TechSpot

Is your graphics card fast enough?

By Nic
Jul 14, 2003
Topic Status:
Not open for further replies.
  1. For those that think todays top end graphics cards will be able to cope easily with up coming DirectX9 games, take a look at this benchmark of the already available Gun Metal game.

    Notice the less than 20 fps at a screen resolution of 1280x1024 (typical for a 17" TFT monitor) for Radeon 9800 Pro with 4x FSAA turned on and no anisotropic filtering to prevent the speed from dropping too low.

    Nvidia's FX5900 Ultra does little better.

    You can bet that newer games will start to place significant demands on current graphics cards once DirectX9 features start being used to the max.

    [​IMG]

    Full article can be found over here ...

    MSI NBox N5900 ULTRA - Review
     
  2. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

    [​IMG]

    I don't really pay much attention to Demo's, especially when it come to Nvidia hardware currently, more importantnly, I ook at feature tests...

    Certainly, in 2 or 3 years when DX9 games are everywhere, older hardware such as the FX5900 and R9800 will be *much* slower than the new stuff, but I'm sure they will still run quite well in newer games, albiet w/ fewer features activated.

    The DX7 GF4MX series runs every game out there just fine, with tech. that's 3 years old........:grinthumb
     
  3. MrGaribaldi

    MrGaribaldi TechSpot Ambassador Posts: 2,802

    Oh, and by the way, from what I've read GunMetal isn't exactly using much in the way of DX9 functions...
    I seem to recall that it's mostly used for the hud, but I could be mistaken about that...

    On a similar note, YetiSoft (the creators of the game) released a demo about 4-6 months ago, which would only run on Nvidia hardware, due to a hardware check... This was "fixed" by 3dAnalyzer, and the demo worked as it should on ATI cards..

    There was also some problems with the benchmark released and ATI cards (image corruption I believe), but was fixed with a patch...

    But the history of the game isn't exactly untainted...

    Anyways, interesting to see the performance of the cards...
    Too bad they didn't do a test with AF on... Would have been nice to see in what way the scores would have changed (aside from dropping even more in speed)...
     
  4. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

    Change to: "... The DX7 GF4MX series runs many games out there just fine, as long as you don't mind cranking down the screen resolution, and turning off some of the graphics options ..."
     
  5. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

  6. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

  7. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

    The aim of this thread was to provide insight into future trends in game design, that require faster graphics hardware, and so help users to make an informed decisions when contemplating the purchase of a new graphics card.

    The fact is that currently Nvidia 5900 Ultra *IS* the fastest graphics solution out there at the moment (across the majority of benchmarks). However, those of us that look at the overall picture when making a purchase (i.e. cost, performance, available budget, etc.) are not concerned with ATI vs Nvidia, and can make our own decisions without bias.

    Graphics card companies are only out to make as much money from us as they can get away with, and so providing them with free promotion is not something to consider, unless qualified and objectively evaluated.

    I have both Nvidia AND ATI cards in my home PCs, and chose to go with ATI for my last PC because of the better cost/performance trade off currenty the case with ATI Radeon 9800 Pro. It's slower than Nvidia's FX5900 Ultra, but considerably cheaper, and that's what did it for me (despite the previous problems that I have had with ATI hardware).

    It is entirely possible to misrepresent the respective performance between ATI vs Nvidia solutions, and I would hope that users are aware of this. Each of us will have our own criteria with which we make our buying decisions, and so pooling the views of many users will help to clarify our own preferences.

    There is little point in stating irrelevant facts, such as brand X makes the fastest graphics cards (model Y), when someone is looking to purchase model Z, as that would require an entirely different analysis.

    Selective posting can be misleading, but this depends entirely on the point being made. If the point of a post was to influence a buying decision, then selectively posting in a way that appears to show item X outperforming item Y, when this is not always the case, then that is misleading. Selectively posting to illustrate a particular technical point, is different and obviously requires a suitable selection to be made, and irrelevant data to be kept out.

    Let's just hope the fanboys keep their opinions to themselves. ;)
     
  8. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

    From what I've read, the two cards are nearly identical as far as actual fps are concerned, (decent resolutions w/ some level of FSAA and AF) perhaps the R9800 winning the majority of GAME benchmarks, albiet it's a very close call. The thought once the NV35 debited was that is was faster than the R9800PRO, but since the cards have actually shipped, and product reviews abound, that is simply not turned out to be the case. I suggest you do a bit more reading up on the plethora of reviews out there....there are some to be taken w/ a grain of salt....yet the majority seem to show the two cards vey even, with a few more that paint he R9800as the faster choice....maybe you should read them so yo don't sound like a fanboy when you say such things as
    Happy reading:)
     
  9. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

    Bollocks. ;)

    Take a look over here ...

    tomshardware

    quote:
    ... Currently, the FX 5900 Ultra can safely be called the fastest card on the market ...

    and here ...

    Hexus

    quote:
    ... Performance was largely as expected. Running with the same 256-bit memory interface as a 9800 Pro, the NV35 uses a faster core and greater memory bandwidth. The stumbling block comes with the use of only 4 pixel pipelines for the common coloured pixel rendering approach. Balance it all out and you can see how and why the 5900 Ultra manages to match or surpass the 9800's benchmarks in the majority of cases. ...

    and here ...

    X-Bit Labs

    quote:
    ... Well, it is evident that NVIDIA GeForce FX 5900 Ultra is not only a wonderful replacement for GeForce FX 5800 Ultra, but also the today’s most powerful gaming accelerator. The launch of NV35 helped NVIDIA to win back the title of the gaming 3D graphics leader, because the fastest solution from ATI - RADEON 9800 Pro – appeared slower than NVIDIA GeForce FX 5900 Ultra in most tests ...

    Pick up a pc game magazine (I subscribe to two pc magazines) from your local newsagent, and you'll find a similar story. I own a Radeon 9800 Pro, and yet I couldn't care less if it's slower.

    PS: Now see if you can find even one review that says the Radeon 9800 Pro is faster than the Nvidia FX5900 Ultra. You might find some that say they are on par, but none that say it is faster, unlike the quotes provided above.
     
  10. acidosmosis

    acidosmosis TechSpot Chancellor Posts: 1,574

    Please remind yourselves that a few FPS can not be "seen" by the human brain.
     
  11. JSR

    JSR Banned Posts: 730

    yes, nic

    and, being a nvidia owner, i'd like to think that is what's important......but, in retrospect, the radeon gains ground, as the bells and whistles get turned on.........therefore, it seems that the issues involving visual quality, still rests with ati......so, when you consider the price differential, in relation to your performance priorities, i'd probably go with the card you currently own......
     
     
  12. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

    Nic

    How about GamePC?
    From http://www.gamepc.com/labs/view_content.asp?id=3x5900u&page=14


    I assert that it is not at all black and white as to whch card is faster. They each do well. To call the FX5900U the fastest card out there is simply, at the very least, debatable, and not at all clear.....
     
  13. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

    I think if you take a close look at the tests carried out in that review you posted, you'll notice something rather strange. There seems to be no consistency between the results presented, which appear to have been specially selected to highlight tests in which the Radeon was better. This is immediatly apparent because each of the benchmarks do not present the same range of resolutions and settings, and tend to show only 1600x1200 with FSAA results, with some tests also showing 1024x768 with no FSAA.

    These results are definitely biased, and I would be surprised if you didn't spot that yourself, especially with all the inconsistancies and the fact that in not a single test did the Nvidia card win. The review has specially hand picked the games that are used in the benchmark (vice city, freelancer, nolf2, warcraft2, ut2003), no doubt as the ATI card performs best in those games at the resolution and settings that were selected.

    I don't think there is any doubt that the Nvidia card is faster overall, but on average it isn't that much faster, it costs considerably more, and the FSAA quality isn't as good. That was enough to swing my decision in favour of the ATI card, but for those that want that little bit extra speed, the 5900 Ultra takes it.

    PS: I have to give you credit though, as you did find a review that favours the the Radeon 9800 Pro over the Nvidia FX5900 Ultra. I didn't think you'd manage to do that, but you did. Well done :=).
     
  14. acidosmosis

    acidosmosis TechSpot Chancellor Posts: 1,574

    Oh and btw, WHEN oh WHEN are you guys going to STOP relying on benchmarks? Please wake up, smell the burning BGA chips, and begin to understand how to benchmark hardware with your own brain and actual usage (if possible).

    Benchmarks are only a slight basis to work from and can NEVER be relied on. If you want to know how your video card will run Half Life 2 then you will just have to use common sense, and if you have a Geforce 3 YES your gaming experience isn't going to be a good one though it will actually run the game. These types of things are common sense.

    When it comes to actually finding out how well your card performs with that game you cant rely on someone elses opinions, benchmarks or actual FPS the person achieved. There are way too many details to consider such as the motherboard they are using, amount of RAM, speed of RAM, quality of RAM, processor speed, processor manufacturer, BIOS updated?, sound/video/etc drivers updated?, what programs are running in the background and how much resources are they using?, does your sound card cause a lot of stress on your processor?, what resolution are you running at?, what is turned on or off or what is the level of detail settigns in ATI or Nvidias control panel settings, is the PC overclocked, etc etc etc.

    Yes benchmarks help, but they are really only a small basis to work from, and most of all for the general public to be completely fooled by into thinking they haev the "l33t3st friggin machine ever" just because 3DMark says so, in turn selling more PC's because the general public is sadly the majority. Those of us that know enough about PC's to know better are the minority.

    No offence, I am just trying to put some sense into your brains, and am annoyed at the amount of opinion that is weighed on benchmarks of any kind.
     
  15. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

    There is nothing wrong with benchmarking, the problem lies in how you interpret the results. I am certain that an Intel P4 will benchmark higher than an old Intel 486DX2, and this also shows in real world performance.

    We all know that statistics can be made to say almost anything. It is only when we look deeper and analyse what they mean, and how this translates into what we want to know, that we can reliably rely on, or discount, their value.

    All of those benchmarks that were posted are true and accurate. The problem is that some were selective and so did not show the reality of the situation that is most relevent to users considering buying one of those cards.

    Each user will have their own values that will influence their decision based on how they will use that particular piece of hardware. If enough results are presented, then users can make objective decisions that are right for them.
     
  16. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

    :grinthumb

    I agree, my point isn't that card (A) is faster than card (B), just that they each have their strength and weaknesses, and IMO it's impossible to declare one "faster" than the other, since each may win and lose the same benchmark, depending on the conditions invloved, ie resolution, AF, and FSAA settings.....
     
  17. JSR

    JSR Banned Posts: 730

    nic, is right

    for sheer speed............everything off........the 5900s' gonna win..........
     
  18. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

    Re: nic, is right

    I don't disagree w/ that, bit that isn't what Nic said.....If he had, I wouldn't of disagreed.

    But what Nic said was,
    And that's why it bothered me. It isn't that simple anymore, and is a cloudy issue to say the least.....
     
  19. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

    I still stand by that statement, which also seems to be the general consensus across most of the hardware review sites and pc magazines.

    I do agree that it isn't all black and white as there are indeed cases in which the 9800 will be faster than the 5900. This is much like the argument over P4 vs Athlon. I have an Athlon, but the P4 is faster across the majority of benchmarks, and that doesn't bother me one bit.

    The thing to keep in mind is that the differences are not large, and I would choose a Radeon 9800 Pro over the Nvidia FX5900 Ultra any day, and in fact, I did.

    There is more than raw speed to take into account when making buying decisions, and as I already outlined previously, each person can make their own decisions based on what they value most.

    PS: I do remember that you were in fact making similar statements saying that the 9800 was the fastest overall, but without the evidence to back it up. My point of view is not 'static', unlike some, and I am always ready to change opinion if there is sufficient evidence to back it up.
     
  20. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

    Nic, according to PC magazine, a source you seem to think accurate,
    http://www.pcmag.com/article2/0,4149,1078540,00.asp?kc=PCAV10209KTX1K0100360

    According to HardOCP, the R9800 again is faster when any AA or AF is applied, the 5900U coming to a grinding halt at anything above 4xFSAA...http://www.hardocp.com/article.html?art=NDg5LDQ= , and go on to say in their conclusion, "
    I don't ever remember not "backing up" my statements without some evidence, as you suggest, especially if someone asks for some proof....

    In light of providing you with just a small sample, I fail to see how the 5900U is the fastest graphics solution out there at the moment

    ;)
     
  21. Nic

    Nic TechSpot Paladin Topic Starter Posts: 1,928

    Fair enough, lets just conclude that both cards are good in different scenarios and leave it at that. They are both top cards and come pretty close in most benchmarks. Enough said.
     
  22. JSR

    JSR Banned Posts: 730

    preserverd

    just go to tomshardware.com............break out the scale.......done deal........yet.........i am interested in the scope of differing opinion.......too many issues need an explanation...........it really comes down to what direction one should take.......and, what is indeed an important attriute or set of attributes to obtain............yet, again, one might buy a card for the specific advantage it has over a different design by how it would affect him in some individual game title he spends most of his time on...........so, i think it's a kick, that this is such a ticking timebomb :grinthumb
     
  23. Phantasm66

    Phantasm66 TS Rookie Posts: 6,504   +6

    These things will NEVER be fast enough.
     
  24. PreservedSwine

    PreservedSwine TS Rookie Posts: 375

    LOL :grinthumb

    Ughh, Tom's is a decent site, but the VGA chart is just that, a fairly vague reference chart w/out alot of specific..
     
  25. JSR

    JSR Banned Posts: 730

    the reality is this

    that ati couldn't, for some reason, improve performance of it's 256 version card..........while on the other hand, nvidia improved itself in the 256 card, ............this is the slight advantage........so, when evaluating purchase, one has to realize, why bother with ati's 256 card, look at the prices of the 128 pro version (because there is a non-pro version that will fool you as to price), check out the spread between the fx 128/256 versions and decide.......... for the most part, i'd think the best value, would be the 128 pro version of ati's card
     
Topic Status:
Not open for further replies.


Add New Comment

TechSpot Members
Login or sign up for free,
it takes about 30 seconds.
You may also...


Get complete access to the TechSpot community. Join thousands of technology enthusiasts that contribute and share knowledge in our forum. Get a private inbox, upload your own photo gallery and more.