GTX 300 Specs Leaked - Page 5

BBN · 10-01-2009, 12:59 | posts: 584 | User is Offline

def over 500 euros , more close to 600 at launch

thats for Europe

NaturalViolence · **Memory**: Corsair DDR2 800 5-5-5-15

Quote:

Originally Posted by A M D BugBear

160 TMU'S??? GAWD DAMN! the 5870 had 80, the ROP'S is also doubled over the 5870. DAMN. This card is going to be fast man, twice as fast as 5870 like one guy said, I would say um no, but we have to wait and see. If the specs are true then this card will be very very nasty.

You realize the GT200 architecture has twice as many ROP's and TMU's as the R770 right? Yet they are still on par. I will argue that fermi will be much faster than cypressxt, but that is not the reason. It's been this way since as long as I can remember, nvidia having twice as many TMU's and ROP's in competing cards that is. It's just been something that I have come to expect, especially since they both double it every gen, which is why it stays that way.

BTW

First picture of the new beast. Nvidia is currently holding some kind of conference or event to show it off, they are going to post video from it afterwards.

Edit: Video is up http://www.nvidia.com/object/fermi_architecture.html you can see it for yourself. So apparently they alreayd ahve a working demo model for fermi.

thecake90 · *Last edited by thecake90; 10-04-2009 at 08:42.*

Quote:

Originally Posted by NaturalViolence

BTW

First picture of the new beast. Nvidia is currently holding some kind of conference or event to show it off, they are going to post video from it afterwards.

Edit: Video is up http://www.nvidia.com/object/fermi_architecture.html you can see it for yourself. So apparently they alreayd ahve a working demo model for fermi.

This has been proven to be non-functioning dummy

Squall Leonhart · 10-02-2009, 01:43 | posts: 1,331 | Location: Australia | User is Offline

i'll just steal it for ...myself, and see just how non functioning it is xD

Mineria · *Last edited by Mineria; 10-02-2009 at 02:11.*

Yet it already does 5x faster double precision rendering, but that is quadro.
Will be interesting to see the gaming generation build over this chip.

From the whitepaper

Quote:

The key architectural highlights of Fermi are:

• Third Generation Streaming Multiprocessor (SM)
o 32 CUDA cores per SM, 4x over GT200
o 8x the peak double precision floating point performance over GT200
o Dual Warp Scheduler simultaneously schedules and dispatches instructions
from two independent warps
o 64 KB of RAM with a configurable partitioning of shared memory and L1 cache

• Second Generation Parallel Thread Execution ISA
o Unified Address Space with Full C++ Support
o Optimized for OpenCL and DirectCompute
o Full IEEE 754-2008 32-bit and 64-bit precision
o Full 32-bit integer path with 64-bit extensions
o Memory access instructions to support transition to 64-bit addressing
o Improved Performance through Predication
• Improved Memory Subsystem
o NVIDIA Parallel DataCacheTM hierarchy with Configurable L1 and Unified L2
Caches

o First GPU with ECC memory support
o Greatly improved atomic memory operation performance
• NVIDIA GigaThreadTM Engine
o 10x faster application context switching
o Concurrent kernel execution
o Out of Order thread block execution
o Dual overlapped memory transfer engines

Squall Leonhart · 10-02-2009, 02:44 | posts: 1,331 | Location: Australia | User is Offline

They are exactly the same chip just binned differently.

NaturalViolence · **Memory**: Corsair DDR2 800 5-5-5-15

Quote:

Originally Posted by thecake90

This has been proved to be non-functioning dummy

.....Did you not just watch them run demos using it to compare against gt200? It very obviously is functioning, and yes it is quadro but it's still a functioning fermi gpu!

If I understand correctly from what nvidia has been saying over the last couple of days this card is going to be the first to effectively function as BOTH the next card series in quadro and geforce rather than them simple being based off of each other like it usually is. Of course I could be wrong but I'm pretty sure they are leaning towards one card for both, possible with different drivers depending on what purpose you have for it.

LedHed

I hope to God this isn't the final design of the power connections:

You can see one facing towards you (6pin) and the other is on the end of the card (8 pin), in the article though it says the card looks like it was literally cut at the rear of the PCB. I also think that is a picture of the non-working GT300 card, so hopefully this is just a fluke on the non working GT300 model.

http://www.xbitlabs.com/news/video/d...r_Q4_2009.html

Quote:

The location of PCI Express Graphics power connectors do not match actual layout of their soldering points. In fact, there are no soldering points for 6-pin connector and the 8-pin one is located incorrectly.

Denial · *Last edited by Denial; 10-02-2009 at 14:15.*

Quote:

Originally Posted by LedHed

I hope to God this isn't the final design of the power connections:

You can see one facing towards you (6pin) and the other is on the end of the card (8 pin), in the article though it says the card looks like it was literally cut at the rear of the PCB. I also think that is a picture of the non-working GT300 card, so hopefully this is just a fluke on the non working GT300 model.

http://www.xbitlabs.com/news/video/d...r_Q4_2009.html

Nvidia already confirmed that the sample they showed was a mock up. The engineering sample they had running was a mess of wires and stuff, so they didn't want their CEO to show that to the audience. But I agree, if that was the actual design it would be terrible.

GC_PaNzerFIN · 10-02-2009, 14:20 | posts: 888 | Location: Finland | User is Offline

Found this pic of bit outdated GPU testing PCB. Its a lowend part testboard. No doubt something complex as GF100 has huge protoboard atm and full of jumpwires, solderings and such.

NaturalViolence · **Memory**: Corsair DDR2 800 5-5-5-15

Quote:

Originally Posted by LedHed

I hope to God this isn't the final design of the power connections:

You can see one facing towards you (6pin) and the other is on the end of the card (8 pin), in the article though it says the card looks like it was literally cut at the rear of the PCB. I also think that is a picture of the non-working GT300 card, so hopefully this is just a fluke on the non working GT300 model.

http://www.xbitlabs.com/news/video/d...r_Q4_2009.html

Ok I was wrong. But thank you for linking proof rather than just saying it! I also hope they don't use that design....that's just stupid. It's a big enough card to begin with, minus well help your mid tower or small full tower users out a little bit by at least putting the power connectors on the top.

Funky Doodle · 10-03-2009, 16:08 | posts: 513 | User is Online

Fermi GX2 - 649$
Fermi GTX - 399$
Fermi GT - 299$
took it from here, but it's on russian!
http://gpunews2.se-ua.net/home/1/24/
nice prices, if those will be final. I think AMD will be having difficult time. I was thinking about ati hd 5850, but right now im interested in Fermi GT also!)

Squall Leonhart · 10-03-2009, 16:35 | posts: 1,331 | Location: Australia | User is Offline

That would be highly aggressive prices if its true.

Redemption80 · 10-03-2009, 16:37 | posts: 2,967 | Location: Glasgow | User is Offline

No idea the credibility of that site, but they seem far too low to me, would be a first to see Nvidia coming in at the same price point as ATI.

If it is, then definetly, my 5850 will be sold and a GTX bought

GC_PaNzerFIN · 10-03-2009, 16:50 | posts: 888 | Location: Finland | User is Offline

I would take those prices with lots of salt. Served chilled.

Its obvious that the card is beast, expensive beast. More like $499 and I doubt there will be GX2 anytime soon.

Denial

It will probably launch at whatever price the 8800GTX launched at, it has the exact same die size pretty much.

waway625 · 10-03-2009, 21:17 | posts: 6,505 | Location: Singapore | User is Offline

Quote:

Originally Posted by GC_PaNzerFIN

Found this pic of bit outdated GPU testing PCB. Its a lowend part testboard. No doubt something complex as GF100 has huge protoboard atm and full of jumpwires, solderings and such.

this:

Quote:

oh dont get me wrong, i completely trust tim on this!!!
its just that he said himself, he couldnt really tell much from the picture he saw, and it sounds like he wasnt 100% convinced that the picture he saw was the REAL REAL fermi card... otherwise he wouldnt even have mentioned that it was a blurry cell phone pic of lots of wires coming off a pcb... if hed really believe he had seen fermi in that moment, he would have written just that, that nvidia then DID show him a picture of the actual card...

btw, a good friend apparently managed to get a copy of the picture in question and mailed it to me

behold! this little puppy right here... this is fermi!

NaturalViolence · **Memory**: Corsair DDR2 800 5-5-5-15

I have a question. Aren't ROP's and TMU's special function units? Earlier I stated that fermi has 160 TMU's and 64 ROP's, I said this because in the white papers it says that fermi has twice as many special function units per SM (it has the same number of SM's as GT200). But I am seeing many articles popping up that say fermi has 128 TMU's and 48 ROP's. Who is right?

skinnyq · 10-04-2009, 00:51 | posts: 18 | Location: Israel | User is Offline

If Fermi GTX really is going to cost $400, its my next card, but I doubt that as the production costs of this product seem to be VERY high.

If not... well, depends on how the GT is going to perform.

Peerless · 10-04-2009, 03:16 | posts: 986 | Location: Seattle, WA | User is Offline

Yea if those are the prices, count me in for the GX2 version. Hopefully I won't have any issues or I would be all about the GTX!

TyPe-ZeRo · 10-04-2009, 03:26 | posts: 220 | Location: California | User is Offline

Quote:

Originally Posted by waway625

this:

gee that looks really close to production

Ieldra · **Memory**: 6GB KVR 1400MHz 7-7-7-14

Quote:

Originally Posted by TyPe-ZeRo

gee that looks really close to production

That's the 'Super Special Nerd Edition' you have to assemble it yourself

GC_PaNzerFIN · 10-04-2009, 08:42 | posts: 888 | Location: Finland | User is Offline

Quote:

Originally Posted by Ieldra

That's the 'Super Special Nerd Edition' you have to assemble it yourself

Manual sold seperately.

Cobrass2 · **Memory**: A-data 1600MhZ CL8-8-8-24

Its so fake rofl
WHAT DO YOU DO when you have a major conference planned to introduce a card, but you don't have a card? You fake it, and Nvidia did just that.

no dvi

lol? Nice mio

Best cooling solution ever

The first one is from the always great PC Watch, specifically here, but we cropped it a bit for size. Some things to note on it. First is that the second set of digits on the first line says 0935A1. A1 is for first silicon, something that, when coupled with a direct quote from Jen-Hsun of, "You are currently looking at a few day old silicon" (from here select the video "See video of Jen-Hsun Huang announcing Fermi") kind of blows the whole 'silicon in Santa Clara last spring' story out of the water. I wonder if anyone will retract that, or just change the article retroactively?

To make matters worse, the other digits are a date code, 0935 means 2009, work week 35. If Nvidia starts its work week on the first full week of the year like everyone else, that would put WW35 at August 28 to September 5, 2009. Where have I heard that date before? Oh yeah, here.

Last but not least, you have the funniest part, the number seven hand written on the GPU. While this isn't proof of anything, normally when you have small lots of samples, you write numbers on them to identify which one is which as an aid to debugging. Where did we hear that number seven again? Oh yeah

Redemption80 · 10-04-2009, 09:03 | posts: 2,967 | Location: Glasgow | User is Offline

Talk about turning up late to the party lol.