https://twitter.com/tmfchipfool/status/892459586199879681
14 replies (most recent on top)
Stratix 10 is not in high volume production so most probably they will not be selling it.
HVM is planned before end of this year with availability in firs half of 2018.
This is kind of bad, because that means, that they will have to directly compete against 7nm Xilinx devices.
Last public Stratix-10 sightings were last October - 10 months ago.
One of many links: https://www.extremetech.com/computing/237338-intel-launches-stratix-10-altera-fpga-combined-with-arm-cpu-14nm-manufacturing
Right - where, exactly, IS Stratix-10? Even Intel's own site doesn't have it available to buy yet.
https://www.altera.com/buy.html
@dib Intel simply has inferior process.
They have better gate-metal-pitch dimension (well actually today with 14nm++ it is close to what other calls 20nm while others improved their density with each subgenerations) but this is at a price of hard limitations in design rules. It makes almost impossible to effectively design any complicated structures.
This is why AMD achieves almost 2x more transistors per mm^2 than Intel.
Altera moved to Intel foundry and they had such a pain with this, so they wanted to go back for TSMC at 10nm. And results? Original plan was to release Stratix 10 in 2013 (earlier than Intel's own processors because they believed that it will be simple to manufacture and good enough in margins to cover yields loses). Reality is that it is still not ready and latest informations are that they can go into HVM before end of this year (if last respin was succesfull).
ICF has zero customers despite being "leading edge".
After 7 years of Infineon within Intel, there is still no modem from TMG.
After years and years of delay Altera got something on 14nm.
Why does TMG s--- at attracting and retaining customers? Maybe the process isn't such hot sh1t after all.
TSMC is bigger than Intel now.
@jqd LOL - I wouldn't even be surprised if you were the real Ashraf.
The crazy thing is: Tomorrow, a startup could claim to have 1nm processor technology, and Ashraf would believe it and claim that they are ahead of Intel, because he has no clue what it means, and he doesn't realize that this kind of term is not standardized in any way, so any company can claim whatever they want.
And there are people who actually buy or sell stocks based on that guy's recommendations. Despite his terrible track record over the years.
@OzDjfvq-fki So how is 10nm yields, do you even know? If you knew how it was and the improvements in yields compared to 1264 or 1268 or even 1270 or1272, you'd shut your a$$ and figure out how to get the BoD to get rid of the clown and the idiots running and ruining TMG
@fki - Do YOU even know what 10nm means, azzhole?
Wake me up when TMG has a credible story as to why they are late and nobody wants to use ICF.
So people trust some guy on Twitter who has no clue what "10nm" even means to give them an assessment of Intel's future? The same guy who three years ago said that AMD was just about to go bankrupt?
Canon and Ice baby... oops we get Kaby and Coffee on 14nm+ not just 14 but the plus baby!
I guess 10nm dang healthy ain't it!
😂😂😂😂😂😂😂😂😂😂😂😂😂😂
HR please collect your $0.0125 and return to circuit.
“yields are continuing to improve pretty much right in line with [traditional] ramps rates,” said Krzanich. “It’s a new technology so there are always some problems to solve, but we’re pretty comfortable with where we are at now,” he added.
There is no mess, right where they wanted it. Ask anyone in Hillsboro or all them seed engineers, right where they expected it and predicted based on plans five years ago, 😂😂😂😂😂😂