2020 Retrospective. Innovation in VerificationPaul Cunningham (GM, Verification at Cadence), Jim Hogan…Read More
The Spartan flow for custom silicon: when losing is NOT an option.Every so often a custom silicon socket comes…Read More
Automotive SoCs Need Reset Domain Crossing ChecksWhen the number of clock domain crossings (CDCs)…Read More
Achronix Goes Public | The Only Independent Leading Edge FPGA Silicon and IP ProviderAchronix, a high end FPGA provider is going…Read More
Siemens EDA is Applying Machine Learning to Back-End Wafer Processing SimulationThere’s a lot to unpack in the title…Read More
This last feature is worth special mention. Matt indicated, “HFSS supports virtually disturbing the mesh for variation analysis. Designers can identify a set of parameters in the system model, and readily see how the electromagnetic analysis results change with manufacturing variations, for a small overhead in simulation time, far more efficient than running full simulations on different parameter samples and unique meshes with small dimensional changes.” This feature provides great insights into where designers could focus on cost versus manufacturing tolerance tradeoffs.
Ansys has prepared an informative demo of how designers can quickly visualize the response to parameter sensitivity – link.
Algorithmic Enhancements for Mesh Generation
Matt identified three key HFSS enhancements of late related to mesh generation.
Adaptive Meshing
The introductory section above described the importance of the 3D mesh to the resulting accuracy. An initial mesh is solved for the fields – a calculation of the electric field gradient is indicative of where local mesh refinements are appropriate. (The basis functions for representing the local fields could also be updated.) A new mesh is solved, and the process iterates until successive passes very less than the convergence criteria.
HFSS recently extended this capability to adapt the mesh each iteration using multiple frequency solutions, over a user-specified range, to enhance the results accuracy when a broad range of spectral energy is present.
3D Components
Traditionally, it has been difficult for designers to build a comprehensive (“end-to-end”) model of even a single long-reach, high-speed signaling interface. The PCB trace S-parameter model generation from the stack-up was relatively straightforward, but obtaining a model for connectors and cables from the vendor was typically difficult.
Ansys realized that enabling link simulation, and ultimately, system-level analysis required a novel method, and developed the “3D Components” methodology:
HFSS Mesh Fusion
Of the steps in the electromagnetic analysis of an electronic system:
Summary
The traditional method for electromagnetic analysis in electronic systems focused on PCB designs and high-speed signaling. The board stack-up and materials properties were defined, and the signal traces were simulated. S-parameter response models for signal loss and (near-end/far-end) crosstalk from adjacent traces were generated, and incorporated into subsequent circuit simulations to measure the overall transmit/receive signal fidelity. However, the complexity of current electronic systems necessitates a more comprehensive approach to electromagnetic coupling simulation, as compared to concatenating individual S-parameter models. Systems will be integrating a broad range of signal frequencies from audio to mmWave, with advanced packaging present in aggressive volume enclosures.
The HFSS team at Ansys has focused on numerous technical advances – both computationally and in the critical area of mesh generation – to enable this analysis. Designers can now evaluate and optimize models of a scope that was once unachievable, with manageable computational resources.
For more info on these Ansys HFSS features, please follow these links (and don’t forget to sign up for the Mesh Fusion webinar):
Broadband Adaptive Meshing – link
Ansys cloud HPC resources – link
3D Components – link1, link2
Mesh Fusion webinar – link
-chipguy
[post_title] => System-level Electromagnetic Coupling Analysis is now possible, and necessary
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => open
[post_password] =>
[post_name] => system-level-electromagnetic-coupling-analysis-is-now-possible-and-necessary
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-26 10:55:22
[post_modified_gmt] => 2021-01-26 18:55:22
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?p=295283
[menu_order] => 0
[post_type] => post
[post_mime_type] =>
[comment_count] => 0
[filter] => raw
)
[1] => WP_Post Object
(
[ID] => 295257
[post_author] => 13
[post_date] => 2021-01-26 06:00:37
[post_date_gmt] => 2021-01-26 14:00:37
[post_content] => GDS and LEF/DEF each came about to support data exchange in different types of design flows, custom layout and place & route respectively. GDS (or stream format) was first created in the late 1970s to support the first generation of custom IC layout tools, such as Calma’s GDSII system. Of course, the GDS format has been updated over the years to support the capabilities of newer IC design tools. LEF/DEF (library exchange format & design exchange format) came along later to support the larger but somewhat simpler data found in P&R flows. Yet there has always been a dichotomy of support for these formats among various design tools. DFM has tended to fit in between custom layout and P&R, with the general assumption that reading LEF/DEF and writing GDS made sense because at or near tape out the flow would move to GDS. However, this has always been problematic because P&R systems have grown more powerful and designers want to see the final layout in them as opposed to custom tools at the end of the flow.
DFM (design for manufacturing) tools, such as Siemens EDA’s Calibre YieldEnhancer, provided a path to DEF so DFM changes could be brought into P&R tools. But the flow was cumbersome and often required steps not in foundry approved rule decks. Siemens EDA now has a white paper that discusses the addition of a fully bidirectional DEF flow for their DFM tools. The paper, titled “Optimizing the Integration of DFM and P&R” by Armen Asatryan and James Paris describes the limitations of the previous extra conversion step to get to DEF and then talks about their new direct write to DEF.
[caption id="attachment_295261" align="aligncenter" width="234"]
Siemens EDA DFM integration with P&R[/caption]
The previous DEF conversion process involved a utility called fdiBA which ran as a separate operation. It involved extra steps and needed inputs for attaching net names to geometry. It had limitations on geometry shapes and did not deal well with multiple orientations of vias. With larger design sizes it tended to face capacity limits and long runtimes.
The new direct write DEF feature in Calibre can create a DEF format database without the need for an intermediate GDS (or OASIS) database conversion. To take advantage of this, only one rule deck option needs to be added – “DFM RDB DEF”. Along with the obvious simplicity of the flow, there are a number of added benefits. Direct write to DEF supports all-angle metal shapes, including metal extension end-caps and via caps. It recognizes multiple orientations of via instances. To reduce size, it performs via array detection. Direct write DEF also provides automated via repair.
One area that will make a big difference is the treatment of fill. Compressed fill is handled properly with direct write to DEF so Calibre YieldEnhancer SmartFill can efficiently transfer fill to P&R tools.
Having been in this industry as long as I have, I am often amazed at the longevity of the GDS format. I actually knew Sheila Brady, the woman who created the GDS format. Even back in the 80’s she would comment on how it had taken on a life of its own and was initially really just intended as a quick and reliable back up for GDSII databases. Yet here we are with complicated flows that rely on legacy formats such as this. Making rational the movement of P&R data between essential tools in the flow with DEF is vital as design size and complexity increase. The full white paper is available for download here.
[post_title] => Calibre DFM Adds Bidirectional DEF Integration
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => open
[post_password] =>
[post_name] => calibre-dfm-adds-bidirectional-def-integration
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-26 08:48:33
[post_modified_gmt] => 2021-01-26 16:48:33
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?p=295257
[menu_order] => 0
[post_type] => post
[post_mime_type] =>
[comment_count] => 0
[filter] => raw
)
[2] => WP_Post Object
(
[ID] => 294860
[post_author] => 11830
[post_date] => 2021-01-25 10:00:12
[post_date_gmt] => 2021-01-25 18:00:12
[post_content] => 
Last August I detailed a webinar about HCL Compass, a tool that provides low-code/no-code change management capability for enterprise scaling, process customization and control to accelerate project delivery and increase developer productivity. There is a lot of activity these days to migrate various enterprise applications to the cloud for better scalability, access and performance. I know from first-hand experience that, when it comes to cloud migration, the devil is in the details. So, a comprehensive guide that explains how to do this is quite valuable. If you’re considering such a move, read on and you’ll discover that HCL provides an on-ramp to the Amazon Elastic Compute Cloud for HCL Compass.
First, a bit about Amazon Elastic Compute Cloud (EC2). Provided by Amazon Web Services (AWS), EC2 is a web service that provides resizable computing capacity—literally, servers in Amazon's data centers—that you use to build and host your software systems. Features of EC2 include:
EC2 has quite a global footprint that includes nearly 400 instances for virtually every business need. There are 24 regions and 77 availability zones globally. EC2 is the choice of Intel, AMD, and Arm-based processors and they are the only cloud provider that supports macOS. Also, both EC2 instances and HCL Compass web server support Windows and Linux OS versions. So, a complete HCL Compass installation using EC2 instances is possible in AWS.
Virtual resources such as those provided by EC2 remove the capital expense of procuring and maintaining equipment as well as the expense of maintaining an on-premises data center, for example, cooling, physical security, janitorial services, etc. In AWS Cloud, AWS provides EC2 instances as the servers in Amazon's data centers, to build and host HCL Compass.
The whitepaper provided by HCL is quite comprehensive and covers general guidance for cloud installation and migration from on-premises ClearQuest or HCL Compass to AWS HCL Compass. Note HCL Compass is the next generation of IBM Rational ClearQuest. The intended audience for the whitepaper is administrators of HCL Compass and administrators of ClearQuest intending to migrate and deploy to the cloud. The document includes strategy for a fresh installation of HCL Compass and migration from ClearQuest to HCL Compass in AWS. HCL Compass 2.0.0 supported platforms include 64-bit Windows and Linux as follows:
Database support includes:
Browser support includes:
All other required software is detailed as well. The document covers a lot of other topics, including:
Pretty much everything you’ll need to know. Additional documentation on virtualization is also provided. You can get your copy of the white paper, entitled HCL Compass in AWS here. You can learn more about HCL Compass here. If you’re considering a move to the cloud, you’ll be happy to know that HCL provides an on-ramp to the Amazon Elastic Compute Cloud for HCL Compass.
[post_title] => HCL Provides an On-Ramp to the Amazon Elastic Compute Cloud for HCL Compass [post_excerpt] => [post_status] => publish [comment_status] => open [ping_status] => open [post_password] => [post_name] => hcl-provides-an-on-ramp-to-the-amazon-elastic-compute-cloud-for-hcl-compass [to_ping] => [pinged] => [post_modified] => 2021-01-26 08:52:14 [post_modified_gmt] => 2021-01-26 16:52:14 [post_content_filtered] => [post_parent] => 0 [guid] => https://semiwiki.com/?p=294860 [menu_order] => 0 [post_type] => post [post_mime_type] => [comment_count] => 0 [filter] => raw ) [3] => WP_Post Object ( [ID] => 295042 [post_author] => 14 [post_date] => 2021-01-25 06:00:59 [post_date_gmt] => 2021-01-25 14:00:59 [post_content] => -Intel good results had a little extra help to be great -New CEO commits to remaining an IDM versus fabless -Claims of strong progress on 7NM fuel optimism inside -Outsourcing to TSMC will not go away but will increase
A good quarter but with some silicon enhancements from ICAP
Intel reported Revenues of $20B and EPS of $1.52, which were well above consensus and expectation. Expectations were for revenues of $17.5B and EPS of $1.10.
But if you look below the surface there was a bit of "artificial enhancement" from ICAP (Intel Capital). $1.692B of the $7.488B pre tax income (or almost 23%) was from ICAP and not operations. This compares to $212M in ICAP benefit for the prior 9 months total and versus $617M in ICAP benefit in the year ago quarter. Backing out this unusually large one time gain would result in a good beat but not the large beat that was surprisingly printed prior to market close Thursday.
Guidance was also a beat with Q1 guide of $17.5B (down 12% YoY) and EPS of $1.10 (down 24% YoY) versus analyst expectation of $16.4B and EPS of $0.87
More importantly Intel staying "Inside"
Intel's new CEO, Pat Gelsinger, was on the call, along with the chairman of the board. Pat made it clear that he thought that 7NM was on the road to recovery and that Intel will continue to produce the majority of its chips inside as an IDM.
This is perfectly in line with what we had predicted in our last two notes on Intel. We obviously also have a strong bias to see Intel remain a true IDM and "own its own fabs". We had also predicted that the final decisions would not be made until Pat was there for a while, which they also said on the call as they did not give full year guidance.
It is also clear that, as we had suggested, Intel has no choice but to continue to outsource to TSMC and in fact will increase the outsourcing as even if they fix 7NM they still will not be able to ramp capacity fast enough.
Pat even referred to the same view we have about Intel being somewhat of a national treasure and key to the US's technology infrastructure.
Get ready for numbers to look ugly and get sandbagged
Although Intel did not give full years guidance, we would hold onto our seats as we have suggested that the dual costs of increased outsourcing to TSMC added to increased spend on Internal efforts to regain Moore's Law pace will be high and pressure margins in the short term which we would view as at least the next two years and maybe more.
If Pat is smart he will sandbag strongly and lower expectations to numbers that can be beaten easily and overestimate the costs of fixing 7NM and ramping capacity while still paying TSMC to make chips. He doesn't want to put out numbers he will miss in his first couple of quarters on the job.
We would hope that this would include a large jump in Capex and R&D to give the engineers and manufacturing the latitude they need.
It may not be the $28B or $30B of TSMC and Samsung but it should be generous.
So what was the problem with 7NM?
While Intel did not explicitly say what was wrong, they did say that the fix required them to re-design a significant number of steps in the process flow to fix the problem. This clearly means that it was not one step or tool or material or even a design.
Intel obviously drove down a dead end from which there was no escape. It meant that they had to back up quite a bit and start over which obviously accounted for the extra time.
Backing that far up means literally going back to the drawing board (EDA tools) and re laying out all the designs and layers. It means a new set of masks, new and additional tools. It can get ugly and out of control quickly.
As with plane crashes its never just one thing that went wrong nor has to be fixed.
Bringing people out of retirement
We had also mentioned that we thought that Intel had lost some of its brain trust through RIFs and retirements etc; We are happy to see that Pat is already bringing back Intel's prior stars to help bail them out. Maybe they should buy out Jim Keller's AI chip start up to get him back to Intel. Intel did mention AI a significant number of times on the call, maybe there's a coded message there.
The Stocks
Intel jumped before the close on Thursday as numbers were somehow released while the market was still open. Intel was up 6.5% during market hours largely due to the leak and down 1.5% in after hours.
We don't know how many analysts or investors dug deeper into the numbers to find that they appeared better due to the Billion dollar plus ICAP benefit.
As we suggested in our last note we think there is a potential opportunity to make some money on a near term pop in Intel's share price due to the double whammy of new CEO and unusual beat. We would likely want to be out of the stock prior to Pat's resetting of numbers a projections for the full year which will likely show higher expenses.
We view this as a neutral to a positive for AMD as the underlying strength of demand is a good thing and not being n the same TSMC lifeboat with Intel gives AMD some room.
Some investors took it as a negative for TSMC which we disagree with. TSMC has more than enough demand to deal with and heaping even more demand would likely strain them in the short run and piss off even more customers on the lower end who won't get serviced or be pushed out by the big boys. TSMC has way more than enough potential business and profits and doesn't need to drown in demand.
Equipment companies should see this as a positive, which we already suggested, as Intel will likely not only continue to spend but likely spend more to catch up and ramp up. So it should be positive across the board. Perhaps ASML may getter a bigger slice of the benefit as Intel will have to ramp EUV much as TSMC already has and Samsung is trying to do. Perhaps it would make up for the recent slowing of a couple of EUV customers.
Long Live "Intel Inside!"
[post_title] => New Intel CEO Commits to Remaining an IDM
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => open
[post_password] =>
[post_name] => 295042
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-25 06:05:08
[post_modified_gmt] => 2021-01-25 14:05:08
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?p=295042
[menu_order] => 0
[post_type] => post
[post_mime_type] =>
[comment_count] => 1
[filter] => raw
)
[4] => WP_Post Object
(
[ID] => 294987
[post_author] => 19
[post_date] => 2021-01-24 10:00:55
[post_date_gmt] => 2021-01-24 18:00:55
[post_content] => 
A study published in ScienceAdvances has shown computer simulations of the movement of virus-laden airborne droplets in cars. The objective of the study appears to be to assess the degree of driver-to-passenger and passenger-to-driver exposure given different open-or-closed window configurations in the context of modeled airflow around and within a four-windowed car.
"Airflows inside passenger cars and implications for airborne disease transmission" - ScienceAdvances -
The study does not recommend any particular window open or closed configuration. This would be difficult as a configuration that might be advantageous to the passenger may not be optimal for the driver. The ideal configuration, not surprisingly, is to have all windows open. In fact, the more windows that are open, the lower the exposure of either driver or passenger.
The study is important for shining a light on the question of driver-to-passenger (and vice versa) exposure in the context of the unique air pressure gradients surrounding the car and the counter clockwise movement of air within the car and the prevailing front to back movement of air when air conditioning or heating is in use. The study also looks at the number of air changes per hour (ACH) per different open-closed window configurations.
What the study does NOT consider is the impact of introducing an in-vehicle partition. The study also does not take into account multiple passengers. What the study really highlights, though, is the reality that no configuration is a guarantee of safety for either passenger or driver. It also fails to determine whether the driver or the passenger is at greater risk - although the driver is presumably at greater risk given the number of potential exposure opportunities during a working day.
The study also highlights the lack of research into COVID-19 infections among taxi, limousine, and ride hailing drivers - and frequent users/passengers. In fact, a broad study of transmission on busses, subways, and taxis - including driver/operators and passengers is overdue.
It would seem that the only real takeaway from this study is that there is no safe window open-or-closed configuration when sharing a vehicle with a stranger or family member. The lack of any testing of partition-equipped cars is unfortunate.
Too many ride hailing companies have adopted Center for Disease Control and Prevention guidelines such as mask wearing and sanitation as sufficient. The CDC guidelines for taxi, limousine, and ride-hail operators can be found here:
https://www.cdc.gov/coronavirus/2019-ncov/community/organizations/rideshare-drivers-for-hire.html
The CDC guidelines have a single reference to vehicle partitions: "If you work for a company that offers a large fleet of vehicles, ask company management for a car/taxi (when applicable) with a partition between driver and passengers, if available." The agency makes no public policy recommendation - such as requiring partitions.
Some ride hailing companies – like Alto, Bolt, Didi Xuching – get it. These operators have required and added partitions to their vehicles.
If partitions are deemed effective at restaurants, grocery stores, gyms, and in schools and on busses, it stands to reason they will be effective in taxis and ridehail vehicles. Somehow this message has not penetrated senior management levels at Uber – now obsessed with shedding assets and shoring up profitability. Lyft makes partitions available to its drivers, but does not require them.
The study discussed here was conducted by Dr. Varghese Mathai, a physicist at the University of Masssachusetts, Amhert, and three colleagues at Brown University — Asimanshu Das, Jeffrey Bailey and Kenneth Breuer. The study has spurred consumer advocates to recommend that the windows opposite the driver and the passenger each be rolled down for the optimal and safest airflow in the vehicle.
Still, it seems that the only really safe means of offering shared vehicular transportation is with a partition installed in the car. Maybe a follow-up study can take a look - and maybe next time the scientists will keep public policy in mind. Passengers and drivers shouldn't have to calculate their odds of getting a COVID-19 infection based on how many windows are rolled up or down in their vehicle.
[post_title] => Playing Pandemic Roulette in Cars
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => open
[post_password] =>
[post_name] => playing-pandemic-roulette-in-cars
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-24 19:47:47
[post_modified_gmt] => 2021-01-25 03:47:47
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?p=294987
[menu_order] => 0
[post_type] => post
[post_mime_type] =>
[comment_count] => 0
[filter] => raw
)
[5] => WP_Post Object
(
[ID] => 294966
[post_author] => 14
[post_date] => 2021-01-24 06:00:52
[post_date_gmt] => 2021-01-24 14:00:52
[post_content] => - ASML has good quarter driven by DUV & Logic (@72%)
- SMIC & other major customer slow EUV plans
- Logic (read that as TSMC) remains key demand led driver
- We are happy memory remains muted given cyclical potential
A very solid quarter with a continued road to growth
The quarter came in at Euro4,254B and net income of Euro1,351B on gross margin of 52%. The company reported year total sales of Euro14B and Euro3.6B in income. EUV fell from last quarters 66% of sales to 36% of sales. The company increased dividends by another 15% showing strong intention to return excess cash to shareholders.
Bookings were even more biased towards logic with 78% of bookings being logic driven. Six EUV systems were booked.
Outlook is for sales of Euro4B +- Euro100M which is somewhat flattish
EUV demand variations cause slowdown ripple through supply chain
Making EUV scanners is a very complex business requiring an impressive global supply chain of technologically complex parts.
ASML has done a very good job of helping out or acquiring and supporting key supply chain manufacturers, starting with Cymer a long time ago.
Even with all the effort put into the supply chain there is still a limit of how fast production of these key components can be ramped up or down given the complexity which means long lead times.
However, customer demand seems to vary a lot more than the flexibility of the supply chain such as we are now seeing. SMIC was an unpredictable issue. We would expect some natural digestion periods and annual cyclical behavior in order patterns as EUV continues its rollout.
We might compare this problem to the whiskey industry where the global demand may increase or decrease in the short term due to global economies or politics the long term supply of 15 year old single malt scotch really can't be changed that quickly.
DUV was a very good fallback
Falling back on DUV tools wasn't all that bad as evidenced in the 52% gross margins so its not like ASML is suffering due to the variability of EUV orders and shipments.
We think it also shows the huge strength in the overall semiconductor market not just at the bleeding edge.
There have been numerous recent articles about shortages of chips for the auto industry. We would point out that most cars that we know of don't use a lot of 5NM or 7NM chips made with EUV tools but have hundreds or thousands of chips made on 200MM and 150MM relatively ancient fabs and toolsets.
DUV will be around for a very long especially given the memory market and trailing edge demand.
Execution remains solid
We think management has done a good job of managing the many issues and complexities especially in light of Covid. Shifting gears from EUV to DUV and dealing with customer demand changes have been handled relatively transparently to the overall sales and earnings impact.
This is not easy to say given the complex global nature of the tool supply chain and organization.
Financial execution remains slid and the company continues to transition to returning more and more cash to shareholders as it throws off more cash from operations.
The Stocks
We think we will see a somewhat muted response in the stock. While the results were good, the outlook is flat. Investors will likely be a bit less happy with the air pocket in EUV sales that may concern them more.
Logic at 72% is not a problem and likely more of a positive given the variability of memory which can be scary.
All in all a good quarter but some complications.
The impact on collateral stocks is also likely somewhat neutral to slightly negative as we didn't have overly positive results nor definitive recovery coming from the memory area.
Given that logic is so strong we would expect to see better results from KLAC than LRCX as they report their quarters given their respective concentrations.
About Semiconductor Advisors LLC
Semiconductor Advisors is an RIA (a Registered Investment Advisor),
specializing in technology companies with particular emphasis on semiconductor and semiconductor equipment companies. We have been covering the space longer and been involved with more transactions than any other financial professional in the space. We provide research, consulting and advisory services on strategic and financial matters to both industry participants as well as investors. We offer expert, intelligent, balanced research and advice. Our opinions are very direct and honest and offer an unbiased view as compared to other sources.
[post_title] => ASML - Strong DUV Throwback While EUV Slows- Logic Dominates Memory
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => open
[post_password] =>
[post_name] => asml-strong-duv-throwback-while-euv-slows-logic-dominates-while-memory-waits
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-24 06:27:40
[post_modified_gmt] => 2021-01-24 14:27:40
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?p=294966
[menu_order] => 0
[post_type] => post
[post_mime_type] =>
[comment_count] => 0
[filter] => raw
)
[6] => WP_Post Object
(
[ID] => 295033
[post_author] => 28
[post_date] => 2021-01-22 10:00:29
[post_date_gmt] => 2021-01-22 18:00:29
[post_content] => Dan and Mike are joined by industry expert Robert Maire for a discussion on China based semiconductors. Robert is an internationally recognized expert on all aspects of semiconductor manufacturing. He joined SemiWiki in 2015 and his blogs have garnered more than 2 million views and many pointed discussions. Robert founded Semiconductor Advisors in 2007. Before that he held executive level positions with Morgan Stanley, Donaldson, Luftkin & Jenrette, Bear Sterns, and Needam & Company.
Official Bio: Robert Maire is President of Semiconductor Advisors, a consulting firm that provides financial and strategic advisory services to technology companies as well as investors specializing in the semiconductor, semiconductor equipment and electronic technology sectors. Prior to that, he was among the first Wall Street analysts to cover semiconductor & semiconductor equipment companies for Morgan Stanley, DLJ, Bear Stearns and Needham, spanning a period of over 20 years. Mr Maire was responsible for raising billions of dollars, more than any other financial professional in the industry, through IPOs and other financing activity for companies such as Applied Materials, KLA, ASML, Sandisk, Agilent, Veeco, Cymer, Rudolph, MKS and many others. He has provided critical insight and advice on many M&A transactions. He has been cited by both the Wall Street Journal and Institutional Investor for his analysis work. He publishes his acclaimed “Semiwatch” newsletter, which is a leading source of information and opinion for investors in the sector. He is frequently quoted in financial media such as The Wall Street Journal, Barrons and numerous online sites . He holds BSEE and BSCS degrees from the State University of New York at Stony Brook.
[post_title] => Podcast EP4: Can China Really Become Self-Sufficient in Semiconductors?
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => closed
[post_password] =>
[post_name] => podcast-ep4-can-china-really-become-self-sufficient-in-semiconductors
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-26 13:15:42
[post_modified_gmt] => 2021-01-26 21:15:42
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?post_type=podcast&p=295033
[menu_order] => 0
[post_type] => podcast
[post_mime_type] =>
[comment_count] => 10
[filter] => raw
)
[7] => WP_Post Object
(
[ID] => 294894
[post_author] => 28
[post_date] => 2021-01-22 06:00:21
[post_date_gmt] => 2021-01-22 14:00:21
[post_content] => 
Lee-Lean Shu co-founded GSI Technology in March 1995 and has served as President and Chief Executive Officer and as a member of the Board of Directors since its inception. In October 2000, Mr. Shu became Chairman of the Board of GSI Technology. Mr. Shu was has held various management positions in SRAM and DRAM designs at Sony Microelectronics Corp and AMD from July 1980 to March 1995. Shu holds a B.S. degree in Electrical Engineering from Tatung Institute of Technology and an M.S. degree in Electrical Engineering from the University of California, Los Angeles. Mr. Shu is the recipient of the award of Inventor of the year, 2017 from SVIPLA.
Tell me more about GSI Technology?
GSI Technology was started in 1995 and quickly became a leader in the global high performance SRAM market. In 2015 we acquired a very early-stage startup in the AI space, which enabled us to combine their technology and software with our advanced hardware design team to create and deliver our Associative Processing Unit (APU) chip. The first-generation device is called Gemini-I.
What challenge is your APU solving?
At a high level, the APU is addressing the challenge that Von Neumann architectures present when attempting to increase compute performance on big data workloads. Processing cores have been increasing in speed, but they continue to use the Von Neumann architecture, so the limits of Moore’s law, power dissipation, and even more importantly the I/O limits brought about by the need to constantly move the workspace data in and out provide reduced system level benefits. The design philosophy in place now to address the Von Neumann issues is to concentrate more of the same processing in less space rather than truly increasing native processing capability – essentially putting more cores to the chip. However, this just miniaturizes the big server problem. It does not eliminate the Von Neumann I/O bottleneck because you still have to get external data into and back out of those cores. The current treadmill of reducing the size of a core and memory combination and then massively duplicating it provides more of the same compute in a smaller space with only limited power savings and ultimately not very much efficiency in an end-to-end system improvement.
Why is the GSI APU important to data processing?
By removing or reducing the I/O cycles used at a system level, the clock speed can be reduced, providing a dual benefit of faster results and lower power. This increases processing efficiency.
What are the benefits of using Gemini?
The Gemini technology results in significant increases for inference workloads (reduction of big data processing times from seconds to milliseconds, for example). The Gemini is akin to having a memory with RISC Boolean processing capability, (note: NOT a number of RISC processors with memory). It is cycle-by-cycle programmable on its memory compute. Due to this flexibility, it also has performance improvement in workload functions that can be efficiently decomposed into Boolean operations. These benefits avail themselves particularly to very large data set problems.
What is the roadmap for Gemini?
Gemini-I is available today for shipment. We will be offering more options on the Leda boards in the next couple of quarters. The next-generation chip, Gemini-II, will be available in 2022. In this new chip, we have increased the L1 memory by 8x and will also double the clock frequency. This will allow us to further penetrate the big data market.
What applications do you see Gemini powering in the next 5-10 years?
Currently, much effort across the industry is being expended on improving training. This is because training is not only done on the initial off-line evaluation of the problem, but also whenever completely new information arrives. Data is only increasing in volume, and the results of searches are only increasing in need as more users are brought on. The Gemini currently is well suited for the inference operation as opposed to the one-time training effort. Also, the goal of all this massive data collection is multi-faceted: data can provide information, which can provide insights, which can be used to improve a criterion or change a behavior. The last output of improvement is being sought after in all industries. The Gemini can accelerate getting this information in real-time for those markets where the fast response from a huge data store enables improvements. Some examples of this includes medical research, personal medical assistance, facial recognition, NLP, e-commerce, space-based environmental monitoring, actual prevention systems (cybersecurity, IIoT, physical security), and traffic intelligence systems.
ABOUT GSI TECHNOLOGY
Founded in 1995, GSI Technology, Inc. is a leading provider of semiconductor memory solutions. GSI's resources are focused on new products that leverage the strengths of its legacy SRAM business. The Company recently launched radiation-hardened memory products for extreme environments and the Gemini APU, a memory-centric associative processing unit designed to deliver performance advantages for diverse AI applications. The APU's architecture features massive parallel data processing with two million-bit processors per chip. The massive in-memory processing reduces computation time from minutes to milliseconds, even nanoseconds, while significantly reducing power consumption with a scalable format. Headquartered in Sunnyvale, California, GSI Technology has 172 employees, 114 engineers, and 92 granted patents. For more information, visit gsitechnology.com.
[post_title] => CEO Interview: Lee-Lean Shu of GSI Technology
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => open
[post_password] =>
[post_name] => ceo-interview-lee-lean-shu-of-gsi-technology
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-19 21:19:14
[post_modified_gmt] => 2021-01-20 05:19:14
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?p=294894
[menu_order] => 0
[post_type] => post
[post_mime_type] =>
[comment_count] => 0
[filter] => raw
)
[8] => WP_Post Object
(
[ID] => 294349
[post_author] => 32540
[post_date] => 2021-01-21 10:00:05
[post_date_gmt] => 2021-01-21 18:00:05
[post_content] => 
Chart 1
Smartphone manufacturers continue to add features to smartphones, such as multiple cameras or more memory, even though these have not been able to incentivize subscribers to shorten upgrade cycles. One feature that smartphone manufacturers have changed is the type of display, migrating from LCDs to rigid OLED displays to Flexible OLED displays to Foldable OLED displays.
Table 1 illustrates this transition in display type. While global smartphone shipments will exhibit a CAGR (Compound Annual Growth Rate) of -0.4% between 2017 and 2022, LCD display shipments will exhibit a CAGR of -6.7% while OLED displays will exhibit a CAGR of 11.9%.
In addition, within the OLED display sector, rigid OLED displays will exhibit a CAGR of 0.2%, while flexible OLED displays will exhibit a CAGR of 26.9%, according to The Information Network’s report entitled “OLED and LCD Markets: Technology, Directions and Market Analysis.”
I expect stronger growth in flexible OLED displays in 2021 coming from Apple (AAPL), which in 2020 its iPhone 12 models all have flexible OLEDs, as sales of these models, which shipped late in 2020 will continue in 2021, along with iPhone 13 models.
Foldable smartphones will grow from just 500,000 units in 2019 to 19.9 million in 2022. A total of 2.8 million foldable smartphones were sold in 2020, as Samsung’s Galaxy Z Flip smartphone Galaxy Z Fold 2 accounted for 73% of them.
Chart 2
For flexible OLEDs, a glass seal is not suitable, and a thin-film encapsulation (TFE) is required that is flexible yet provides a robust hermetic seal. Chart 3 illustrates the utilization of TFE for a flexible OLED display. Since it uses a thin film instead of glass for encapsulation, the overall thickness of a display panel is also decreased.
Chart 3
I refer readers to a comprehensive analysis of OLED encapsulation in my May 24, 2016 Seeking Alpha article entitled “Applied Materials Stock Bounced On Display Orders - But Is It Sustainable?”
Since that article was written, the utilization of flexible OLEDs into smartphones has increased from less than 25% to more than 35% (Table 1). As a result, the structure and deposition processes for the TFE are increasingly being scrutinized. Microscopic pinholes or microcracks that may form during deposition of the TFE or when bending the device result in rapid device failure.
The increasing number of foldable smartphones adds another level of complexity in the design of the OLED and TFE due to the extreme strain and stresses placed on the OLED structure. Add to that the plethora of applications such as automotive, tablets, smart home, and TVs, and the number of TFE applications substantially increases.
Remember that foldable smartphones are made with modified flexible OLEDs. There are numerous OLED manufacturers with flexible OLED facilities located in Asia with a total capacity of 675,000 Gen 6 panels per month with a motherglass measuring 1500mm ×1850mm, as shown in Table 2.
To put things in perspective,
Come work for the Apple’s OLED Encapsulation team comprising of amazing engineers who make the best OLED displays in the world including the Apple Watch and the iPhone. We make the world’s most reliable thin film encapsulation for OLEDs with worlds best electrical, optical and mechanical characteristics while achieving the small display bezels. In this position, you will play a critical role in pushing the boundaries of OLED encapsulation for next generation of products. To do so, you’ll be working on core-technologies behind Thin Films deposited though Low temperature ALD, PECVD. With this core know how as a springboard, you will be creating next generation OLED displays with unique product design and the best optical and mechanical characteristics. You will be working on these technologies with Apple’s OLED display vendors across the globe to bring them from prototyping to mass production.Depending on whether BOE is accepted as a supplier for the iPhone 12, the three suppliers are:
The PVD films exhibit relatively low film quality, with many defects and pinholes in the film, and the PECVD films can cause plasma damage and high process temperature issues. In contrast, ALD enables the thin film to be stably deposited at a low temperature in a vacuum chamber, and to be almost defect-free, but is 1000 times slower than the other processes.
Single-layer TFE has been attractive because it has a simple fabrication process compared to other encapsulation methods. But multilayer technologies address the aforementioned issues by inserting an interlayer. This can improve the WVTR by increasing the lag time. The flexibility can also be improved by the newly inserted layer. The inorganic layers (such as SiNx, SiOx, Al2O3) serve as the main barriers to block the moisture. Since defects such as pinholes and other particles unavoidably occur, organic layers (such as epoxy resin, phenolic resin, PET, PBT) effectively block the moisture infiltration paths.
Applied Materials, as mentioned above, is used for TFE for Samsung’s Gen 6 flexible OLEDs. It’s Enflexor Gen 6H PECVD deposits a range of buffer barrier films. AMAT’s TFE process is proprietary, but we can learn a bit from its patentfiled in July 2018. The encapsulant can include a first barrier layer, a buffer layer, and a second barrier layer.
The first barrier layer can include a dielectric film, such as SiN (silicon nitride). The buffer layer can be an organic layer, such as a hexamethyldisiloxane (HMDSO) layer. A second barrier layer comprising SiN is then deposited over the buffer layer.
AMAT’s Enflexor Gen 6H system is shown in Chart 4, showing the size and complexity of the cluster tool. This description of multiple layers using multiple process chambers explains the need for such a large and complex AMAT system, and can be explained in further details in the patent:
Chart 4
“A method of encapsulating an organic light emitting diode OLED device, comprising: generating a first plasma comprising silicon and nitrogen in a first process chamber; depositing a first portion of a first barrier layer comprising silicon and nitrogen over the OLED with the first plasma; generating a second plasma comprising silicon and nitrogen in a second process chamber; depositing a second portion of the first barrier layer comprising silicon and nitrogen over the first portion of the first barrier layer with the second plasma, wherein a density of the first plasma and the second plasma differ by a factor of at least 100; depositing a buffer layer over the first barrier layer in a third process chamber; and depositing a second barrier layer over the buffer layer in a fourth process chamber, wherein the first process chamber, the second process chamber, the third process chamber, and the fourth process chamber are arranged around a single transfer chamber.”Using a unique technology for the display market, privately held Denton Vacuum (Cherry Hill, NJ) uses ion beam technology to deposit a Diamond-like nanocomposite organic layer sandwiched between two SiOxNy dielectric layers to form the encapsulant. But the equipment is priced at $3 million compared to an estimated $25 million for AMAT's Enflexor. To achieve a throughput of 1 Gen 6 motherglass panel per minute, the Phoenix In-Line PIB-CVD Deposition System is configured with 6 chambers with 2 linear sources in each chamber. Three chambers for each of two materials (inorganic and organic), PIB-CVD of Inorganic SiOxNy and Organic a-C:H DLN
But based on my analysis, Apple's scrutiny of Samsung's encapsulation, which uses AMAT TFE equipment, and BOE Technology, which also uses AMAT TFE equipment, is a headwind for AMAT from two standpoints:
Why is electromagnetic coupling important? Consider the simple example illustrated above – three examples of a microstrip line on a board are shown, all the same length, but with varying serpentine properties. Due to the electromagnetic self-coupling present between different segments of the line, the frequency response (e.g., the insertion loss) of each varies significantly –a discretization mesh of the lines with meandering segments is needed to accurately calculate the behavior.
Now consider the example below, where the detailed field distribution is infinitely more complex. Matt provided this electronic system as a representative example of the types of models for which designers are seeking to analyze the electromagnetic behavior.
Matt chuckled, “When I first starting working with HFSS over 20 years ago, we were solving systems with maybe 10K to 40K matrix unknowns. Now, we are routinely solving models with more than 100M matrix elements. The ongoing advances in electromagnetic analysis have dramatically expanded the types of designs that are able to be simulated.” Matt elaborated on some of those advances.
Computational Electromagnetics
Several algorithmic enhancements have been incorporated into HFSS, to enable the use of HPC resources.
This last feature is worth special mention. Matt indicated, “HFSS supports virtually disturbing the mesh for variation analysis. Designers can identify a set of parameters in the system model, and readily see how the electromagnetic analysis results change with manufacturing variations, for a small overhead in simulation time, far more efficient than running full simulations on different parameter samples and unique meshes with small dimensional changes.” This feature provides great insights into where designers could focus on cost versus manufacturing tolerance tradeoffs.
Ansys has prepared an informative demo of how designers can quickly visualize the response to parameter sensitivity – link.
Algorithmic Enhancements for Mesh Generation
Matt identified three key HFSS enhancements of late related to mesh generation.
Adaptive Meshing
The introductory section above described the importance of the 3D mesh to the resulting accuracy. An initial mesh is solved for the fields – a calculation of the electric field gradient is indicative of where local mesh refinements are appropriate. (The basis functions for representing the local fields could also be updated.) A new mesh is solved, and the process iterates until successive passes very less than the convergence criteria.
HFSS recently extended this capability to adapt the mesh each iteration using multiple frequency solutions, over a user-specified range, to enhance the results accuracy when a broad range of spectral energy is present.
3D Components
Traditionally, it has been difficult for designers to build a comprehensive (“end-to-end”) model of even a single long-reach, high-speed signaling interface. The PCB trace S-parameter model generation from the stack-up was relatively straightforward, but obtaining a model for connectors and cables from the vendor was typically difficult.
Ansys realized that enabling link simulation, and ultimately, system-level analysis required a novel method, and developed the “3D Components” methodology:
HFSS Mesh Fusion
Of the steps in the electromagnetic analysis of an electronic system:
Summary
The traditional method for electromagnetic analysis in electronic systems focused on PCB designs and high-speed signaling. The board stack-up and materials properties were defined, and the signal traces were simulated. S-parameter response models for signal loss and (near-end/far-end) crosstalk from adjacent traces were generated, and incorporated into subsequent circuit simulations to measure the overall transmit/receive signal fidelity. However, the complexity of current electronic systems necessitates a more comprehensive approach to electromagnetic coupling simulation, as compared to concatenating individual S-parameter models. Systems will be integrating a broad range of signal frequencies from audio to mmWave, with advanced packaging present in aggressive volume enclosures.
The HFSS team at Ansys has focused on numerous technical advances – both computationally and in the critical area of mesh generation – to enable this analysis. Designers can now evaluate and optimize models of a scope that was once unachievable, with manageable computational resources.
For more info on these Ansys HFSS features, please follow these links (and don’t forget to sign up for the Mesh Fusion webinar):
Broadband Adaptive Meshing – link
Ansys cloud HPC resources – link
3D Components – link1, link2
Mesh Fusion webinar – link
-chipguy
[post_title] => System-level Electromagnetic Coupling Analysis is now possible, and necessary
[post_excerpt] =>
[post_status] => publish
[comment_status] => open
[ping_status] => open
[post_password] =>
[post_name] => system-level-electromagnetic-coupling-analysis-is-now-possible-and-necessary
[to_ping] =>
[pinged] =>
[post_modified] => 2021-01-26 10:55:22
[post_modified_gmt] => 2021-01-26 18:55:22
[post_content_filtered] =>
[post_parent] => 0
[guid] => https://semiwiki.com/?p=295283
[menu_order] => 0
[post_type] => post
[post_mime_type] =>
[comment_count] => 0
[filter] => raw
)
[comment_count] => 0
[current_comment] => -1
[found_posts] => 7462
[max_num_pages] => 747
[max_num_comment_pages] => 0
[is_single] =>
[is_preview] =>
[is_page] =>
[is_archive] =>
[is_date] =>
[is_year] =>
[is_month] =>
[is_day] =>
[is_time] =>
[is_author] =>
[is_category] =>
[is_tag] =>
[is_tax] =>
[is_search] =>
[is_feed] =>
[is_comment_feed] =>
[is_trackback] =>
[is_home] => 1
[is_privacy_policy] =>
[is_404] =>
[is_embed] =>
[is_paged] =>
[is_admin] =>
[is_attachment] =>
[is_singular] =>
[is_robots] =>
[is_favicon] =>
[is_posts_page] =>
[is_post_type_archive] =>
[query_vars_hash:WP_Query:private] => e64def1ba05c75a0d2020dce1b92d1d6
[query_vars_changed:WP_Query:private] => 1
[thumbnails_cached] =>
[stopwords:WP_Query:private] =>
[compat_fields:WP_Query:private] => Array
(
[0] => query_vars_hash
[1] => query_vars_changed
)
[compat_methods:WP_Query:private] => Array
(
[0] => init_query_flags
[1] => parse_tax_query
)
[tribe_is_event] =>
[tribe_is_multi_posttype] =>
[tribe_is_event_category] =>
[tribe_is_event_venue] =>
[tribe_is_event_organizer] =>
[tribe_is_event_query] =>
[tribe_is_past] =>
[tribe_controller] => Tribe\Events\Views\V2\Query\Event_Query_Controller Object
(
[filtering_query:protected] => WP_Query Object
*RECURSION*
)
)
With the increasing density of electronics in product enclosures, combined with a broad range of operating frequencies, designers must be cognizant of the issues associated with the radiation and coupling of electromagnetic energy. The interference between different elements of the design may result in coupling noise-induced… Read More
GDS and LEF/DEF each came about to support data exchange in different types of design flows, custom layout and place & route respectively. GDS (or stream format) was first created in the late 1970s to support the first generation of custom IC layout tools, such as Calma’s GDSII system. Of course, the GDS format has been updated… Read More
Last August I detailed a webinar about HCL Compass, a tool that provides low-code/no-code change management capability for enterprise scaling, process customization and control to accelerate project delivery and increase developer productivity. There is a lot of activity these days to migrate various enterprise applications… Read More
-Intel good results had a little extra help to be great
-New CEO commits to remaining an IDM versus fabless
-Claims of strong progress on 7NM fuel optimism inside
-Outsourcing to TSMC will not go away but will increase
A good quarter but with some silicon enhancements from ICAP
Intel reported Revenues of $20B and EPS of $1.52, which… Read More
A study published in ScienceAdvances has shown computer simulations of the movement of virus-laden airborne droplets in cars. The objective of the study appears to be to assess the degree of driver-to-passenger and passenger-to-driver exposure given different open-or-closed window configurations in the context of modeled… Read More
– ASML has good quarter driven by DUV & Logic (@72%)
– SMIC & other major customer slow EUV plans
– Logic (read that as TSMC) remains key demand led driver
– We are happy memory remains muted given cyclical potential
A very solid quarter with a continued road to growth
The quarter came in at Euro4,254B… Read More
Dan and Mike are joined by industry expert Robert Maire for a discussion on China based semiconductors. Robert is an internationally recognized expert on all aspects of semiconductor manufacturing. He joined SemiWiki in 2015 and his blogs have garnered more than 2 million views and many pointed discussions. Robert founded Semiconductor… Read More
Lee-Lean Shu co-founded GSI Technology in March 1995 and has served as President and Chief Executive Officer and as a member of the Board of Directors since its inception. In October 2000, Mr. Shu became Chairman of the Board of GSI Technology. Mr. Shu was has held various management positions in SRAM and DRAM designs at Sony Microelectronics… Read More
I remember it like it was yesterday. I was sitting at lunch with good friend Scott Santandrea, explaining my struggles to get traction with the sales channel for the Analog to Digital Converter product line I was managing. My business line had been spending a lot of resources developing high-performance 24-bit delta sigma and 20-bit
Smartphone shipments have been dropping over the past few years, as shown in Chart 1, as a result of several factors, but primarily the slowdown in smartphone innovation while at the same time prices have kept increasing. Even with the much anticipated 5G in 2020, unimpressive speed gains coupled with a Covid-19 backdrop, smartphones… Read More
Podcast EP4: Can China Really Become Self-Sufficient in Semiconductors?