9700 benchmarks

[epicstruggle]

Just saw this over at gamespot, and thought someone might find it interesting.
http://gamespot.com/gshw/stories/flat/0,12880,2877609,00.html
the interesting pages are:
http://gamespot.com/gshw/stories/flat/0,12880,2877609-7,00.html
and
http://gamespot.com/gshw/stories/flat/0,12880,2877609-8,00.html

 

[Pete]

Not too shabby.

@ the 10x7 4xAA #'s.

I'm curious if they used a 9700 or a 9700 Pro. Alienware's site doesn't even offer the 9700 as an option, but Best Buy advertises a new Alienware system with a 9700--maybe it's time to swing by and do some snooping. Where would I find the 9700's core and clock speeds? Display Prop.'s? Device Manager? DXDiag?

Ah yes, and the article seemed rather poor. They lambast the Alienware for not having USB2 ports, but fail to mention the biggest frickin' perk: the 9700, which on occasion doubles 4600's AA scores. I wonder if the 4600's loss of a TMU with FSAA (so it comes down to 4x1 vs 8x1) or the 9700's double bandwidth factor more into the disparity FSAA scores.

 

[MPI]

Damn, 10500 3dmarks with 4x FSAA... That's pretty damn good. But only 25% faster in UT2003 at 4x FSAA? Not quite CPU-bound either.

No word on FSAA quality yet? Screenshots?

 

[Doomtrooper]

Wait till Monday...

 

[RussSchultz]

Wait till Monday...

Story of my life.

 

[epicstruggle]

any suprises

was anyone suprised by any particular benchmark result for the 9700 good or bad?

 

[Doomtrooper]

There is lots of so called reviews/previews slipping out, some are accurate..some are not. The one above is not fully accurate..something is wrong with that setup.

Here is another 'preview'..engineering sample..beta drivers

 

[Sabastian]

Possibly they had driver conflicts.......

"The one exception is that we tested the Alienware system as it shipped to us, with the Radeon 9700, and we also tested it with the graphics card the rest of the systems had, a GeForce4 Ti 4600."

 

[Fuz]

Didn't read the whole article, but one thing that stood out was this.

It's interesting to note that the Radeon 9700 actually doesn't do that well in some of the lower-resolution tests, but since it wasn't shipping when we received the Alienware unit, there's no telling whether the drivers were final

The 9700 was scoring the same as a Ti 4600 at lower resolution with out AA (1024x768).
Any one with a half a brain would realise that the Ti 4600 and 9700 are CPU limited at low res.

 

[Nagorak]

Didn't read the whole article, but one thing that stood out was this.

It's interesting to note that the Radeon 9700 actually doesn't do that well in some of the lower-resolution tests, but since it wasn't shipping when we received the Alienware unit, there's no telling whether the drivers were final

The 9700 was scoring the same as a Ti 4600 at lower resolution with out AA (1024x768).
Any one with a half a brain would realise that the Ti 4600 and 9700 are CPU limited at low res.

Well, no one would make the mistake of thinking the people at Gamespot were technical geniuses, I guess.

 

[g__day]

17,421 3d Marks 2001SE == Hello Radeon 9700

And this with unoptimised drivers:

http://service.madonion.com/compare?2k1=4202539

120 FPS in the nature test seems...amazing

 

[MuFu]

LOL, bets on macci will break 20k with this card?!

 

[KimB]

Just one quick comment:

Unoptimized drivers? You can bet that 3DMark2k1 is the first program that ATI is optimizing for.

We'll only know in another six months just how optimized those scores are right now...

 

[Doomtrooper]

What are you saying Chalnoth, ATI is the only company that optimizes its drivers

 

[KimB]

I'm just saying don't expect 3DMark2001 performance to increase as much as, say, Alien vs. Predator 2 performance with new drivers.

 

[Geeforcer]

For all I care, 3dmark performance can decrease by a 1000 points if gaming performance goes up.

 

[SvP]

don't expect 3DMark2001 performance to increase as much as...

ExtremeTech did a comparison of various drivers of 8500.

http://www.extremetech.com/article2/0,3973,447739,00.asp

 

[KimB]

Which pretty much confirms my statement. Did you notice the massive increase in performance with later drivers on a couple of the actual games? The 3DMark2k1 score changed only about 10%.

 

[Fuz]

For all I care, 3dmark performance can decrease by a 1000 points if gaming performance goes up.

Hang on, I love playing 3dmark, its so much fun.
All joking aside, I couldn't agree more with Geeforcer.

 

[jb]

I'm just saying don't expect 3DMark2001 performance to increase as much as, say, Alien vs. Predator 2 performance with new drivers.

3dmarks != real gaming perfromance

Never has, never will. Only use 3dmarks to gauge and verify your own system. Its useless for anything else.