[Solved] There is no GT300. The new GF100 is rumored

Already posted, do try and keep up.

Plagiarism!, you copy and pasted! Well on that blurry jpeg I see a cheeseburger and hotdog or 'other' same as in the other thread posted earlier, however the line

does make it count as rumour but then if you saw the thread "need rumors on GT300 please" why not just post your contribution there?

"-"

reasons:
1.) The following is a THF thread and is an example of "WHY NOT".
10-14-2009 at 11:37:50 AM | BBCode | Report

Youngster
More Information
5870 vs. GT300 "Fermi"

http://www.youtube.com/watch?v=10xKBSEEvwQ

2.)"where how who...
from the horse's ass
.
"Nvidia GF100: 128 PBC and 48 ROPs

Before yesterday, Nvidia has announced its next-generation chip GF100. He sits on the so-called Fermi-architecture, is about 3.0 billion transistors difficult to have 512 scalar shader units and a 384-bit wide memory interface that can manage up to 6 MiB GDDR5 memory.
Has been concealed until now, will have as many ROPs PBC and the chip, so started in the last hours and the big guessing game. We now have available exclusive information.

The GF100 has - as many expected - a total of 128 PBC. Puzzled had last made a turned up at Beyond3D document that suggested 256 units. This is - as it turns out now - not wrong, but simply reflects the number of interpolators, is the 256th As for texturing, however, two interpolators are necessary, one can say that the GF100 effectively owns 128 PBC.
In general, the PBC Traderjob fall from more flexible than the G80 and GT200. This allows, for example, instead of a billinearen samples, two load vectors on a TMU.



As we were also able to find out, the number of ROPs of 32 increased to 48, although the memory interface has been reduced by 25 percent, from 512 to 384 bits. It is obvious that Nvidia simply has pro Storage - GF100 has six memory controller with 64-bit width - the number of ROPs of 4 doubles to 8. Accurate information is not before us, however.
ROP Pro 8 Z units are to be installed so that a total of 384 Z-tests can be performed simultaneously.

Already GF100 chips to run executable with 650/1700/1000 MHz - compared to the A1 stepping 700/1600/1100 MHz so no great change, if it is really one and the same chip.
Later this year, it will probably stage a launch event and can bring a little luck cards on the market. Good availability but it seems to be only from 2010.

GTX 380 W/1 and 2 Gb GDDR5
.
Nvidia G300 has mastered Tape Out
GT300 is on 512-bit GDDR5 interface



About a week ago we reported in association with Bright Side of news that Nvidia's possession the next graphics
chip over massive 512 shader units, Now we can present in a renewed collaboration with Bright Side of news first,
exclusive information about the used memory and memory interface.
Accordingly, Nvidia is back with the GT300 or the temporary flagship of the GeForce GTX 300-series of GDDR5 memory.
In addition, there have been adjustments to the memory interface, which will continue, however, have 512 data lines.
Earlier this year we reported an improved memory management in the GT300.
The smallest available GDDR5 memory is 0.5 nanoseconds, and thus quickly specify to 2000 MHz I / O clock.
In combination with a 512 bit wide memory interface, this would mean whopping 256 GB / s memory bandwidth.
For comparison, the GeForce GTX 280 just 140.8 GB / s memory bandwidth has.
What is the built-GDDR5 memory later on the presumably
under the name GeForce GTX 380 graphics card will be calling at bars, but is still open.
Only following successful tape-out will be subject to initial statements can make.
The same applies to the storage size will be the 1024 and / or 2048 MiB." 512 scalar shader units,
which is now titled Traderjob cores as Nvidia CUDA." http://resources.vr-zone.com/image [...] 24f540.jpg
.
The rumor by: Florian Spooren
.

#2
page 2
HD 5870x2 runs like butter


.
HD 5870 X2 läuft auf CryEngine 3 butterweich
http://resources.vr-zone.com/image [...] 24f540.jpg

#3
page 3
.
Source: The Source: (no-source-rumor): (KKT) and (KS)

Here today we have the opportunity to review first hand the source, (who), at work on Hardware-infos, (where).
Just how KonKorT and Florian work is the subject of this thread and it should be abundantly clear that
the definition of no-source-rumor = fake is documented and it
starts here for Nvidia and GF.

Will the Queen and the Borg fulfill KKT's and KS's expectations...? or are KKT and KS one and the same?

The mystery of the source has finally been put to rest. The fakes are to be continued.

.-"-
. y

3.)Thank you for the opportulnity.
The chart is translated and clipped.
That is the only way to correctly represent the home made chart and yes it is KonKorT.
It is the makings of the infamous hardware-infos no-source-rumor, the fake.

The no sourced article is by Spooren F (FS).

...................... So there you have The Source of the Nvidia, GF, no sourced rumors. Not only the source but where how and who...

...................-"-
............. TY y
HD 5870 X2 läuft auf CryEngine 3 butterweich

Edit: 18Oct09 unfortunately the link above is not available today. ? as often happens with no-source-rumor=fakes...

...................

You are not the only one whining about my style so the following is presented as comparison, no-source- rumor-fake, for you to study at your leisure.
http://s1.bild.me/bilder/9750992Sn [...] 31_004.jpg

The following will certainly fill some of the gaps in your virtual memory... but first my supper.


.

What you read is verbatum. It is sourced, translated where possible and cllipped as complete including photo/s.

It is the source of a rumor-fake... you should understand that. If you do not understand so what it is only virtual meaning occupying neither space nor time. lol

There is no GT300, there is no spoon and there is no new or unknown facts being divulged so there is no point to this thread.