Larrabee matches a GTX285?

[Idontcare]

Personally I like to hope that Intel took some lessons-learned to heart over the Merced and i740 situations and if 45nm Larrabee doesn't dominate when time of release comes up then they just skip it altogether and delay Larrabee debut for 32nm shrink, repeat as necessary until they come up with a winning combination of architecture and process technology plus driver maturity to snatch the halo at time-zero.

Really they don't need any negatism to surround the release, and they don't need profits/revenue from the chip in the meantime (should 45nm be skipped) unlike AMD and Nvidia that need to sell everything they've invested the resources into developing.

Its just a hope of mine, not meant to be grounded in reality of business of course. But I'd like to see them X-25M/X-25E the GPU market whenever they do finally release Larrabee if you know what I'm saying.

(or else just not bother because Intel isn't going to be happy fighting NV and AMD for mainstream GPU margins in the 20-30% realm, which is where those GM's are now and will be when a third player is added...Intel needs a path to 50% GM or else they abandon the market segment...ala HDTV and early cellphone efforts that failed circa 2004-2005)

 

[Keysplayr]

Originally posted by: Scali

Originally posted by: Keysplayr
You know, I'm still not quite sure what Larrabee is, from a physical standpoint. Is it a discrete PCI-e card? Is it a CPU? Will a specialized motherboard with a new socket be needed? Anyone know? I haven't really been keeping up. It seems like Larrabee has been in development for 30 years now. It got boring.

If I'm not mistaken, Larrabee is the codename for Intel's GPGPU architecture (such as G300 or RV870), which is a massively parallel in-order x86-based architecture, and a derivative of the Terascale project.
However, this GPGPU is going to be placed on a PCI-e card with dedicated video memory, just like your average discrete graphics card from AMD, nVidia and others.
People often refer to the card as a whole as Larrabee aswell, because we lack a better name (just like how the codenames for future AMD and nVidia GPUs float around on the web, but we don't really know if the cards they will be on will be called Radeon HD5870 or GeForce GTX385 or such...)

After the initial launch of Larrabee as a discrete card, Intel will also integrate the technology in their future IGPs, which will be integrated into the CPU package, rather than the chipset (a CPU with integrated GPU will probably be launched later this year, but this is still based on their current grahpics technology, not Larrabee). There will be a new socket introduced for CPUs with integrated graphics, but there isn't going to be a second socket on the motherboard (CPU + GPU), if that's what you were thinking.

So "Larrabee" discrete PCI-e graphics card. System still requires a CPU (whatever platform is out by then: SandyBridge, Granite Park, Gelatinous Boardwalk) < hehe.

 

[Scali]

I'm not sure if Intel can afford to fail with Larrabee. There seems to be a trend of moving ever more processing tasks from CPU to (GP)GPU. Those are especially the most CPU-intensive tasks, in other words, the tasks that people buy faster CPUs for (folding, video encoding, image processing, 3d rendering, physics, AI etc).
I think that is actually the sole reason why Intel bothers to get into the GPU game now. They are probably already seeing demand for their high-end CPUs dropping, because people no longer need them for gaming. They'd rather buy a relatively simple dualcore CPU and team it up with a big $300+ GPU. As GPU-accelerated physics, video encoding, PhotoShop etc also become more mainstream, we'll see the same trend expanding to a large part of the PC market.
Intel has no problem with people buying expensive GPUs, but ofcourse they will have to be Intel GPUs.

 

[Scali]

Originally posted by: Keysplayr
So "Larrabee" discrete PCI-e graphics card. System still requires a CPU (whatever platform is out by then: SandyBridge, Granite Park, Gelatinous Boardwalk) < hehe.

Yup. Although Larrabee has x86-roots, it wouldn't be very suitable for running conventional x86 software (just like running linux on the PS3 was possible because of the Cell's PowerPC heritage, but it performed poorly). So a conventional CPU is teamed up with Larrabee. At first in the form of a discrete PCI-e card, and later as the CPU and GPU married together in a single package on the same socket (much like AMD's Fusion plans).
Discrete cards will probably be around for the high-end though, as dedicated high-bandwidth videomemory has huge advantage over an IGP sharing its memory with the CPU.

 

[Zstream]

Originally posted by: BFG10K

Originally posted by: Zstream

If we are discussing "game performance", then I say who cares!

Considering gaming is by far the biggest market these parts compete in, anyone that wants to be a viable discrete GPU competitor should care. Remember that Larrabee is a discrete PCIe part, not some kind of new CPU socket.

The meaning of this chip is not to dominate games or have real any impact in that segment.

Then it can?t really compete with ATi?s or nVidia?s GPUs.

I would have to argue that specialized science development is more important. Schools can may millions of dollars unlike the small retail vendor. I do not have any proof of this but based off of company's such as HP/Dell they tend to look for the business deal.

Yes, I agree it will not compete with ATI or Nvidia's GPU's for some time. It seems to be more of a PC on a PCIe card.

 

[Scali]

I wouldn't be so quick to write off this card as a gaming card.
I mean, if they do reach the goal of GTX285 performance in games at introduction, then they have a very good mainstream gaming performance level (and the answer to the ubiquitious question "Can it run Crysis?" would be: "Yes, abso-freaking-lutely!"). Most people probably don't have GTX285-like performance in their systems yet, so it would be a nice upgrade (I know I don't...).
All Intel then has to do is sell them at the right price. Then they can indeed dominate the market. They don't need to make a profit on them. The CPU market is much bigger and more profitable than mainstream GPUs at this point, so Intel can easily compensate the GPU division with the CPU sales while trying to get a 'foot in the door' on the GPU market.
Then they'll just tick-tock the GPU architecture into perfection over the following years, as they've done with their CPUs.

 

[ilkhan]

Originally posted by: Keysplayr
You know, I'm still not quite sure what Larrabee is, from a physical standpoint. Is it a discrete PCI-e card? Is it a CPU? Will a specialized motherboard with a new socket be needed? Anyone know? I haven't really been keeping up. It seems like Larrabee has been in development for 30 years now. It got boring.

It's an add-in graphics card, just with Intel silicon under the heatsink rather than nVid or AMD silicon.

 

[LOUISSSSS]

Originally posted by: Scali
I wouldn't be so quick to write off this card as a gaming card.
I mean, if they do reach the goal of GTX285 performance in games at introduction, then they have a very good mainstream gaming performance level (and the answer to the ubiquitious question "Can it run Crysis?" would be: "Yes, abso-freaking-lutely!"). Most people probably don't have GTX285-like performance in their systems yet, so it would be a nice upgrade (I know I don't...).
All Intel then has to do is sell them at the right price. Then they can indeed dominate the market. They don't need to make a profit on them. The CPU market is much bigger and more profitable than mainstream GPUs at this point, so Intel can easily compensate the GPU division with the CPU sales while trying to get a 'foot in the door' on the GPU market.
Then they'll just tick-tock the GPU architecture into perfection over the following years, as they've done with their CPUs.

good point, intel doesn't need to make (much) profits on larabee. all they need to do is steal some profits from NV to hurt them a little bit each quarter. intel has a great tick/tock schedule to keep NV on their feet.

Intel HAS the cash in hand to lets say "PWN" nvidia. they already have i7 researched and developed so they dont need to devote any $$ into that for the next 4-5 years. and i7 (i assume) is bringing in loads of profit fromthe cpu department.

with regards to the GMA series, i dont think it should be compared. the GMA is NOT for gamers, nor for heavy GPU tasks. it is just there to output some video for desktop/business users. admit it, you dont use your 3/4 series GMA to play any 3d games. i know i dont on my laptop. Intel has been in the market and have studied/watched nvidia long enough to be able to do as well or better than them.

i would like to see larabee do better than nvidia right now since they dont seem to do anything if they dont have competition. all they do is release re-badged old cards to trick un-educated consumers. intel doesn't do that. without any competition from AMD, they release the high end i7 with pretty good prices.

 

[ilkhan]

Originally posted by: Scali
a CPU with integrated GPU will probably be launched later this year, but this is still based on their current grahpics technology, not Larrabee). There will be a new socket introduced for CPUs with integrated graphics, but there isn't going to be a second socket on the motherboard (CPU + GPU), if that's what you were thinking.

I just want to say that the CPU+GPU socket is the standard s1156 coming for lynnfield and clarkdale. It'll take a chipset with FDI to pass the graphics data, but not a "special" socket.

Originally posted by: LOUISSSSS
Intel HAS the cash in hand to lets say "PWN" nvidia. they already have i7 researched and developed so they dont need to devote any $$ into that for the next 4-5 years.

No, instead of working on Nehalem they're working on sandy bridge/ivy bridge haswell/beyond. Researching never stops for them.

 

[LOUISSSSS]

Originally posted by: ilkhan

Originally posted by: LOUISSSSS
Intel HAS the cash in hand to lets say "PWN" nvidia. they already have i7 researched and developed so they dont need to devote any $$ into that for the next 4-5 years.

No, instead of working on Nehalem they're working on sandy bridge/ivy bridge haswell/beyond. Researching never stops for them.

which is GREAT! they dont starve the market like a monopoly/oligopoly would.
right now, i honestly feel the software is the bottleneck. 99% of PC users are fine on an AMD X2 / Intel C2D cpu. i know my cpu is overkill for things i do and i have a 2-3 year old Q6600.

i am willing to sacrifice some cpu/ram resources for more powerful software even if its just stuff like firefox/HW monitoring/music and video playing. most of the time i'm only using about 25% of my ram and <10% of my CPU.

 

[AmdInside]

My problem with Intel graphics is that their drivers suck. It isn't brought up much because no one uses Intel integrated graphics when they need to do 3D so quality, bugs and performance are never looked into.

 

[Hacp]

Originally posted by: LOUISSSSS

Originally posted by: ilkhan

Originally posted by: LOUISSSSS
Intel HAS the cash in hand to lets say "PWN" nvidia. they already have i7 researched and developed so they dont need to devote any $$ into that for the next 4-5 years.

No, instead of working on Nehalem they're working on sandy bridge/ivy bridge haswell/beyond. Researching never stops for them.

which is GREAT! they dont starve the market like a monopoly/oligopoly would.
right now, i honestly feel the software is the bottleneck. 99% of PC users are fine on an AMD X2 / Intel C2D cpu. i know my cpu is overkill for things i do and i have a 2-3 year old Q6600.

i am willing to sacrifice some cpu/ram resources for more powerful software even if its just stuff like firefox/HW monitoring/music and video playing. most of the time i'm only using about 25% of my ram and <10% of my CPU.

You can transcode 1080p 5+mbit h264 realtime?

 

[lopri]

"Larrabee matches GTX285" sounds like "Radeon 4890 has 1.2 Gflops of processing power", or "GeForce GTX 280 has 150GB/s bandwidth" or something in that vein.

 

[SickBeast]

Originally posted by: lopri
"Larrabee matches GTX285" sounds like "Radeon 4890 has 1.2 Gflops of processing power", or "GeForce GTX 280 has 150GB/s bandwidth" or something in that vein.

It is all PR.

That said, it does make sense that the Larrabee will be a midrange/upper midrange part.

 

[poohbear]

Originally posted by: WelshBloke

Originally posted by: Scali

Originally posted by: Rusin
Larrabee is huge and very expensive chip to make. One 300mm wafer could hold only about 64 Larrabees when for comparison one 300mm wafer could hold 94 Nvidia's 65nm GT200 chips. Which means that Larrabee couldn't fight against GTX 285 price wise.

You don't know that.
nVidia has to have a third party like TSMC manufacture their GPUs. Intel does everything in-house, and on a huge scale. Intel's production facilities are also more advanced than TSMC's, and their 45 nm process is very successful and mature, where TSMC is struggling with 40 nm, and 55 nm is still the bread-and-butter of nVidia and AMD GPUs.

Aside from that, yes you may get 94 GT200 chips out of a 300mm wafer, but how many of those are GTX285? Probably a minority, as most of them are salvaged as GTX260/275/295.
So it's really hard to make a comparison between nVidia and Intel... There are so many factors involved here.

Intel has proven before that they can beat a competitor with a much larger chip. The Pentium 4/D were much larger than the competing Athlon XP/64/X2 processors. Instinctively you might think that AMD would get the lowest prices and/or the highest profits, but in fact it was Intel undercutting AMD's prices with the Pentium D series, and Intel had the better business results during that era, so they would have had the higher profit margins.
I think Intel will take advantage of their production capabilities in a similar vein here.

Yep. If Intel can compete with a GTX 285 (I personally dont think they will) they have the ability to price it so low that it'll sell very well.

AMD/Nvidia have no way in hell of competing pricewise against Intel (if Intel want to muscle in), they both need to be much faster to stay in the market.

that's what i was thinking, if we have a gtx285 class card in 6 months for $150, then its.... game over man!!!! game over!!!!!

 

[IntelUser2000]

Originally posted by: LOUISSSSS

Intel HAS the cash in hand to lets say "PWN" nvidia. they already have i7 researched and developed so they dont need to devote any $$ into that for the next 4-5 years. and i7 (i assume) is bringing in loads of profit from the cpu department.

LOL I wouldn't be sure about "loads of profit" for i7. Currently its only taking 1% of Intel's total market share, which is NOTHING.

Originally posted by: LOUISSSSS

with regards to the GMA series, i dont think it should be compared. the GMA is NOT for gamers, nor for heavy GPU tasks.

Although I believe the next generation GMA will be good enough to be a competitive IGP, its still an IGP. Yea see, Larrabbee, unlike GMA has x86 roots and its what Intel is familiar with.

Sorry but this doesn't sound like Intel PR to me but a real estimation by 3rd party testing Larrabee: "According to one close Intel partner that wished not to be named, this isn't the case. We were told that Larrabee is currently only capable of performance levels similar to Nvidia's GeForce GTX 285."

Depending on the clock tested by the "source", the maturity of the drivers, and the SKU tested, it might as well turn out to be competitive. We are approximately SIX months from release!

 

[evolucion8]

Originally posted by: IntelUser2000
LOL I wouldn't be sure about "loads of profit" for i7. Currently its only taking 1% of Intel's total market share, which is NOTHING.

But it should be still profitable, after all the Nehalem architecture is not a brand new architecture, for sure they didn't invest a lot of money create the Nehalem unlike the Pentium 4 architecture or the Pentium M which is also a derivative of the P6 architecture like Nehalem.

 

[Keysplayr]

Originally posted by: evolucion8

Originally posted by: IntelUser2000
LOL I wouldn't be sure about "loads of profit" for i7. Currently its only taking 1% of Intel's total market share, which is NOTHING.

But it should be still profitable, after all the Nehalem architecture is not a brand new architecture, for sure they didn't invest a lot of money create the Nehalem unlike the Pentium 4 architecture or the Pentium M which is also a derivative of the P6 architecture like Nehalem.

On die memory controller
monolithic die (native quad core)
hyperthreading
Shared L3 cache

Nothing about these things (short list) rings new architecture in your mind?

 

[Scali]

Yea, Nehalem is a "tock" -> new architecture.
See this PDF for more info: http://www.intel.com/pressroom...whitepaper_nehalem.pdf

 

[Idontcare]

Originally posted by: Keysplayr

Originally posted by: evolucion8

Originally posted by: IntelUser2000
LOL I wouldn't be sure about "loads of profit" for i7. Currently its only taking 1% of Intel's total market share, which is NOTHING.

But it should be still profitable, after all the Nehalem architecture is not a brand new architecture, for sure they didn't invest a lot of money create the Nehalem unlike the Pentium 4 architecture or the Pentium M which is also a derivative of the P6 architecture like Nehalem.

On die memory controller
monolithic die (native quad core)
hyperthreading
Shared L3 cache

Nothing about these things (short list) rings new architecture in your mind?

I expect the space shuttle to cost a dime a dozen too, its just a derivative of the wright brothers aeroplane after all. And I should knowz, I wiki like a mother!

 

[evolucion8]

I'm talking about execution engine level, of course it has more features and performance. The Nehalem isn't any wider than the Core 2 architecture, they probably share the same pipeline lenght. Since great part of the chip like the front ends went underutilized quite often, adding Hyper Threading without widening it's execution engine allowed performance gains. Using the same logic as you I should call the Phenom a new architecture, when really is an optimized K8 processor.

http://forums.anandtech.com/me...GE=3&parentid=31138578

Originally posted by: Idontcare
I expect the space shuttle to cost a dime a dozen too, its just a derivative of the wright brothers aeroplane after all. And I should knowz, I wiki like a mother!

I also expect that you cost a dime a dozen too, after all you are just a derivative of the Pan paniscus

 

[Obsoleet]

Interesting development.
In sum, goodbye Nvidia.

 

[Idontcare]

Originally posted by: evolucion8
I also expect that you cost a dime a dozen too, after all you are just a derivative of the Pan paniscus

Chronologically, Pan paniscus parallels Homo sapiens, we are not derivatives of Pan paniscus. Rather, we are derivatives of Homo rhodesiensis.

Eventually something will be a derivative of Pan paniscus, in another 0.2-0.4M yrs, provided they don't go extinct in the same timeframe.

The geographic distribution of Bonobos is rather limited, and in a region of frequent civil unrest, so their longevity is certainly not a foregone conclusion.

 

[evolucion8]

Originally posted by: Idontcare

Originally posted by: evolucion8
I also expect that you cost a dime a dozen too, after all you are just a derivative of the Pan paniscus

Chronologically, Pan paniscus parallels Homo sapiens, we are not derivatives of Pan paniscus. Rather, we are derivatives of Homo rhodesiensis.

Eventually something will be a derivative of Pan paniscus, in another 0.2-0.4M yrs, provided they don't go extinct in the same timeframe.

The geographic distribution of Bonobos is rather limited, and in a region of frequent civil unrest, so their longevity is certainly not a foregone conclusion.

I know I just wanted to give you the curiosity to research the meaning of that word Hope they don't extinct.

 

[Kakkoii]

I think Larrabee will be a good thing for the PC Gaming industry if it's something OEM's decide to pick up. It would have the potential to create a much larger install base of gaming capable PC's, which would in turn help grow the PC gaming market, of which has been in a bit of a slump with developers choosing the consoles more often these days. If we didn't have Steam the industry would probably be a lot worse.