[Personal test] Effect of CPU Freq (Q66 from 2.4 GH - 3.9 GH ) on VANTAGE and CRYSIS

[Gaul]

Q 66 ( 2.4 GH - 3.9 GH )
4870 ( def ) 750/900


Q 66 @ 2.4 GH



Q 66 @ 3.2 GH



Q 66 @ 3.4 GH



Q 66 @ 3.8 GH



Q 66 @ 3.9 GH

[Gaul]

Q 66 @ 2.4 GH, AVG = 30 FPS



Q 66 @ 3.2 GH, AVG = 30 FPS



Q 66 @ 3.8 GH, AVG = 30 FPS



WELL, any comment ?

[Leeghoofd]

Can you compare FarCry 2 ? Never liked Crysis anyway

[m0da]

we can draw two possible conclusions regarding your system:
1. vantage is cpu-limited (higher cpu freq -> higher score)
2. crysis is most likely gpu-limited (higher gpu freq -> higher fps)

if you repeat the tests, but vary your gpu frequency, we can see whether our statements are wholly true, or if we need to alter them.

[RPGWiZaRD]

Well I still see Crysis as among the most GPU performance limited games, there are always people telling you CPU will also matter but perhaps it does but for Crossfire HD 4870x2 or GTX 285/295 SLI and alike setups only in that case. For one GPU, it'll be GPU limited for sure.

Of course that doesn't mean you can use some old single core Pentium 4 CPU, but say E8400 @ stock already is prolly enough to be severely GPU limited with a single HD4890 or GTX 275 already.

[blaxtr3m3]

I c that your gpu is limited to 37 FPS max
While your minimum nearly increase 4 FPS thats 13% improvement in your minimum FPS.

[NaMcO]

Try disabling two cores on a QUAD and see Crysis/Far Cry/FSX (yes, FSX)/Whatever perform the same. Not valid for benchmarks obviously. I did that a while back, and guess what, i'm running with two cores since then: Less power draw, less heat and same performance on what i use it for. I'll enable them when i need them in the future, lol.

[Gaul]

@ all = thanx 4 u'r comment ! that was my old 4870, i've already sold it, i can't test anything now, now i have only 4890.....

[Gaul]

we can draw two possible conclusions regarding your system:
1. vantage is cpu-limited (higher cpu freq -> higher score)
2. crysis is most likely gpu-limited (higher gpu freq -> higher fps)

if you repeat the tests, but vary your gpu frequency, we can see whether our statements are wholly true, or if we need to alter them.

point 1. = yes....but only for CPU AND OVERALL SCORE, not GPU, even i used i7 @ 4.5 GH the gpu still 9000 !

point 2 = with q66, at 2.4 gh its i call bottleneck, but with 3.2-even q66 @ 4GH still have smiliar perform i think !

[Chumbucket843]

in real gameplay in crysis it becomes more cpu dependent when there are many enemies and physics.

[Gaul]

in real gameplay in crysis it becomes more cpu dependent when there are many enemies and physics.

I think Crysis not support PHYSIC !!

[Spork!]

What am I missing?
I have a lowly e5200 that gets 21255 (cpu points) @ 3.5GHz and 29703 @ 4.0 on what appears to be the same benchmark (3d Mark Vantage) My fsb is also lower than yours, @ 1333. My gfx card is a 9800GT so that shouldn't be giving me any advantage either, esp in CPU test...

[tool_462]

Just to show what Far Cry 2 results are like, I wrote a short article on the scaling of cores in FC2 using my "normal" gaming settings and then very low settings to negate GPU bottleneck.

"Normal" settings.



"Low" settings.




Source: http://www.xcpus.com/reviews/121-Far...ng-Page-1.aspx

[NaMcO]

Looks like a waste of power for the increase in performance, specially when you play at 1920x1200 like me.

[Gaul]

What am I missing?
I have a lowly e5200 that gets 21255 (cpu points) @ 3.5GHz and 29703 @ 4.0 on what appears to be the same benchmark (3d Mark Vantage) My fsb is also lower than yours, @ 1333. My gfx card is a 9800GT so that shouldn't be giving me any advantage either, esp in CPU test...

oh man !, don't u know nvidia use physic !

[Spork!]

It makes THAT much difference? I can understand my GPU keeping up with, or marginally beating your 4870 in a nVidia biased bench - Haven't been tracking my GPU scores in vantage, but they are around the 9-10k mark from memory. (core clock @ 667, shader @ 1667, mem. 900[ish]) So the GPU (very strongly) influences CPU score in CPU tests in Vantage? They should call CPU test a "combined" test then...
If it is indeed the power of the green card making such a big difference does it also make gaming twice as good in physX games? Is it possible to have a nVidia card AND a Redeon card (and their drivers) in the same PC and switch between the 2 for different applications/games?
Still a n00b, so please forgive me if my questions seem stupid - I won't learn the answers if I don't ask them. Thx for your patience.

[Gendo]

Spork I don't think you understand what he's trying to say: with a nvidia card and physx enabled your gpu calculates the cpu test, not the cpu. Try dsabling physx in the nvidia control panel and run the test again, you'll understand then

[randa]

Yups, it's clear enough that a game like Crysis is mostly a GPU limited game, From the 3 SS we can see the ave FPS, except for Min FPS it really gets a bit higer for a 3,8 Ghz clock. Could it be that CPU only help to keep the game runs smoother than that of a lower clock ...CMIIW .

@GAUL : See u in your next personel test

[Jamesrt2004]

It makes THAT much difference? I can understand my GPU keeping up with, or marginally beating your 4870 in a nVidia biased bench - Haven't been tracking my GPU scores in vantage, but they are around the 9-10k mark from memory. (core clock @ 667, shader @ 1667, mem. 900[ish]) So the GPU (very strongly) influences CPU score in CPU tests in Vantage? They should call CPU test a "combined" test then...
If it is indeed the power of the green card making such a big difference does it also make gaming twice as good in physX games? Is it possible to have a nVidia card AND a Redeon card (and their drivers) in the same PC and switch between the 2 for different applications/games?
Still a n00b, so please forgive me if my questions seem stupid - I won't learn the answers if I don't ask them. Thx for your patience.

ignroe that bench it uses the gpu to calculate it as has been said, turn OFF physx in the control panel then try doing the bench again adn youll see a real score thats comparable buddy

[Gaul]

It makes THAT much difference? I can understand my GPU keeping up with, or marginally beating your 4870 in a nVidia biased bench - Haven't been tracking my GPU scores in vantage, but they are around the 9-10k mark from memory. (core clock @ 667, shader @ 1667, mem. 900[ish]) So the GPU (very strongly) influences CPU score in CPU tests in Vantage? They should call CPU test a "combined" test then...
If it is indeed the power of the green card making such a big difference does it also make gaming twice as good in physX games? Is it possible to have a nVidia card AND a Redeon card (and their drivers) in the same PC and switch between the 2 for different applications/games?
Still a n00b, so please forgive me if my questions seem stupid - I won't learn the answers if I don't ask them. Thx for your patience.

hope now u understood ! peace !

@ gendo,jamesrt2004 = thanx 4 the answer
@ randa = well ...i'll do my best !