I actually have a I7 system,so get off the glue nut bar.
Not trying to trick anyone.
I just tested a CPU E8500 with a 480GTX maybe some people will say can I upgrade my video card without there being too much off a bottleneck.
I say yes,its ok to do this.
Core I7 maybe faster sometimes,post your benchmark and prove me wrong.

gerardfraser
07-31-10, 11:26 PM
I would like to see some benchmarks from games that really need good CPU... ArmA 2 and GTA IV for example where 4 cores are welcome.
Sorry man I do not have them games,I am sure a quad core CPU would be better ,but by how much I do not know,I played arma 2 with 3 x285 GTX and it sucked and 5970+5850 and it sucked on a I7 920 @ 4.1 GHZ.Sucked as in bad FPS.

onmikesline
07-31-10, 11:29 PM
really this guy did a great thing to help others. that how i feel. but anything lower then a 8500 will bottleneck for sure... my poor old core2quad6600 :( didnt make it in life

JasonPC
07-31-10, 11:56 PM
I think arma II was not even optimized for quads until more recently so performance might have changed.

ViN86
08-01-10, 08:12 AM
Well the tittle of thread is misleading.Its not a true CPU scaling test.Basically bought rig to test some games and see if they play well on an older machine.

For the comparing part Iam just going from what I done with an I7 920 machine,but even that I always run trisli/trifire.Also compared from other peoples settings/reviews.

Maybe someone can try some benches same settings on I7 system and compare with the benches I done.

That was what I was trying to say. :o

Really, I appreciate all the work you did and I know what you meant. But it's not a true bottleneck test, in the literal meaning of the word.

Still great work though. It'd be nice to see the same tests done with your i7. :thumbsup:

Eltigre88
08-01-10, 08:14 AM
Don't mind them they have way too much time on their hands.....
I'm not very impressed with the new or at this point in reference to the 460 old Fermi cards. My 8800GTX was pretty close to my 470's stats. Still hoping as the drivers mature I will see a big difference. Next time just put JFA in your title and they won't get so wound up...