It depends.

You need to research what RAM is fitted to the card, it's the first important consideration. Not the amount of RAM, but the SPEED of the chips themselves.

Some budget manufacturers use el cheapo components on their cards, this may not make any difference, however, if the RAM fitted is a slower type, it makes a lot of difference. The RAM chips quality is based on it's timings specification, the faster the timings, (measured in nanoseconds ... ns) the better the chips and the faster the card will perform.

The nanoseconds (latency) should be as low as possible, the lower the number, the faster the RAM and therefore the card.

The difference RAM timings make is substantial, and a card with half the RAM, yet faster timings will outperform a card with more RAM and slower timings.

A cheap trick used by manufacturers is to stack on a heap of cheap RAM to fool the common man into believing he has something special.... so, when I hear people spouting off how much RAM they have on their video card, I know straight away they don't know what they are talking about...... if they did know, they would be quoting the RAM timing in nanoseconds instead.

If you look at the advertiser's specs at pricewatch.com you will often see the /ns specified, and it should be taken into account when purchasing or making price comparisons on video cards.

I hope that clears one or two things up for you.

It's this knowledge that has kept me from posting my PC spec in my Avatar, you see, it seems pointless to me to post a specification without a detailed configuration description as well. This will help to explain why 2 computers of apparent same specification will perform very differently in hardware alone, let alone software configuration.

While on the subject of configuration, You should also know that each video card has a "reference" specification. This spec is often modified within certain limits depending on the manufacturer of the hardware.

Certain architecture and components used will either make the card above or below spec. If the card is sufficiently above spec, the manufacturer may safely alter the GPU and RAM timings into an overclocked state without fear of stability or overstress problems.

Other manufacturers may experience stability problems and therefore underclock their cards to compensate.

There are just a few manufacturers who have a track record of producing consistently fast cards because they use the best components and strive to serve a niche market of enthusiasts. Gainward is one of those, XFX is making their mark, others have produced some very nice cards.

Video card selection for most is but a hit and miss affair, most advice is based on ignorance at best. There is actually a lot of research and current knowledge required to make a good purchase decision when you realise what is involved.

But, then comes along a product line like the NV6800 which just blows everything else away and it all becomes easy again........      

If that's you sitting in that Angel's Phantom, I'm jealous! 

Quote:


True to an extent, most graphics cards follow the manufacturers "reference" design, in other words, look at the layout of say an MSI card, compare it to a Gainward (Assuming they are both the same chipset, for example a 5900XT) you'll find the tracks on the pcb are more or less in the exact same position, chips in the same spot etc.

Hence, normally theres little differance between two 5900XT's from different company's indeed some are identical bar the name!

Comparing my Asus 9560 FX5600 to a Gainward FX5600, the two boards were very different and my Asus was about 600marks ahead on 3D Mark 2001SE - hardly a great margin of differance! -  a friend of mine has an un-branded FX5600, it looks the same as my Asus, they score identical benchmarks on my pc!!!

Brundlefly, theres another obvious bottleneck in your system - your processor - I upgraded an old Celeron 1.7ghz Sct 478 cpu for a 1.8 p4 you'd think an extra 100mhz would'nt make much differance and it did'nt, what did was the p4's 256k of L2 cache memory (current p4's now have 512k L2 cache RAM) compared to the Celeron's 128k L2 memory, the performance differance was amazing, and my 3D Mark 2001 score ,at the time, jumped from about 5000 to 8500 FS2002 was greatly improved.....

The same 1.8 p4 now runs FS9 very well with an Asus 9520 FX5200(!) and 512MB of PC2100 DDR ok its not as good as my own machine (specs below ) but for all those nay sayers who think an FX5200 can't cut it with FS9 my experiences with my #2 PC tell me different.

Ultimately, a £20 Geforce 4 MX will be an improvement over your current graphics set up, as ever its a case of how much money can you afford to spend? £40 will get you by with FS9 and an FX5200. Multiply that by 10 and you'll have a nice shiny Geforce 6800 Ultra - a card more than upto FS9 at whatever settings you like - trouble is, the rest of your system would'nt be enough to get the best from it 

Food for thought eh? 

Paul.