Nvidia GeForce 600 serija

[djalex]

* The GeForce 600 Series will use a new Kepler style architecture, successor to the Fermi style architecture used by the 400 and 500 series.

* Will be the first Nvidia graphics card series to support the new PCIe 3.0 interface, which delivers a 8 GT/s bit rate (effectively doubling the PCIe 2.0's bandwidth).

* Manufactured using a 28 nm fabrication process opposed to 40 nm.

* Used XDR2 memory.

 

[Rapture]

XDR2???, ajd i to da vidimo , kompletna tabela je malo kontradiktorna dosadasnjoj tradiciji nv.

 

[MaTRaX]

a sta je XDR2 ?

 

[Lukija]

Bilo bi fer da se navede i izvor odakle su uzeti podaci.

 

[Rapture]

a sta je XDR2 ?

Naslednik gddr5, navodno duplira protok, ali i dosta skuplja od 5-ce, ma tabela je 99% fake.

 

[Steamroller]

Tabela izgleda kao da je sa wikipedije...

 

[djalex]

Da sa Wikipedia-e, postavljena je jos 24.12.2011. god. Sacekao sam ovih nedelju dana da vidim da li ce nesto da je menjaju ali ostala je ista. Sacekacemo jos koji dan/nedelju da vidimo predstavljanje nove 600 serije.

 

[djalex]

a sta je XDR2 ?

Changes from XDR DRAM

Signaling

In addition to a higher clock rate (up to 800 MHz), the XDR2 differential data lines transfer data at 16 times the system clock rate, transferring 16 bits per pin per clock cycle. This "Hexadecimal Data Rate" is twice XDR's 8× multiplier. The basic burst size has also doubled.

Unlike XDR, memory commands are also transmitted over differential point-to-point links at this high data rate. The command bus varies between 1 and 4 bits wide. Even though each bit requires 2 wires, this is still less than the 12-wire XDR request bus, but it must grow with the number of chips addressed.

Micro-threading

There is a basic limit to how frequently data can be fetched from the currently open row. This is typically 200 MHz for standard SDRAM and 400–600 MHz for high-performance graphics memory. Increasing interface speeds require fetching larger blocks of data in order to keep the interface busy without violating the internal DRAM frequency limit. At 16×800 MHz, to stay within a 400 MHz column access rate would require a 32-bit burst transfer. Multiplied by a 32-bit wide chip, this is a minimum fetch of 128 bytes, inconveniently large for many applications.

Typical memory chips are internally divided into 4 quadrants, with left and right halves connected to different halves of the data bus, and top or bottom halves being selected by bank number. (Thus, in a typical 8-bank DRAM, there would be 4 half-banks per quadrant.) XDR2 permits independently addressing each quadrant, so the two halves of the data bus can fetch data from different banks. Additionally, the data fetched from each half-bank is only half of what is needed to keep the data bus full; accesses to an upper half-bank must be alternated with access to a lower half-bank.

This effectively doubles the number of banks and reduces the minimum data access size by a factor of 4, albeit with the limitation that accesses must be spread uniformly across all 4 quadrants.

 

[gx-x]

koji ce im ocin XDR2 na icemu osim na flagship modelima?! Samo poskupljuju celu pricu bez potrebe. Poslednje 4 graficke iz srednjeg segmenta koje sam overklokovao (sa gddr5) su imale jedva neku primetnu korist od veceg memory bandwidtha. Verovatno bi na velikim rezolucijama razlika bila veca nego kad se testira 1080p ali sta ce jedan gtx650 na 2048 ili 2560 rezolucijama?! (retoricko pitanje )

 

[M_M]

Crveni već predstavili HD7970, sad još samo mi zeleni da dočekamo 680-ticu...

 

[fr0wn3r]

Ali suzise magistralu sa memorijom na svim modelima, dok je ATI sa GDDR5 memorijom svoje povecao. Kanda je NVIDIA pouzdana da ce XDR2 dovoljno brzo nahraniti GPU, price cut odradila na magistrali i stavila XDR2. Cene pariraju ATI-ju kanda su i tu pouzdani da imaju sta da pruze, da je i njihov GPU zvercica. Svidja mi se... Cak, malo su i nize cene. Btw, a i da nisu smanjili magistralu i ustedeli i ako nasi jesu mozda nasi, uvek su imali tendenciju da makar malo dignu cenu sa necim.

 

[djalex]

Mali update tabele: prepravili su config core i OpenCL iz 1.1 u 1.2

Izvor wikipedia

 

[Uruk]

http://www.fudzilla.com/component/k2...nsen-on-kepler
...po ovima za sada bez nekih opipljivijih informacija, a po najavama mozda bude nesto konkretnije krajem aprila ili pocetkom maja.

 

[neSSa]

I kao i obicno krece "RAT CRVENIH I ZELENIH", samo da nabavim kokice....

 

[Uruk]

NVIDIA GeForce GK-104 (GTX 770/760) Specification Leaked


...po ovima:

Launch is sooner than you (and me) thought ...

 

[gx-x]

morace da stave 3+ GB rama ako oce da se takmice sa onom ati zverkom u visokim rezolucijama

 

[Alk0]

Nece na 660 (760) stavljati toliko memorije. Mora 680 (780) necim da se izdvoji.

 

[Rapture]

Na 660 2g - puna kapa.

 

[gx-x]

naravno da sam mislio na 680/670 i jace

 

[Lukija]

NVIDIA GeForce GK-104 (GTX 770/760) Specification Leaked


...po ovima:

Nije ovima nego njemu. U pitanju je OBR, covek sa izvesnim problemima u glavi. Potrefilo se da je u pravu sto se tice Bulldozera ali takodje i lupeta previse. Recimo ovako:

http://www.obr-hardware.com/2011/12/amd-radeon-hd-7000-fiasco-as-bulldozer.html

 

[Ximi]

Pročitao sam danas da NV izbacuje GF 680 12 marta , zna li neko koliko je ovo tačno ?

Linkovi:
http://tinyurl.com/6pfqcwe
http://tinyurl.com/83rrk49
http://tinyurl.com/6o4wh69

 

[FantomX]

Mda, najjaca je prva recenica sa prvog linka

Kepler is real, people have seen it, editors are getting cards and even notebooks based on it as we speak, and the launch happens later this month.

 

[Grodos]

Review

Standardno za nV - dobre performanse po visokoj cijeni...

 

[mipko]

omg, ovo nije flagship GPU a pokidao je ATI kao mache muskatlu...

 

[M_M]

Review

Standardno za nV - dobre performanse po visokoj cijeni...

Ali je ipak GTX680 u proseku brži, manje troši i jeftiniji od HD7970, dakle definitivno bolja kupovina. Za veliku većinu nas naravno preskup, ali doći će uskoro valjda i GTX670, tik nešto sporiji a nadajmo se značajno jeftiniji sa očekivanim performansama višim od GTX580 a po ceni postojeće GTX570.