I still bet we a 9800GTX type card before end of year

Zorachus

[H]F Junkie
Joined
Dec 17, 2006
Messages
11,308
nVidia has not released a brand new high end card since November 2007, the Ultra doesn't count because it is just a OC GTX, nothing too new.

While AMD/ATI did NOT put up much of a fight in the Spring, and now with the new 2950XT due out in the next month or so, it will be faster than a 8800Ultra, which is not saying much because now the 8800 series is almost one year old. So ATI will hold top dog spot even though it will be held by just a few %, the 2950XT will just barley beat the Ultra, I would think nVidia would want a high end part to gain the top spot again A.S.A.P.

Also with new games due out this fall like Crysis, UT-3, Frontlines-Fuel of war, there is a real need for a high end card to run these games at default rez of 2560x1600 that most [H]ardcore games run on their LCD's, or lower end rez of 1920x1200

So I still think we will see a 9800GTX or at the least a 8950Ultra or something come November. It is the calm before the storm, and no news yet means something good, look at this time last year in September 07 did we all know the 8800GTX was coming and was going to be such a beast ??


My point is come Nov/Dec there will be a nVidia VideoCard out that is much faster than the 8800Ultra
 
Ok. Cool. You couldnt have posted your deductions in another thread already dealing with this? I think theres already 172 or so.
 
nVidia has not released a brand new high end card since November 2007

I think you mean November 2006.

and now with the new 2950XT due out in the next month or so

Most sources believe it'll most likely be a 2900pro (downclocked 2900xt). 2950Xt and the 2950pro probably won't see the light of day until Q1 2008 according to DAAMIT's release schedule. DAAMIT isn't concerned with the High End right now, they simply want the mid range, in order to bring in something called, "money."
 
Sorry I just don't agree. As much as I might like to see one, I really doubt we'll get a new flagship card before the end of the year. Just because your spidey sense suspects we might, I just don't see it. Nvidia is just making too much money as it is.

Also you think most hardcore gamers run their games at 2560X1600. Uh, no. I don't think so. You're talking about a very very select few. I consider myself pretty hard core and I don't run at that resolution.
 
It's a possibility. nvidia has been very secretive in the past and has been giving plenty of misdirection during the year.

2 clues to the truth are:
1) CUDA documentation mentions a double precision shader part is due in "late 2007". http://developer.download.nvidia.co..._CUDA_SDK_releasenotes_readme_win32_linux.zip
and
2) nvidia (Andy Keane) said DP shaders were expected on Tesla by the end of the year. http://www.hpcwire.com/hpc/1625213.html Tesla is built on high end Quadro cards (e.g. Tesla C870 = 8800GTX).

I'm going to call that as 1+1=2. ;) (Separately I have read that DP shaders is only going to be a high end feature. It's a rumor at this point and a very disappointing one if true.)
 
I think you mean November 2006.



Most sources believe it'll most likely be a 2900pro (downclocked 2900xt). 2950Xt and the 2950pro probably won't see the light of day until Q1 2008 according to DAAMIT's release schedule. DAAMIT isn't concerned with the High End right now, they simply want the mid range, in order to bring in something called, "money."

My bad I mean November 06, and yes the more I read on the top secret stuff my friend knows about the new 2950XT will be a better cooling and faster mid range card for a good price
 
Sorry I just don't agree. As much as I might like to see one, I really doubt we'll get a new flagship card before the end of the year. Just because your spidey sense suspects we might, I just don't see it. Nvidia is just making too much money as it is.

Also you think most hardcore gamers run their games at 2560X1600. Uh, no. I don't think so. You're talking about a very very select few. I consider myself pretty hard core and I don't run at that resolution.


I think it's more wishful thinking on some people's parts. If we actually see one then hooray for commerce. If we don't then really what have any of us lost? I'm sure the real morons of this board will complain and 'lose faith' then jump ship to ATI because Nvidia never really was a quality company to begin with. :rolleyes:
 
Give us some examples; 9800ProGTX or 9700Ultra :D

I believe quite firmly that the next REAL upgrade for the NVidia line will be sold under the name of:

"X" or "Gen X".

Some shit like that.

They're going to skip past the 9's and go straight to ten.

Really, mark it down.
 
Guessing wouldn't make you look like an idiot. That's what rumors are for. :p
 
haha..I can totally imagine ebay sellers selling it and then you buy it and it comes and you get an old ati 9800 128mb..lol..

In geek talk " cool look what I just bought off e-bay a brand new 9800Pro, I think it's the new nVidia card and it was only a hundred bucks what a steal, lets open the box and check this bad uber card :eek: WTF is this
 
All the fan boys are starting to get screwed. If ATI had put up a better fight, I can guarantee that we'd see GTS 640s as low as 250$ right now, and possibly a whole lot more initiative from Nvidia to get a faster card on the market. They're sitting on their rears, milking the low end and midrange with high yield parts because lack of competition allows them too, and because it's more lucrative than cranking out better high end cards ASAP. The stupid GTS 640 has actually gone up in price recently. It's ridicules.
 
If Nvdias R&D has kept doing their homework NV should be able to put out a card that will womp the new thing from ATI (If its not another assumption pulled straight out of a nerds pimply ass.). They have a chance once agian to dominate the market like they did this year. Not releasing a card would lose them alot of $$$ and respect.
 
I really hope ATI gets a die-shrunk R600 with functional AA hardware out soon, if only go force nVidia to get off it's ass and stop charging $500 for year-old cards.
 
2950XT faster than the Ultra? Yeah, right.

It'll be exactly the same speed or slower, just like the r600 was the same as/slower than the 8800 GTS.

It's not a walk in the park to create a component that beats the most advanced video ever made. The GTX still carries that title now, and may well do so for another half a year.
 
I'm going against the grain here....

I'm glad to see there isn't some new high end card or a refresh out already. Makes it feel like people are getting more for their hard-earned money with some sort of longevity. Upgrading every 6-8 month can be a chore for some people (not all).

There really hasn't been a game within the past year and a half that can't be played at high settings and res with last gen cards (x19xx & 79xx cards) especially in SLi/Crossfire. Most newer games included.

Sure competition lowers prices and provides a catalyst for better hardware but sometimes its just not necessary. Many enthusiasts buy a new video card simply for the sake of having one--and that's fine for them, just not the majority (even in the enthusiasts circle). All 6-8 month GPU cycles are doing is fattening manufacturers wallets without even being necessarily needed in most cases.

I know I'm in the minority but I kind of like where this is going, however, I do have a problem with the prices being static.
 
I really hope they launch a faster card. Its time for my upgrade cycle at the end of this year and I just can't force myself to pay an arm and a leg for an old card (a great card non the less). I usually buy the best card available to hold me off for two or so years. I really wish I bought one last year...
If I was nVidia I would just get the mid range refreshed as thats where the money is (despite what gamers think). They should keep any new stuff underwraps and continue to refine it as I think Amd and Intel are going to give them a run for their money soon, not to mention Sony and Microsoft cutting in on gaming hardware sales (people moving console only). And no, a year or two really isn't that long.
 
Either way you slice it , the 2900xt is a under preforming design. It just cant handle AA well enough (as to why that is with a massive 512-bit memory interface and even DDR4 ram ..not sure) to even come close thus far to a 8800 gts (which is the market they aimed for) with AA turned on. 8800 gts doesnt have near the memory bandwith yet achieves betters fps through design. A die shrink wont suddenly enable Ati to pull way ahead in benchmarks and then suddenly change the minds of thousands of 8800 users into ebaying there cards for the brand new "2950XT" .

As it stands right now Nvidia has no reason to release a faster 8800 GTX variant until they hear Ati streaming something worthwhile to challange them down the pipe.
 
Either way you slice it , the 2900xt is a under preforming design. It just cant handle AA well enough (as to why that is with a massive 512-bit memory interface and even DDR4 ram ..not sure) to even come close thus far to a 8800 gts (which is the market they aimed for) with AA turned on. 8800 gts doesnt have near the memory bandwith yet achieves betters fps through design. A die shrink wont suddenly enable Ati to pull way ahead in benchmarks and then suddenly change the minds of thousands of 8800 users into ebaying there cards for the brand new "2950XT" .

As it stands right now Nvidia has no reason to release a faster 8800 GTX variant until they hear Ati streaming something worthwhile to challange them down the pipe.

you dont know why? its because there is no dedicated AA hardware in the 2900. this is a first. all AA is done via shaders. ATI just made dumb decision.

IF ATI manages to add dedicated hardware to the 2900, even if only as current as the x1950 series, and call it the 2950, it WILL reach/exceed Ultra performnace levels. but im not holding my breath.
 
you dont know why? its because there is no dedicated AA hardware in the 2900. this is a first. all AA is done via shaders. ATI just made dumb decision.

IF ATI manages to add dedicated hardware to the 2900, even if only as current as the x1950 series, and call it the 2950, it WILL reach/exceed Ultra performnace levels. but im not holding my breath.

Ya but that would require a major redesign and 65nm die shrink to say the least .. considering how much power and heat the 2900xt outputs I dont see a feesable method to make that happen.
 
Ya but that would require a major redesign and 65nm die shrink to say the least .. considering how much power and heat the 2900xt outputs I dont see a feesable method to make that happen.


i dont think semi conductor are your speciality and i also dont think you are in any position to assess feasibility. Further, your belief that the 2900's AA issue was attached to its 512 bit bus and saying you are "not sure" about the real real reason for its AA faults, honestly just proves your full of sh*t.
 
All the fan boys are starting to get screwed. If ATI had put up a better fight, I can guarantee that we'd see GTS 640s as low as 250$ right now, and possibly a whole lot more initiative from Nvidia to get a faster card on the market. They're sitting on their rears, milking the low end and midrange with high yield parts because lack of competition allows them too, and because it's more lucrative than cranking out better high end cards ASAP. The stupid GTS 640 has actually gone up in price recently. It's ridicules.

Agreed.


If Nvdias R&D has kept doing their homework NV should be able to put out a card that will womp the new thing from ATI (If its not another assumption pulled straight out of a nerds pimply ass.). They have a chance once agian to dominate the market like they did this year. Not releasing a card would lose them alot of $$$ and respect.


Uhm...where are you living? Nvidia is still king of the hill. That they may not have a new part ready means little to nothing. The smartest thing to do in this case is exactly what they may be doing. Wait on a new part, refine and tweak it to insane capabilities...when the competition releases their new part that will probably under perform compared to yous (assuming it's in the same architecture class as their recent one), blow them out of the water again with a new part.

Will that strategy hurt their bottom lines? Probably not. Will it cause them to lose "respect". LMAO...definitely not.
 
Uhm...where are you living? Nvidia is still king of the hill. That they may not have a new part ready means little to nothing. The smartest thing to do in this case is exactly what they may be doing. Wait on a new part, refine and tweak it to insane capabilities...when the competition releases their new part that will probably under perform compared to yous (assuming it's in the same architecture class as their recent one), blow them out of the water again with a new part.

Will that strategy hurt their bottom lines? Probably not. Will it cause them to lose "respect". LMAO...definitely not.

I must say I am shocked, I had to respond to this post. I cannot believe someone here actually has some business sense (one of the few here imo). Everyone thinks being first matters, you can be first with a rushed product and fail. Timing is everything, being first at something is just the shallow part of the statement, there's things like quality control (releasing when it's right/ready). Your competitor can catch up anytime they make a killer product or when you screw up. It's doubtful that ATi can match G80/9x's arch (possible though), but one thing Nv has at it's control is to not screw up by rushing a product because .05% of the market cries about it :rolleyes:.
 
I must say I am shocked, I had to respond to this post. I cannot believe someone here actually has some business sense (one of the few here imo). Everyone thinks being first matters, you can be first with a rushed product and fail. Timing is everything, being first at something is just the shallow part of the statement, there's things like quality control (releasing when it's right/ready). Your competitor can catch up anytime they make a killer product or when you screw up. It's doubtful that ATi can match G80/9x's arch (possible though), but one thing Nv has at it's control is to not screw up by rushing a product because .05% of the market cries about it :rolleyes:.

What we dont know though is....is the part right/ready or not... It could very well be ready for a November release.
 
Well i would love to know who runs at 2560x**** resolution? If anyone can find me a descent priced monitor that supports that I would gladly appreciate it. I'm at the crossroads, thinking of upgrading my gts 640 to gtx/ultra. Which one you guys think I should get? It will be xfx, heard they have the highest clocks. Or should I wait and see if they will come out with another graphics card within a month? If ati came out with something that beats the ultra and under $600 bucks I would def. be interested, as my IP35-Pro only supports crossfire ; (

Thanks, Mike
 
Sorry I just don't agree. As much as I might like to see one, I really doubt we'll get a new flagship card before the end of the year. Just because your spidey sense suspects we might, I just don't see it. Nvidia is just making too much money as it is.

Also you think most hardcore gamers run their games at 2560X1600. Uh, no. I don't think so. You're talking about a very very select few. I consider myself pretty hard core and I don't run at that resolution.

Yup, pretty much it. Why would Nvidia want to release a new card right now when the 8800's are still selling great? I believe I was watching a show on Nintendo's portable game systems once... They said that the GameBoy Advance SP was ready well before they released it, but the only reason they didn't release it, was because the GameBoy Advance was still selling great. Nvidia is using the same concept right now.
 
Ya but that would require a major redesign and 65nm die shrink to say the least .. considering how much power and heat the 2900xt outputs I dont see a feesable method to make that happen.
No, it wouldn't require a major redesign (a redesign, yes). It isn't that much logic. I'd think it's entirely possible on 90nm.

And it's 'feasible'.
 
I must say I am shocked, I had to respond to this post. I cannot believe someone here actually has some business sense (one of the few here imo). Everyone thinks being first matters, you can be first with a rushed product and fail. Timing is everything, being first at something is just the shallow part of the statement, there's things like quality control (releasing when it's right/ready). Your competitor can catch up anytime they make a killer product or when you screw up. It's doubtful that ATi can match G80/9x's arch (possible though), but one thing Nv has at it's control is to not screw up by rushing a product because .05% of the market cries about it :rolleyes:.

Amen brother. :)

What we dont know though is....is the part right/ready or not... It could very well be ready for a November release.


Indeed it could. Alternatively if Nvidia senses that their part may be lacking or could use some definite refinements they could wait on the part and release later. Ideally they'd want to catch the holiday season for maximum sales however they'd be more hurt by pulling an ATI/AMD and releasing a product that quite simply is not ready and does not perform as anticipated. THAT would hurt 'reputation'.
 
From a fellow forum member :) :)
According to the information given at the same conference discussing the 1TFlop "beast," nvidia has adopted a yearly release cycle. The high-end should appear in Q4 of the year, with the mid-range parts coming the following spring.

I agree that yes by the end of the year for sure there will be a brand new high end VideoCard released to be able to play Crysis, UT-3 and other new games the way they were meant to be played
 
i dont think semi conductor are your speciality and i also dont think you are in any position to assess feasibility. Further, your belief that the 2900's AA issue was attached to its 512 bit bus and saying you are "not sure" about the real real reason for its AA faults, honestly just proves your full of sh*t.

I thought I'd read the AA logic on R600 was "broken". They couldn't solve the problem because it was already so late to market and they ran out of time and money for the project. If they fix that issue, I could see a die shrunk revision making substantial gains.
 
Also you think most hardcore gamers run their games at 2560X1600. Uh, no. I don't think so. You're talking about a very very select few. I consider myself pretty hard core and I don't run at that resolution.

Agreed. Almost nobody has a 30inch display. 1680 x 1050 is might be the most common widescreen resolution. Most gamers are probably at 1280 x 1024.

The install base for the displays just isn't there yet, but I do think higher resolutions are where the hobby is headed. Part of the problem is that the response times aren't as good on the really really big high-def displays.

I'd love to go up to 2560 x 1600 though. My personal opinion is that 1680 x 1050 looks really good but there is room for improvement.

By the way, I read the marketing of higher resolutions by developers as code for "we are codeveloping games for PC and console now, so if you want to get a better experience on PC, you are going to have to just up the resolution, because we aren't going to push the hardware that hard anymore." That is more or less what Epic Games and Valve have led me to believe.
 
Back
Top