GeForceFX 5200 Image
Quality
These days games cannot just be 'fast', they also have to look
good. Image quality is a very subjective and opinionated point,
but we still tested the image quality with 4xS and 8x AF enabled and compared
that to a Radeon 9700 Pro with with 4x AA and 8x AF enabled. Yes I'm
well aware that the Radeon 9700 Pro has higher AA and AF settings, but with both
cards set the same, the Radeon 9700 Pro's image still looks sharper. nVIDIA has
gone a long way in their image quality but it's still not up to par with ATi's
offering.
For some examples of GFX5200 Vs. Radeon 9700 image quality comparisons please
click here. The visual difference between reference-designed GFX5200 cards is
not distinguishable in most cases.
Overclocking:
Even though the MSI GFX5200 is clearly a budget videocard, this
doesn't mean it will be spared from our overclocking tests. The Albatron GFX5200 wasn't
the greatest overclocker, so we were really not sure what to expect from the
MSI FX5200-TDR128.
We've usually had pretty good luck overclocking MSI videocards in the past,
and were hoping that this little card would keep the tradition running.
We first set out to see how high we could overclock the core.
Using the coolbits registry hack we began to raise the core speed.
Going up 5 MHz at a time. We easily passed 275, 280 and
even 290 MHz. At 295 MHz we ran into a few stability problems, the card would
run through both 3DMark tests with no problems, however, it would lockup while running the CodeCreatures
benchmarks.
Lowering the core speed to 293 MHz eliminated all problems. Honestly this was
a bit disappointing considering the Albatron FX5200P was able to reach 333 MHz
with just a passive cooler. I guess it goes to show you overclocking is
still luck of the draw.
With the 4ns Samsung memory rated for 500 MHz, I decided to be a bit more
aggressive with our memory overclocking. Going up 25 MHz at a time, 475 MHz, 500
MHz and even 550 Mhz was easy - quite a contrast with our Albatron card which
maxed out at 410 MHz memory speed. In the end we settled with a 578 MHz memory
speed, anything higher and Artifact tester would start registering artifacts.
Visible artifacts were not present until we got up to 585 MHz but as always,
we benchmark with zero artifacts on the screen visible or not.
Test System Specs:
pcstats test system
specs: |
|
computer hardware: |
|
processor: |
intel pentium 4 2.8
ghz |
clock speed:
|
21 x 133
mhz |
motherboards: |
msi gnb
max-fisr* |
chipset: |
intel
e7205 |
videocard: |
ati radeon 9700
pro (325/620) msi g4ti4600-td (300/650) msi g4mx440-t8x (250/400) albatron geforcefx 5200p (250/400) msi fx5200-tdr128 (250/400)
msi fx5200-tdr128
(293/578) |
memory: |
2x 256mb corsair xms3500 cas2 |
hard drive:
|
20gb maxtor
diamondmax+ |
cdrom: |
nec 52x cd-rom |
floppy: |
panasonic 1.44mb floppy drive |
heatsink: |
avc
sunflower |
powersupply: |
vantec 470w psu |
software
setup: |
windowsxp intel inf's 4.30.10 Catalyst
3.2 Detonator 43.45 |
Benchmarks: |
3DMark2001SE 3DMark03 AquaMark CodeCreatures Commanche 4 Quake III Arena Ver
1.17 UT2003
AA Test, AF and AA+AF
Test 3DMark2001SE Quake III Arena |
* - 8x AGP was enabled and functioning properly during all
testing.