Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Intel SNB 17" laptop with backlit kbd? Anyone have a source with Intel IGFX? Any thing I've found so far has AMD/NVidia discrete crap if backlit keyboard.
tnx
gb
Chuckles positively gushing over EP on S|A. See front page articles. Two so far...
S|A now running on EP box.
Fanboys, aka deer in headlights, seem strangely quiet... I'm sure they'll come up with snappy rejoinders and break out a fresh set of pompoms eventually.
gb
Given the amount of activity in this segment with the big cloud providers I suspect Intel has their own thing going on which likely made Seamicro receptive to a buyout offer from AMD.
Intel has made no secret of their direct involvement with Google etc on specialized servers.
gb
Amd charge:
http://finance.yahoo.com/news/amd-amends-wafer-supply-agreement-013000548.html
Plus Seamicro purchase
I wonder where this leaves thei balance sheet?
gb
You've got to wonder why with a free choice of processors SealM chose Intel over AMD. Perhaps they were best suited and were what the market for their products wanted.
This is going to be fun to watch.
gb
No idea. But then why did amd buy them. If it was to get some jump on intel it would seem that intels work with sm in the past along with the patent xl would obviate any advantage.
Btw, Chuckles claims sm "fabric" is nothing more than a clever use of pcie.
gb
With the amd/intel patent cross license in place wouldnt intel have access to sea micro patents if any?
gb
they're clearly betting on bobcat/shared nothing/low power strategy that i suggested they might several months ago when rr announced the layoffs and the forthcoming new strategy.
imo they've given up on chasing intel's architecture and process tech lead and doing what they can with what they have, i.e. designing around slower process tech and putting lipstick on that pig. the shade-of-the-month is seamicro fabric.
i'm sure they'll continue to sell opterons and pretend that line has a future but the future will come when opteron can replace bobcat at the lowend on a basic volume foundry process rather than some boutique soi crap.
i wonder how the seamicro stuff would stand up to a mic chip from intel?
gb
I wonder how they'll pay for this? It's not like AMD is awash in cash.
gb
and if i understand the capsule summaries of tsmc conf calls all four together amount to squat in their financials.
gb
While I wouldn't want Intel to lose money on the phone chips I think something well below traditional margins may be just fine since the amount of wafer starts they'll consume should be minimal.
I think if Intel becomes competitive/successful then volume learning and process shrinks that well outpace the foundries will allow them to grow the margins that when blended with traditional PC chips should do just fine.
First order of business is win the business to squash the squawks about Intel not being in the game.
gb
I have to agree.
But look on the "bright" side. If Intel is having this much trouble think what the other guys are going to go through.
gb
I think arguement is that eliminating the need/benefit for all but the highend 250w sku will kill the volume needed to make that 250w monster profitable. The Professional/Business workstation market chips are nothing but repackaged highend gamer SKUs as well. So Haswell is likely to kneecap the whole discrete market economics by taking over not only the lowend but all of the midrange as well.
I'm sure there will continue to be the professional cards from someone but their price is likely to rise and the frequency of updates like to decline as the market dynamics change.
gb
If that's true then perhaps it allows for an architectural change in the gfx engine which may have presumed BW limitations in the past. Of course committing to that means committing to having that memory around on all the SKUs. It will be interesting to see when it's disclosed. They must have silicon already so perhaps leaks will begin. Sept IDF no doubt will have a lot of details.
gb
Chuckles is actually claiming that Haswell will all but eliminate the market for discrete gfx.
Rather unusual from him.
gb
I was surprised to see this on 32nm. I wonder if 22nm with the FD reached by the trigate structures might show even more promise due to less substrate coupling?
gb
Actually the first slides I saw on LRB showed the bus as CSI. I wouldn't bet on pcie being the interface.
gb
I liked Elmer's response:
"When you're out of bullets, throw the gun"
AMD's out of bullets...
gb
in a word: outdated.
gb
actually 5.26%. 267873158 shares. it's the $7b+ number. they're serious investors.
gb
Yup. 5% position in intel by blackrock.
gb
Dunno. Just got a 13D notice via email.
gb
Blackrock ownership of Intel now >5%.
gb
I strongly suspect transactional memory and AVX are on the Intel SW priority list above anything resembling HSA.
Keep in mind that AMD has a lot of time before they can demonstrate a working implementation of HSA. Even their own roadmap showed 2014 at the earliest. And we know how roadmaps go...
Given the competitive situation with Intel and even ARM I expect them to continue to lose unit and revenue market share over those next two years. Meanwhile there will be Haswell and its shrink to 14nm by then.
gb
i know what they're trying to do. i don't think it will be successful for reasons of relevance, volume and competition for resources by intel initiatives that address performance. unlike the athlon days intel is not complacent.
gb
So if you're a developer with limited resources do you commit those to HSA and volumes avaialable there or to the 50x volumes of Haswell and beyond with AVX and transactional memory?
Not talking national labs as they'll do one offs of a lot of things, but real commercial developers.
gb
WoA app and driver distribution only via MSFT app store.
Win8/x86 otoh is have it your way as before.
http://tech.slashdot.org/story/12/02/10/1238247/microsoft-details-windows-8-for-arm
gb
Win8 at MWC. I wonder what the demo vehicles will be?
http://mobile.bloomberg.com/news/2012-02-08/microsoft-to-unveil-windows-8-test-version-on-feb-29-in-spain.html?cmpid=yhoo
gb
I can imagine two scenarios that might have occurred.
"I can't sell this crap." "You're fired"
"This stuff is so good it sells itself." "What do I need you for? You're fired."
gb
Read's style at the presentations he's made make me think he's got a pretty big ego. Since he's firing anyone who is high paid and experienced its all going to land on him when it fails.
His style of speaking reminds me of the lead on "Burn Notice."
gb
more of read's emphasis on "execution." he is from tx afterall...
gb
Socket compatibility may be a constraint driven by the odms who won't redesign for amds typically low volumes.
gb
i'm beginning to think that intel has a much better handle on workloads and simulation ability which combined with a longstanding cache advantage gives them a significant and insurmountable lead in x86.
amd appears to have realized this and is going for their 'hsa' hail mary.
gb
i managed to earn a badge of honor for stating the obvious over there.
gb
i liked the financial analysis combined with the techie. put it in perspective.
gb
Good summary. Note authors bio.
http://www.forbes.com/sites/patrickmoorhead/2012/02/05/amds-risky-ultrathin-notebook-strategy/?partner=yahootix
gb
from what i've seen over the years such details are often omitted/redacted and when disclosed later by accident or legal action appear to be somewhat embarrassing to one or both of the parties in some way which was likely the reason for the secrecy in the first place.
gb
Yup. No phone chips from AMD.
gb