Anything new in that? Don't feel like reading through stuff that doesn't include actual game benchmarks.Quote:
Originally Posted by GoldenTiger
Printable View
Anything new in that? Don't feel like reading through stuff that doesn't include actual game benchmarks.Quote:
Originally Posted by GoldenTiger
Nothing much, tessellation performance as previously rumored, but no actual gaming perf. numbers.
What a surprise :eek:Quote:
Originally Posted by annihilat0r
Realy not much there. Its just slides of stuff that's kinda known and slides of what they say the performance will be. No real benchmark test done by reviewers.
The image linked is pretty much all there is in that article other than an off-the-cuff remark mentioning probably 20% better all-around performance than the 5870. There's also a video showing Far Cry 2 in action comparing it to the GTX 280 right off next to eachother on two monitors.Quote:
Originally Posted by kgk
Including architecture scaling, GAMING PERFORMANCE and a whole wack of other stuff.....
RIGHT HERE!!!!!
http://www.hardocp.com/image.html?im...8yNV9sLmdpZg== One of the slides shows a few nv-provided performance #'s.
EDIT:You rock!Quote:
Originally Posted by SKYMTL
EDIT: Wow... great article, and the gaming results are insane...
P.S. 5870 for sale!!! :p:
EDIT2: The HardwareCanucks article is by far the best, most of the others just showed the slide deck and some generic commentary. Yours has actual stuff written out, some educated guesses, and the exclusive performance info :D.
GF100 is getting like Duke Nukem Forever. We keep seeing teasers but never anything useful.
ok im waiting fermi as much as the rest of you but where the hell are the numbers?
it has been shown running benchmarks so why not get the numbers out? there not doing a good job trying to stall ati's sales...
im not sure guys im not liking this secrecy.
edit: pk so we have some number :P
:up:Quote:
Originally Posted by SKYMTL
Thanks for the compliment.Quote:
Originally Posted by GoldenTiger
People here at XS would also do well to take a gander at the Architectural Scaling section for some hints as to what COULD be coming. ;)
As expected: Something in between 5870 and 5970 considering performance...
Ever since I first discovered your site last year, it's really started to grow on me. Solid reviews and well written articles.Quote:
Originally Posted by SKYMTL
Keep it up!
Wow, it looks really promising! I'm glad that I didn't immediately hop on the ATi boat. But let's see if NV can put out a reasonably priced card.
Fermis are here at PDXLAN :)
:) Read through a lot of the article now, comment still holds ;). See edit2 above.Quote:
Originally Posted by SKYMTL
Agree wholeheartedly. Hardware Canucks is by far the best article. :up:Quote:
Originally Posted by GoldenTiger
Looks like they did a very thorough job of extrapolating real world figures based on the live demo.
The minimum FPS on FC2 @2560x1600 is pretty impressive. Where does the 5970 fit on that graph out of curiosity?
CUDA Cores = 512 scalable into 32 core chunksQuote:
Originally Posted by H2O
Texture Units = 64 scalable into 4 unit chunks
Memory: 384-bit scalable into 64-bit partitions
ROPs: 48 scalable into 8-unit groups
Memory interfaces, ROPs and L2 Cache must be scaled together.
Don't know about the other articles but it's in mine...
Here's the section about architecture scaling: LINK
you mean he wasnt reliable in the past? :confused:Quote:
Originally Posted by Boissez
IF fermi really performs as well as a 5970, that will be pretty impressive...
if you think of efficiency, thats way better than cypress... to deliver the same perf as 550mm2 big fermi, ati actually needs two 333mm2 cypress cores, which is 550 vs 666. with perfect yields that is... ;)
even with all the broken part recycling aka 360s im sceptical on how much money nvidia can make with gt300 parts, at least on 40nm... i guess thats not what fermi is meant for though, nvidias cash cow seems to be the 350... i cant wait to hear more about THAT one! :)
hahahah awesome! :DQuote:
Originally Posted by annihilat0r
i wonder how biased the slides are... if nvidia claims too much its going to backfire on them... big time... if you raise the expectations too much, especially with pc enthusiasts, youll run into issues even if you have a great product at a great price :DQuote:
Originally Posted by annihilat0r
So NVIDIA ran the benchmarks for GF100? Just clarifying, as that's what it sounded like.Quote:
Originally Posted by SKYMTL
Not really exclusive. Many people in my session were busy recording it on their cameras. Me, I bolted for a piece of paper to write down the results. :up:Quote:
Originally Posted by GoldenTiger
Ah, well that answers the question :DQuote:
Originally Posted by SKYMTL
Yes but it was the standard Ranch Small test in the Far Cry 2 benchmark. In addition, the program itself wasn't running in the background in any way.Quote:
Originally Posted by randomizer
They booted the benchmark, changed the settings in front of our eyes and then ran it. Simple as that and like anyone else would do. The only difference is that none of us was holding the mouse. Basically, it doesn't make one lick of difference whether it was an NVIDIA system or not; they stated the specs, ran the same tests 99% of the other review sites usually do and showed results. ;)
Me, I replicated their test system as outlined by NVIDIA and popped up the HD5870 result. They weren't BSing either since I was able to replicate their GTX 285 results down to a few percentage points.
Review up @ AT too.
http://www.anandtech.com/video/showdoc.aspx?i=3721