So it looks like the Black Edition will replace the current Titan
http://hexus.net/tech/news/graphics/...etail-listing/
pricing seems to be near identical to the current titan with the 790 coming in at the pricing of 'a lot'
Printable View
So it looks like the Black Edition will replace the current Titan
http://hexus.net/tech/news/graphics/...etail-listing/
pricing seems to be near identical to the current titan with the 790 coming in at the pricing of 'a lot'
I still question the legitimacy of that listing and that article. If its true, I hope this card will at least have HDMI 2.0 because I really need a GPU that is capable of running 60Hz at 4K. If not, huge epic fail from NVIDIA.
The Titan is over, not sure how it slots into the lineup now that the 780ti is out. I think a 790 now would be the only reasonable possibility, but as time goes on it would be too close to a GTX 800 series launch.
Not really. The titan could never be justified on price simply for gaming. However given that the 780 and the 780ti still have their GPU compute crippled and Quadro/Tesla cards still cost more than Titan it has a niche.Quote:
Originally Posted by ZX2Slow
For a card that straddles between high end consumer and professional both on price and features, it's actually a really great value.
May as well wait for Maxwell at this point.
Is there any official word on when it will launch?Quote:
Originally Posted by zalbard
Nvidia must be happy, now it can charge USD 1000 for a single consumer GPU once again... yay!
Official - no, but it's reasonable to assume that it'll happen within the next 6 months.Quote:
Originally Posted by AliG
Blowing $1000 on a video card only to watch it become obsolete in less than half a year doesn't make much sense to me...
Rumors so far point to Maxwell-based GTX 750 Ti and 750 next month and nothing high-end until late this year at the earliest.
The age of Kepler is over, I think all new releases will be Maxwell from here on out.
Maxwell is still based off Kepler, and its still 28nm.
Just the lower-end chips. There's no space for growth for high-end chips on 28nm.Quote:
Originally Posted by kuroikenshi
Official launch time of TITAN BLACK is 02/18/2014, enjoy!
Same as 750 and 750Ti ? :confused:
yes, exactly!
How will it be obsolete? Again, the value in Titan is cheap GPU compute, not playing games.Quote:
Originally Posted by zalbard
Gtx 790 6gb = us$1,400 :eek:
i think will be 1499!
yes but can it play crysis?
So what are the differences between the original Titan and the "new" Titan Black ?
I have noticed 3GB being a barrier in multi monitor setups and in effect 4K gaming so a Titan Black might be useful in such scenarios. But then again it does have two problems one being Maxwell and the other being vfm 290X OC non reference.
The new Titan comes in black. :DQuote:
Originally Posted by Gunslinger
Sometimes I'm blown away by my brilliance. :D
More efficient cores and same clocks as the 780Ti.
Maxwell will have almost double the FLOPS at the same TDP... And unified virtual memory support as well.Quote:
Originally Posted by crash5s
DP, not SP. Next card with significantly more DP performance is 1.5 years away if I had to guess. Titan came 2013 in February, so Maxwell Titan should arrive in 2015.Quote:
Originally Posted by zalbard
The original Titan should have had the fully enabled GK110 and subsequent cards be harvested versions of it. NVidia has basically flipped the bird to high-end enthusiasts by doing this and will turn the same trick again with Maxwell this year. I know I won't consider the Titan Black Edition, and the performance increase will be slight at best.
I will buy a GTX 790 though, to go in my ITX gamer that I'm building...if they maintain the rumored 10GB of memory...otherwise, I'll just move my GTX 690 out of my server...;)
New info: launch date is not 18th of february, but later, next month!
18 of February was always only 750 (512 cuda cores) and 750Ti (640 Cuda Cores)...
I am certain I will buy a Titan Black for doing my compute work. I have been waiting for it for ages. :welcome:
So they're not 768 and 960 as rumors have claimed?Quote:
Originally Posted by KiSUAN
640/512 = 960/768 = 1.25 so those numbers would still imply 5 and 4 (or a multiple of such) SMXs, each with 128 (or a divisor of) CCs/SMX.
launch is third week of MarchQuote:
Originally Posted by umvue
Nope, if 750Ti had 960 it would be replacing the GTX660, they are replacing the GTX 650Ti to compete better with the 260 and 260x which, at least where I work, are selling like hot cakes...Quote:
Originally Posted by iMacmatician
There are two version of 750 Ti, one has 960 CUDAs :)
And about this "maxwell 0.5", the most important info is about power consumption, 750 Ti has performance above GTX 650 Ti, but without power connector, and with significantly lower power consumption.
So is one a Maxwell part and another a Kepler part?Quote:
Originally Posted by PedantOne
Also, when would the 880 and 870 be coming out?
kepler
That's why you watch forums / craigslist until you find someone selling it for a price you're comfortable with.Quote:
Originally Posted by zalbard
When the 700 series was coming out, I snagged a GTX 690 for $650 & added a water block. Still high, but much cheaper than two 680s + 2 blocks. I think that's where the deal w/ these cards lie; Cheaper on the used market if you compare the cost of water blocks thrown in & the ITX market.
As for people who don't mind dropping $1400, by all means. I'd like to be in that position that $1400 is no biggie. I'd like to be in those shoes, lol!
That's a strategy I can agree with! :up:Quote:
Originally Posted by [XC] Synthetickiller
You need to contact Faye Turner. According to "her" dozens of posts on the ESPN forums she makes $10k a month just by working at home lolQuote:
Originally Posted by [XC] Synthetickiller
and my silverstone ft03, i can really only fit the single gpu into it.Quote:
Originally Posted by [XC] Synthetickiller
The higher end Maxwell cards will have much higher SP performance. It's the lower end cards that are crippled.Quote:
Originally Posted by boxleitnerb
[/quote]Next card with significantly more DP performance is 1.5 years away if I had to guess. [/quote]
where are you getting this information? Because it doesn't mesh with any of the facts. like how TSMC has moved from 20nm risk production to volume production about 3 weeks ago and are way ahead of schedule.
no way on earth is nVidia going to release a few low end maxwell cards next month and then wait till next year to bring out the big cards. they've never spaced out their cards even remotely that far apart, and i highly doubt they'll start doing it now.Quote:
Titan came 2013 in February, so Maxwell Titan should arrive in 2015.
these :maxwell: cards are not real maxwell, but only Kepler refreshed, real maxwell will be launched this summer, and BIG maxwell maybe next january!
while everybody thinks the opposite i think it is better for everybody that things gets a little slower so nice to hear this.Quote:
Originally Posted by PedantOne
What's the difference? If it's a Kepler refresh wouldn't it be a Kepler part instead of a Maxwell part? (I don't seem to recall, say, GK104 being called a GF114 refresh, for instance.)Quote:
Originally Posted by PedantOne
So what happened to this then?Quote:
Originally Posted by PedantOne
Quote:
Originally Posted by PedantOne
The only thing I know/knew is that most of the information available to insiders last year was that a successor of GK104 was going to be launched in February with high end performance. And now it's summer? Or will something else still come this quarter?Quote:
Originally Posted by PedantOne
I would say: because he is guessing without any facts behind those statements...Quote:
Originally Posted by Tim
I know that information did not originate from him but from trusted sources. There was supposed to be a high performance 104 maxwell variant launched Q1, Feb was likely. Since he probably has access to one or more sources, I thought maybe he can shed some light on this apparent change.Quote:
Originally Posted by Dimitriman
imo it is because of amd 290x and especially 290's price performance.
Could be, I think that information surfaced around the 290x launch date, but I would of imagined nvidia knew about the 290x performance well before launch.Quote:
Originally Posted by kromosto
Yeah, PedantOne wasn't the only one claiming a GK104 successor for Q1 2014.
If we trust SemiAccurate's articles and their tags, then from what I've gathered from the stuff that isn't restricted to subscribers, my guess is that some sort of "GM104" was planned for release at some point in time but was then canceled some time between then and now. (PedantOne and others feel free to comment on my guess.)
GM107 (or at least stuff consistent with that type of part) has been in the rumor spotlight in the last few weeks but that name has been floating around since last year.
performance yes they knew, not exactly but close, price i guess they knew it close enough too but what they couldn't calculate was impact on market. also mantle gets noticeable attention too.Quote:
Originally Posted by Tim
They likely got blindsided by the mining market more than anything.Quote:
Originally Posted by kromosto
The mining market has opened up an entirely new market for the high end that was not earlier available. People who didn't game much, started to buy AMD super high end and as a result, not nearly enough cards were made.
If AMD knew this mining thing was going to happen, they would have likely priced their cards 100 dollars higher to reap the benefits instead of retailers. Even at 100 dollars higher, the r290 is a good card. A r290x at 650, not so much but it would still sell with miners scooping up supplies.
The black edition seems like such a strange move. Its coming out so late and cost so much like the titan. Generally these Titan cards seem like Nvidia way of giving themselves an excuse to rape earlier adopters and sell super cut downs at high prices. The original titan should of always been a full chip with the gtx 780 having atleast one more smx of shaders.
A 650 dollar card, should never have 20 percent of its shaders cut off. That's ridiculous.
I don't think with this B revision, Nvidia even has to cut off any shaders. Some of the gtx 780 ti's being shipped to consumers have an asics qualities in the 50's.
I think we are going to sell a significant drop in Nvidia's 4th quarter results and I expect the stock to fall as a result.
Just saying, any claims without proper sources are just claims, and saying its from "trusted sources" doesn't make it better unless you are Charlie or Jen-Hsun Huang's old classmate. If the user runs a hardware review site like SKYTML, they might have insider info. But remember that even AMD's Marketing Manager John Freue (JFAMD) spread information on this site that ended up being false.Quote:
Originally Posted by Tim
So with no sauce, I tend to take information with a grain of salt...
things changed, nvidia called this maxwell, but it is not REAL maxwell. And yes, ive seen card with some chip ong time ago, but this chip was cancelled.Quote:
Originally Posted by Tim
So how is the efficiency of this "new" not-real Maxwell compared to Kepler? Do you have some idea as to why the original GPU was cancelled?
efficiency is great, same chip as kepler, better performance, but without additional power connector
by the way someone here told us bull:banana::banana::banana::banana:s about 640/512 CUDAs etc, iT was shiiit as i said before, First "Maxwell" cards has CUDAs like we saw on Chinese leaks :)
960 and 768, enjoy
ps. i can post gpuz screens, pics of gpu chip etc.
Oh dear :lol:
So PendantOne is OBR?
Rofl.
-PB
why? obr
guru3d agrees:Quote:
Originally Posted by PedantOne
http://www.guru3d.com/news_story/nvi...in_the_uk.html
hey PedantOne...
will a Liquid Block for cut titan / 780Ti fit Titan Black pcb?
Here you have a photo of an actual product that's not an engineer sample, not a fake GPU-Z SS, it wasn't leaked until now (it also contains a suprise) and it has 640 cuda cores.Quote:
Originally Posted by PedantOne
http://s3.postimg.org/8cdq0rvb5/php_Rpkyw_P.jpg
Now tell me, who is full of bull:banana::banana::banana::banana:?
Is the "surprise" an ARM core? :D
640 CCs for the GM107 seems reasonable to me, at least number-wise.
I think the surprise he is talking about is the 6-pin connector on the image...Quote:
Originally Posted by iMacmatician
reference card has no psu connector, who cares about some OC models?Quote:
Originally Posted by KiSUAN
so it has a huge oc potential that is very good.
and actually i don't care about power connectors but what about cuda cores?
look at gpuz, reading it has 960 CUDAs, but i will check documents.
GPUz is wrong, maximal number of CUDAS is 640.
ty now the last unknown left is what will be the price of card at my country.
http://abload.de/img/750tim8x5b.jpg
no psu connector ..
http://abload.de/img/750ti-264bsw.jpg
and again, no psu connector
*** edited ****, you are right, CUDAS are 640, there is problem with number of CUDAs in SMX, it is different then kepler, and manufacturers/GPUz are reading it wrong!
So, if 960/192= 5 SMX; 640/5 SMX = 128 cuda cores for each SMX for maxwell arquitecture?
Well, it should give better SMX binning options... Or not... lol
yes, one SMX has 128 CUDAs and GM107 has 5x SMX.
No need for name calling here PedantOne. Let's knock off the personal attacks/comments.
That's interesting.Quote:
Originally Posted by PedantOne
So these are indeed Maxwell cards? And not another Kepler rebrand?
28nm Maxwell, very nice indeed.
In lieu of this I'd like to see what 20nm Maxwell will bring in terms of power efficiencies too, but if this is a preview of what's to come then I'm impressed
Have had 2x 780s die in the last few weeks, might just sell the replacements and get Titans lol.
-PB
They will apparently sell for 999, though, what a rip off.Quote:
Originally Posted by paulbagz
These cards are first maxwell generation, not another kepler. Real maxwell for enthusiast will come later.
PS. Titan Black will be silently on shelves next week.
Does GM107 has an ARM core?Quote:
Originally Posted by PedantOne
no arm core in first maxwell generation
Ugh pass.Quote:
Originally Posted by zalbard
-PB
http://abload.de/img/kokotiskocyu5q.jpg
Titan Black Edition, has new GPU version, look at Picture.
lol :D (text screenshot)
Oh, you're Czech like OBR and you don't know him ? You're a complete noob or do you think people are dumb ?Quote:
Originally Posted by PedantOne
Sorry, i am not Czech. Why?
PS. Titan clocks are: 889 MHz/980 MHz Boost
Are you Slovakian then :DQuote:
Originally Posted by PedantOne
Anyway we all know you are OBR in disguise but it doesn't bother me at all. At least sometimes you have some legit info and I've learned to look through your bias ;)
while i remember obr i can't remember what happened?
I barely remember but I remember thinking it wasn't something too terribly important. I couldn't honestly care much less who the info comes from if it sounds reasonable and oftentimes is correct :D.Quote:
Originally Posted by kromosto
I'm gonna go off the deep end here with a detailed guesswork at what we might see with GM200 at 20nm. THE RESULTS AND MATH IS THEORETICAL ONLY; THIS IS NOT BASED ON ANY SECRETLY-KNOWN INFO FOR UNRELEASED PRODUCTS. So rumor sites, don't post this up as anything if it's even half-sensible like I think. Feel free to use for a base of your own speculations the idea here, of course.
http://imageshack.com/a/img203/1222/8v51.jpg
This slide says the case for Maxwell's first generation is reportedly, for the same performance level we should be seeing half the wattage used even on 28nm, and that per-core we are going to see a 35% gain (by their claims, at this point of course). So a card that would take Kepler 300w can be done in a 150w or less envelope is their claim, on 28nm even. That leaves a huge amount of extra power to continue stuffing in more cores and increasing clocks. Now add in that the Big Maxwell is going to be (GM200/204) on 20nm, and that improves power efficiency and die size usable even further. Now add in that this is Maxwell FIRST generation, for the GM107, and that GM200/204 are SECOND generation by all known info.... and we have a recipe for easily doubling the performance at least, if their claims are anywhere near true.
So, time for some math based on the theoretical and what is a real product launching in a couple of days (GTX 750 Ti).
GM107 has 5 SMM units containing 128 cuda cores, in each GPC. One GPC is what GM107 is using. It uses 60 watts for performance that is only beaten out by a GTX 660 by about 12% per the leaked benches. A GTX 660 has 960 Kepler cores, while a GM107 has 640 Maxwell cores. A GTX 660 uses a TDP rating of 140 watts. See an interesting number here?
960 is 50% more cores than the 640 ones Maxwell is using for the 750Ti. Now, it is around 12% slower there.... see the magic number close by? Nvidia claims a 35% performance increase PER CORE and additionally that the cores there will use only around half the power overall in TDP. Now extrapolate on some napkin math: the GM107 only has a 128-bit bus. That means the bandwidth efficiency is greatly improved because a GTX 650 Ti Boost is a huge amount above a GTX 650 Ti (192-bit vs. 128-bit bus widths and basically the same otherwise).
So it's safe to say that with 50% more bandwidth and 50% more cores, a GTX 660 only performing ~12% higher than a Maxwell part with slightly higher clocks is incredibly impressive. Now let's scale! We know at 28nm lithography that GM107 is around 148 square millimeters for the die size. Let's COMPLETELY IGNORE 20NM for a second here on the power savings and size! Forget about it for a minute. It would be extremely easy to see, since even on 28nm the reticule size is around 570mm2, that they could use 3 GPC units on 28nm taking around 420-430mm2 with this imaginary chip that would have 15 SMM units. 15 SMM units times 128 per unit would mean 1920 Maxwell cores.
Pretend their scheduler is great and the performance scales well with core count and clocks, and that they kept the idea of triple everything there in this hypothetical, non-existent card that is an illustration only. So we'd have a 384 bit bus with 1920 Maxwell cores, probably 7ghz memory speed of GDDR5 like Kepler does at least, and a TDP that fits inside of 200 watts. Now let's say that you only get about 75% scaling from core count here, which is reasonable even though Kepler scales pretty linearly, but it's a new architecture with Maxwell, so let's make the safe assumption. So a GTX 660 performs 12% better than a GM107 with 640 cores. Triple the core count there with our rough napkin math again with everything else and you would have a card performing around the same as GK110 fully unlocked by that theory-crafting, at least, and it has better potential for higher clocks thanks to the lower power usage.
However, in reality, we know they are going 20nm. This allows for even more power savings. This also allows for many more transistors per square mm on the die. So pretend they want to go for a 520mm2 chip on 20nm, keep costs down a tad for Big Maxwell and improve yields per wafer. According to released documents such as this: http://www.cadence.com/rl/Resources/...ew/20nm_qa.pdf we can expect to see transistor counts possible of 8-12 billion. GK110 is 7.1 billion transistors. Using that as a point of reference, let's scale 28nm GK110 to 28nm Maxwell GM107: we need less memory controllers and pad space, so we can safely come up with a number in the neighborhood, considering the 148mm2 die size compared to GK110's 551mm2 size. At 20nm, you will be able to fit upwards of 11-12 billion transistors for a high-end part. For 28nm the 148mm2 die size indicates roughly 1.7-1.8 billion transistors with the 128-bit bus.
So now we have a decent number here: we know that Maxwell at 28nm in GM107 form is taking about 2 billion transistors to perform at a level about 89% as fast as a GK106. Let's use this as a base for the next part of this thought exercise :).
So 2 billion.... it's safe to say they could fit 7 GPC's at 20nm easily since 20nm should provide roughly a 2x density shrink in die size used per transistor, very easily, and we wouldn't need to duplicate memory controllers beyond 3x of the 28nm design's if we went for a 384-bit bus. At 20nm let's say they went for a 384-bit bus, to feed 7 GPC's worth of cores since the architecture is more bandwidth-efficient than Kepler, clearly by far. That'd give us 4480 Maxwell cores which already are much more bandwidth-efficient, so it probably would be plenty well-fed on that end of things.
Power-wise we'd be looking at an envelope of, 250 watts since Maxwell GM107 at 28nm takes a full-card power of 60w for 640 cores. That means probably around 45w for the GPU itself, allowing 10w for the GDDR5 and 5w for the fan and other circuitry. The move to 20nm will improve power efficiency drastically, due to the shorter gate lengths. So it's fair to say they could fit seven of those inside of that envelope, more than easily.
Again, let's go for a linear scaling factor of approximately 75% for the core count improvements... so we have a GTX 660 GK106 card we will use and compare to a GK110 780 Ti. A GTX 660 at 1080p is able to pull about 50% of the performance of a fully-unlocked, 780 Ti (source: http://www.techpowerup.com/reviews/N...780_Ti/27.html).
Now we know each SMM will provide around 90% of that performance, and revision 2.0 (second gen) on 20nm will probably be closer to a 100% figure. So let's conservatively say each SMM-based GPC results in a real-world performance of a GTX 660. In other words, around half of a 780 Ti. Now let's conservatively also say we only see the benefit of 75% of the cores when scaling it to 7 GPC's or 35 SMM's. 35 SMM's would be, as you recall 4480 maxwell cores. Multiply the performance of the 50% card by 7 and we'd have 350% (or 2.5x faster than) of the performance of a full GK110. However, let's now apply the 75% rough rule and we come up with a much more reasonable 262.5% of the performance, or 2.6x as fast as (1.6x faster than) a GTX 780 Ti.
My predictions, therefore, are that we will see a Big Maxwell on 20nm with at minimum twice the performance of a GTX 780 Ti, and by current rumors it is due this year. Add in that Maxwell will be able to clock higher at 20nm (I based those power figures off of the numbers above which were of a card at 28nm with a 1085mhz GPU clock, and realistically they can probably fit a stock core speed of 1100-1150mhz of this big a chip in then. Thus my predictions are we would see a 450-460mm2 die size for this hypothetical card, at 20nm, with a 250-260w TDP rating and thanks to the 2x transistor density, approximately 9.2b transistors.
Napkin math, for sure, but worth thinking about, eh?
WHen i read this type of theory, i understand why Nvid?a launch a low end range product first with maxwell lol .. ( 1st generation ofc ) ...
But i think you are a little bit in the sad side.. 300% performance gain on the minimum..
Quote:
Originally Posted by Lanek
??? 300% gain would mean 4x the performance. I said at minimum we will probably see twice the performance of GK110 full (780 ti) in a Big Maxwell on 20nm. Please actually read before replying with completely wrong info about what's written very plainly in the post you're replying to :). I'm not sure if you were reading another post but if so you should refer to who you're talking to. Personally, I wrote the below:
Quote:
My predictions, therefore, are that we will see a Big Maxwell on 20nm with at minimum twice the performance of a GTX 780 Ti
I think you are being overly optimistic there, GoldenTiger. Performance doubling with every new generation is a thing of the past...
I don't know. Nvidia seems to have a track record for surprising us over the past 10 years - especially for new architectures. I'm personally expecting a pretty significant jump in performance with the Flagship card.Quote:
Originally Posted by zalbard
Yeah, they've managed to at least match (usually surpass) the previous generations dual gpu card with their single gpu flagships.Quote:
Originally Posted by EternityZX9
If they lead with the mainstream gpu again I'll wait though.
No! This will be a weak overpriced release.....the price is validated so high from the first titan....then next titan will again kick ass...like a tick tock effect.
but i'm just making up shart out of the thin air like most of you :P
Does anyone know if they have a hydro copper waterblock designed for the titan black?
Yep EVGA has em on the website.Quote:
Originally Posted by Zaxis01
-PB