Tom Petersen, Director of Technical Marketing la NVIDIA, descrie performanta GeForce GTX 480 si o compara cu solutiile existente pe piata [ATI Radeon HD 5870]. Tom pune in evidenta puterea cipului GF 100 la partea de tesselation si ofera un scurt preview al tehnologiei 3DVision Surround.
 

Mult asteptatul GTX4xx a fost surprins in cateva poze la CEBIT, intr-unul din sistemele pregatite de nVidia pentru demonstratii. PC-ul rula tech demo-ul Supersonic Sled.  PC Games Hardware [care este de fapt si sursa acestor poze] a reusit sa faca cateva poze a ceea ce se presupune a fi FERMI. Din ce se poate vedea, placa are 2 iesiri DVI si 1 HDMI. Pentru alimentare se folosesc 1 conector cu 6 pini si 1 conector cu 8 pini. Pe stickerul placii se putea observa: GF-100-375-A3. Acest lucru ne spune ca avem de-a face cu editia a 3-a a cipului grafic [respin-ul de silicon no 3].
 
Se mai observa ca nu a fost updatata nici denumirea placii catre 4xx [GF-100-375-A3 inseamna GeForce 100 Model 375 Revizia A3]. Din moment ce avem de a face cu un GTX 375 ne putem astepta ca nu acesta sa fie varful de lance bazat pe FERMI.
  

 

 

 

 

 

 

Cea mai notabila prezenta la CES 2010 din Las Vegas este noua placa grafica de la nVidia bazata pe arhitectura FERMI [GF100 fiind numele de cod pentru varianta consumer].  

 

Placa poate fi vazuta la lucru de catre toti participantii, ruland in timp real benchmark-uri sintetice si de jocuri. Performanta speculata este de 100% peste GTX 285 si 36% peste ATI 5870. Aceste cifre nu au fost inca nici confirmate nici infirmate de catre nVidia.

 

Pentru cei care nu au putut ajunge la CES 2010, PCGamingXtreme aduce o parte din CES la voi. FERMI, ruland in timp real Unigine Heaven:

 

 

 

 

 

 

Problemele cu yield-urile de 40 nm la TSMC’sau accelerat dezvoltarea si reorganizarea Globalfoundries, viitorul partener al ATI/ADM in ceea ce priveste producerea cipurilor pe 32 si 28nm.

 

Parteneriatul intre ATI/AMD si Globalfoundries va incepe in primul sfert al lui 2010 iar primele cipuri care vor beneficia de noul proces de fabricatie vor fi procesoarele proiectate de AMD. Dupa ultimele discutii se pare ca ATI doreste sa faca un pas inainte si sa faca trecerea direct la producerea cipurilor pe 28nm in cazul procesoarelor grafice [viitoarea serie 6000], trecand peste procesul de fabricatie de 32nm.

 

Nu ar fi prima data cand se intampla asta, seria 2900 a folosit un proces de fabricatie de 80nm iar seria 3000 a trecut direct la 55nm, trecand peste procesul de fabricatie pe 65nm.

 

Puteti discuta acest subiect si pe forum:

http://www.pcgamingxtreme.ro/forum/viewtopic.php?f=18&t=75

 

 

 

 

 

 

  

 

 

 

 

 

In timp ce nVidia intampina tot mai multe probleme cu Fermi / GT300, ATI anunta astazi lansarea Radeon HD 5970, prima solutie dual GPU din seria 5000 si in acest moment cea mai puternica placa grafica de pe planeta.

 

Puteti discuta aceasta tema pe thread-ul dedicat de pe forum: http://www.pcgamingxtreme.ro/forum/viewtopic.php?f=18&t=40

 

ATI Radeon HD 5970 are o putere de calcul de aproximativ 5 TeraFLOPS, full suport in hardware pentru DirectX 11 si este construita pe un proces de fabricatie [litografiere] de 40nm [0.04 microni].

 

Specificatiile ATI Radeon HD 5970 sunt urmatoarele:

4,3 miliarde de tranzistori

2 GB DDR5 la 1 Ghz

725 MHz frecventa GPU

3200 procesoare shader / stream processors

4.64 TFLOPs putere de calcul

160 unitati de texturare

64 unitati ROP 

 

Imaginile si benchmark-urile au fost preluate de pe TECHSPOT, unde de altfel puteti citi un review complet: http://www.techspot.com/review/221-ati-radeon-hd-5970/

 

 

 

 

 

 

 

 

Si o serie de teste cu placa:

 

 

 

 

 

 

 

 

 

 

 

 

 

Pe langa faptul ca datorita problemelor intampinate in procesul de fabricatie la TSMC s-a amanat lansarea GT300 pentru aprilie-mai 2010,  se pare ca FERMI mai are si alte probleme. VR-Zone a facut o analiza detaliata. Numele de cod al Fermi este GF100, contine 3 miliarde de tranzistori si are o dimensiune de 500 mm^2. In comparatie cu Cypress [ATI 5870], care are nu mai putin de 2.15 miliarde de tranzistori, cip-ul de la nVidia ar parea un monstru. Se pare insa ca specificatiile “on paper” impresionante nu ofera in practica performanta la care ne asteptam.

 

In primul rand, desi performanta unitatii de calcul in dubla precizie este impresionanta, GF100 nu are nici o unitate de calcul in simpla precizie – care are un rol extrem de important in ceea ce priveste performantele in jocuri. Mai mult, cipul se pare ca nu functioneaza la frecventele pe care si le-a propus nVidia, acest lucru fiind alt factor care penalizeaza performanta. Frecventa tinta a fost de 1.5Ghz pentru procesoarele shader iar frecventa reala – calculata in functie de performantele masurate ale placii este prognozata la 1015 – 1230 Mhz.  In momentul in care nVidia are toate aceste probleme cu Fermi, ATI detine coroana celui mai rapid GPU si a umplut planeta de 5870/5850 si versiunile mainstream ale seriei 5000.

 

Intreg articolul de pe VR-Zone il aveti mai jos:

 

We all know it is constantly being delayed – but at the end of all those delays, what we were expecting was a stellar product from Nvidia. However, with today’s press release, certain inconvenient details are revealed. Let’s forget about the delays for now, and just consider the product itself.

 

The first Fermi GPU – GF100 – as we know for a while now, is a 3 billion transistor giant, taking a die size of around 500 mm². Compare this with the 2.15 billion transistor, 330 mm² Cypress on the same 40nm TSMC process, and you would be expecting a different class of product. Unfortunately, the details revealed today about cast an uncertain shadow over this basic assumption.

 

The first thing worth noticing is a complete and total absence of Single Precision performance figures or any comparison to direct competition – i.e. ATI’s GPGPUs. It is clear that Fermi’s real performance advantage would be Double Precision performance – had it hit the right clock speeds.

 

However, today’s press release suggests Nvidia have missed target speeds by a lot. To be fair, Tesla products do clock lower, though not by much. In fact, GTX 280 and Tesla C1060 were clocked the same. Even taking a generous increase for Geforce products, things are still uncertain. As a result, DP performance is rated at between 520 GFlops and 630 GFlops. Suddenly, ATI Radeon HD 5870 – which wasn’t even supposed to be a direct competitor – is performing right on par with 544 GFlops against Fermi’s supposed strong point.

 

Consider Single Precision – far more important for gaming graphics, and things turn rather ugly. GF100’s target speeds were reported to be 1.5 GHz for the shaders. Based on the 520 / 630 GFlops figures, the shader clocks can only be estimated at 1015 MHz and 1230 MHz respectively.

 

The SP theoretical performance from 512 CUDA cores? Between 1.05 TFlops and 1.26 TFlops. Even less than 1.05 TFlops, considering the lesser part is likely to have units disabled. Now, no amount of overclock can bridge the enormous gap to the smaller, already available HD 5870, which stands pretty at 2.72 TFlops. Even the mainstream HD 5770 clocks in at 1.36 TFlops! Barring a different clock speed for SP units, or other technology we are unaware of, this is a dismal performance from the Fermi shaders. Sure, Nvidia’s shaders are much more efficient, but this is just too massive a gap to claw back.

 

Then comes the price. A previous-gen C1060 released at $1699, falling to $1199. Compare this with fellow Geforce model, GTX 280, released at $649, falling quick to $500, and finally $300. The price of a next-gen C2070 is a whopping $3999. Nearly double the price as the previous generation C1060. Clearly, these are expensive products to make, so how much can Nvidia sell a Geforce version of Fermi for? Even the cheapest Tesla 20 variant, the C2050 costs $2499, nearly 50% more than the GT200 based C1060 flagship. Can Nvidia sell the $3999 Tesla product at $399 as a Geforce product?

 

So far, we are comparing GF100 to Cypress. Where, in reality, GF100 should be compared to Hemlock. 4.64 TFlops vs. 1.26 TFlops is not much of a comparison at all, however, CF limitations, ATI’s less efficient shaders aside.

 

The other, much less common rumour is the possibility of a dual-GPU Fermi product. Well, considering “Typical” power of Tesla 20 is 190W, this will be highly unlikely, at least for a while. Not to mention, Geforce products might end up clocked higher. A HD 5870s peak TDP is 188W, lower than GF100’s “typical” power! TDP is expected to be 220W, at least, and that is just too hot for a dual GPU product.

 

And we have not factored in the fact that GF100 is nowhere to be seen, and are unlikely to be on shelves in quantity for at least 4-5 months. Any further delays, and we will be looking at new products from AMD.

 

In the end, AMD have a solid product already available that is efficient, economical, scalable. Nvidia have ink on paper – and even that is not looking as promising as we may have hoped for. At this moment, we can only hope for “hidden” or “magical” gaming features which might bring about a revolution in how GPUs work. Short of that, all signs point to Fermi being in real trouble.

ATI HD5970 a aparut deja pe oferta unor magazine online. Pretul de lista pentru Sapphire ATI HD5970 se situeaza intre 505.39 si 514.46 de euro. Acceleratorul grafic de la Sapphire este identificat de un part number 21165-01-50R.

 

Specificatiile lui ATI HD5970 sunt urmatoarele: 2 GB DDR5 [2x1GB], 2 GPU-uri la 725Mhz si 2 x 1600 stream processors. Procesul de litografiere este de 40 nm, este oferit full suport DX11 si este prima placa dual GPU de la AMD din seria 5xxx.

 

Producatorul american EVGA a lansat o noua placa grafica care foloseste chipset-ul Nvidia GeForce GTX275 si nu orice fel de placa, ci una cu doua procesoare, unul dedicat procesarii grafice si al doilea pentru accelerarea in GPU a motorului de fizica nVidia PhysX. Printre aspectele care deosebesc PhysX de alte tehnologii prezente pe piata amintim simularea in timp real a fluidelor, ilustrata si in acest clip:

 

 

De asemenea, aceasta placa grafica dispune de suport complet pentru DirectX 10 si tehnologia Nvidia SLI care permite instalarea a pana la 3 procesoare grafice intr-un singur sistem, pentru un spor de performanta considerabil in jocuri.

 

EVGA Geforce GTX 275 CO-OP PhysX utilizeaza o interfata PCI Express 16X si dispune de o memorie totala de 1280MB de tip DDR3. Pe partea de conectica putem enumera un port HDMI (prin adaptor), doua porturi DVI-I si un port VGA (prin adaptor).

 

Sporul de performanta obtinut de aceasta placa grafica raportat la o placa grafica de referinta Nvidia GTX 275 (fara procesor dedicat pentru PhysX) poate fi vizualizat in graficul de mai jos:

 

 

 

 

 

 

 

 

 

 

EVGA Geforce GTX 275 CO-OP PhysX este asteptata in magazine in cursul lunii curente la un pret recomandat de catre producator de 349,99 USD.