GeForce 3Ti 200 or GeForce 4mx 440

azezil

I know Kali posted a similar topic, but this time its a GeForce 3 Ti

Anyway the choices:

Sparkle Geforce3 Ti 64MB

Nvidia Geforce3 Titanium 200
64MB DDR-RAM, AGP4x/2x
Vertex shader & pixel shaders 3D technol.
Advanced lightspeed memory architecture
DirectX8, TV-out
Windows 9x/ ME/ NT/ 2000/ XP, Linux

Euro 189.00 incl. VAT 21%

---

PNY Geforce4 64MB

64MB DDR-RAM, GeForce4 MX440, AGP4x
Dual Ramdac 350Mhz
Video Processing Engine -Enable the highest-quality
full-frame rate, full screen HDTV and DVD
DirectX8.1, TV-out
Windows 9x/ ME/ 2000/ XP, Linux

Euro 187.00 incl. VAT 21%


Only 2 euro in the difference but from what i've read they're pritty much the same but the GeForce 4 has some groovy new featurers

Kali

just on the prices you have there:
that seems to be a horrible price for the geforce4.. when i was checking the prices for the other thread there, the mx440 is €128 on komplett.. including vat... theres also a version with slightly quicker memory (4,0ns instead of 5,0ns).. and even that was cheaper than your price there (it was €165 or so)...
but on the other hand that seems to be a very decent price for a geforce3 Ti.. whereabouts are you getting these from?

[edit]
there was quite a decent url posted in that other thread:
http://www.tomshardware.com/graphic/02q2/020418/index.html
.. basically a roundup of a good twenty or so different cards, the conclusion of which puts the mentioned geforce3 slightly ahead of the mx440 (note that the mx440 in that test had memory at a speed of 400Mhz.. unlike the 350Mhz you have quoted.. again this could be the better 4,0ns version?)

Jelvon

umm, I bought I geforce3 ti200 last week for £90 (sterling) and I had to make the same choice as you.

Basically what swayed me was the fact the gf4mx440 was based on the nv17 chipset (not sure if that is the right number) but it DOESN'T have the nfiniteFX II engine that the gf4 ti series has which means no per pixel shaders, etc.

Whereas the gf3 has got the nfiniteFX I pipeline also the gf3 ti200 tends to come out ahead in benchmarks...

I only bought it as a stopgap before the gf5 comes out in augest anyway, at 90 quid from dabs.com it didn't exactly break the bank

Gerry

remember that the geforce4 mx is missing most of the hyped features of the geforce4. Its memory controller is not as good as on the geforce4 ti either. If you read the other thread, you would see that the 440 is a little bit faster than a geforce2 pro, which means that a geforce3 ti 200 is a fair bit faster than it. So go buy the ti 200, or save up for a geforce4 ti 4200 (all the features of the 4600, and can be clocked up to almost the same speeds ) €260 on komplett.

PPC

Also sparkle generally make good cards.
I've heard the GF4MX's are just a little more powerful the the GF2MX's
And a GF DDR is ment to be more powerful than a GF2 MX.
Get the Ti.

azezil

Originally posted by PPC
Also sparkle generally make good cards.
I've heard the GF4MX's are just a little more powerful the the GF2MX's
And a GF DDR is ment to be more powerful than a GF2 MX.
Get the Ti.

theGF4 mx420 is fair shit alrite, the 440 doesn't seem too bad.

Kali prices quoted from marx-computers

WhiteWashMan

go for the gf3.
the gf4 is just an upgraded gf2

azezil

lol okeee thanks for that wwman.



/am actually loling

Quigs Snr

No doubt about it. Geforce 3 Ti 200 all the way. One of the guys in work here upgraded from a GF2 GTS to a GF4 MX440. His frame rate (based on an XP1800) IN 3DMARK 2001 was as follows:

(Approxitmate figures)

GF2 GTS 3900
GF4 MX440 4900
GF3 Ti 200 6700

The GF3 Ti was significantly better. Of course if you can stretch to the GF4Ti 4200, those babies clock up very nicely. I have seen a GF4 Ti4400 clocked up to above the rate of a 4600 and score about 10,000 on 3d Mark. I would expect in the order of 9000 from a clocked up 4200.

c0y0te

Don't bother with either of those cards.!!!!!!

If you can wait until middle of June, the TI 4200 cards will be available both here (Ireland) and in the UK. Scan will have them for 7th I think, and Komplett are already pre-selling them as far as I know.

These 4200 cards knock the **** outta everything else out there, and are great value for money.

If you think I'm overstating this - go check some hardware reviews for the TI 4200... you won't be dissappointed.

c0y0te

JustHalf

Quigs, it's not such a good idea to rely on 3DMark2001 scores to compare DX8 compatible and non-DX8 compatible cards, particularly when most games don't use DX8 features (such as pixel shaders).

For example, there is no way the GeForce3 Ti 200 will run Quake 3 anywhere near to 33% faster than the GeForce 4 MX440.

It will only be when DX8 games are widespread that this will be relavent.

Quigs Snr

No it's not a good idea. However, 3DMark is as good a real world test as any to check the horsepower of a card. If one card scores 5000 and the other scores 4000 you can bet the farm on the one that scores 5000 being the faster card.

3DMark although not perfect, is the main standard upon which graphics cards are measured these days. If anything quake 3 is a
poor example as it does not benefit at all from newer features built into the cards. Wolfensteins tweaked version of the Quake 3 Engine is a far better test.

In any case, DX8 compatible or not, any modern card will run older games like quake and unreal as fast as you will ever need, so you should buy your current card with whats on the market now and whats coming in mind. In this regard, DX8 and hence 3DMark are as good a way as any of testing these cards as with DX9 still a bit away, and games for it probably a year or more away, DX8 is the way to go right now.

I have tested many many cards with 3DMark and I have to say that if I want to fire up a game of medal of honor on a card that scores 4000 it plays a dream as does quake. If I fire it up on a card that scores 2000, MOH plays like a turd, Quake 3 still great, almost no difference. 3DMark stresses the card, you can if you like eliminate the pixel shader demo's etc.... and just test the cards for polygon pushing abilities too if you want a more 'pure' test.

JustHalf

I'm simply questioning the relevance of the 3DMark2001 scores for testing the performance of video cards for games.

Ultimately, the best way of testing the gaming performance of video cards is with the games themselves. As I said, 3DMark2001 scores includes several pixel shader tests, which on their own SIGNIFICANTLY alter the final result. Given that pixel-shaders are in very few (if any) games released it's not a good test for games at the moment. When more and more games come out which use DX8-only features, then it will be more relevant.

Quigs Snr

Fair enough, when comparing cards I normally just run the polygon pushing tests.

I think the only game I have seen which uses a lot of directX8 features by the way is Commanche4. Apparently I uses a lot of that stuff for the water effects etc........

Have to say the water looks great, the rest of it still looks like Delta Force Land Warrior and I could probably run that on a digital watch never mind a GF4Ti (slight exaggeration but you get the idea).

Twill be interesting to see how Halo shapes up, it's supposed to take advantage of a lot of DX8 features.

PPC

Originally posted by Quigs Snr
No doubt about it. Geforce 3 Ti 200 all the way. One of the guys in work here upgraded from a GF2 GTS to a GF4 MX440. His frame rate (based on an XP1800) IN 3DMARK 2001 was as follows:

(Approxitmate figures)

GF2 GTS 3900
GF4 MX440 4900
GF3 Ti 200 6700

The GF3 Ti was significantly better. Of course if you can stretch to the GF4Ti 4200, those babies clock up very nicely. I have seen a GF4 Ti4400 clocked up to above the rate of a 4600 and score about 10,000 on 3d Mark. I would expect in the order of 9000 from a clocked up 4200.

I got 7884 on my GF3 with a slight OC and hit 300FPS in Q3.

You can probaly get a cheap GF3 somewhere

Quigs Snr

That would be about right with a very good processor. The machine I did those tests on was a TBird 1200 and processor plays a fairly important role in 3DMark tests.

Thats pretty good all the same. Is yours the 200 or 500 ?

PPC

Standard GF3.
I've a XP1900+ running at 1.7 with a gig of DDR so that might be the differnece too :rolleyes:

JustHalf

I wonder how you have you're computer set up? Considering I got pretty much the same result with an overclocked GeForce 3 (O/C'd to Ti 500 speeds) on a 1GHz Axia @ 1333MHz and with 256Mb of SDRAM, it seems a bit low.

Quigs Snr

The machine in question wasn't the greatest. It was a 200FSB Tbird as far as I can recall with 256MB PC100Ram (set with safe timings as there was some sort of bizzare problem with it otherwise if I recall). The HD was a 5400rpm ATA66 Seagate thingy, and the Mobo was based on the ALI Magik chipset which doesn't exactley set the world on fire.

The card was not OC'd. Settings on 3DMark were default. PPC your standard GF3 sits between the 200 and 500 (closer to the 500) so it is faster anyway. And I think your machine setup would help the score along a bit too!

Justhalf, thats a good score you're getting there.

JustHalf

I was asking PPC to be honest

Of course it's shot down about 400 points after installing the "PCI Latency" batch.