Sign in with
Sign up | Sign in
Your question

GF4 MX-440 beats my GF4 Ti4200

Last response: in Graphics & Displays
Share
September 27, 2002 12:09:49 PM

Hi,

Today I got my new SUMA 64 Mb GeForce 4 Ti4200.
I tested 3D Mark 2001 and got 7018 3d marks. You guys said that I should be expecting around the 8000 barrier.

I then told him to test MOHAA on his computer. He gets around 30-90 fps. In my computer I only get 10-45 fps. My brother is beating me with a GF4 Mx-440. What can be wrong with my system?

We then tried putting my GF4 Ti4200 on his computer and running MOHAA and I got around 20-70 fps. This still means that my Ti4200 is worst than his GF4 Mx-440. Could this be true?
Or is something just wrong with the Ti4200.

GeForce 4 Ti4200: Core: 250 Mhz Memory: 500 Mhz
GeForce 4 Mx-440: Core 270 Mhz Memory: 400 Mhz

My card has a faster memory but I still don't beat him. This is making me very angry as I paided more for my card and get less performanace than his.

The only thing my Gf4 Ti4200 beats him so far is in 3D Mark 2001, but that's not really a game.

We also tried Quake 3 Areana Demo001 and on my computer I get 131 fps and on his computer with his GF4 Mx-440 he gets 170 fps.



Below are the systems.

My System:
AMD Duron 1.2 Ghz
Soltek SL-75DRV5 Kt333 Motherboard Socket A
Suma GF4 Ti4200 64 Mb
Seagate 40 Gig 7200 rpm ATA 100
256 Mb DDR-RAM

My Brother's System:
AMD XP 1900
Gigabyte GA-7VRX KT333 Motherboard Socket A
Leadtek GF4 Mx-440
Mator 40 Gig 7200 rpm ATA 133
2 x 128 Mb DDR-RAM
September 27, 2002 3:24:14 PM

Can the fact that he has a faster processor play into this? Maybe your computer is one of those full of crap with all old drivers etc, and his is nice and clean? do you have all the new drivers? Is V-sync off? is AA off? Is anistropic filtering off?

My sig's faster than yours, and it overclocks better too....
September 27, 2002 3:38:09 PM

this is a mix of issues for this.
one, his pc is indeed faster. the speeds people say one can achieve with the card you have are based on faster computers than what you own.

The other reason is simply that the MX card was made to be cheap and fast. Your card is a full blown GF4, that is underclocked. the MX card is a suped up GF2!!! it has no real directx8 support, and lacks so many features that the geforce 4 has that to buy the MX card, youd be making a terrible mistake. be happy with the card you have. for the money you have a great product, your numbers will improve dramaticly if you upgrade your cpu/mobo/memory type, but the mx card should not even be something to think about, its a bad product that should only be used in workstations, not gaming rigs.
Related resources
Can't find your answer ? Ask !
September 27, 2002 4:11:51 PM

Hmm ... That is an odd situation. My guess would be more along the lines that you're using older software to test each cards performance.

If I remember correctly, the MX cards are designed to run a lot of pixels, but minimal effects per pixel. So if your software has little or no special effects being rendered on a pixel, then it'll look lightning fast. However, if you have numerous effects, then the MX will have to loop that pixel through the same pipeline repeatedly until it gets all of the effects in, which slows it down badly.

Plus the MX doesn't even have programmable vertex or pixel shaders which are used in DirectX 8 and higher. So any special effects from new software that use these won't even be available to the MX cards.

The GeForce4 however is designed to layer numerous visual effects on the same pixel in just one pass through the pipeline. So in software with littel or no effects, the GF4 doesn't perform amazingly because most of the processing power is being wasted by uncomplex software. However, in newer software that uses complex rendering to put many effects per pixel into the rendering to make for some stunning grapgics, the GF4 will show some real muscle and completely blow away a MX card. Also any software that uses DirectX 8 or higher special features (AKA programmable vertex and pixel shader effects) will run just fine in a GF4.

So I think the situation that you are seeing is simply because you aren't running software that actually makes any real use of the GeForce4's true potential. Try more modern games. Then your GeForce4 will beat up on your bro's MX.

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
September 27, 2002 5:16:31 PM

I think part of difference is bus speed. The XP runs 133 Mhz FSB while the Duron only 100 Mhz. Slow bus limits memory throughput and game performance.

Your processor is also 333 Mhz slower than your brother's.

<b>[Edit</b> - Make that 400 Mhz slower. (I have trouble calculating real clock speeds from AMD PR ratings).<b>]</b>

You could unlock the Duron and run it at 9*133.

<b>I have so many cookies I now have a FAT problem!</b><P ID="edit"><FONT SIZE=-1><EM>Edited by phsstpok on 09/27/02 01:36 PM.</EM></FONT></P>
September 27, 2002 8:00:21 PM

Quote:
We then tried putting my GF4 Ti4200 on his computer and running MOHAA and I got around 20-70 fps. This still means that my Ti4200 is worst than his GF4 Mx-440.

So the results are same even when the GF4 is run from bro's PC. Not likely to be the processor then.

<pre><A HREF="http://www.nuklearpower.com/comic/186.htm" target="_new"><font color=red>It's all relative...</font color=red></A></pre><p>
September 27, 2002 11:51:42 PM

See but thats the thing. We tested my video card on his computer already and as I stated Medal of Honour Allied Assult went only 20-70 fps with my GF4 Ti4200 and then he tried his GF4Mx-440 and he got 30-90 fps.

How do you turn on and off AA thing?
September 28, 2002 12:40:11 AM

Simply putting your card in his computer doesn't give an accurate result. His computer is optimised for his GF4 MX.

Any DirectX 7 game will run faster on his card until your card's GPU matches his in Mhz(not a hard feat, considering yours is currently underclocked). Recomendation PowerStrip.

You can't possibly get the full effect from your card until you up your fsb from 100 to 133(or higher). Your GPU may very well be waiting on your processor.

3DMarks runs faster on your card (as will any DX8 game), because his has to run multiple passes to produce the effects your's can produce in a single pass.

Pain is the realization of your own weakness.
September 28, 2002 2:01:05 AM

10-45 fps in the Duron based machine, 20-70 fps in the XP based machine, seem to show a significant diffence but those numbers alone aren't accurate. MOHAA is not really a reliable program to use for benchmarks because framerates aren't a summary of performance just instananeous values. For example, in the Omaha Beach landing scene, the most 3D intensive scene, 70 fps would be outstanding but 20 fps would be horrible for the XP w/ Ti4200 if they were both average values but in that scene 70 fps might be a reasonable peak framerate and 20 fps a reasonable minimum framerate. The range would be different than an MX-440 and could be lower if the Ti4200 was trying to produce effects that the MX-440 can't. (I'm not really sure). We need more information from more commonly used benchmarks like Quake III, Serious SAM, Aquanox, etc (just examples I only uses 3DMark2001 and Quake3 to guage my own system). However, I belive that Q3 is good for showing CPU peformance these days, Aquanox is good at show video card speed (I think), Serious Sam is pretty balanced with CPU/video card. Maybe I'm wrong.

If you are only using one benchmark 3DMark2001 isn't that bad because it utilizes three different 3D engines (or so I here) and everyone uses this benchmark so it's easy to compare. Plus, 7000 isn't terrible for a Duron w/ Ti4200 but it isn't great either.

Please remember that we are trying to determine if we can improve performance on the Duron system or at least determine if something is wrong. It's not really important how the Ti4200 did in one-time trial in the brother's computer.

Taking a quick look at Madonion's ORB, I see one score for 7760 for a Duron system and a Ti4200 with memory at 520. I see other scores in of 8000 range but most of those video cards have the memory clocked at 600+ mhz. All of this kind of shows that Alexi02's 7000 isn't too far off the mark.

<b>I have so many cookies I now have a FAT problem!</b>
September 28, 2002 3:28:22 AM

Anyway, I tried MOHAA again and set everything to low detail and everything else and in the first mission is still kind of goes slow. I am thinking that MOHAA can't be too compatible with my video card.

In 3d Mark 2001 on my computer I get 6300 - only with the first 3 games enabled and on his computer with my Geforce ti4200 I get 8000 - only with the first 3 games. So there is a difference of 1700.
September 28, 2002 3:50:37 AM

MOHAA kills with a Ti4200 card. It's awsome. At least with that game I wish I hadn't bought the Radeon 8500. Though, UT2003 demo works well, as do most other games, so go figure.

OK, I don't know why you are running only the first 3 games in 3DMark2001. Please try the default benchmark so we can actually compare your results.

If you still get 6000ish scores I think I know what is wrong. AGP isn't running or isn't running at full speed.

I forget but I thought you said you have a Via KT333 chipset motherboard. Try reloading the 4-in-1 drivers. (Get the latest <A HREF="http://www.viaarena.com/?PageID=2" target="_new">here</A>). A "clean" install is best. Run the installer once but choose the uninstall option for each component. When prompted don't reboot (saves time and prevents Windows from auto-loading the older driver which it sometimes wants to do). Run the installer a second and let the components load. This time re-boot when prompted.

This should fix your performance problems. You may need to reload the nVidia drivers and possibly DirectX for some minor problems later but a clean install of the 4-in-1 drivers are the most important.

If you happen to use Windows 9x/ME I can help you with clean installs of nVidia drivers and DirectX which do help but to a much lesser degree than the clean install of the 4-in-1's.

<b>I have so many cookies I now have a FAT problem!</b>
September 28, 2002 9:17:01 AM

I will try installing the VIA 4 in 1 and I will format my hdd and try installing everything again.
Now in 3d mark 2001 with all four games I get 7018 3d marks(I haven't formated yet)

Anyway I run Si Santa 2002 and powerstrip and they both say that my AGP transfer rate is 1x. I have the bios set on 4x but it still says that it is on 1x. How do I fix this????
September 28, 2002 2:04:26 PM

I'm pretty sure it's your CPU holding you back. You probably could've saved money and got a R9000 PRO/R8500.

...And all the King's horses and all the King's men couldn't put my computer back together again...
September 28, 2002 3:49:04 PM

In 3DMark2001 the default benchmark runs 17 tests (I think). You don't need to select anything to run the default benchmark, which is 1024x768, 32-bit color.

As I said, your AGP isn't working correctly. To enable AGP 2x/4x you need to select the "Turbo Mode" option while installing the 4-in-1 drivers. The later drivers (4.38 and later, I think) automatically select Turbo Mode. With earlier drivers you had to click the option.

I wouldn't bother reinstalling Windows just yet. The clean install of the 4-in-1 drivers is easy enough to do. It will at least fix your performance problem.

<b>I have so many cookies I now have a FAT problem!</b>
October 9, 2002 6:09:14 PM

Thanks.

I wondered what a stock Ti4200 could do in XP 1600-2000 class machines.

<b>I have so many cookies I now have a FAT problem!</b>
October 10, 2002 12:55:46 AM

Why not try the MX440 in your machine and see if it performs better on both machines?
!