CPU bottleneck in games or not?

P

Pitbull

Limp Gawd
Joined
Aug 21, 2004
Messages
361
So i'm still running a e8400 ocd to 4250 mhz (8.5x 500) with 2x 2g of ocz reapers. Its water cooled since day 1 on a dual 120mm rad with 4 fans so temps are fine. I stuck a 6950 modded to 6970 in the rig recently. I play my games at 1920x1200 and im wondering if im bottleneck'ed by my cpu and by how much. I mean real world performance in games not artificial benchmarks like 3dmark or alike. Thinking of a sandy bridge build in the future, but not sure if i should hold of for the x78 chipsets.

Thanks for the input.
 
yes it will limit you in many cases but for the most part it is nothing to worry about. to check, all you have to do is lower your res and if the framerate remains the same then your cpu is the limitation.
 
I suppose technically you could say its a bottleneck cause something like an i7-950 would give you some more frames out of your games. But, a 4 gig Core 2 Duo aint nothing to sneeze at and would be more than capable of handling any game at your resolution with a 6970. Thats not to say you shouldnt upgrade. Sticking a Core 2 Quad in there would definitely help especially with a mean overclock like the one youve got now.
 
Thanks for the replies guys! I doubt i could find a quad these days that will hit 4.25g.
Is there a way to download a time demo for Black Ops/Battlefield BC2/Left 4 dead 2/home front or crysis 2 so i can bench at different resolutions? and how would i do it? I havent benched since cs:s and Quake 3.
 
you dont even have to run a benchmark. just fire up a game with fraps and look at the framerate. then lower the res and see if the framerate changes. I can tell you right now that you will be 100% cpu limited in many games such as BC2, GTA 4, Black Ops, Crysis 2 and such. also in most of those games I just listed you will not be able to average 60 fps because of your cpu. in most games though, especially like L4D 2, your framerate will be so high that it will be of no consequence.
 
nxcess said:
It'll depend on the games you play.
Click to expand...

Very True.

I run Second Life on a new 2600K \ GTX 570 system and SL brings the system to its knees, spits out the 2600K and says "Give me more power!" especially when your draw distance is set to 512 meters and a ton of people are in the same simulation.

That's why I wish we could double the 2600K in a system, but we can't. :(

And now, SL has added avatar physics in the Project Snowstorm viewer, so you can see bouncing breasts in the night clubs. :) Takes more CPU power to handle that kind of thing I can assure you.
http://www.youtube.com/watch?v=rt-zIMaMD24

Poor 2600K doesn't stand a chance.
 
Last edited:
cannondale06 said:
yes it will limit you in many cases but for the most part it is nothing to worry about. to check, all you have to do is lower your res and if the framerate remains the same then your cpu is the limitation.
Click to expand...

Another way to gauge cpu bottleneck is with taskmgr and msi afterburner (or whatever has an ingame gpu osd). If your gpu usage isn't 90% and up, then you're probably cpu limited. You should be able to look at taskmgr and see almost 100% cpu utilization
 
LittleTinyScooby said:
Very True.

I run Second Life on a new 2600K \ GTX 570 system and SL brings the system to its knees, spits out the 2600K and says "Give me more power!" especially when your draw distance is set to 512 meters and a ton of people are in the same simulation.

That's why I wish we could double the 2600K in a system, but we can't. :(

And now, SL has added avatar physics in the Project Snowstorm viewer, so you can see bouncing breasts in the night clubs. :) Takes more CPU power to handle that kind of thing I can assure you.
http://www.youtube.com/watch?v=rt-zIMaMD24

Poor 2600K doesn't stand a chance.
Click to expand...

Second life? Are you for real? Maybe it's a sign of a poorly design/shit game maybe?

But yes it really does depend of what game you play and whether or not it's designed for multicore processort as well. I'm generally fine with an overclocked i7.
 
Well im kinda thinking about going Q9450 and sell my Q6600, I play at 4320x900 and im sure my CPU is limiting me in games......
 
Timmy121 said:
Well im kinda thinking about going Q9450 and sell my Q6600, I play at 4320x900 and im sure my CPU is limiting me in games......
Click to expand...
unless you are getting it for almost nothing then don't even bother with such a miniscule upgrade.
 
one way to check do the settings drop to minimum and then setting it all back to high again test and see if the fps stay the same
 
Spooony said:
one way to check do the settings drop to minimum and then setting it all back to high again test and see if the fps stay the same
Click to expand...
no that is a silly and very inaccurate way to check. many settings impact the entire pc while some settings may only impact the cpu or gpu. as I already mentioned the EASIEST way to check is to lower the res and see if the framerate goes up. if it does not then the cpu is the limitation in that scenario.
 
Rossi~ said:
Second life? Are you for real? Maybe it's a sign of a poorly design/shit game maybe?

But yes it really does depend of what game you play and whether or not it's designed for multicore processort as well. I'm generally fine with an overclocked i7.
Click to expand...

Indeed, SL is making a comeback this year. Mesh imports are happening very soon and a new CEO means SL isn't fiddling around. You'll be hearing alot more about it this year.

But yes, the programming is bad, however, it is multi-core and has been for many years now, so it will use whatever you have, 8 cores or more. It was one of the first multi-core games released I believe.
 
cannondale06 said:
no that is a silly and very inaccurate way to check. many settings impact the entire pc while some settings may only impact the cpu or gpu. as I already mentioned the EASIEST way to check is to lower the res and see if the framerate goes up. if it does not then the cpu is the limitation in that scenario.
Click to expand...

well how can it be silly by dropping AA and all the gpu related settings and then push them up again? If they stay the same you know its on the cpu side. how can that effect the whole system? Does your whole system got to do with AA? No
 
Best purchase i made was from all the guys here at HARD convinced me to buy a 2600K or 2500K ( i went with the 2600K i got a sweet deal on it ) and Overclock it and save money for other gear. Well the 2600K overclocks to 5Ghz, im running 24/7 at 4800Mhz on 1.3 V, it blows my previous chip Q6600 outta the water.
 
cannondale06 said:
unless you are getting it for almost nothing then don't even bother with such a miniscule upgrade.
Click to expand...

Well yeah i know, i sold my E6300 for $55 and trying to sell my Q6600 atm for $95 shipped then use that money for a Q9450
 
Spooony said:
well how can it be silly by dropping AA and all the gpu related settings and then push them up again? If they stay the same you know its on the cpu side. how can that effect the whole system? Does your whole system got to do with AA? No
Click to expand...
because other than AA, many settings also impact the whole system and sometimes just cpu. so again why in the heck would you go through the trouble for an inaccurate result when lowering the resolution takes 2 seconds?
 
Just upgraded from a 9450 @ 3.6 to a 2500K at 4.1. Just got it, haven't really pushed it yet. Anyway, yeah. MUCH faster. My 5850 gpu runs 99% all the time now. Fps jump was huge. As much as a 50 fps avrg difference in bfbc2. That is with higher settings too.

Can't wait to see what my incoming 480 GTX SOC does with this chip.

You could get a 2500k and a cheap gpu like the 5850 and still play any game maxxed out at 1920x1080 all day. Skip the 9450 imo. Just save up for a few months like I did.
 
cannondale06 said:
yes it will limit you in many cases but for the most part it is nothing to worry about. to check, all you have to do is lower your res and if the framerate remains the same then your cpu is the limitation.
Click to expand...

+1

This way, you can check in all games as well to see if any are CPU limited.

I doubt a game like L4D 2 would be CPU limited, but some of the newer titles out will be hard on a CPU, and this method of testing is bar-none the easiest way to tell.

@Spooony: Changing in-game settings like that won't help as it will vary from game to game. Checking on a hardware-level is much easier.
 
spaceman said:
Just upgraded from a 9450 @ 3.6 to a 2500K at 4.1. Just got it, haven't really pushed it yet. Anyway, yeah. MUCH faster. My 5850 gpu runs 99% all the time now. Fps jump was huge. As much as a 50 fps avrg difference in bfbc2. That is with higher settings too. .
Click to expand...

Thanks for verifying this. I've been telling others a C2Q will bottleneck a 5850 / 5870 in BFBC2. Even I feel it with my system. Bulldozer needs to hurry up so I can decide.
 
Unless your playing wack ass games, then no you arent being CPU limited 'very' much. Its not going to be worth upgrading, it would be better to hold out for ivy bridge. Most of the pc ports from consoles are running on hardware thats several generations beyond what the consoles run on. Beyond that, its only going to be RTS games or lame second life type games that your being limited by. In FPS your almost always being limited by your GPU, i'd say the same for RPG's nowadays as well.

/flame on
 
Just ran Crysis 2 bench looks like im limited a bit but nothing too horrid. I guess i'll holdout for ivy bridge, or if i get some better cash flow i'll grab a 2600K.

Setting "User/Graphic Options/resolution" => "1024x768"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 25.818079
FrameCount = 2000
AverageFps = 77.465096
MemoryWorkingSet = 1176
MemoryPageFile = 1291
MemoryPageFaults = 546249
Benchmark ended

Setting "User/Graphic Options/resolution" => "1280x960"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 26.203230
FrameCount = 2000
AverageFps = 76.326469
MemoryWorkingSet = 1206
MemoryPageFile = 1305
MemoryPageFaults = 955812
Benchmark ended

Setting "User/Graphic Options/resolution" => "1920x1200"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 35.675610
FrameCount = 2000
AverageFps = 56.060707
MemoryWorkingSet = 1234
MemoryPageFile = 1318
MemoryPageFaults = 1365731
Benchmark ended
 
Last edited:
Tried now with maxing AF at 16x and adding AA. Previous settings were extreme with modified AA to off and AF to 4x.

Setting "User/Graphic Options/resolution" => "1024x768"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 25.419010
FrameCount = 2000
AverageFps = 78.681274
MemoryWorkingSet = 1173
MemoryPageFile = 1289
MemoryPageFaults = 540596
Benchmark ended

Setting "User/Graphic Options/resolution" => "1280x960"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 26.197399
FrameCount = 2000
AverageFps = 76.343452
MemoryWorkingSet = 1201
MemoryPageFile = 1306
MemoryPageFaults = 950277
Benchmark ended

Setting "User/Graphic Options/resolution" => "1920x1200"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 35.100719
FrameCount = 2000
AverageFps = 56.978886
MemoryWorkingSet = 1228
MemoryPageFile = 1323
MemoryPageFaults = 1368098
Benchmark ended
 
Last one. Full extreme settings.

Setting "User/Graphic Options/resolution" => "1024x768"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 24.337000
FrameCount = 2000
AverageFps = 82.179398
MemoryWorkingSet = 1167
MemoryPageFile = 1289
MemoryPageFaults = 533685
Benchmark ended

Setting "User/Graphic Options/resolution" => "1280x960"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 26.241219
FrameCount = 2000
AverageFps = 76.215973
MemoryWorkingSet = 1194
MemoryPageFile = 1312
MemoryPageFaults = 949976
Benchmark ended

Setting "User/Graphic Options/resolution" => "1920x1200"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 35.195850
FrameCount = 2000
AverageFps = 56.824883
MemoryWorkingSet = 1212
MemoryPageFile = 1323
MemoryPageFaults = 1363597
Benchmark ended
 
Pitbull, is there any chance you could run the full extreme settings benchmark at 1920x1200 and 1680x1050 with your cpu at stock 3.0 or at least close to that speed? techspot had the 3.16 E8500 only averaging 23fps even with a gtx590 at 1920x1200 and think their results are bogus.
 
Hedonsim Bot said:
Could I ask were you got the crysis 2 benchmark from so I can compare my system?
Click to expand...

in crysis directory there is benchmark.bat run that :D

It kept giving me an error about not being able to access benchmark_result.txt so i just created the file in the dir (win 7 requires admin rights to create folders/files in program files dirs.)
 
with c2d at default memory at 800mhz
Heh not much of a difference. That's quite odd.

Setting "User/Graphic Options/resolution" => "1024x768"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 32.729649
FrameCount = 2000
AverageFps = 61.106674
MemoryWorkingSet = 1168
MemoryPageFile = 1289
MemoryPageFaults = 532653
Benchmark ended

Setting "User/Graphic Options/resolution" => "1280x960"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 32.589600
FrameCount = 2000
AverageFps = 61.369270
MemoryWorkingSet = 1186
MemoryPageFile = 1302
MemoryPageFaults = 941684
Benchmark ended

Setting "User/Graphic Options/resolution" => "1920x1200"
****************************************************
Benchmark started: benchmark.cfg
****************************************************
PlayTime = 39.299587
FrameCount = 2000
AverageFps = 50.891117
MemoryWorkingSet = 1209
MemoryPageFile = 1318
MemoryPageFaults = 1349707
Benchmark ended
 
You must log in or register to reply here.
Back
Top