Rant @ Nvidia

[rant]
Yeah, ATI came out with some mean cards this year, so far they have been all but crushing Nvidia. Good thing they have the production of NForce2/3 chipsets to fall back on. Hopefully the deal with the 5900 and the NEWEST drivers is not another way that Nvidia has found to cheat. This is rather hard to believe though since the scores for the card did not go up in any benchmark tests other than 3DM2003. Nvidia just does not have the credibility it used to, they have lost the trust of thousands of their customers and should redeem themselves right away. I think that it would have been better had they apologized for cheating right after they got caught rather than deny what was obvious and possibly go on to create more ways to cheat that we cannot as yet detect. If Nvidia wants people to buy their cards rather than ATI's, they should get their asses in gear and make some improvements. Yeah, yeah, I know that the technology that the FX line of cards is built on is state-of-the-art but that does not seem to be cutting it. Why is it that ATI Radeons can get such excellent performance at substantially lower clock speeds (heh, sounds like AMD) and the technologically superior FX cards can only do the same by cheating? Nvidia has no doubt taken ATI's cards and torn them apart in order to try to find a solution to this dilemma of theirs, maybe to some avail but that is yet to be determined. Nvidia's first step to regaining the market though should be to regain their rather damaged credibility and the respect of their former devoted customers. NOT by cheating on benchmarks, by actually improving performance. I guess all I'm trying to say is that I'm rather disappointed with Nvidia at the moment. Though I used to be a devoted NVidiot, as Ed calls us, my next upgrade will most definitely be ATI unless Nvidia shapes up and fixes its errors. (Same thing again, just replace NVidiot with AMDroid, ATI with Intel and Nvidia with AMD.)
[/rant]

ooohh lordeh!!! bring down the poweh to nvidia!!!



no but really... ah crap i forgot what i was going to say, really. No i was gonan say something. Honestly.

oh yeah now i remember!!! What i was going to say is since the FX is new tech, people will buy. Theres always that guy that buys something because its new, like the 5200 for instance, i wouldnt buy that ever. Also nvidia has all teh big (gay) companies like dell and compaq to buy all their stuff cause of contracts and such. You know how it goes, some people dont buy some products cause they think it is, how should i say, "dirty." some people just think that AMD is a bad company cause Dell or Compaq doesnt use them and they dont have commercials. ATI might have the same kind of problem. People that dont know better wont buy them cause nVidia has a reputation (that is getting destroied by the way).

But once ATI comes out with new tech... watch out everyone...

If the 2003 tests can be 'taken off the rail' and still deliver the same results, then nVidia will have successfully implemented a good, honest application optimization.

What worries me is that while they may have proved they can do it (Hey, by doing thisThisNthis we improved performance, legit!) they WON'T do it for real DX9 titles that come out, unless they're used for benchmarking. Obviously, Doom III WILL be a benchmarking app of choice, and if you didn't see that one coming, turn in your keyboard for a cluemeter. Doom III will recieve targetted optimizations via driver updates, we can all guarantee.

How about other games that sell well, but aren't used regularly for benchmarking?

That's what worries me, IF the latest improvements turn out to be legit.

MoonWolf said:

Good thing they have the production of NForce2/3 chipsets to fall back on.

Click to expand...

The nForce chipsets are nothing they can fall back on...period. The occupy about 5% of a 30% market...definatly not a money maker...and certainly nothing that they can "fall back on"

Same here, these latest drivers that Nvidia has released appear at first glance to be legit... the camera was taken off the rail and there was no visual difference in the scene, it appeared to have been fully rendered. The thing that makes me apprehensive is the fact that the scores of ONLY 3DM2003 improved, some of the other benchmarks actually decreased by 1% or so. If Nvidia just keeps making optimizations for benchmarks and does not try to make any real improvement in day to day applications and actual games, the performance boosts are worthless.

but that market is getting increasingly bigger. People will more and more be wanting one. Hell, im gonna get one with my next board.

Re: Re: Rant @ Nvidia

Evnas said:

The nForce chipsets are nothing they can fall back on...period. The occupy about 5% of a 30% market...definatly not a money maker...and certainly nothing that they can "fall back on"

Click to expand...

Sorry, I did not make myself clear. I meant that they could fall back on NF simply to keep their company alive, not to be able to continue production of expensive video cards. And maybe not, with the coming of Opterons and Hammers (if those are not completely worthless also) NF3 may be in high demand.

well i certainly wouldn't call 1700 tbred b's worthless. They may not get the highest scores but a 1ghz overclock on 1.467 processor is certainly nothing to scoff at. then to add to that it only cost me 50 bucks. try to get an intel running at 3.06 ghz for 50 bucks. not this year.

thats a very good point. We have to look at cost levels here. Cost is a very important factor to me as i dotn have money growing out of my ***. I rely on AMD to provide me a solid processor at a resonable price. 50 bucks for a 2.4 ghz processor is NOT BAD AT ALL. you barley have to push those things to get it to 2.4.

Re: Re: Re: Rant @ Nvidia

MoonWolf said:

Sorry, I did not make myself clear. I meant that they could fall back on NF simply to keep their company alive, not to be able to continue production of expensive video cards. And maybe not, with the coming of Opterons and Hammers (if those are not completely worthless also) NF3 may be in high demand.

Click to expand...

Well right now, their money making in video cards is really whats fueling their chipset production

InThrees said:

If the 2003 tests can be 'taken off the rail' and still deliver the same results, then nVidia will have successfully implemented a good, honest application optimization.

What worries me is that while they may have proved they can do it (Hey, by doing thisThisNthis we improved performance, legit!) they WON'T do it for real DX9 titles that come out, unless they're used for benchmarking. Obviously, Doom III WILL be a benchmarking app of choice, and if you didn't see that one coming, turn in your keyboard for a cluemeter. Doom III will recieve targetted optimizations via driver updates, we can all guarantee.

How about other games that sell well, but aren't used regularly for benchmarking?

That's what worries me, IF the latest improvements turn out to be legit.

Click to expand...

There is no good or honest optimisation in 3dMark . NO optimisations are allowed , and any present are cheating .

Furthermore they still aren't rendering off the rail in gametest 2

I don't think nv needs to worry... their FX cards, whether good or not (I'll leave that to the many other threads which discuss this matter, HINT HINT) are definitely selling.

Heck, three weeks ago I stepped into a CompUSA and found a TNT2 on sale for $70. That has to mean something...

Cowboy X said:

There is no good or honest optimisation in 3dMark . NO optimisations are allowed , and any present are cheating .

Furthermore they still aren't rendering off the rail in gametest 2

Click to expand...

I'm sorry, I don't agree with that.

DX9 is still relatively new, and since 3dMark03 is a DX9 benchmark (and currently the only one I'm aware of, but it's not like I actually care enough to look for more) the optimizations could be generally DX9 oriented, in which case 3DMark03 would benefit, while 2001SE would not.

I don't know how you can claim bona fide 'optimizations' are not allowed - if they can increase efficiency, there shouldn't be some mystical 'rule' or 'industry standard' preventing them from doing so. Note, I am not lumping in the clipping fiasco into 'optimization' - that was a pure and simple cheat, which would give misleading results.

I think we won't really know about these new drivers until a good hardcore DX9 stress-test title is released, and tests can be run with the new drivers (or the current at that time drivers) and some of the initial drivers.

The Futuremark rules state that no 'optimisations' which haven't been approved by them are to be allowed ( they haven't approved any by the way ). End of story . That's why ATI took theirs out of CAT 3.5 , even though they didn't decrease image quality or decrease the work done by the card .

Most developers agree that putting optimisations and application detection in drivers for benchmarks is wrong especially when you sneak them in and they decrease quality .

The purpose of a synthetic benchmark is to have a standard where you control and understand the variables . This makes the results valid as hardware stressors and sets a level playing field . When these hidden routines are put in place the results can no longer be trusted and are close to useless .

Nvidia needs to leave optimisations for games ! Not synthetic benchmarks or timedemos .

Edit : spelling

Personally I would much rather see them spend all of their time optimizing games, not 3dmark2003. While the new optimizations may very well be legit dx9 optimizations, isn't it kinda fishy that they haven't said something? Saying that your new optimizations are dx9 would give some reasurrance that you are not only optimizing 3dmark2003, and since they haven't done this I doubt that it is something like that.
Even if whatever they have done is legit, it is completely worthless in my opinion. I don't shell out $500 to look at 3dmark run all day long, and bask in the glow of the numbers at the end. I might run a benchmark once, maybe even twice, but I am really shelling out the $500 to play games. So to me, image quality and game optimizations are much more important that 3dmarks.

Sterculus ................. precisley , and their current set of drivers improves everything but image quality and games !

I'll just do the smart thing and wait till there is some really good DX9 games out there before makeing the jump to a DX9 card. By then the performance gain MIGHT be worth the insane asking price.

snvpa said:

well i certainly wouldn't call 1700 tbred b's worthless. They may not get the highest scores but a 1ghz overclock on 1.467 processor is certainly nothing to scoff at. then to add to that it only cost me 50 bucks. try to get an intel running at 3.06 ghz for 50 bucks. not this year.

Click to expand...

I never called them worthless, I was talking about the Hammers/Opterons that AMD seems to be having trouble with. I love the current prices for AMD processors, it fits right into my budget, along with that of many other people. The quality of the XP line of processors is not being questioned, I'm just wondering if Hammers will be worth it and whether or not the performace will justify they price you are likely to pay.

Sounds like a plan. I want to see HL2 in DX9 but I don't want to pay out of the *** for new technology that is still maturing. When all the kinks have been worked out and the price drops significantly(day after thanksgiving anyone?) I will buy one, but I'm sticking with my ti4200. I haven't ever had a Radeon, but right now I'm leaning towards a 9800pro .

To me, there should be no optimization at all. I use 3d Mark as a way to test my overclock, to see how much greater I have goten or how great a product is. I do not use 3dMark, however, for normal everyday use. Its the normal everyday use stuff that matters. If nVida is sitting there optimizing the 3dmark benchies to get higher and higher, then thats pathetic. If they cant make it to where the whole aspect of games and applicatons is improved then its sad.