Larrabee will be massive?

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
THG article

"According to current known information, our source indicated that Larrabee may end up being quite a big chip--literally. In fact,we were informed that Larrabee may be close to 650mm square die, and to be produced at 45nm. "If those measurements are normalized to match Nvidia's GT200 core, then Larrabee would be roughly 971mm squared," said our source--hefty indeed. This is of course, an assumption that Intel will be producing Larrabee on a 45nm core"

"Our source also indicated that Intel is looking to ship Larrabee two years later, putting us in summer of 2011. Of course, by that time, we will have GPUs that are 2 to 4 times faster than current GPUs from both AMD/ATI and Nvidia. However, at that time Larrabee may not be what it is today either."

To be clear, I don't really trust "sources" much, and THG less...
I am posting this as a "rumour only"...
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
I think your title is inaccurate.

Larrabee will be 650mm at 45nm; it would only be 971mm if it were manufactured on the same process as the GT200 (55nm?).
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: SickBeast
I think your title is inaccurate.

Larrabee will be 650mm at 45nm; it would only be 971mm if it were manufactured on the same process as the GT200 (55nm?).

No mate..."normalized" means that once you take the 650mm core and make the whole package (I/O, etc...), it's a 971mm2 chip
 

sandorski

No Lifer
Oct 10, 1999
70,130
5,658
126
Methinks the End Result will not live up to the Hype. Hell, wouldn't surprise me if it never comes out, other than just another Mediocre Onboar Video implementation. Actually I'll be generous: A Competitive Onboard Video implementation.
 

Keysplayr

Elite Member
Jan 16, 2003
21,209
50
91
Originally posted by: Viditor
Originally posted by: SickBeast
I think your title is inaccurate.

Larrabee will be 650mm at 45nm; it would only be 971mm if it were manufactured on the same process as the GT200 (55nm?).

No mate..."normalized" means that once you take the 650mm core and make the whole package (I/O, etc...), it's a 971mm2 chip

Is anyone sure that this thing is going to need a CPU? Or is it in fact the CPU and GPU all rolled into one? No CPU necessary as previously conceived?
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
Originally posted by: Keysplayr
Originally posted by: Viditor
Originally posted by: SickBeast
I think your title is inaccurate.

Larrabee will be 650mm at 45nm; it would only be 971mm if it were manufactured on the same process as the GT200 (55nm?).

No mate..."normalized" means that once you take the 650mm core and make the whole package (I/O, etc...), it's a 971mm2 chip

Is anyone sure that this thing is going to need a CPU? Or is it in fact the CPU and GPU all rolled into one? No CPU necessary as previously conceived?

I suppose with an updated BIOS, you could "boot from PCIe".

It would really be a bombshell if intel came out with a platform where you could just drop in a Larrabee as if it were a CPU, then just use your system memory for vram. If they do this, it might actually make Larrabee cost and performance competitive with NV and AMD. In fact, if people can use Larrabee instead of a CPU, then it makes it that much cheaper in a sense.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Keysplayr
Originally posted by: Viditor
Originally posted by: SickBeast
I think your title is inaccurate.

Larrabee will be 650mm at 45nm; it would only be 971mm if it were manufactured on the same process as the GT200 (55nm?).

No mate..."normalized" means that once you take the 650mm core and make the whole package (I/O, etc...), it's a 971mm2 chip

Is anyone sure that this thing is going to need a CPU? Or is it in fact the CPU and GPU all rolled into one? No CPU necessary as previously conceived?

I can't see anything like that running on Windows for example, and (as Itanium has taught) any form of emulation makes things unworkably slow...
 

Scali

Banned
Dec 3, 2004
2,495
0
0
Are they talking about final silicon though? As far as I know, Intel wanted to create prototypes on 65 nm first (that would be the dies that are available now for development). 650 mm^2 on 65 nm seems reasonable.
 

Viditor

Diamond Member
Oct 25, 1999
3,290
0
0
Originally posted by: Scali
Are they talking about final silicon though? As far as I know, Intel wanted to create prototypes on 65 nm first (that would be the dies that are available now for development). 650 mm^2 on 65 nm seems reasonable.

Well, the article specifically said 45nm...but as I said, this is all from "sources" at THG. I'm never inspired to confidence with them...
 

BFG10K

Lifer
Aug 14, 2000
22,709
2,980
126
I thought the same site was saying it?ll be here in six months and it?ll be as fast as the GTX285? Now they?re saying 2011?

Like I said, I don?t have high performance hopes for this thing, and I?ll only believe such wild claims when I see it.
 

Idontcare

Elite Member
Oct 10, 1999
21,118
59
91
Originally posted by: Viditor
Originally posted by: Idontcare
Laughably unbelievable.

Which part?

All of it...in fact it is so much so that Intel actually broke their usual "we don't comment on pre-released products" and stated this about the "information":

We contacted Intel for comment in regards to the above information. Intel denied that any of the above is true.

But to first-order, the physical limits of a die size is right around 1000 mm^2, i.e. the size of the litho shot itself. That would make it the largest die fabbed in history, and then to argue it is going to be sold in a product that retails at $300? Chips that sell for $10k don't have enough margins to be fabbed at >900mm^2.

It is unbelievable that anyone would actually make a rumor or speculation so bad, so technically inconceivable, as that. They might as well said Larrabee is going to be 30,000 mm^2 monster! :laugh: Equally ridiculous. Laughably so.

Look at the margins on Dunnington and the die-size for it...there's a working boundary condition for an upper limit of a practical die-size for Larrabee IF it could be sold at Dunnington's gross margins. (which it can't/won't so dial down the die-size upper limit correspondingly)
 

Rusin

Senior member
Jun 25, 2007
573
0
0
Intel could squeeze 64 Larrabees on 300mm wafer (45nm). For comparison Nvidia squeezed 94 GT200 on 300mm wafer (65nm) and GT200 65nm was 575mm^2

At least on wafers Larrabee looked HUGE
 

Scali

Banned
Dec 3, 2004
2,495
0
0
Originally posted by: Rusin
At least on wafers Larrabee looked HUGE

Well, that was my point. If Intel stayed true to their claim of making prototypes on 65 nm, then the Larrabee wafers we've seen so far are 65 nm, not 45 nm.
In that case, any size calculations based on the wafer shots should be adjusted to 65 nm, because the assumption of 45 nm would be wrong.
 

Idontcare

Elite Member
Oct 10, 1999
21,118
59
91
Originally posted by: Scali
Originally posted by: Rusin
At least on wafers Larrabee looked HUGE

Well, that was my point. If Intel stayed true to their claim of making prototypes on 65 nm, then the Larrabee wafers we've seen so far are 65 nm, not 45 nm.
In that case, any size calculations based on the wafer shots should be adjusted to 65 nm, because the assumption of 45 nm would be wrong.

Otellini said the wafer shown to the public was an "extreme" variant of Larrabee, not the mainstream version. He said that in an interview some time after the wafer viewing.

I don't doubt the Larrabee wafer shown was 45nm, but I do doubt whether the wafer shown (with the 680mm^2 die) is actually the same 32 core larrabee depicted later in Intel's slides.
 

M0RPH

Diamond Member
Dec 7, 2003
3,305
1
0
Originally posted by: Viditor
Originally posted by: SickBeast
I think your title is inaccurate.

Larrabee will be 650mm at 45nm; it would only be 971mm if it were manufactured on the same process as the GT200 (55nm?).

No mate..."normalized" means that once you take the 650mm core and make the whole package (I/O, etc...), it's a 971mm2 chip

I agree with SickBeast, I think you're misunderstanding what they mean by normalized. They are trying to compare it to the Nvidia part so they're saying that it's equivalent to 971mm in a 55nm process.
 

Rusin

Senior member
Jun 25, 2007
573
0
0
Originally posted by: M0RPH


I agree with SickBeast, I think you're misunderstanding what they mean by normalized. They are trying to compare it to the Nvidia part so they're saying that it's equivalent to 971mm in a 55nm process.
"This is of course, an assumption that Intel will be producing Larrabee on a 45nm core"

 

ilkhan

Golden Member
Jul 21, 2006
1,117
1
0
My unlearned opinion is that they'll launch at the same time as clarkdale/arrandale in order to take advantage of the 45nm fab space that opens up. But they'll probably switch to 32nm pretty quick with the amount of money they are putting into getting 32nm fabs online.

650mm with Intel's tech would be doable for the extreme part. But not for the mainstream version, which will be the reduced core count.
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
If the Larrabee core is too big, it will be easy enough for intel to pare it back to less cores.

I was always under the impression that intel was going to manufacture Larrabee on a 32nm process. In fact, doesn't intel have to be designing it for a specific process right now?

Until intel releases some kind of verifiable benchmarks in a real game like Crysis, everything we read about Larrabee will be FUD, speculation, and PR (and will probably be blatantly inaccurate).
 

Snooper

Senior member
Oct 10, 1999
465
1
76
Originally posted by: IdontcareChips that sell for $10k don't have enough margins to be fabbed at >900mm^2.

Your joking, right? With 300mm wafer costs (in HVM) at or below the $3,000/wafer range, do you REALLY think they can't make money on a device above 900mm^2?

Assuming 80% utilization of the silicon (varies depending on die size, shape, major flat or notch, etc), you should be able to get right at 100 die on there. Heck, if they can only hit 50% die yield, you are still talking about 50 good die per wafer.

Now, even if this IS a $300 die (I REALLY doubt it), then you are looking at 50 * $300 = $15,000 PER wafer! More than likely, you are looking at a die cost of around $60 per good die. Through in another $60 for the rest of the card and another $60 for profit (perhaps a bit low), you are looking at a possible $180 mid range card with the performance of today's top of the line single GPU card. Not bad. ESPECIALLY for Intel.

The whole margin thing on uber expensive die comes in when you are only building a few lots of wafers at a time. If I only expect to sell 10,000 chips and I can get 50 good die per wafer and I am making 25 wafers in a lot, you are talking about a run of only EIGHT lots to make all the die I expect to need for this high price project. That leads to the simple problem that you never hit HVM, so you never get the bugs worked out. Your die yields generally are in the trash, so you have to make a lot more wafers to make up for it and that costs money.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
Look guys. GT200 is on a 65nm process.

The bolded part specifically states that Larrabee is 650mm2 at 45nm. How did we conclude that its 900mm2??

"According to current known information, our source indicated that Larrabee may end up being quite a big chip--literally. In fact,we were informed that Larrabee may be close to 650mm square die, and to be produced at 45nm. "If those measurements are normalized to match Nvidia's GT200 core, then Larrabee would be roughly 971mm squared," said our source--hefty indeed. This is of course, an assumption that Intel will be producing Larrabee on a 45nm core"
 

Kuzi

Senior member
Sep 16, 2007
572
0
0
The +900mm^2 Larrabee was just speculation for it's size "if" produced on a 55nm process, that will never happen of course, and Larrabee will be produced on 45nm or 32nm.

I still don't see Larrabee being able to beat high end NV and ATI offerings in 3D/gaming performance when released. Even if it had a larger die size.
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Ananda already covered this die size thing . in a rather good article here at AT. If you read it you will find that at 45nm on a 650 die size 64 larrabee cores fit . But until Intel actually release larrabbee. We won't know. Which die people are referring to the 65nm test die . The production 45nm die or the upcoming 32nm. This ones to hot to touch . Intel could do many things here.

I just hope its on time and plays todays games at playable frame rates and has great graphics. and brings new game play ops and visuals to the programmer/ Artist.

Than if it can do other workloads ontop of graphics all the better. I not going to get excited about todays games. But they need to be playable frame rates . Its the new releases DX11 games and Intels game studeo games. Hopefully the Project offset Title will release with Larrabee. Si intel can showcase a properly built game on there system . ..

Not long to wait and see now. 6 months. By than well have cards or know a whole lot more.
 

bunnyfubbles

Lifer
Sep 3, 2001
12,248
3
0
Originally posted by: Viditor
Originally posted by: Idontcare
Laughably unbelievable.

Which part?

What set off my BS detector was combining 2011 and 45nm. 32nm is supposed to be seen as early as late this year. With delays, early 2010. By 2011 32nm will certainly be the norm for intel and we'll be talking about the next step...

Heck, even even other GPUs will be going 40nm by the end of this year (if not already as seen with the RV740). There is no way Intel will ship Larrabee in 2011 on a 45nm process.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |