Teveel bezoek bij Semiaccurate, de website is bereikbaar maar je moet geluk hebben. Eerder op de dag kon ik de pagina's nog laden maar het duurde ongeveer een minuut voordat het geladen was.
Hier de tekst. De analyse ontbreekt helaas, want daar moet je minstens $1000 voor betalen. Mijn gedachte hierover:
- Al het R&D-budget, werknemers en de overige R&D-faciliteiten beschikbaar stellen voor iets wat wel haalbaar is
- Misschien op die manier ook meer capaciteit vrijmaken voor hun 14 nm. node aangezien je nu bepaalde ruimte en machines die voor de 10 nm. node werden gebruikt hiervoor kan inzetten om zodoende het tekort te beperken en de prijsverhoging te beperken. In de USA zijn de prijzen nog redelijk maar in Nederland moet je goed gek zijn om nu nog een 8400/8700k/9900k te kopen, of zwemmen in het geld en er niets om geven.
SemiAccurate has learned that Intel just pulled the plug on their struggling 10nm process. Before you jump to conclusions, we think this is both the right thing to do and a good thing for the company.
Quotable Quotes:
Before we explain why knifing 10nm is good for Intel, lets start with a few quotes:
“We hear that internally Intel is quite worried about making the launch of Cannon even though it is still about a year away. This is no ordinary early silicon issue, it is a serious and unexpected problem. Coffee lake being added at the last minute between Kaby and Cannon should shed some light on the depths of Intel’s 10nm woes, things are a mess. More when we get it, but for now, not a merry Christmas for those singing from hymnals D1C and D1D.” -SemiAccurate Dec 22, 2016
“Doing the same job on a 14nm process results in cheaper chips that perform better. Intel has to put out 10nm parts because they are expected to by Wall Street. Don’t for a second think the top brass isn’t aware of this and actively making product decisions because of that pressure.” -SemiAccurate Dec 28, 2016
“All of these things back up the three financial graphs Intel presented multiple times at Manufacturing day. Although costs are going up, with hyperscaling the net result is a great benefit for Intel, not the expected losses from node transition slips. This is all quite true if yields are the same or better for newer nodes. Intel normalized the graphs for some reason so you can’t tell what the actual yields are, nor their effect on the financial outlook. If they are far worse on newer nodes as some have intoned, the rosy financial conclusions may not smell quite so sweet. If yields on 10nm are better than 32, all is well. Feel free to draw your own conclusions based on yields for current processes.” -SemiAccurate April 4, 2017
“This says that Intel has roughly zero confidence in their current roadmap for 10nm, another insertion of a 14nm product undoubtedly means another delay to the above schedule. What does it mean? It means don’t be surprised when the Q1/2019 conference call delays things yet again.” -SemiAccurate May 21, 2018
“The 10nm Cannon Lake parts aren’t real and never will be viable, financially or technically speaking.” -SemiAccurate May 29, 2018
“Told ya” – Authors Note, Today
Don’t Mourn:
For several years now SemiAccurate has been saying the the 10nm process as proposed by Intel would never be financially viable. Now we are hearing from trusted moles that the process is indeed dead and that is a good thing for Intel, if they had continued along their current path the disaster would have been untenable. Our moles are saying the deed has finally been done.
This isn’t to say the road to this point has been easy or straightforward, and the road ahead is even less solid. Intel has continually moved the public bar on 10nm back, incrementally, while singing a different song internally. In their Q1/2018 earnings call they moved the timetables and spun it in a curious way but were telling partners a different story.
Don’t ask questions, just enjoy the message
Nothing however tops the masterful “Hyperscaling” stunt where Intel brought in press and analysts to a ‘manufacturing day’ in early 2017 to explain how the crippling slide of 10nm was not actually a slide, it was a good thing and not a delay at all. SemiAccurate laughed and stopped just short of calling Intel liars.
The company redefined terms well past the breaking point to show that scaling was ‘on track’ even if node cadence was ‘intentionally’ longer. As you can see from the above graph, all was good publicly, internally SemiAccurate was hearing a very different story. (Note: Intel was on track to miss that graph by 1+ year and sliding before 10nm was killed.)
In their own words…..
The knifing of 10nm shows that Intel is finally willing to do the right things for the right reasons even if it costs them some short term pain, it is the first adult decision we have seen from the company in several years. Let us walk through the reasons why it is a good thing, from cost to timetables to competitiveness to management changes to potential product roadmaps. It is not a clean, easy or pithy story to pull a sound bite from but it is interesting.
Geef het nog een paar weken tot maanden.
“Doing the same job on a 14nm process results in cheaper chips that perform better. Intel has to put out 10nm parts because they are expected to by Wall Street. Don’t for a second think the top brass isn’t aware of this and actively making product decisions because of that pressure.”
Dit citaat vind ik erg verhelderend, het past bij het beeld wat ik al een tijdje had. Waarom de hele tijd ontkennen en steeds maar weer in stapjes erkennen dat die 10 nm. node volledig flopte? Hoe vaak hebben ze nu de officiële plannen aangepast na ontkenning? 2 of 3 keer? Hoe vaak hebben ze de nieuwe generatie (Canon Lake of zo) uitgesteld na het te ontkennen? Ze zijn bang voor een stock-'crash' als ze in 1 keer het nieuws melden. Begrijpelijk maar kan je op een gegeven moment niet beter in 1 keer het slechte nieuws brengen dan elke paar maanden nieuw slecht nieuws? Dat laatste geeft geen vertrouwen bij investeerders en ze hebben het al een paar jaar gedaan, met name het laatste jaar was het het ene slechte nieuws na het andere. Nu moet Intel het in ieder geval even ontkennen omdat ze het niet kunnen hebben dat het bekend wordt voordat zij zelf officiële mededelingen hebben gedaan. Ze willen natuurlijk graag zelf bepalen hoe het nieuws wordt gebracht (de nodige 'context' en spinning) als het waar is. Mijn vermoeden: voor het einde van het jaar heeft Intel officiëel medegedeeld dat ze die node (hebben) laten vallen. In het beste geval - als je erg optimistisch bent - is dit een signaal dat het nog slechter gaat met die 10 nm. node dan dat al werd erkend want we kunnen er zeker van zijn dat die Charlie niet uit zijn nek kletst gezien zijn klanten.
@
Deathchant
Nvidia creëert geen schaarste maar komt weg met hoge prijzen doordat AMD in de high-end van de graphics
voor gaming geen competitie biedt en al 2,5 jaar Polaris heeft. Dat komt overigens wel weer goed bij de eerste generatie na Navi, ze kunnen nu weer meer investeren in de GPU dankzij het succes van Zen...
Samsung, Hynix en Micron creëren met opzet een schaarste, zij komen ermee weg omdat ze samenwerken. Intel kan zich dat niet volgens mij niet veroorloven want de Ryzen-CPU's zijn goed genoeg dat met dat prijsverschil voldoende mensen die anders een Intel-CPU zouden kopen nu een AMD-CPU kopen. Bovendien komt er binnenkort Zen2 uit die waarschijnlijk qua performance zeer competitief gaat zijn met Coffee Lake, als het niet beter is, en dit voor een veel lagere prijs. Nu kan Intel alsnog die prijs verlagen maar we weten dat Intel niet al te ver gaat met het verlagen van de prijzen, zeker niet bij reeds bestaande modellen. Intel kan het zich niet permitteren, hoe gek dat ook klinkt aangezien Intel een reus is.
Ik geloof dus niet jouw complottheorie. Intel zou er zeker toe in staat zijn maar ik geloof niet dat het Intel baat om dit nu te doen.
[
Voor 45% gewijzigd door
Verwijderd op 22-10-2018 20:18
]