Time is what keeps everything from happening at once - Karamazov Brothers
|
|
|
Nvidia reveals Fermi GPU architecture
THE GREEN GOBLIN has been showing off its Fermi architecture at its GPU Technology Conference in San Francisco.
In what looks like a first wave of chip details, Nvidia has been praising Fermi's GPU-compute capabilities.
Fermi has a number of computing features never before seen in a GPU, which according to Tech Report should enable new applications for GPU computing and create some new horizons for Nvidia's GeForce and Tesla products.
Fermi has 16 streaming multiprocessors and 512 discrete cores, more than double the number of CUDA cores of the GT200.
Nvidia's next generation GPU uses 64-bit interfaces and six DRAM interfaces, which means that Fermi has a total path to memory that is 384 bits wide. This is fewer than the GT200, but Fermi more than makes up for that by delivering nearly twice the bandwidth per pin via support for GDDR5 memory.
There is not enough information available to tell if Fermi will give ATI's Cypress a kicking yet. So far Nvidia has not revealed the details about its graphics resources or what clock speed the GPU will achieve.
Tech Report seems to think that the GPU clock speed will be about 1500MHz, a reasonable frequency target for Fermi's stream processing core and about the same as that of the GeForce GTX 285. If we assume Fermi reaches that speed, its peak throughput for single-precision math would be 1536 GFLOPS, or about half of the peak single-precision floating point speed of the ATI Radeon HD 5870.
It also thinks that if Nvidia uses the same 4.8Gbps data rate for GDDR5 memory that AMD has for Cypress, Fermi's peak memory bandwidth should be 230GBps, roughly 50 per cent higher than that of the Radeon HD 5870, which has a memory bus width of 256 bits. µ
Share this:
Share
Ricky Bobby, not only is the GPU a Fake the whole nvidia company is a fake!!!!!
check it out at semiaccurate.com
They don't even have the card...
How many beautiful beaches did they destroy to make that card?
I think I'll keep my $90 HD4770, on account of the fact that spending $400 on a video card is ridiculous.
erick.mendes I don't think you need to wait for nvidia, they have nothing to offer. They only produce crappy products and that's a fact!
Dave, the Radeon 5000 series may use more power on FULL LOAD, but on idle the power usage has really dropped like 50 w... So, considering the feature set and power, the 5000 series from ATI are a great product. Let's wait and see what Nvidia can put in the table.
[and I hope they bring something really wicked, cause I want to see prices coming down, so I may put my hands in a Fermi/Cypress faster ^^]
ATI makes the best video cards in the world, PERIOD. It is second to none. How do I know this you ask. Because I do! Nvidia is just wasting time and money building these second rate video cards which burns out, catches on fire, and rapes and kills women and children. I wish Charlie was here, he would support me. I'm sorry to all the Nvidia fanbois out there the truth hurts!
dont fight guys... i spent whole day finding an article which pulled GT300's pants down and peed in its pants.
cuz of late finding my post is gtting burried deep below in chain.
Charlie Demerjian
http://www.theinquirer.net/inquirer/news/1137331/a-look-nvidia-gt300-architecture
Vs
Nick Farrell
http://www.theinquirer.net/inquirer/news/1556946/nvidia-reveals-fermi-gpu-architecture
round 1 ...fight!
I tried to warn Rector Huiz and commissioner Gordon that this maniacal Jensun Huang was not going to go away.
Now the world + dog on his terms?, and cheap as chips?
Not likely to see the back of him. In the old Matrox, Neo would have had a fighting chance. Agent Smith likes to play with his nvision.
Quick! Hand me that airtight nano-composite shell with electromagnets on each end.
Rumour Is That Nvidia Is Too Bloated to Work DX11 & Slower than GT2XX Series, at this Time. Yet, How Many Experts Have Gushed Real Swill? Heres Idea. Join FaceBook & My Page Here:
http://www.facebook.com/home.php#/profile.php?id=583087688&ref=name
So Far Captured Mike & Charlie & Mr. MoorHead & Quit Few Computees', plus Few bulldogs Whom Arn't Straight. Yet Slowly Making Nest of Pundits & Some Intresting Links, You Could of Read Todays Article, Yesterday Morning from Theo Valich, There. One Person: tomas lee, got so angered Quit My friendship, About half Dozen Dropped only to find could of Hidden Their Comments & been Better Off.
Track; Lucky Loos, CEO & even Intel Vpro. Some are amazingly straight Forward. Thrustee'. EveryWhere there. Wouldn't even Consider this Lout. So GO Bang Mikes Drum....
DRASHEK
Uh, the last i checked, they were huge and used more power than the last generation. Bashing Nvidia because they use more power is like criticizing a Hummer because its bigger than a Suburban. They both suck (gas).
it really should be faster than the 5870 by a fair margin, at a minimum it has ~50% more memory bandwidth, and it is much faster at some math operations (though slower at others).
It will probably be slower than a 5870x2 though, I think really it's just a repeat of the current generation, except nobody knows when it will be released.
Nvidia is trying to push it's GPU into the highly parallel CPU arena. The company is betting this arch can be used for CPU and HPC applications. Note the incresed number of floating point units!
Nvidia having pushed CUDA and jumped onto OpenCL earlier than most, this should be obvious to anyone who can see beyond the surface of their monitors.
Semiguy, if your supposed to compare the 5870x2 to the GT300 then what the heck does ATI have to compare a GT300x2?
Come on give your head a shake there laddie.
ya nvidia card is just to big and powerful. I will stick with my ati card.
Sorry nvidia I don't need your crappy power.
Long live Charlie! Nick you suck balls, for not bashing nvidia for their crappy products.
Semiguy
- 1.5G is not unreasonable, as their current generation products run at about this speed in 55nm
- so they should be able to hit this speed, or thereabouts in 40nm
- ATI was able to up their clocks going from 55nm - 40nm
- ATI's architecure runs its 'cores' at a little less (~1GHz), but they have more ALUs.
Fermi can never target 1.5G speed. Look at the size of die and transistors. Keeping thermals in mind, I think they will target for 750 - 900MHz. Anything beyond that will be nightmare on thermals.
Secondly Fermi should be compared to ATI Radeon HD 5870 X2.....not 5870.
I think Nvidia will have tough challenges on yield. Look at the die size man...its huge....
For gaming purpose, I would rather go for 5770 than....
dont know how to make good graphics cards, just big ones.
Muscle man and his muscle card have gone mad. I don't want a super-sized GPU to play games. Where is nvidia going? Will they introduce something that can compete with say 5770 and when?