View Full Version : GeForce FX dominates R9700Pro at high CPU speeds!
surfhurleydude
08-Mar-2003, 18:51
http://www.bit-tech.net/review/187/1
A little short review of the GeForce FX using a standard stock AMD Athlon 3000+ CPU and... a 3.72 gHZ Pentium 4! It yields some interesting results...
In 3dmark2001, the GeForce FX is only 500 points ahead of the Radeon 9700 Pro using the Athlon based system (16081 vs. 15546)... Then when using the Intel system, the gap is dramatically widened the GeForce FX scoring 18439 and the Radeon scoring 16846.
Then during the 1280x1024 UT2003 Abestos flyby, the Radeon 9700 Pro scores 176.72 FPS on the AMD system while the GeForce FX scores 214.52 FPS! Then on the Intel system, the Radeon 9700 Pro gets 178.23 FPS, and the GeForce FX scores 240 Frames Per Second...
These are some pretty odd scores, but it seems as if in every test with the Intel CPU, the GeForce FX clearly dominates the Radeon 9700 Pro..
Any thoughts?
Personally I'd wait for more reviews on higher CPU's before making a conclusion based on one website's results. I'm just wondering if maybe the drivers are highly optimized for taking advantage of Intel CPU's, or if CPU limitation has been a factor all along. If it's limitation, then my P4 2GHz probably won't help much.
Are we going to see the day when high-end video cards can only be evaluated properly on super fast CPU's that most people can't afford? I hope not. :?
demonic
08-Mar-2003, 19:00
Yes, how about using AA and AF with those benchmarks...
We all know, because of the higher Core Mhz.. of the GFFX vs the R9700, its gonna with, with vanilla settings. :roll:
The funiest thing about that review is the situation in which Nvidia comes out on top -- pair an unobtainable video card with an unobtainable cpu. That makes perfect sense. :roll:
surfhurleydude
08-Mar-2003, 19:08
Yes, how about using AA and AF with those benchmarks...
We all know, because of the higher Core Mhz.. of the GFFX vs the R9700, its gonna with, with vanilla settings.
Check the link, there are AA (2x and 4x) and AF (8x) benchmarks too, and the GF FX wins by a fairly large margin there as well.
And another thing, a "bleeding edge/brand new" video card better damn well beat a 7 month old video card. If it doesn't, it's just laughable.*
*Oh wait a second, the situation is laughable when you enable the image quality features such as AA and AF. [2x AA doesn't count.]
THe_KELRaTH
08-Mar-2003, 19:19
hmmm.. they do seem rather odd. I've got an Aopen Granite Bay mobo with a P4-2.4b with 9700Pro 324/310
3Dmark2001 Standard = 14594
CPU o/c (only) to 2.7 = 15705.
3.70Ghz = 16842 (His score using a 1Ghz faster CPU)
*edit* just quickly tested at 160fsb
2.88Ghz = 16105
I don't see how 480Mhz increase gives a 3Dmark score increase of 1511 yet an increase of a further 820Mhz and faster memory only adds 737
His scores seem odd to me, as on my AMD XP2400+ o/c to 2.3Ghz with a small overclock on the 9700Pro [339/330 from 325/310] on the old Cat 2.5 drivers I scored slightly over 17.1K. (http://service.futuremark.com/compare?2k1=5694705) I would think a Barton 3K with the extra cache and P4 with its extra memory bandwidth would be closer to my score. Something seems odd indeed.
surfhurleydude
08-Mar-2003, 20:38
Why was this thread moved when there is a bunch of Radeon 9800 Pro review threads in the other forum???? Hmm...
hmmm.. they do seem rather odd. I've got an Aopen Granite Bay mobo with a P4-2.4b with 9700Pro 324/310
3Dmark2001 Standard = 14594
CPU o/c (only) to 2.7 = 15705.
3.70Ghz = 16842 (His score using a 1Ghz faster CPU)
*edit* just quickly tested at 160fsb
2.88Ghz = 16105
I don't see how 480Mhz increase gives a 3Dmark score increase of 1511 yet an increase of a further 820Mhz and faster memory only adds 737
even more so take into accont that i manage 13132 with my little xp@1636. that is just a mere 136mhz fsb on a kt400 board with mem@408mhz. heck, i even found a few people beating the tech-bit score with ti4200s (http://service.futuremark.com/servlet/Index?pageid=/orb/projectsearch), so their numbers don't sit very well with me.
ATI Catalyst 3.1
NVIDIA Detonator 42.68
Both set to balanced performance/image quality.Too vague, especially considering the disparity in rendered quality between the two cards. The settings may be "equivalent," as in all sliders set to the middle, but the output may not be (a single set of 3DM03 comparison screens is not enough to base a IQ comparison on). Thus the scores are hard to trust.
But it's interesting to see the FX U stretch its legs, despite its bandwidth disadvantage. I'm curious to see if the 9800P would offer performance improvements beyond its clock speed bumps.
Sharkfood
12-Mar-2003, 07:08
I'm really curious how the 9700 Pro would stand-up if it were overclocked to a clock speed that it could only maintain for several minutes at a time before overheating...
We know that several outfits are selling heatpipe 9700 Pro's that run full-time at 400mhz. I'd imagine if you overclocked the 9700 Pro so as it can only run for a few minutes without overheating (and in the case of the NV30, downclock down to 30%+ lesser performance) it would be very cool to see.
Tagrineth
12-Mar-2003, 15:14
Mmm... I thought 3DMark2001 was generally agreed to be too platform-dependent on high-end video cards anyway?
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.