Pics of gt300

[STEvil]

OMG, a Nvidia driver thread gets posted here that has some changes that might be considered newsworthy and it gets moved quick. Someone posts a thread for AMD's monthly driver in the News section and it takes forever to get moved. How long did that thread stay in the news section touting like 100 points in Vantage and a couple fps in the RE5 benchmark? I'm sorry I love this forum but it is the ATI clubhouse, at least in the news section.

Edit: It never got moved.

moved now

its a fanboy clubhouse if anything

[MadMan007]

Where do I demand a refund for the last fifteen minutes of reading this thread? :p It's probably just a mockup for show, big deal, companies do mockups all the time.

[Cybercat]

"LEAVE NVIDIA ALONE! ;_; "

So yeah. No negative speculation because it could actually imply that Nvidia would have really messed up big time this time?

We know for a fact that bigger chip means bigger costs.
We know for a fact that 384-bit memory controller means more complex PCB and more costs.
We know for a fact that Nvidia's CEO used FAKE card to show the upcoming card.
We know for a fact that Nvidia's product is late to the market, so it has been having some problems already.

Expensive card AND competitive pricing means less profit for the company.

So, based on this, what positive can you say about the upcoming product? If nothing, would you propose people to be quiet for the sake of "being fair"?

Wait, are we as consumers supposed to care how profitable a card is for a company? I know I don't.

384-bit is less than the 512-bit they've been using for over a year now, so I don't know what the hell your point is there. There's plenty of positive things to say about this architecture.

[saaya]

Why don't you all just stop freaking out and have some patience until it actually releases?

lol, why dont you tell that to nvidia?

i completely agree... show us a real gt300 card or stfu :P

[annihilat0r]

lol, why dont you tell that to nvidia?

i completely agree... show us a real gt300 card or stfu :P

Have no idea why people are so psyched to see a real GT300 and are talking trash because Nvidia is yet to show one.

Sorry but when was the first time we saw a 5870 card shot? One month before its release, right?

Then what makes you think there must be something wrong with Nvidia if they can't show a working card right now? At most it only shows that we won't have GT300 in the next month. Big deal.

[randomizer]

The trash talking is not because they haven't shown us a real card, it's because they showed a fake card. I'd have had more respect for them if they showed nothing and just said "it's not ready yet." But Jen-Hsun rarely commands respect.

[LordEC911]

Have no idea why people are so psyched to see a real GT300 and are talking trash because Nvidia is yet to show one.

Sorry but when was the first time we saw a 5870 card shot? One month before its release, right?

Then what makes you think there must be something wrong with Nvidia if they can't show a working card right now? At most it only shows that we won't have GT300 in the next month. Big deal.

Cypress and Juniper and then Hemlock were shown way before a month before launch.
ATi showed working silicon way back at Cebit and a wafer.

[saaya]

Have no idea why people are so psyched to see a real GT300 and are talking trash because Nvidia is yet to show one.

Sorry but when was the first time we saw a 5870 card shot? One month before its release, right?

Then what makes you think there must be something wrong with Nvidia if they can't show a working card right now? At most it only shows that we won't have GT300 in the next month. Big deal.

just think about it... would any of us be reading this thread or a similar thread if nvidia wouldnt have shown gt300 at all? this thread wouldnt even exist...

its like driving to work and then yelling "hey everybody, check out my new car" and its just a normal toyota or nissan city car... how can you complain about people ripping on you for that car? yes, theres nothing wrong with it, but obviously people expected something better when you tried to get their attention...

The trash talking is not because they haven't shown us a real card, it's because they showed a fake card. I'd have had more respect for them if they showed nothing and just said "it's not ready yet." But Jen-Hsun rarely commands respect.

exactly

well i doubt jensen is actually involved in business decisions at nvidia... but i must say, i have the greatest respect for nvidias engineers and their business managers... i mean the REAL business guys you never hear of or see in a meeting or presentation etc... nvidia has been run very very well from a business point of view in the past... and their engineering, gpu wise, has been excellent most of the time...

nvidias biggest problem atm is that the top guys are not really in touch with the rest of the company, mostly their engineering guys... in theory, jensens visions are great and he really knows whats going on and is swimming in the ether and all... but what he wants to do and what their engineers and tsmc etc CAN do is quite diferent and he seems to ignore it and think itll get solved by itself or if he insists long enough things will suddenly start to work miraculously

[saaya]

Cypress and Juniper and then Hemlock were shown at least 2 months before launch, if not 3.
ATi showed working silicon way back at Cebit and a wafer.

well who knows if those cards werent mockups and empty inside as well... and if the wafer was something else... but they were at least smart enough to fake stuff properly... we will never know how much of what we see in product launches is actually fake... sometimes there are small hints or something doesnt sound or look right... but this fake nvidia card is truly epic... ill remember that for the rest of my life

[Sly Fox]

Saaya, you say what I'm too inarticulate and lazy to say.

Props.

[largon]

show me where a 1.1tflop 4870 beats a 1 teraflop 285.

Furmark.
OCCT GPU test.
Two completely shader dependant apps in which RV770 utterly trounces G200.

[DilTech]

That's assuming it doesn't crash entirely largon... Remember, ATi 48x0 cards have a problem with both of those specific applications, because they weren't made to handle that kind of power draw and load.

Rename furmark's exe file and watch...

[saaya]

That's assuming it doesn't crash entirely largon... Remember, ATi 48x0 cards have a problem with both of those specific applications, because they weren't made to handle that kind of power draw and load.

Rename furmark's exe file and watch...

same with nvidia, but they implemented a throttle mechanism afaik

[DilTech]

Exactly why you can't really consider them good benchmarks to use Saaya. Not like you can really "play" furmark, no?

[mibo]

That's assuming it doesn't crash entirely largon... Remember, ATi 48x0 cards have a problem with both of those specific applications, because they weren't made to handle that kind of power draw and load.

Rename furmark's exe file and watch...

This is not true since catalyst 9.8. ATI now detects "uncommon" load on the different GPU parts and throttles the chip to not overload the GPU power supply.
See Anandtech 5870 article -> power consumption

[gamervivek]

show me where a 1.1tflop 4870 beats a 1 teraflop 285. just because ATi's theoretical flops are higher doesnt mean the card is faster in the real world. hell, a 5870 has double the specs of a 4870 but that doesnt show. i dont see how 2.3tflops is slow either. it will be competitive as a gaming card considering it has 2x more shaders, bigger caches and faster memory.

A GF100 doesn't have 2.3TFLOP SP performance.end of story.




Hu Ha!
and then there is also 8xAA and dense matrix multiplication.

[DilTech]

I read the review, and noted the 4890 and 4870x2 still crashed mibo.

The 5xxx series throttles it in the hardware, while the 4 series and previous did it in the drivers.

Either way, we're all now WAY off topic here.

[DilTech]

tesla card presentes on keynote seems to be a fake

http://forum.beyond3d.com/showpost.p...&postcount=262

Welcome to yesterday...

[Jamesrt2004]

its nice to see a mod getting in the action... as in debating in threads and still being part of community :/ i feel some others kinda worry and seemed to post less and less and JUST only enforce the rules which is a pain as theres a lot of knowledge on these forums...


anyways....

mid range ati going unchallenged for a lot of months (Even said my nvidia) surley thats a bad thing.. it would have been better for nvidia to send out the mid for now atleast, then you can kinda guess and get more excited about the bigger/better products anyway :S

just meh seems silly as obv its the midrange that makes the most money

[LordEC911]

Charlie's rampage continues...

[netkas]

Welcome to yesterday...

oh sry, looked last two pages and didnt found it.

[randomizer]

Confirmed: Fermi is faster in games than the HD5870

It's official, hard data and all.

[Newblar]

Confirmed: Fermi is faster in games than the HD5870

It's official, hard data and all.

yeah marketing data

[mibo]

Confirmed: Fermi is faster in games than the HD5870

It's official, hard data and all.

This is so LOOOL
It seems they have very few working chips (maybe even none).
From very few chips you can not see what frequency to expect from later mass produced gpus.
Of course Nvidia has to make their chip faster than 5870. The question is: how many chips per wafer can reach the needed frequencies? If many -> cheaper cards. If very few -> more expensive.
These fuad-news are just logic. There is no new info!
Also, the power draw will be comparable to 285 and 295 for single and dual gpu - what a surprise. 1. There is not much room for more power draw and more heat output (see the 295 lower clocking arguments). 2. The power draw will be defined by the frequency needed to beat ATI and the voltage needed for doing so.

[RaZz!]

I read the review, and noted the 4890 and 4870x2 still crashed mibo.

The 5xxx series throttles it in the hardware, while the 4 series and previous did it in the drivers.

Either way, we're all now WAY off topic here.

it's not the 4800 series in general. only 4870 based cards are affected. 4850s for example run furmark and occt without any problems.