was thinking like 5x stronger though, and not nvidia bs like disabling physx when ati card present. let their products become a standard, not an optionQuote:
Originally Posted by Cybercat
Printable View
was thinking like 5x stronger though, and not nvidia bs like disabling physx when ati card present. let their products become a standard, not an optionQuote:
Originally Posted by Cybercat
The more I think about it, the more it makes sense that hemlock is actually a 5850x2.
The 5890 is not closed to confirmed at this point and yeilds haven't been the best at 40nm. The pricing of a 4870x2 at launch was 550 dollars. The price of a 4870 1gb was 280-300 dollars.so doubling the price you get somewhere near the 550 price point. Look at the price of the 5850, its $259 and the price of the hemlock is under 500. If the launch price of hemlock was somewhere around 479-499, wouldn't it make more sense to use 5850s if we look at past pricing?
It makes no sense if the 5870 is 379 and the double version of that is a little over a hundred bucks more, since it a more expensive chip to make than the 4870(larger die, pcb, worse yields), yet it dual version cost less.
Additionally, a dual 5850 would be 40 percent faster than a 5870 and would fall in line with the rest of AMD line up quite nicely. Also gains from using 5870 in cf over using 5850 in cf are marginal most of the time to a max of 20 percent better in rare instances. AMD could make a 5850x2 and launch it as hemlock.
5870x2 exists and you know that.Quote:
Originally Posted by tajoh111
http://www.xtremesystems.org/forums/...d.php?t=235302
I think 5850x2 is for countering nvidia when GT300 is launched, just in case 5870x2 is too expensive.
EDIT:
Are they trying to let AMD win the consumer GPGPU market?Quote:
Originally Posted by Chumbucket843
Since 5870 and 5850 is the same chip, AMD would not really have a problem with going 5870x2, provided the yields are good enough. Due to the price difference between 5870 and 5850, AMD should not have a significant problem in allocating all salvaged cores there, as the demand would be much, much higher. So using full Cypress cores for Hemlock is logical, especially if it has to compete with a good 40%+ better than 5870 Fermi.Quote:
The 5890 is not closed to confirmed at this point and yeilds haven't been the best at 40nm. The pricing of a 4870x2 at launch was 550 dollars. The price of a 4870 1gb was 280-300 dollars.so doubling the price you get somewhere near the 550 price point. Look at the price of the 5850, its $259 and the price of the hemlock is under 500. If the launch price of hemlock was somewhere around 479-499, wouldn't it make more sense to use 5850s if we look at past pricing?
Also, I think that the card is confirmed, and @ full clocks anyway....
From what ive read there is gonna be both a 5850x2 and 5870x2. The 5870x2 clock speed was originally specd to 950 core but tdp problems arose from that clock. Instesd they are gonna offer 725 or 750 clock speeds to keep the tdp down to keep pcie standard, but as soon as its overclocked whats the difference then. Its just to keep the standard pcie certification label. 5850x2 will go head to head with gt300. A dual gt300 card is gonna have the same problems as ati staaying under the tdp, if they do release a dual gt300 it will be a cut down version and Im guessing it will only be marginally faster than 5870x2.
I'd believe this. The already mentioned a dual card running at 750Mhz and I somehow doubt it was a 5850x2.Quote:
Originally Posted by To(V)bo Co(V)bo
Hemlock might refer to both the 5850x2 and the 5870x2 those similar to how cypress is both the 5850 and the 5870. Especially if the 5850x2 really does exist and is a official card from AMD, this is extra likely.
AMD might just be putting 500 dollars out there to look good on paper and might just be the entry price point into the hemlock card series, similar to you see cars being sold at a starting price. I still don't see why AMD is only charging 100 dollars more for the x2 version of the card, especially if NV gt300 card is a paper launch, they don't need to drop prices yet.
Considering the price points of both cards currently and hemlock launches in october 22, when there is certain to be no gt300 available, wouldn't it make more sense for AMD to charge 499 for the 5850x2 and 599 for the 5870x2? Especially considering this is the enthusiast market anyway which is typically not that scared of price.
If the 5850x2 does exist, where does it price fit in into the 259$ $379 and $499 pricing scheme? If a 5890 comes out at $379, is it even relevant if a 5850x2 is out and is slotted between the $379 to $499 pricing scheme. The 5890 would be an awful card if the 5850x2 was $439.
If the 5850x2 does exist and is an official card from AMD, the only place I can be it priced is between 479 and 499, which would make it the card in AMD charts.
There are both 5850x2s and 5870x2s, the 5870x2s are obviosuly ready to go. The 5850x2s are still using 2 zalman hsfs operating at fullspeed so I'm not sure when/if they are coming out.
IMHO, this is actually the best time for ATi to win marketshare with nVidia retreating a bit from competition -since they have nothing for competition against Evergreen cards, so decreasing the price as yield improve would be very reasonable, margin would still be healthy while stealing lots and lots of marketshare before Fermi and its lower end derivatives arrive.
December pricelist:
HD 5870X2 4 GB @499 US$
HD 5850X2 2 GB @399 US$
HD 5890 2 GB @349 US$
HD 5870 1 GB @279 US$
HD 5850 1 GB @209 US$
One can only dream for Xmas, hahaha. :D
the 4850x2 was trading blows with the 4890, i hope we can see the same kind of performance difference between them for the 5000 series.Quote:
Originally Posted by tajoh111
what kind of clocks would the 5890 need to match the perf a 5850x2
This is the fermi thread?
Because all I see is speculation about hemlock.
probably 1.15GHz if the 5850X2 is 750MHz, although if they release better CF drivers then that number could go up quite a bit.Quote:
Originally Posted by Manicdan
lol mate :DQuote:
Originally Posted by Jowy Atreides
go to 5xxx series thread :P lots of gt300 talk there
and lots of ati talk here
it's truly quite amusing :p:
I know man lol :cool:Quote:
Originally Posted by Jamesrt2004
It makes a lot a sense if the chip get smaller, if not well... why? Thermals?Quote:
Originally Posted by Chumbucket843
Going by Nvidia numbers, Fermi will have a huge advantage in performance\cost, in supercomputing, so why the 500mm\3 Billion transistors?
Nvidia could have done something smaller, still be unbeatable in the supercomputing area, still have the fastest single card against ATI, making a X2 solution a lot easier, and maybe it could have been out by now... but hell no, let's make another 550mm beast, we can't show to people that ATI was right with R770.
those numbers sound spot on :)Quote:
Originally Posted by spursindonesia
and this will effectively force gt300 to be 499$ or less from day1 i think... as i highly doubt a single gt300 can beat a 5870x2...
do you know if there will be a 5850 2gb? or 5850 4gb?
cause price/perf wise a 5850 is really the sweet spot, but only 1gb isnt really cutting it if you plan to use the card for a year or more...
200$ for a 5850 is really nice :slobber:
right now? not much cause xfire of 58xx is kinda broken lol :DQuote:
Originally Posted by Manicdan
overall i dont think a 5890 will be able to beat a 5850 unless the game doesnt have propper xfire support...
i didnt see, did anybody find something about the comment from nvidia that ecc supposedly hurt perf 5-25%?
so desktop cards wont have ecc right?
oh and people talking about 58xx here and about gt300 in the 58xx thread... dont read too much into that.
i wouldnt say its cause of fanboys, most is naturally mixing cause people talk about what highend vga to upgrade to, and right now many still wait for 58xx to get cheaper, or they wait for the x2 cards or the 5890...
so its natural that x2, 5890 and gt300 discussions overlap... :)
SAAYA, bro, i don't know, but i dream, A LOT, these days, LOL. :D
heheheh :D
well i hope ati doesnt restrict its aibs from making 5850 cards with 2gb and a 5850 with 4gb...
a 5850 with 2gb doesnt even make sense imo... that horse power can only be used for huge resolutions and there you will notice that 1gb per gpu is slightly limiting already today from what ive heard...
man, isn't 4 gigs of GDDR5 around $120 itself... I still cant believe we'll see a 5870x2 with 4gigs around $499
Oh Shi-
Sorry bro, no news for you, apart from the fact that Nvidia's current strategy seems to be quite dead, at least the execution of it is. Looks.. BAD.Quote:
Originally Posted by Jowy Atreides
It's saddening :(Quote:
Originally Posted by Calmatory
Something about this launch seems ....traumatic. I sometimes wonder if fermi will be released in it's original form.
The GT200 respins were canned and the current cards aren't economically viable now.
It seems like that hatchet job Fermi shown was sheer desperation.
ATi pre-demowed their cards on a ship with hundreds of demo units.
Nvidia show 1 demo on a card that's questionable and a fake display unit.
Really, look at the evidence. GTX 380 launch? Is it REALLY going to happen in the timeframe claimed?
Guess I'll have to settle with my 5870s for now, at least that's a great comfort
I'm gonna wait tell the end of Nov and see what happens if i don't like what i see I'm going with HD5870 hack driver and use my 8800GTX for physxQuote:
Originally Posted by Jowy Atreides
i think this launch of nvidia cards are not going to be that profitable overall for desktop owners.
why,
their only card sofar is going to be a monster, good for people who want to break records, and can let them offer a nice premium when a 3x gt300 beats a 4x of any ati. so those who want the best, will let them get ok margins. sure its not going to give them huge number of sales, but sales none the less.
this strategy isnt that bad, whats really bad is how much of a failure 40nm silicon was. we just got real 40nm chips from ati, and we already hear how 32nm will be out within a year. sure we can count the 4770 as the first 40nm chip, but that was supposedly too low of yield rates to continue.
basically, this is a very bad time to be in the market of building big fing chips.
the good, is nvidia tried to expand their business to professionals. not quite sure what kind of markup they can get by selling the fermi chips, or what other alternatives professionals had, but it looks like a very large opportunity when a company can spend a few thousand dollars, and increase performance by unmeasurable numbers. ill have to see how this turns out for them in the next year.
also nvidia separates themselves from ati by offering tools and features. ati just plays games. CUDA is taking its time to mature, but i think in 3-5 years, it will be doing things we never thought possible.