Sign in with
Sign up | Sign in
Your question
Closed

Do you think the CPU matters in games?

Tags:
Last response: in CPUs
Share
a b à CPUs
December 29, 2010 11:05:15 PM

Okay I ran 3Dmark06
setup is
Pentium D 945 3.4 ghz
2gb PC2-5300
Sapphire 1gb HD 4650 (shows as 4670 in 3DM06)

Here is link to my score (score 5189)
http://3dmark.com/3dm06/15221536?show_ads=true&page=%2F...

here is link to other scores with I5-860s,C2Qs etc
using a HD4670
http://3dmark.com/search?resultTypeId=14&linkedDisplayA...

Looking at the scores an I7 or C2Q or Athlon x 3 etc only score about 200 points or so (roughly 5400) above my
score of 5189.

Now OBVIOUSLY any of these CPUs are far superior to mine in performance YET in a gaming benchmark using the same GPU barely made any difference.

In fact inferior cpus wih better GPUs slaughter the I7s,Athlon x 3 with the inferior GPUs etc

And really the I7-860 only made about a 300 point difference in 3DM06 (which in that scoring system is a minimal difference)

Something to think about isnt it?




More about : cpu matters games

a b à CPUs
December 29, 2010 11:26:53 PM

Some games are more dependant on a good CPU than the Graphics Board, a good example is ArmA2!

Best solution

a c 78 à CPUs
December 30, 2010 12:28:16 AM
Share

Actual, you are seeing two things,

1) the bias of 3DMark06. It favors higher clock speed of the CPU versus the number of cores. I would try the test with 3DMark Vantage or 3DMark11 and you'll see a much bigger difference. When 3DMark06 was released, Tri & Quad cores weren't really out or just hitting the market, so it wasn't optimized for them.

2) bottlenecking of the GPU to the CPU. The weaker GPU is holding back the much more powerful CPU. You would see a similar result with running your CPU with a GTX580.OT would perform worse than one driven by an i7 860. Your CPU is bottlenecking your GPU.
Related resources
December 30, 2010 12:31:39 AM

Of course CPU matters - without it, your computer wouldn't even be able to turn on, let alone run a game.
a b à CPUs
December 30, 2010 12:55:40 AM

tecmo34 said:
Actual, you are seeing two things,

1) the bias of 3DMark06." It favors higher clock speed of the CPU versus the number of cores""
. I would try the test with 3DMark Vantage or 3DMark11 and you'll see a much bigger difference. When 3DMark06 was released, Tri & Quad cores weren't really out or just hitting the market, so it wasn't optimized for them.


2) bottlenecking of the GPU to the CPU. The weaker GPU is holding back the much more powerful CPU. You would see a similar result with running your CPU with a GTX580.OT would perform worse than one driven by an i7 860
."""" Your CPU is bottlenecking your GPU"""".


1) Really then two cores of an I7-860 are not much better than 2 cores of a Pentium D 945 3.4ghz?!
remember the scores with the Pentium D and the I7-860 were close with the SAME GPU (HD4650 1gb)
I know that a I7-860 KILLS a Pentium D easily.


2) Definitely not true.
I ran the HD 4650 at stock speed and then I OCd it.
My score got much better with the OC settings.
That would tell me that the CPU wasnt a bottleneck.
a b à CPUs
December 30, 2010 12:56:48 AM

GoldenI said:
Of course CPU matters - without it, your computer wouldn't even be able to turn on, let alone run a game.


Seriously?!
Face Palm
LOL :D 
a c 78 à CPUs
December 30, 2010 1:21:28 AM

king smp said:
1) Really then two cores of an I7-860 are not much better than 2 cores of a Pentium D 945 3.4ghz?!
remember the scores with the Pentium D and the I7-860 were close with the SAME GPU (HD4650 1gb)
I know that a I7-860 KILLS a Pentium D easily.

2) Definitely not true.
I ran the HD 4650 at stock speed and then I OCd it.
My score got much better with the OC settings.
That would tell me that the CPU wasnt a bottleneck.

1) Your comment just proved my point... They higher native clock speed of the Pentium plays a bigger factor than the amount of cores of the i7 860 (4-cores). You overclock the i7 860 to the same speed and you should notice a bigger difference.

2) It just shows that you see an increase in performance with a "faster" GPU. This would be natural, just like you see an increase by the faster CPU (i7 860). You would see a difference in performance with a GTX 580 with the two CPU's, as well. There is a bottlenecking happening between the i7 860 & ATI 4650, believe me or not :) 
December 30, 2010 2:00:58 AM

king smp said:
Seriously?!
Face Palm
LOL :D 

Was being a bit of a smart-arse. ;) 
a b à CPUs
December 30, 2010 2:07:47 AM

GoldenI said:
Was being a bit of a smart-arse. ;) 



I know :D 
I was LMAO
a b à CPUs
December 30, 2010 2:20:40 AM

tecmo34 said:
1) Your comment just proved my point... They higher native clock speed of the Pentium plays a bigger factor than the amount of cores of the i7 860 (4-cores). You overclock the i7 860 to the same speed and you should notice a bigger difference.

2) It just shows that you see an increase in performance with a "faster" GPU. This would be natural, just like you see an increase by the faster CPU (i7 860). You would see a difference in performance with a GTX 580 with the two CPU's, as well. There is a bottlenecking happening between the i7 860 & ATI 4650, believe me or not :) 


Oh OK
I thought that you meant the HD4650 was bottlenecked by the Pentium D
I guess that the I7-860 is capable of sending more data than the HD 4650 can handle?

I just cant believe that if you take one core of the I7-860 and matched it up to one core of the Pentium D 945 3.4
that the I7 wouldnt slaughter the Pent D even with the speed difference.
I own the PentD and while it doesnt deserve all of the negativity it gets it is still way inferior architecture (sp) to the I7 obviously.
It tells me IMO that in gaming benches that the CPU doesnt affect the rendering as much as I would of thought.
I am sure in a synthetic CPU intensive bench that one core of the I7-860 would easily outperform the the Pentium
D 3.4 one core.
Now of course when you get into games that are optimally threaded for multicores that the I7-860 would shine.

The point of this exercise was to show that if you have a reasonably capable chip (E6800 or higher) that when it is time to upgrade for gaming purposes definitley spend the money on GPU instead of CPU unless it is a seriously CPU bound game.
a b à CPUs
December 30, 2010 3:36:54 AM

Quote:
It matters a lot in bad ports like GTA4 and games we will use as a benchmark for a year or two.
Then you get games like Crysis where the developers had a dream about hardware that would max out their game in the distant future.


^+1 True
But overall in most games made up until 2009 and even most games in 2010 with the exception of GTA4, MS Flight Sim if you
have a reasonable CPU then it is all about the GPU.
If the primary purpose is gaming then going from a E6800 to a I3 (for example) or a Q9550 to a I5/I7 in most games will probably have little impact for the money spent.
Take a Q8400 or E6800 with a HD 6970 or GT 580 and you can play ALMOST anything cranked way up.
Before the AMD guys get mad I just want to say I am most familiar with Intel.
I am sure a 965 BE with a HD 6970 will crank it out too. :) 
December 30, 2010 6:14:56 AM

omg... these dipsh**s are now spamming in our forums?!?!

we need better security here....
anyway i hope your right about having a better gpu is more efficient then getting a new cpu since i just got a E8500 for my gtx 460 and it can run anything i throw at it, hopefully this will last at least 1 more year b4 i have to buy a brand new pc.
Anonymous
a b à CPUs
December 30, 2010 7:32:45 AM

omg... these dipsh**s are now spamming in our forums?!?!

we need better security here....




Yes, the spammers are still here but I think in decreasing numbers as we have a team of Moderators who are working tirelessly to zap them. Thanks to all who report spam -- keep it up as it makes life easier for everyone.
December 30, 2010 12:33:01 PM

this thread is a fail.

Now OBVIOUSLY any of these CPUs are far superior to mine in performance YET in a [b said:
gaming benchmark using the same GPU barely made any difference.]Now OBVIOUSLY any of these CPUs are far superior to mine in performance YET in a gaming benchmark using the same GPU barely made any difference.
[/b]

synthetic benchmarks != gaming benchmarks
a b à CPUs
December 30, 2010 2:32:08 PM

Depends really.Yes and no.Frame rates are influenced such that a celeron cannot beat an i7 in modern games.
a b à CPUs
December 30, 2010 3:36:50 PM

ghnader hsmithot said:
Depends really.Yes and no.Frame rates are influenced such that a celeron cannot beat an i7 in modern games.


Of course true
But if the CPUs are reasonably close in architecture (sp) then GPU is the primary factor in determining framerate.
If the game utilizes two cores and you have a fast C2D and I3 then the FPS difference between the two
CPUs is minimal if you have the same GPU especially at higher resolutions..
a b à CPUs
December 30, 2010 3:41:42 PM

wh3resmycar said:
this thread is a fail.



synthetic benchmarks != gaming benchmarks


I believe I was talking about the difference between CPU specific benches vs gaming benches (CPU/GPU)
There is a difference between 3DMark06 and PCMark Advantage
You can have a system that isnt so hot at a gaming benchmark but slaughter a CPU benchmark.
Dual Xeon/Opterons come to mind as examples.
They could destroy an encoding bench with alot of threads but get beaten by a really fast dual core in
a gaming bench
a b à CPUs
December 30, 2010 7:39:19 PM

Best answer selected by king smp.
a b à CPUs
December 30, 2010 8:06:09 PM

See your point King smp.
Sadly as i See it talking about a game, or raw cpu bench.
Alot of programs are not coded for sole use of one Cpu or Gpu.
How code is optomized, relates to the Arcutecture, and how each cpu/gpu executes.
But to create a bit of code that is optomized for each cpu, or a Gpu.
That will be the day !
Most coders go for middle of the road. All format.
a b à CPUs
January 28, 2012 2:47:12 PM

This topic has been closed by Mousemonkey
!