[Solved] GT300 specs, gtx380,360 gts350,340

This site is generally pretty reliable in regards to new hardware. Looks like NVidia has some serious firepower coming our way. Hopefully, info is accurate and prices aren't obscene! Based on just the numbers they give, theoretically gts 350 and above will outperform 5870. Of course, these cards aren't available yet and we don't know what real world performance will actually be..... but still something to look at and drool over!


Main page
http://www.techarp.com/showarticle [...] =88&pgno=0


Nvidia

//www.techarp.com/article/Desk [...] _4_big.png



For the AMD guys

http://www.techarp.com/article/Des [...] _4_big.png


It's hard to imagine; 280+ gb of bandwidth form a single chip card. It says max board power for the 380 will be 225w. I wonder if that means they snuck in under the limit using 2-6 pin connectors or have to have a 6+8 pin for OC headroom. I wish it was out NOW!

The fillrates of the two cards are virtually the same, but the bandwidth is about 15% higher on the 350. I would guess that the bandwidth with the help of the larger buswidth would mathematically push the 350 to a slim advantage.

Well I wouldn't doubt that Nvidia will take the single gpu crown back, IN 3 MONTHS.

Memory bandwidth has has very little to do with overall performance.

You're right about bandwidth having little to do with overall performance, but everything else being "mostly" equal, tiebreak (theoretically) would be in the bandwidth.

I guess we will see, IN 3 MONTHS. If its faster than the 5870, it doesn't matter what price it is, ATI will just lower prices.

2 months actually for the 380 (nov 27 is listed as release date). Besides, as far as AMD goes, and this is not meant as a slight at all, being the value leader- which I think they have been for the most part- is still a good position to be in. The midrange is the vast majority of sales anyway.
That being said, I still am seriously excited about the gt300 if these numbers are factual. With the 380 being out for the holiday season and around th windows 7 launch is a huge plus for NVidia. If it had fallen to the first of the year, it would have looked much worse.

what a bullshit . Whats the date on that table?

~2 Bil transistors is just no way. Ati went >2x to double up everything. Nvidia is more than doubling so I guess min is ~3 bil and more.

Also that table must have been pulled up their arses, checking out the 4 cards from 1 chip (which is actually super waste of silicon for the 340 having half of everything but still HUGE chip.

total crap

They can't. That will choke off their previous gen cards, and remember, AMD is still in major financial trouble. They need a profit, now, not market share.

Lower the 5870 to $350, and you have to lower the 5850 to $275 to compensate. This kills off the 4870X2, which would need to drop into the ~$250 range, and the 4890, which would need to be priced down around $199. Nevermind what that does to the 4870, 4850, 4770, and the 4670 as a result. AMD simply can't afford to do that right now to cater to such a small group of customers.

Wasn't the die size for the 285 massive as compared to the 48xx series. Even with the shrink to 40nm, a 50% increase in transistor count might not be as unrealistic as it sounds. Plus with the 512 buswidth they probably had some power restraints.

from:http://www.pcgameshardware.com/aid [...] d/Reviews/

If the die size was already double the 4800, I think doubling the transistor count would be difficult. The yield would likely be so small nobody could afford what they would likely have to charge.

Assuming that is all true, Nvidia have given up on having the fastest single card, because there is no way to X2 a chip of this size without a die shrink.

It should be interesting to see what AMD do here. If the 5870x2 is released first, it will put a real dampener on the g380's release because there is no way the g380 can be as fast. Another alternative would be to wait and just release it straight afterwards, sorta for shock value. Give Nvidia a day or two of having the crown back

Either way, ATI has the fastest card this round. The 5870x2 cannot be beaten by any single gpu, and this nvidia cannot be doubled up without a die shrink.

I don't think that is necessarily true. I could see them multichipping perhaps the gtx360 to make a 395. I'm not sure if this is any more feasible, but still easier than 2x 380. I'm quite sure NVidia has not given up on a multichip card. It's too bad we have to wait until November for the 380. It's going to be very interesting when this thing hits the market.

If this is all true... (including that release date of the gtx380) I will be hopefully stepping up to a GTX380 from a GTX295 through EVGA it seems. That would definitely be pretty sweet if all these numbers are true.

I have no doubt the card will be fast....very fast. I'm just afraid of what it will cost. If you think NVIDIA will sell that for $450 then you probably have your head in the clouds.

If Nvidia mange to double the performance of the GTX280 (and that's a big if) then Nvidia won't be any better off then they were in the last generation.

Why not? AMD doubled the 4870 with the 5870, why can't Nvidia do the same? Actually, if they went from 240 to 512, they more then doubled the shaders. And by moving to faster GDDR5, they won't have any memory bandwidth problems either. Even if this is nothing more then G(T)200 doubled, then it should still be fast enough to be faster then the 5870. They will have their die size problems, but that won't take away the faster then 5870.

See above.

Why would it have to be?, surely if you wish to remain fair as it's something Nvidia never does, if so then the g380 should only have to match or beat the 4870 x2 just as the 58570 was supposed to beat the GTX295, next gen single card versus last gen dual card no?

Well I for one still only rank single GPU cards against single GPU cards, hence why I was taking the pee when people were posting the power figures for the 5870 & 295 giving it all "ooh and ahh it uses so much less power, that's amazing!!" when if it's only got one GPU on a smaller process that was only to be expected really.

Stranger things happen at sea, apparently.

Oh I don't expect or believe that they will take the lead (unlike some), it would be nice sure but it hardly tops my list of 'things that must happen to keep me going' I'm far more interested in the (hopefully) plethora of cards that may be available to me later next year when I come to replace my gaming rig. By the way, have you got your 5870 yet?

Fair 'nuff. There is however a shop in London where you could get one over the counter...instantly!, just saying.

I would take this "data" with a grain of salt, there are a lot of inconsistencies. But lets say for argument sake Fill rate and bandwidth is correct, then GTX350+ indeed has a chance of beating 5870, real games will show if its so. Still I have some doubts Nvidia vastly improved AA back-end as did ATI (100% outperforming GTX295 at 2560 AA 8x!), so its very much possible GTX350-360 will lose in ultra high-end configs, but GTX380 should win IMO. Still I see very little Nvidia can do about 5870 X2, 6+ bln. transistors is a bit too much on one card at 40nm.

Lets not forget multiple displays benchmarks will start soon, and I have little doubt Nvidia has no answer to eyefinity either.

Raven makes a good point that I haven't brought up yet.

ATI were held back by TSMC on 40nm, and are still being held back if reports are true. However, just because they couldnt get cards released didn't mean they stopped innovating or progressing.

The 5870 could have been ready and waiting on TSMC getting their *** sorted out. If that is true, ATI are in an incredibly commanding position, with x2 probably ready already and a 5890 not far off. It could be possible that a 5890 might beat the gt380 quite quickly if required.

Also, this 5870 is considerably undervolted at reference. Not a lot has been said about this yet, but expect to see 1ghz 5870's before the end of October.

I know I like to paint rosey pictures for AMD, but I don't think I'll be too far out in the end.

Ahh, but when did they actually start?

Pretty much.

I'm a bit narked tbh because GTX275 stocks do not seem to be getting replenished nor does their price seem to be dropping, so personally I'm not a happy camper.

Yes, jenny is an AMD fanboy.

Some of this is big, others not so much. New DX level? Its not like that came out of left field. They new it was coming, and what they'd have to add to support it. I'm sure they can also build a tesselator. The big problem they'd have with that is where to put it on the chip. They also don't have to develope the new NM process, TMSC is doing that. They do have to get it all working together, but that should be a "simple" process for Nvidia.

Their big problem is increasing the SP, ROPS, etc. Adding a tesselator. Adding all the new things on the chip, its going to be huge. Even at 40nm. It will be fast yes, but it will cost, probably be hot, and expensive. AMDs 5870 will perform slower, but will be the better buy. And as a gamer, I'd much rather have eyefinity then GPGPU.