Friday, October 15th 2010

NVIDIA to Counter Radeon HD 6970 ''Cayman'' with GeForce GTX 580
AMD is undertaking its product development cycle at a breakneck pace, NVIDIA trailed it in the DirectX 11 and performance leadership race by months. This November, AMD will release the "Cayman" GPU, its newest high end GPU, the expectations are that it will outperform the NVIDIA GF100, that is a serious cause for concern, for the green team. It's back to its old tactics of talking about GPUs that haven't even taken shape, to try and water down AMD's launch. Enter, the GF110, NVIDIA's new high-end GPU under design, on which is based the GeForce GTX 580.
The new GPU is speculated to have 512 CUDA cores, 128 TMUs, and a 512-bit wide GDDR5 memory interface holding 2 GB of memory, with a TDP of close to that of the GeForce GTX 480. In the immediate future, there are prospects of a more realistic-sounding GF100b, which is basically GF100 with all its 512 CUDA cores enabled, while retaining its 384-bit GDDR5 memory interface, 64 TMUs, and slightly higher TDP than that of the GTX 480.
Sources:
3DCenter.org, PCGH
The new GPU is speculated to have 512 CUDA cores, 128 TMUs, and a 512-bit wide GDDR5 memory interface holding 2 GB of memory, with a TDP of close to that of the GeForce GTX 480. In the immediate future, there are prospects of a more realistic-sounding GF100b, which is basically GF100 with all its 512 CUDA cores enabled, while retaining its 384-bit GDDR5 memory interface, 64 TMUs, and slightly higher TDP than that of the GTX 480.
195 Comments on NVIDIA to Counter Radeon HD 6970 ''Cayman'' with GeForce GTX 580
And in my honest opinion, your statement of AMD being cheap ass's using a 256-bit bus+32 ROPs vs the GTX 480's 384-bit bus and 48ROP's, really only states that AMD's design is more efficient. The GTX 480 has 50% more bus width and 50% more ROP's not to mention transistor count and die size, yet the HD 5870 is only 11% slower overall. Let's not get started on the GTX 470 which has 25% more bus width and 25% more ROP's....and fall's below HD 5870 perf.
I must say i completely disagree with your logic. Afterall, when there are source's like this showing AMD's new 960SPU card beating AMD's 1440SPU 5850, and almost going neck and neck with the 1600SPU 5870....i really fail to see how you figure Nvidia is going to bounce back.
This also means(if all the rumors for the past 2months r true) that AMD has increase their shader efficiency by roughly 34%. So imagine the performance of one of AMD's new chips at 1440SPU's. It's possible for Nvidia to make a come-back(which they will with time), but it won't be this gen IMHO
I agree to an extent in terms of transistor count and die size, and it physically does have more grunt, ATi just plain did freakin' well with the 5000 series. they started off fast and just kep getting faster for the following 6 months or so with better and better drivers.
AMD will most likely get a little more energy efficient and just a little more performance than before at a lower price point. Probably also featuring a 5770 rebadge with no 6-pin molex requirement for the HTPC crowd.
Probably a repeat for Nvidia without being quite as late and what we thought Fermi to be all along will be found in the GTX 580. AMD will most likely gain a little more marketshare on the GPU front and continue the trend of trailing Intel by a longshot in CPU tech and sales.
a fully unlocked dual GF104 has the potential to be 50% faster than a GTX480, it just seems logical to me to go down that path first, and get a new GPU completely right.
I think the dual GPU full 460 will be competing with the HD 6900 on the high end. 490 might be just that. I think the 580 is a silly rumor. but then again who knows.
WIN!
512-core GTX 485 on quick marketing here we come! :laugh:
shader don't directly affecting framerate and they aint the filling unit. a 10000sp with just 16rops will most definitely stuck at maximum frame rate while 128sp with 64rops will have 4x theoretically frame rate under same clock rate. more shader or powerful shader only apply the change of graphical quality and heavily detail loading. the reason why criticizing being cheap ass because they sell low spec gpu for higher price and all they doing are just clock up and stuff more shader and then another new price tag....it feels like it's not worthy to spend 300+dollar on something you can just swipe bios/overclock it in you house...while nvidia offer more feature and spec rich (cuda/physx/3Dvision/tessellation, 40/48rops and larger bus). in the same price gtx 470 offer far better performance/feature than what cypress can do but cost the same! the only criticism is the power consumption but WHO CARE? i just want my money to be spend in the right way then worry about this dirty ugly planet. want me to buy an amd card to save the planet?...i'd rather watch the planet die then pay for overprice low spec card like this.
unless amd can bring out some serious spec card with decent price i would never spend any penny just for the sack of power efficiency and saving the god damn planet....
That's what i have still yet to understand, these are top tier cards you're paying for, now of course this doesn't excuse inefficiencies, but if you're paying for a card like that at least pay for a sufficient amount of power and cooling for it if you're considering a card that costs $350 and up.
On a differnt note, AMD has definitely got a good performance per watt thing going with the 5 series, and i hope it continues to get even better.(and i hope Nvidia follows suit in that regard)
the lower my bills the more money i have to spend on other things, and if my gpu choice can save me near $150 in a year thats proberley 8 or 10 games that are on sale on steam (my pc is on 24/7 so power adds up pretty quick at $0.22 per kw/h).
But i do see your point, but cooling is a pain in the ass as no air con and its a tiny room with no window :laugh: don't ask :p so heat output means a lot to me.