Nvidia uses GDDR5, Finally !!

WTF? Companies use a lot of tech during their R&D, it doesn't mean that every tech will be used in the future.
 
It only makes sense to adopt GDDR5. It is a revolution over GDDR3 in every way. You can get the same bandwidth as GDDR3 with half the bus, better performance per watt, and we still haven't even come close to the speeds GDDR5 is capable of (Going from 4GHz to 7GHz in only a matter of a year). GDDR3 is old tech.
 
OMG, Nvidia is using GDDR5 now, so it MUST BE DA BOMB!.....lol welcome to early -mid 2008 tech.....
 
Actually, it does make sense.

In doing so, Nvidia can use a lower bus width and get the same bandwidth. The wider the bus, the more PCB layers may be required. (I think - I could be wrong.)
 
Josh_B said:
In doing so, Nvidia can use a lower bus width and get the same bandwidth. The wider the bus, the more PCB layers may be required. (I think - I could be wrong.)
Click to expand...

i don't think you're wrong. however given the cost premium on GDDR5, i don't think it makes as much sense to use them on a "low-end" card. i wouldn't be surprised to see it introduced on a mid-range card. seems like nvidia likes to try out technologies for the first time on the mid range, so they can get it right, before they try it on their top tier stuff.
 
Josh_B said:
Actually, it does make sense.

In doing so, Nvidia can use a lower bus width and get the same bandwidth. The wider the bus, the more PCB layers may be required. (I think - I could be wrong.)
Click to expand...

more pcb layers and/or more die space. this is why nvidia's chips are so huge. look at the ati article on anandtech
 
The more GDDR5 is used, the cheaper it will become. AMD probably already brought the price down with the 4800 series selling as well as they did, in addition to introducing competition from multiple producers. If NVIDIA enters the game as well, it'll only get better. I'd be interested to see how they would implement it into their cards though. A 512-bit bus with GDDR5 would seem like a complete waste.
 
dystopic said:
i don't think you're wrong. however given the cost premium on GDDR5, i don't think it makes as much sense to use them on a "low-end" card. i wouldn't be surprised to see it introduced on a mid-range card. seems like nvidia likes to try out technologies for the first time on the mid range, so they can get it right, before they try it on their top tier stuff.
Click to expand...

It's already closer in cost than you think.
 
RadXge said:
GDDR5 on a low-end card does not seem to make much sense...
Click to expand...

GT214 isn't necessarily low-end. It's the equivalent of the performance segment 9600GT of last generation which used the same top-end GDDR3 as the flagship parts. It's quite possible they stick to high-speed GDDR3 though - that's still good enough for ~ 78GB/s on a 256-bit bus. If the rumours around RV740 are true and it's going for GDDR5 on a 128-bit bus then Nvidia won't be at a disadvantage. They aren't at one now either so I'm not sure why people think ATI had an advantage this time around with GDDR5 - both approaches provide the same bandwidth in the end.
 
As mentioned, GDDR5 has been around long enough now that the cost premium over GDDR3 isn't much and quite likely low enough that the savings from going to GDDR5 are greater than the costs of not doing so, hence the move to it even on low(er) end cards.
 
OMG, Nvidia is using GDDR5 now, so it MUST BE DA BOMB!.....lol welcome to early -mid 2008 tech.....

Lol. I hope this gets off the ground quickly and preferably be followed up by AMD so we can get more competition rolling and the prices come down :)
ecstatic.gif
 
debated, perhaps debated to death. i probably would have laughed myself, but i'm a wee bit hungover so i cringed a little bit.
i don't get excited about new products until i see [H] put them to the test, but i was a little surprised no one had mentioned that "leak" in this thread.
 
dystopic said:
debated, perhaps debated to death. i probably would have laughed myself, but i'm a wee bit hungover so i cringed a little bit.
i don't get excited about new products until i see [H] put them to the test, but i was a little surprised no one had mentioned that "leak" in this thread.
Click to expand...

Look how that "leak" also mentions GDDR5 tests for the G96...*hint-hint*
 
big bang 3? :p

maybe NV should start giving these thing subtitles like in bad hollywood sequels. "Big Bang 3: Return of the J3d1".
 
Atech said:
Posting without reading the link? Get a clue...:rolleyes:
Click to expand...

I read your link and the argument is no different than the one here, so basically your post is useless and unwarranted just like every other time you open your mouth.

On to the ignore list you go.
 
merlin704 said:
I read your link and the argument is no different than the one here, so basically your post is useless and unwarranted just like every other time you open your mouth.

On to the ignore list you go.
Click to expand...

You make sure to put that DDR5 G96 on the ignore list too ;)
 
Why do you care what type of memory they use.

Using GDDR5 will not give them some kind of boost in performance, because the memory bandwidth on their current cards is as high or higher than the ATI cards.

Sure lower power consumption is a plus, and usually narrow bus widths (256 vs 512) can mean smaller die sizes, but this isn't exactly something to cheer about...
 
chinesepiratefood said:
Why do you care what type of memory they use.

Using GDDR5 will not give them some kind of boost in performance, because the memory bandwidth on their current cards is as high or higher than the ATI cards.

Sure lower power consumption is a plus, and usually narrow bus widths (256 vs 512) can mean smaller die sizes, but this isn't exactly something to cheer about...
Click to expand...
Memory bandwidth is always going to be secondary to core speed, as the main goal is to design the subsystem to sufficiently feed the core. That said, how you get the memory bandwidth makes a big difference in final cost to the consumer, which is something we should cheer about. Thinner PCB's, 256-bit busses, lower power consumption, and other effects of using GDDR5 make cards cheaper to produce and will make them cheaper for us to buy :cool:.
 
Mr. K6 said:
That said, how you get the memory bandwidth makes a big difference in final cost to the consumer
Click to expand...

No it doesn't. Competition and market demand determine the cost to the consumer. Chip size, board design, bus width and memory technology used are all costs to the IHV and OEM's. With two major competitors in the market price elasticity is very high...they don't have much control over how much they charge us for their products. What they can control is how much performance those products provide and increasing the value proposition to the end user who in turn will be willing to pay more for that performance.
 
You guys realize that Joe Macri, that sits on the JEDEC board, helps build the GDDR design and ratifies it, is also Senior Director of Graphics Engineering at AMD? And you wonder why ATI always gets the cool GDDR first. ;) Joe also races 90MPH go-karts, can drink a lot of beer, and is just an all around cool guy that is very passionate about his job.

But anyway, that is the real meat of the story.
 
trinibwoy said:
Haha, quality over quantity man.
Click to expand...

If a quality post falls in the forest and no one is there to hear it...... ;) That forum of elitist pricks dug their own grave a long time ago. Used to be a thriving community that had a lot to give back to the community, but that time has passed. Its insight seldom stretches beyond the edges of the PDF.
 
Atech said:
that is why i read there, most post are informative, tehcnical and free of some of the "noise" that occour other places ;)



I disagree, the whole MAD/MUL about the 8800 series and a thread like this gives me great inside, that I have not found here(not that I don't find great info here, but it's seldom so technical):
http://forum.beyond3d.com/showthread.php?t=49327

Sure it's not 3Dmark...or Crysis...but it's educational.
Click to expand...


I think we even get 3 or 4 great posts a year too. :p Surely we can give them a couple. ;)
 
Kyle... sorry for committing thread necromancy in here. :( i hope the fact that i, personally, look at these rumors and discussions with a sense of bemusement and even mild ridicule comes through to some extent.
 
trinibwoy said:
No it doesn't. Competition and market demand determine the cost to the consumer. Chip size, board design, bus width and memory technology used are all costs to the IHV and OEM's. With two major competitors in the market price elasticity is very high...they don't have much control over how much they charge us for their products. What they can control is how much performance those products provide and increasing the value proposition to the end user who in turn will be willing to pay more for that performance.
Click to expand...
Who pays for the cost of the cards if it isn't the consumers? The money doesn't appear out of thin air. Both NVIDIA and AMD are businesses with a bottomline, and having market share doesn't mean much if you're still hemorrhaging money. They create a product to sell for profit against the competition. Were you not here in June? AMD was able to take over NVIDIA because they could produce a competitive card more cheaply and sell it for less. Was this all due to just GDDR5? God no, but it's a part of it.
 
i'm curious about something hypothetical. it doesn't seem like either nvidia or ATI is yet at a point where they need more memory bandwidth than GDDR5 on a 256-bit bus can offer. but if that point comes, is there's something to be said for nvidia's experience implementing wider memory buses? are wider memory buses more technically difficult to produce, or just simply more expensive?
 
dystopic said:
i'm curious about something hypothetical. it doesn't seem like either nvidia or ATI is yet at a point where they need more memory bandwidth than GDDR5 on a 256-bit bus can offer. but if that point comes, is there's something to be said for nvidia's experience implementing wider memory buses? are wider memory buses more technically difficult to produce, or just simply more expensive?
Click to expand...

Let's not forget nVidia isn't the first to implement a wider bus. If my memory serves me, they are the 3rd. Matrox Parhelia was first with AMD 2nd and we see how that worked out.
 
You must log in or register to reply here.
Back
Top