ASUS P5W DH Deluxe vs. Gigabyte GA P965 DS3 - final verdict

[K.I.T.T.]

Originally posted at www.maxitmag.com



A big thanks to Dan at maXit, who sent the board out to me. Without his help, this would not have been possible.

There's been a debate over which board is faster clock/clock. Most say the ASUS is faster, wheras others firmly believe nothing seperates the 2.

I've taken the liberty to run a series of tests on both boards and publish the results for the benefit of the general public.

These are the tests I decided to run:

3D Mark 06 (Default settings)

3D Mark 2001 SE (Default settings)

Aquamark 3 (Default settings)

Super Pi 32M (Realtime priority)

F.E.A.R. Multiplayer Demo version 1.02 (Computer - High ; Graphics - custom : 1152 x 864, Everything on maximum, no soft shadows, 4 x AA, 16 x AF)

All tests were run at 2 settings on both boards:

a) stock - 266 x 7 (1866 MHz), DDR2 667, 3-3-3-10

b) overclocked - 400 x 7 (2800 MHz), DDR2 800, 4-4-4-10

Hardware used was:

ASUS P5W DH Deluxe and Gigabyte GA-P965-DS3 motherboards

Intel Core 2 Duo E6300 ES (Allendale 2 MB L2 Cache)
2 x 1 GB Team Xtreem PC2 5300 RAM 3-4-4-8
256MB Connect 3D X1800XT
Maxtor DiamondMax10 80 GB HDD
NEC ND 3500 AG DVD-RW drive
Antec Truepower 2.0 550W EPS 12v PSU

Stock Cooling was used throughout.

OS: Clean Install of XP Pro SP2 was used with each board
Graphics Drivers: ATi Catalyst 6.6 official

The X1800XT was run at stock speeds for each test.

Results:

Stock Speeds

3DM2001 SE

ASUS


Gigabyte


3DM06

ASUS


Gigabyte


Aquamark 3

ASUS


Gigabyte


Super Pi 32M

ASUS


Gigabyte


F.E.A.R.

ASUS


Gigabyte


Overclocked


3DM2001 SE

ASUS


Gigabyte


3DM06

ASUS


Gigabyte


Aquamark 3

ASUS


Gigabyte


Super Pi 32M

ASUS


Gigabyte


F.E.A.R.

ASUS


Gigabyte





This took me quite some time to do, and I hope it is of use to you.

Cheers,
Ash

PS:

Pic of my testbed

[dinges]

very good stuff, the asus gets spanked the 32m oc'ed result is a bit weird though..

anyway, thnx for giving me another reason not to buy asus

[Dominus]

Very nice, thanks a lot!

Seems the Gigabyte will be the best choice for a gamer, until the RD500/600's and NF590's can be considered.

[celerity]

Interesting results, nice work!

[kman540]

Nice test. So at stock speed the Gigabyte was faster in SuperPI, OC'd the Asus was faster?

[dinos22]

i'm puzzled by that 32M OC result

Anus clearly gets spanked in majority of benchmarks except for that particular one..........can you rerun it pls and also can you play with voltages.......it would also be interesting to see other OC settings and results in 32M......really strange result

i'm really glad i got the Gigacrap as one first allendale/conroe board now

[K.I.T.T.]

Nice test. So at stock speed the Gigabyte was faster in SuperPI, OC'd the Asus was faster?

yep...i found that very strange but i tested twice on each board and got the same results....

Ah well...

EDIT: @dinos...

reran Pi 3 times, with the same result (1-2 sec difference)

[yotomeczek]

When overclocked boot the same bootsrtap and fsb ??

[NoSinder]

Awesome. DS3 FTW.

[K.I.T.T.]

don't know about the bootstrap.

the only thing i changed was the FSB.

[Jumba]

Great tests, thanks.

[Vric]

very nice job you did there.

I found the SuperPi overclocked result strange, but beside that it look to be constant. Gigabyte is faster. (or maybe more the 965P is faster)

[kman540]

Did you run Super PI 1m on both setups or just the 32m test?

[eva2000]

don't know about the bootstrap.

the only thing i changed was the FSB.

see that's where the key maybe at <400fsb the gigabyte might have tighter internal latencies especially at 266fsb look at AM3 cpu score differences

also seems 3dmark06 the key diff is much higher cpu score component on Gigabyte board

shame gigabyte board doesn't work properly with memset to see all the timings

I haven't tried either of these boards so all just guessing heh

[dinos22]

yep...i found that very strange but i tested twice on each board and got the same results....

Ah well...

EDIT: @dinos...

reran Pi 3 times, with the same result (1-2 sec difference)

interesting.......well i'm an SPI freak so i won't rest easy until i figure that part out......i'll do must bit on Tuesday next week

[N00B]

Really good test!

Tks!

[me too]

Really strange 32pi!

[/me too]

[dinos22]

same gigabyte board doesn't work properly with memset to see all the timings

how would you explain the SPI times while OCd if gigabyte had the advantage but lost pretty badly there

[Cooper]

KITT what cooler was on Conroe ?

[eva2000]

how would you explain the SPI times while OCd if gigabyte had the advantage but lost pretty badly there

look in screenies task bar area - alot of background apps open , any apps could attribute to slower times...

probably have to run each test 3 times within same session of windows and average them for better indictation

[Scarlet Infidel]

He says: Stock Cooling was used throughout.

[alpha0ne]

Thanks for you time and effort K.I.T.T.

You have really done eeryone @ XS a great service, thankyou

[kman540]

KITT what cooler was on Conroe ?

I believe he said "Stock Cooling was used throughout"

[N00B]

KITT what cooler was on Conroe ?

Stock Cooling was used throughout.

i think it was stock, unless he was talking about the graf card.

[misteroadster]

Good job kitt

[ArcTan]

nice test

I'm glad I went with the Gigabyte