Is Intel pulling a Fermi?

[JAYDEEJOHN]

We all saw earlier nVidia showing a card that wasnt there, even have pictures of its CEO holding up a fake.
Earlier this week, Intel showed off what some believed to be LRB. Some even wrote about it as such
http://www.theregister.co.uk/2009/11/17/sc09_rattner_keynote/
While others took the time to see the others pulling a "mistake"
http://www.computerworld.com/s/article/9140949/Intel_to_unveil_energy_efficient_many_core_research_chip?taxonomyId=1
Its not hard to see someone was pulling something or other

 

[jennyh]

I think it's pretty apparent that HKMG gave intel more of a lead in certain areas than say, immersion lithography gave AMD.

I think Jaydee is mainly anxious to see just what HKMG will do for graphics. Remember, up till now we've never really seen it. Globalfoundries are making claims of up to 40% increased speeds and I'm sure you can understand what that would mean for ATI suddenly having access to that.

 

[daedalus685]

I am not bitter at all. It just seems that you flip flop more than a politician.

http://i61.photobucket.com/albums/h65/sissneym/flipflop.jpg

http://i161.photobucket.com/albums/t202/lulu121290/mccain.jpg

"flip flop" as a term always irked me. Since when is it a flaw to reevaluate ones opinions? Keeping my nose out of this debate here.. but the worst scientists in the worlds are the ones that stick to their guns and never deviate.

 

[JAYDEEJOHN]

For its part, Intel is working on a GPU product of its own, code-named Larrabee, that is expected to be released in 2010. Numerous Nvidia patents, such as those covered by the 2004 cross-license agreement, would likely apply to Larrabee, Brookwood says. So, without the right to use those patents, Intel may have difficulty getting Larrabee out the door. "Nvidia has a lot of fundamental intellectual property for that kind of hardware and I'm not sure that Intel can work around it," he says.
http://www.businessweek.com/technology/content/dec2009/tc2009122_478796_page_2.htm

 

[yomamafor1]

Btw yomama - maybe you should have a closer look at that top 500 supercomputer list before making a fool of me in your sig?

You've made a fool of yourself. Cray's Jaguar is more than 3x more powerful than the closest intel (which incidentally is half intel half ATI), so I am completely correct in saying that there is no chance whatsoever of intel taking top supercomputer spot for the forseeable future...if ever.

Interestingly, if you take a look at the individual output of each CPU, the Istanbul used in the Jaguar outputs about the same as a Nehalem EP used in NUDT TH1...

...aside from that, Jaguar also has 3x more cores than NUDT TH1...

And unless you have a crystal ball (which I doubt), or come from the future (which I also doubt), there's no possible way for you to make a groundless claim just based on fragmented information.

But again, AMDiot always knows best, who am I kidding?

 

[spud]

I am not bitter at all. It just seems that you flip flop more than a politician.

http://i61.photobucket.com/albums/h65/sissneym/flipflop.jpg

http://i161.photobucket.com/albums/t202/lulu121290/mccain.jpg

Seems??? *whispers* he does it all the time.

 

[jennyh]

Interestingly, if you take a look at the individual output of each CPU, the Istanbul used in the Jaguar outputs about the same as a Nehalem EP used in NUDT TH1...

...aside from that, Jaguar also has 3x more cores than NUDT TH1...

And unless you have a crystal ball (which I doubt), or come from the future (which I also doubt), there's no possible way for you to make a groundless claim just based on fragmented information.

But again, AMDiot always knows best, who am I kidding?

No actually I am making that claim and if you believe otherwise you're quite laughably lacking knowledge in that area.

Jaguar will not be beaten by any intel HPC for many years, if ever. Now shoo off back under the rock you crawled out from under.

 

[IzzyCraft]

No actually I am making that claim and if you believe otherwise you're quite laughably lacking knowledge in that area.

Jaguar will not be beaten by any intel HPC for many years, if ever. Now shoo off back under the rock you crawled out from under.

Too bad if life is hilarious and nvidia venture into HPC could do one better. (That sentence is meant in hypothetical context but i'm unwilling to re word it as such)

 

[BadTrip]

I think it's pretty apparent that HKMG gave intel more of a lead in certain areas than say, immersion lithography gave AMD.

I think Jaydee is mainly anxious to see just what HKMG will do for graphics. Remember, up till now we've never really seen it. Globalfoundries are making claims of up to 40% increased speeds and I'm sure you can understand what that would mean for ATI suddenly having access to that.

I agree with you jenny. I am very pleased with gpus getting HKMG. Looks very promising.

 

[BadTrip]

"flip flop" as a term always irked me. Since when is it a flaw to reevaluate ones opinions? Keeping my nose out of this debate here.. but the worst scientists in the worlds are the ones that stick to their guns and never deviate.

Re-evaluating your opinion when new data is obtained is absolutely necessary. Changing you position because your "savior" decides it is needed is foolish at the very most.

 

[jennyh]

Looking at say Larrabee vs Cypress...

Going to take the worst case for intel here based on what we've learned so far -

If those current figures of 1 teraflops SGEMM for 16 overclocked cores at 150w is on a 32nm process - things don't look so good. Cypress can hit 2.5 gigaflops SGEMM at 188w on a bad 40nm process - overall that's about twice as fast (in pure shader power) for not much more power draw.

If however, this 16 core Larrabee is 45nm, things are looking up. At 32nm there is no reason to believe it couldn't do 2 teraflops SGEMM (due to 32 cores instead of 16, or a mix of faster and more cores)...but that still isn't beating the Cypress die though it would be close on perf/watt.

There is an awful lot more to it than just single precision of course. Nvidia cards texture like nothing else, it has always been their forte. Larrabee needs to get everything right to usurp the throne - shaders, texturing, software...it's asking a helluva lot even from intel.

 

[spud]

Re-evaluating your opinion when new data is obtained is absolutely necessary. Changing you position because your "savior" decides it is needed is foolish at the very most.

Word, playa.

 

[jimmysmitty]

No actually I am making that claim and if you believe otherwise you're quite laughably lacking knowledge in that area.

Jaguar will not be beaten by any intel HPC for many years, if ever. Now shoo off back under the rock you crawled out from under.

They can if the super computer maker puts more PC at work. Thats what its really all about. Hundreds of PCs with multiple CPUs doing work.

If LRB does whats claimed Intel could easily take back the HPC.

Then again who gives a flying duck. That doesn't affect us at all in the end.....

 

[jennyh]

It's a lot more likely that a company like Cray will be hooked up to likely movements in that area though jimmy. They've been at the top for a long, long time and its not like they are even attempting to branch out much. HPC is their business, they won't get it badly wrong but it will be interesting to see if they go the gpu route sooner or later.

 

[yomamafor1]

No actually I am making that claim and if you believe otherwise you're quite laughably lacking knowledge in that area.

Jaguar will not be beaten by any intel HPC for many years, if ever. Now shoo off back under the rock you crawled out from under.

LOL... is that your "defense" to counter my argument? How pathetic LOL... :lol: :lol: :lol: :lol:

What you're effective saying is, "What 3x more cores? You're definitely lacking knowledge in that area!"


Now excuse me while I go laugh myself to death...

 

[daedalus685]

Re-evaluating your opinion when new data is obtained is absolutely necessary. Changing you position because your "savior" decides it is needed is foolish at the very most.

Fair enough, but the term is all to often applied to us boffins by the masses for doing what we are supposed to do. As if the universe only has one correct undeviating answer. Nothing is black and white, not even black and white.

All a politician has to do is support a war one day, then not support it a decade later to be a flip flop as if one is not allowed to changed thier mind in the light of new facts.. It is different when one changes there mind weekly based on the wim of the people they are trying to impress, but that is not the context the term is used in the vast majority of the time.

/Resume this threads flame war.

 

[JAYDEEJOHN]

This is true, otherwise the ones I tried to impress surely still arent impressed LOL

 

[yannifb]

lol sorry wrong thread!

- i had 2 threads opened and i accidently posted on this one when i meant the other one

 

[JAYDEEJOHN]

Anyone read Charlies announcement?
And ummm, waddya think now?
http://www.semiaccurate.com/2009/12/04/intel-kills-consumer-larrabee-focuses-future-variants/

 

[JAYDEEJOHN]

http://news.cnet.com/8301-13924_3-10409715-64.html
Intel said Friday that its Larrabee graphics processor is delayed and that the chip will initially appear as a software development platform only.

This is a blow to the world's largest chipmaker, which was looking to launch its first discrete (standalone) graphics chip in more than a decade.

"Larrabee silicon and software development are behind where we hoped to be at this point in the project," Intel spokesperson Nick Knupffer said Friday. "As a result, our first Larrabee product will not be launched as a standalone discrete graphics product," he said.

"Rather, it will be used as a software development platform for internal and external use," he added. Intel is not discussing what other versions may appear after the initial software development platform product is launched next year.

Larrabee, a chronically delayed chip, was originally expected to appear in 2008. It was slated to compete with discrete graphics chips from Nvidia and Advanced Micro Devices' ATI graphics unit.

Intel would not give a projected date for the Larrabee software development platform and is only saying "next year."
You can say Im for AMD, or against Intel, I say I was right all along

 

[pausert20]

Hi All,

My friend at Intel confirm that Larrabee 1 is now dead as a product. It will only be used for Software development.

 

[JAYDEEJOHN]

Too many hills to climb, no real advantages for gfx, and the target was moving too fast.
Itll be here as a gpgpu component, and do well as I said, and like as Ive been saying all along, and dont care whether some think or dont think, Im excited to see its usage for gfx combined in a Haswell, as then we'll finally see a cpu going against gpu for gfx, and Intels gfx solution will surely be improved, which is a good thing

 

[JAYDEEJOHN]

As for some, claiming this as a flame thread was a poor choice of thought.
This thread brought up real facts, things known, and was constantly being tried to deviate from it, and it was those that tried, not me.

If this thread didnt belong, then LRB wouldve been here by now, and it wouldnt have been written to begin with, some things to think about

 

[yannifb]

Did you guys know about this?!:

http://www.semiaccurate.com/2009/10/06/nvidia-kills-gtx285-gtx275-gtx260-abandons-mid-and-high-end-market/

Things are looking up for AMD

 

[Mousemonkey]

Did you guys know about this?!:

http://www.semiaccurate.com/2009/10/06/nvidia-kills-gtx285-gtx275-gtx260-abandons-mid-and-high-end-market/

Things are looking up for AMD

That has been known since, ooh October.

Hint: check the article date.

 

[jennyh]

How much humble pie can one eat anyway? So soon after the last very large slice, I mean?

Intel fanboys are getting either very fat or very humble, put it that way.