4850 to 5850 w/ e8400 Results!

[Gabkicks]

Some benchmark runs I did. I used 9.9 drivers for the 4850 and latest beta driers for 5850.

3DMark 06 default settings
3.7ghz+ 4850: 3DMarks: 12588 SM2.0: 5147 SM3.0 5860
3ghz + 4850: 3Dmarks: 11849 SM2.0:5015 SM3.0 5852

3.7ghz+ 5850: 3DMarks: 17820 SM2.0: 8093 SM3.0 9514
3ghz+ 5850: 3DMarks: 15039 SM2.0: 6624 SM3.0 8506

Furmark: 1680x1050 4xAA
3.7ghz + 4850: min 20 max 33 avg 23
3ghz + 4850: min 20 max 33 avg 23

3.7ghz+ 5850: min 42 max 72 avg 50
3ghz+ 5850: min 42 max 72 avg 50

Tropics Unengine DX 10.1 1680x1050 8xAA 16xAF max
3.7ghz + 4850: avg 13.6 Score: 342
3ghz + 4850: avg 10.6 Score: 266

3.7ghz+ 5850: avg 46.3 Score: 1180
3ghz+ 5850: avg: 46.3 Score: 1166
.................................................. .................

Resident Evil 5 1680x1050 8xAA 16xAF
3.7ghz+ 4850: avg 58.4
3ghz + 4850: avg 50.4

3.7ghz+ 5850: avg 62.5
3ghz+ 5850: avg 51.3

Street Fighter 4 1680x1050 8xAA
3.7ghz+ 4850: avg 68.04 Score: 6748
3ghz + 4850: avg 67.9 Score 6740

3.7ghz+ 5850: avg 86.1 Score: 7692
3ghz+ 5850: avg 87.01 7761 (***?)

GTA IV 1680x1050 All high, Medium Texture) View distance 50
3.7ghz+ 4850: avg 47
3ghz + 4850: avg 39

3.7ghz+ 5850: avg 47.8 (same w/ high textures and higher settings)
3ghz+ 5850: avg 42.8

Crysis Warhead: train benchmark 1680x1050(veryhigh? AA?)
3.7ghz+ 4850: min 5 avg 13 max 15
3ghz + 4850: min 5 avg 13 max 15

3.7ghz+ 5850: min 19 avg 21 max 29
3ghz+ 5850: min 18 avg 21 max 29

[OnDborder]

Thanks for the comparison :thup:
The 5850 is quite an improvement.

[chevro1et]

Thanks for the comparison :thup:
The 5850 is quite an improvement.

+1 on both comments. Just gotta say, wow, really quite impressed with the difference between the two cards.

[SuperDave1685]

wow, quite an improvement! Although with GTA IV, there doesn't seem to big a huge difference. Any ideas as to why?

[DrSpanky]

wow, quite an improvement! Although with GTA IV, there doesn't seem to big a huge difference. Any ideas as to why?

i dont have the game but from what i have read the game is very cpu intensive.

[Gabkicks]

yeah, i gain more from oc'ing my cpu than from putting in a brand new graphics card. I was thinking of trying toget a q9550 and/or selling my e8400 :) My motherboard is mediocre when it comes to overclocking, so i cant go above 420FSB... I was thinking maybe get a q9550 and run it at 3.5ghz to delay me having to buy new CPU/MOBO/RAM :p.

[Bobnova]

GTA4 is so insanely CPU limited it's not even funny, that's the issue there.

Would you be willing to OC both cards and run it again?

[Rich'[ard]]

GTA4 is so insanely CPU limited it's not even funny, that's the issue there.

Would you be willing to OC both cards and run it again?
mm yeah, you need an i7 to run that game to be called being able to run that game 'well'.
the GTA4 in-game benchmark said i was on 100% CPU usage, 56% GPU usage.
even if i'm on a outdated/measly 3.4ghz e6400 (2mb L2 cache as well)...
i think it's not even clock speed related, GTA4 performs much better with a quad core. just some figures, i remember reading somewhere it performs better on a stock Q8300 (2.8ghz?) over a 3.6-4ghz e8400.

EDIT: look at the min fps difference in Crysis warhead! whoa...lol

[Bobnova]

In a well written multi-threaded game you can add the clocks together for a very rough speed estimate. As such 4x2.3 is along the lines of 2x4.6. More cores makes a LOT more power if they are well used.


Yes, i know this isn't even close to exact, but it's good enough.

[Rich'[ard]]

In a well written multi-threaded game you can add the clocks together for a very rough speed estimate. As such 4x2.3 is along the lines of 2x4.6. More cores makes a LOT more power if they are well used.


Yes, i know this isn't even close to exact, but it's good enough.
mm yeah. even if the game is a PITA to run, at least it scales with better hardware. i'm sure in a year's time it'll run like smooth. people with 4ghz i7 920's and 5870s should run this game fine. it's just us folks with mid-range dual cores and mid-range graphics cards that'll run this game not as well as other games we can run.

ey Gabkicks, how are the temps with your new card?
does it still go up to the 90c range when loaded?

[jason4207]

Very nice comparison! :thup:

It's easy to spot the CPU limited games.

The 5850 is a beast! OC that animal :burn: and report back!!!!

[Gabkicks]

I am trying to figure out how to flash my BIOS so i can OC higher. I was using GPUZ to try and figure out my max temps, but is there a better program? I am confused by which temp to use with gpuz since so many show up.
I think it gets into the mid to low 80's for temps at full load running at 775/1125. its about the same at stock speeds. Sofar, i've just left fan speed at default variable. it runs very quietly.

Is there a program like realtemp for gpu that lists max and min that gpu has hit over a period of time?

[chevro1et]

Tried RivaTuner? Not sure if its been updated to support the 5xxx series yet tho.

[ghost_recon88]

This could make a great review *hint hint* ;)

[Gabkicks]

I'll try to put up a review w/ a few bar graphs and pictures between Tuesday and Thursday. I haven't touched excell in a while, but maybe i can throw together something buy Tuesday :)

I'm a bit scared to flash the card and raise voltages. i'll have to be vewy careful.

[Rich'[ard]]

I am trying to figure out how to flash my BIOS so i can OC higher. I was using GPUZ to try and figure out my max temps, but is there a better program? I am confused by which temp to use with gpuz since so many show up.
I think it gets into the mid to low 80's for temps at full load running at 775/1125. its about the same at stock speeds. Sofar, i've just left fan speed at default variable. it runs very quietly.

Is there a program like realtemp for gpu that lists max and min that gpu has hit over a period of time?
you could try Furmark. leave the test for 10 minutes or so and it should display min/max temps on the go...i think anyway.

[rezy]

this 5850 is insane, perfomanse are doubled than 4850 :eek: