You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
We need to exterminate those outrageous real estate, power hungry sloths (Speedsters if you are a CPU guy). They are taking over the data centers and will soon be in your neighborhood. We don't need these insane EUV machines that emit high 13.5nm radiation levels. Taxpayer funds should not used for these small company killers that are being used for the corporate elites who can afford these $50M+ NREs.
CHIP funds should be diverted immediately towards fast turnaround shuttles (MPWs), cheaper interposers (500nm process) with 80um'ish pitch HBMs (less memory, understood), and funds should be used by $10M DUV machines with simpler finfets (Japan is smarter than us).
I expect you miserable GAA/nanosheet snobs, with your indesipherable acronyms and process numbers that make no sense to attack. I am prepared to sacrifice myself on behalf of my neanderthal brothers (JMS, etc).
FPGAs are indispensable for debugging ASIC designs. FPGAs are also an important component for low-volume applications (meaning an ASIC can't be justified) which can't be performed in software meeting the required performance level with reasonable power consumption. FPGAs are also convenient for certain networking applications where protocol processing of several different protocols can be accomplished with the same hardware. And these are just few applications off the top of my head which can be done best with FPGAs.
Yea yea yea... understood. Truth be told, I do believe in small embedded FPGAs within the ASIC
Seriously, I wish one day there can be more MPWs with a faster throughput and less ambitious interposers. 55um pitch and below with so many pads sounds awfully hard. I don't think Amkor goes below a 90u pitch.
We actually started making an array of programmable LUT5 CLBs with switchable busses several months ago. If you can't beat them... I just avoid that 4 letter word, which I consider an F bomb.
Research suggests they fashioned tools, buried their dead, maybe cared for the sick and even conversed. But why, if they were so smart, did they disappear?
Those of us with Eurasian descent have a percent or two Neanderthal in us. There is more N DNA around now than at any time in the past, so the selfish genes won.
Seriously, I wish one day there can be more MPWs with a faster throughput and less ambitious interposers. 55um pitch and below with so many pads sounds awfully hard. I don't think Amkor goes below a 90u pitch.
Working on it. Not compromising on the resolution, either. But the advanced node IP and EDA costs will likely still make you cry, I can't fix everything.
Are you working flip chipping and connecting to interposers, or producing TSV's, or intermodulating waves/particles of energy, or putting multiple customers on the same maskset, or something else? I am intrigued.
Apparently, we do the stuff that make others cry, but I didn't think so. We just mostly make rectangles and different widths and spacings, and just in abstract form. You guys tell us what you can do and we believe you.
As far as I know throughput of maskless ebeam is in hours/days per wafer rather than wafers per hour; especially for resolutions for which EUV is used.
