Advances In Hardware: Chips To Systems - Ieee Computer Society: Fill & Download for Free

If you still want to pursue a career in the semiconductor industry, focus on the "hardcore" VLSI stuff.Justification for Focusing on Programming Skills for IC Design and EDA Software DevelopmentIf the computer languages that you learn cover a broad scope, hiring managers may think that you are not focused on careers in VLSI design, Electronic Design Automation (EDA) (EDA) software development, computer architecture, and related areas.Yes, arguments can be made for LabVIEW-based FPGA design, UML for VLSI design, integrating databases into EDA tools to cope with the large amount of data that EDA tools have to work with (for performance improvement), GUI-development of EDA tools, and what not.However, think about the core competencies of semiconductor companies. Always ask yourself how are you influencing the bottom line of the company (profit maximization).EDA tools sell (very) well if their back-end engines are great in terms of performance, quality of results (QoR), and memory management/usage. If you can develop good GUI, does that improve the performance, quality of results (QoR), and memory management of the EDA tool? No. That is why EDA developers don't care about GUI development. Grad students in EDA just wanna publish research papers, graduate, create EDA start-ups or work for EDA R&D teams (e.g., Cadence Design Systems, Synopsys, and Mentor Graphics) and research labs (e.g., Intel Strategic CAD Lab and IBM Research, and previously Synopsys Advance Technology Group and Cadence Research Laboratories - now defunct). See Why are GUIs of EDA tools so buggy? Why aren't the big guys working on this?.VLSI design is measured on design metrics like performance (not just delay/latency, but also throughput), power consumption, and thermal distributionAnalog/RF and mixed-signal (AMS/RF) IC design: various metrics like total harmonic distortion, phase noise, power consumption, signal to noise ratio (SNR), ...Computer architecture: performance (not just delay/latency, but also throughput) and memory managementHow is learning Ruby on Rails relevant or Django relevant to the above? A good guide is to use job requirements (and preferred skills in the job descriptions) to guide you. Do this from junior to senior R&D positions in your area, such as EDA software development or VLSI design.Suggested Computer Languages to LearnHere are some computer languages that I suggest that you learn:Verilog and SPICE for VLSI designSystemC-TLM for TLM modeling and embedded system design (i.e., hardware/software co-design), and SystemC-TLM and SystemC-RTL for microarchitecture design. SystemC-AMS for cyber-physical system (CPS) designSPICE (and SKILL/Python) for AMS/RF IC design and custom digital IC designVerilog-A, Verilog-AMS, VHDL-AMS, SystemC-AMS for behavioral AMS/RF modeling, and CPS design. Verilog-A is also used for device/compact modeling.UNIX shell scripts (mandatory) and Perl/Python for process automationTcl, Perl, and/or Python for driving EDA tools (i.e., automating the design process via CAD engineering, as in VLSI CAD)C++ for EDA software development, although some companies want Java developers. I have only seen one EDA company looking for C# developers. In any case, C++ would suffice. Make sure that you learn parallel programming, including concurrent programming, as well.C and/or C++ for embedded system design (i.e., hardware/software co-design), and behavioral or electronic system-level models (for high-level synthesis)MATLAB/Python for statistical analysis, if you need it. Or just C++ libraries for statistical analysis.C++ (perhaps MATLAB and Python as well) for MATLAB for data scientist positions in semiconductor manufacturing, which is automated; see Intel's Automated Manufacturing Technology... Here, C++ is preferred over MATLAB and Python for performance and memory management reasons. See Pasquale Ferrara's answer to Should I use Java or C++ for writing high performance code?. You would need to know some skills in information systems and databases, since you will be handling TBs of data from the test chips and fabricated wafers. In this case, the Map-Reduce/MapReduce topic is relevant.So, yeah, learning PHP may help you get a job as a web designer/developer for a cloud-based EDA company, or as a web developer with a semiconductor company to maintain its web page. However, it will not help you design innovative ICs or develop innovative EDA tools.Likewise, don't bother learning LabVIEW for IC design, since student licenses for such tools are very expensive (>US$1,000). Also, while UML is helpful for modeling the software architecture of EDA tools, and even design VLSI circuits and systems, most EDA developers come from engineering backgrounds and do not know UML. So, more likely than not, EDA R&D engineers and researchers will not ask you challenging questions about UML during interviews for R&D positions and internships in EDA software development.EDA Programming ContestsYes, competitive programming competitions will help, but mostly of the EDA variety. See the following contests/competitions in EDA.CADathlon held before the start of the International Conference on Computer-Aided Design (ICCAD): CADathlon at ICCAD. This is the EDA variant of ACM International Collegiate Programming Contest (ICPC). Solve 6 EDA problems (programming challenges) in 9 hours (8 a.m. to 5 p.m.).CAD contest at ICCAD: 2013 CAD Contest at ICCAD. Turn in binaries for your EDA software that solves at least one of the EDA programming problems.ISPD contests, which are associated with the International Symposium on Physical Design (ISPD): ISPD ContestsTAU contest, which is associated with the International Workshop on Timing Issues in the Specification and Synthesis of Digital Systems: TAU Contest 2013Ad-hoc EDA programming contestsDAC 2012 Routability-Driven Placement Contest, which is organized as part of the Design Automation Conference (DAC)eASIC Placement Design Challenges: eASIC Issues $30,000 Worldwide Placement Design Challenge – Everyone Welcome to ParticipateProgramming challenges for the International Workshop for Logic Synthesis (IWLS): Third IEEE Programming Challenge at IWLSIMPORTANT NOTE about EDA Programming ContestsWinners of these research/programming contests tend to publish their novel/innovative algorithms/heuristics/techniques in EDA research conferences like DAC and ICCAD. National Taiwan University has teams that placed in the top 3 of the CADathlon and ISPD contests regularly, since 2007. If you can win these contests, you will be looked upon favorably in industry... Think about this, if you can't even turn in your binary on time for the contest, what does it say about your EDA software development skills? How are you gonna compete with graduates from Berkeley, UCLA, UT Austin, and University of Michigan? See Pasquale Ferrara's answer to What are the best VLSI CAD research groups in US universities?There used to be a programming contest for processor simulators, which is organized by IEEE Computer Society. It is no longer organized.How do you develop the necessary programming and algorithmic skills?For VLSI design, AMS/RF IC design, and computer architecture (especially microarchitecture), design some ICs.While designing ICs, use scripts to automate boring tasks. Make your life easier. See Someone anonymous's answer to Why do I need to learn scripting language to get a vlsi design job? I know that scripting language is used for automating tasks of different EDA tools but is it really necessary to learn that?Pasquale Ferrara's answer to As an undergraduate student without an adviser, how do I conduct research in the fields of VLSI, computer architectures, and low power systems with the aim of publishing papers in reputed journals/conferences? How do I plan to publish papers?Pasquale Ferrara's answer to How can I find people who are interested to collaborate with me to do some independent research, if I am an engineer from the industry?If you are a student, take part in VLSI design contests. See Pasquale Ferrara's answer to What are some good VLSI design contests? and Pasquale Ferrara's answer to I'm familiar with computer systems, but I want to learn more about architecture and processor design. Where do I start?.Pasquale Ferrara's answer to What is a good book to learn computer architecture? Use these books to help you design simple single-core 32-bit processors, and eventually high-performance multicore processors. Or, Computer Organization + Computer Architecture.Questions regarding algorithm analysis/design and data structures are usually fairly easy for positions (including internship positions) in VLSI design, let alone AMS/RF IC design. A fundamental computer science (CS) class on data structures and algorithms would suffice.However, for positions (including internship positions) in CAD engineering (as in VLSI CAD engineering), you need to know the material in the introductory data structures and algorithms class very well and create algorithms (or algorithmic solutions) to programming problems in technical interviews. Some typical positions in CAD engineering are physical design engineer, DFT (design for testability) engineer, and signal integrity engineers. Ditto for VLSI verification engineers and computer architects.For EDA software development, look at the web pages of the aforementioned contests, and use the information and benchmarks to develop your EDA tools.Pasquale Ferrara's answer to What is the best way to start writing electronic design automation tools?Resources for EDA software development: Pasquale Ferrara's answer to What are some good books on EDA (Electronic Design Automation)?C++ Programming... Advanced C++ programming skills are preferred, especially for mid-level and senior R&D positions in EDA software development. These positions typically require a MS/Ph.D. in EECS (electrical engineering, computer engineering, and/or computer science). Use Cracking the Coding Interview (2011 book) to help you with this.Relevant data structures for specific EDA problems. When you develop EDA software for a specific problem, say combinational logic synthesis, you will use specific data structures; in this case, you should know about binary decision diagrams (BDDs) AND-inverter graphs (AIG).Relevant algorithms for specific EDA problems. When you develop EDA software for a specific problem, say placement (from "place and route"), you will use specific algorithms. For placement, you should know about min-cut placement and force-directed placement (developed/co-developed by Prof. Melvin Breuer at University of Southern California); simulated annealing; quadratic placement (based on Quadratic Programming); and the like. See Placement (EDA).Relevant data structures and algorithms (from related CS areas) for EDA problems. If you have a strong background in electrical engineering and computer science (EECS), preparing for this is easier (but not easy). For example, if you work in physical verification (e.g., layout compaction, layout extraction, parasitic extraction, layout versus schematic check, electrical rule check, and design rule check) and computational lithography, you would probably be asked relevant questions in computational geometry. This is when a good CS background is handy. Likewise, for positions involving electronic system-level (ESL) and front-end design and verification tools, knowing about compiler design can help you answer questions related to parser development for these tools, and high-level synthesis (has multiple similarities to compiler design)... Use Cracking the Coding Interview (2011 book) to help you with this... What you need to know depends on what EDA tool do you want to develop, and which product team in EDA companies and integrated device manufacturers (IDM) do you want to work for.References:Choosing a Graduate Program in VLSI Design & Related Areas: Things to ConsiderPasquale Ferrara's answer to What are the different programming languages an electronics engineer must have in his arsenal?Pasquale Ferrara's answer to As an electrical engineering undergraduate, what are all the software, computer skills, programming languages that I should know?Pasquale Ferrara's answer to What amount of programming is there in electronics and communication engineering?Pasquale Ferrara's answer to What is the best way to start writing electronic design automation tools?Someone anonymous's answer to Why do I need to learn scripting language to get a vlsi design job? I know that scripting language is used for automating tasks of different EDA tools but is it really necessary to learn that?EECS (ECE + CS)P/S: I hope that I am not insulting your intelligence by simplifying the explanation for you. Any decent graduate (BS EE, B.Tech EE, or equivalent, MS/Ph.D. EE) in electrical engineering should be able to figure this out.

A lot... If you really want to be outstanding in electrical and electronics engineering (EEE), otherwise simply known as electrical engineering (EE) in the U.S.. It can also be known as electrical and computer engineering (ECE).I strongly disagree with Razvan Baba. I may be wrong, but he does not seem to have a good grasp of ECE, in terms of breadth across the scope of ECE or depth in any area of ECE.Look at IEEE journal and conference papers. Talk to faculty and graduate (MS/Ph.D.) students in ECE, as well as senior R&D engineers in ECE. Ask them if they can get away with programming and computer modeling.Programming in MATLAB for mathematical modeling, system/filter design and analysis, and simulation is used in control engineering, signal processing, antenna design, and many areas of ECE. Just check out the MATLAB and Simulink toolboxes: Products and Services. Challenging and academically rigorous classes, especially at the graduate level, will require you to implement or even design algorithms for control and signal processing. So, yes, you will need to understand algorithm analysis and design. Think about circuit complexity in VLSI circuit and system design. Isn't that a lot like computational complexity? Does the algorithm run in O(n^3)? Is the complexity of the circuit O(n*log n)?Furthermore, modern/advanced control systems are multi-input, multi-output, stochastic, adaptive, digital, autonomic, and/or nonlinear. CS students will not want to learn about nonlinear dynamical systems. Trust me. That is why the application of control engineering in autonomic computing has very few takers. Why? You need to be good at ECE and CS; that is, EE + CS = EECS.Many, if not most, computer science (CS) programs do not train you to design analog/RF and mixed-signal integrated circuits (ICs), and VLSI circuits and systems. So, all the talk about using hardware description languages (HDLs) for RTL design of ICs that are implemented on FPGA boards or standard cell logic is nonsense. Look, if they do not even teach computer organization classes that require students to design a simple 32-bit processor, do you think they can teach VLSI design or computer architecture effectively? Hell no! Behavioral modeling of AMS/RF circuits and systems with SystemC-AMS, Verilog-AMS, Verilog-A, and VHDL-AMS? Forget about it! If you want to design ICs and VLSI systems, pursue a MS/Ph.D. in ECE at a good research university (preferably in the U.S.). And, yes, VLSI system design may involve modeling with UML, invoking design patterns, using Petri Nets, and using formal/mathematical logic for formal methods and formal verification. VLSI design also involves programming in Perl, Tcl, and Python (or UNIX shell scripts), in addition to C, C++, and SystemC. I consider behavioral modeling in SystemC, Verilog, and VHDL as VLSI design, rather than programming. You are designing electronic systems and ICs, not programming a processor (as in system software, or application software, development).Instead of using a word processor for documentation, you can do that with LaTeX and Doxygen (works with VHDL, too!). Want to draw something? Use Graphics Layout Engine (GLE), Asymptote (vector graphics language), MetaPost and TikZ!Use build automation to compile, synthesize, or run your scripts/tools. Or, typeset your LaTeX source files. E.g., put a UNIX shebang at the top line of your SPICE netlist and run it like a script from the command line.Use revision control to manage different versions of your source files: MATLAB (or GNU Octave), C++, SystemC, Verilog, Python, Tcl, Perl, UNIX shell scripts, LaTeX source files, and SPICE (yes, you can write your own SPICE netlists from scratch and simulate them with a SPICE tool). Markdown works for GitHub, if you use that to commit your source code for MATLAB, Verilog, VHDL, scripts, C++ code, and what not.To work effectively in a UNIX-like operating system, knowing how to write simple UNIX shell scripts quickly helps you work efficiently and effectively. Learn how to use regular expressions.Take an advanced graduate class in antenna design, and you will have the "luxury" of implementing complex numerical methods in C, C++, FORTRAN, or some other programming language to model your antenna, simulate electromagnetic wave propagation, and analyze the system for electromagnetic interference and capability (EMI/EMC).Nanoscale device engineering will inevitably demand computational modeling and numerical computation in C, C++, Verilog-A, or other languages used for device/compact modeling.ECE-based approaches to systems engineering and reliability engineering will involve a lot of computer modeling and programming.Information theory and communication theory basically is based on mathematics and theoretical computer science. Being able to implement your ideas/methods as a computer program allows you to test your ideas and verify/validate them. Think about methods for encoding/decoding, and error detection and correction.Optical engineering and telecommunications allow you to explore different techniques for transmitting information, which is usually modeled in computers (think programming, again!), and examine everything from the performance and energy consumption of routing data packets in a telecommunications/computer network or network-on-chip (NoC). Yes, optical NoCs can use similar packet routing concepts as packet routing in telecommunications and computer networks.Other aspects of telecommunications: multimedia compression? Yes, you can implement them as software in C, C++, or MATLAB, or as VLSI circuits in SystemC, Verilog, or VHDL.Power engineering? The design, modeling, and analysis of electrical machines would involve modeling with LabVIEW, a graphical computer language (or graphical "programming" language, if you like). Smart grid design? Definitely a lot of computer modeling (read: computer programming).Now, which part of ECE does not involve programming? See IEEE Society Memberships and IEEE Technical Councils for the scope of ECE.Bottom line: You can't run away from programming in ECE. If you hate it, remember that a lot of programming and computer modeling is used in financial engineering and computational finance. So, if you wanna make big bucks in investment banking that exploits your ECE skills, think again!Addendum:[Read the last portion to address the question of whether to accept the offer to study EEE, or try to switch to (or join) a CS program.]You can learn much more skills and languages in good internships, where you are thrown into real-world projects and are expected to perform to justify your US$20/hr - US$40/hr pay check. Unfortunately, good internships where you can actually develop/design stuff and learn aren't that easily available in many areas around the world. See Pasquale Ferrara's answer to Which is better to study for a short term winter course, with the hope of a future foreign internship: Embedded systems or VLSI design?.A lot of undergraduate classes in many ECE programs may not involve programming. It all depends on where you go to college. I had to implement some numerical methods for my applied/engineering math classes in vector calculus and differential equations. But, I had a choice of programming languages to choose from. Many math classes in other programs won't require you to. Ditto for many undergraduate ECE classes in electrical machines, analog circuit design, and what not. However, at good graduate programs in the U.S. for MS/Ph.D. students, you will have lots of opportunities to learn and program, since you will have lots of projects to complete (on top of any research for Ph.D. students).Generally, they don't make you learn something for no reason. If you have to learn a bunch of languages and software to use for your engineering projects, it may be because they want to expose you to different design steps for that particular area/subfield. So, you don't have to learn that many skills and computer languages, if you don't want to.For example, for digital VLSI design, I learned Verilog for RTL design, SPICE for cell characterization and circuit simulation, and Tcl for driving/customizing EDA tools. That said, I did not get to learn SystemVerilog nor hardware verification languages, such as e and Vera. In industry, for entry-level jobs and internships in VLSI design, you need to know how to work in a UNIX environment, program in Perl or Tcl (and hopefully SKILL from Cadence Design Systems), and use Verilog (or VHDL) for RTL design. Why? Because that is what you will be paid to do. So, the technical questions are meant to determine if you can work effectively in your job, or if they have to spend a ridiculous amount of time training you. Training employees may be common in some places, such as India. In the U.S., you need to have demonstrated use of your skills in class projects, open source hardware/software projects, and prior work experience. If your source code is on Open Cores and GitHub, evaluating your skill set is so much easier.I use MATLAB as much as possible, so that I don't have to use R and what not.Using LaTeX helps me write documentation much easier. As a freshman, I started writing papers and reports with >40 references, which can be hard to manage in number-based citation/referencing style in Microsoft Word (back in the day)... The notion that ECE documentation and research papers do not involve mathematics (discrete mathematics, numerical analysis, or even abstract algebra), stochastic modeling, and statistical analysis is nonsense. Writing mathematical equations in LaTeX is so much easier than with word processors, especially if your computing environment is well set up and you have a good process for technical writing; hint: macros and templates help a lot.I used to hate UNIX, but when I saw my classmates destroying me in my mandatory CS classes in computer systems (assembly programming projects), data structures and algorithms, and software engineering, I had to pick up more skills from them so that I can be effective in my class projects. Working in the computer lab instead of my dorm room helps me interact with others, and learn how to work more effectively and efficiently. Yes, it can be distracting when others bug you for help and try to chit chat with you. But, you get to learn how the top students (possible rockstar engineers/developers) do something in 3 hours that takes you days to do it. So, you learn tricks that help you work better. Basic stuff like unit testing, test automation, regression testing can be applied to EE projects, too. Ditto for fault isolation, decoupling of modules (to reduce dependencies), and what not. This set of skills can be summed up as computational thinking, which can be learned in any academic major and be applied to any profession. See Computational Thinking and Pasquale Ferrara's answer to What tangible non-platform-specific skills do computer scientists pick up through their undergraduate education?.To paraphrase Michael Jordan, remember that the fundamentals don't change, and all that changes is your attitude/approach towards them. Reference: M. Jordan. I Can’t Accept Not Trying: Michael Jordan on the Pursuit of Excellence. Harper San Francisco, San Francisco, CA, 1994.Basically, programming, like mathematical analysis, and knowledge of physics are basic skills (or "tools") that you employ to solve real-world engineering problems. Don't use more tools/skills than you need, since you want to save time and effort (and $$$). But, if you can't outcompete your competitors with your current skill set or tools, re-tool yourself and pick up more advanced skills. Programming is not the be all or end all of engineering.The technologies come and go. You need to pick up new tool-specific skills, platforms, computer languages, and what not over time. However, the fundamentals in engineering design, verification, validation, and testing do not change. Modularity is modularity, and it exists in software architectures, VLSI architectures, embedded systems, large engineering systems from cars and airplanes to telecommunication systems. Ditto for fault isolation, fundamental concepts in electromagnetic wave propagation, and what not.Yes, not going to a good engineering program to earn your BS ECE (or equivalent) and advanced degree (MS/Ph.D. ECE) affects your choices and opportunities. However, as you realized, there are a lot of learning resources online to help you grasp what you are missing out on. Joining IEEE and ACM helps expose you to what your peers are doing in their free time, for fun (e.g., IC design, and publishing their novel circuits in a research conference), and taking graduate ECE classes for MS/Ph.D. students while they were still undergrads. Studying in universities with good undergraduate programs, such as those that I mentioned in Pasquale Ferrara's answer to When recruiting Software Engineer/Computer Science majors for US companies, what international universities are on par with MIT/Stanford?, helps a lot, too. If you can't study in Politehnica University of Bucharest, fine, ask your friends who may be studying there. Ditto for the other top engineering programs in your region or the world, such as the Technion. It's okay if you can't have 100% of the opportunities that some students have. 70%, or even 30%, isn't that bad. It is better than 0%.Like I said, asking people in industry and monitoring job advertisements as you progress through college and graduate school will you figure out what skills employers (or rather, hiring managers) want.As for why good ECE programs encourage you to build up a broad base of skills spanning ECE (as aforementioned), and yet more skills in critical areas, such as technical writing, technical/engineering management, and intellectual property law, well, it is to enable ECE graduates to explore different career paths. You don't have to study in the U.S. to realize that a B.A./B.S. (or equivalent) is a basic degree in many professions (such as medicine, law, ECE, and CS). Some career paths may not require more than a B.A./B.S.. But, if you want to outcompete others, you need to have a unique skill set that may be obtained in MS/Ph.D. programs.WIth this in mind, since not everybody wants to be a computer architect, you will find ECE graduates (like many other graduates) venturing into different professions. Some NCAA Division I student-atheletes graduating with BS ECE may become professional athletes. Others may go into management consulting. Some may choose to teach science and math in high schools, while others may teach English to non-English speaking people in Europe or East Asia. Some go to law school and pick up their law degrees (e.g., J.D. or LL.B.), and become lawyers in intellectual property law for high-tech companies. Some become start-up entrepreneurs, and/or venture capitalists. This is where the breadth of skills come in handy. Ditto for having intercultural competence. Also, this helps you in interdisciplinary research, if research is your cup of tea.I may be terribly wrong, but I believe that there are much more low-skilled software engineers and web developers than VLSI designers (read: digital IC designers). There are much more VLSI designers than AMS/RF IC designers. The proportion of rockstar engineers/developers is very small. The number of highly-skilled software developers is also very small, especially in niche topics like embedded computer vision and electronic design automation; these people typically have MS/Ph.D.s in CS and/or ECE.The number of not-so-good BS/MS CS programs is huge. It is very hard to differentiate yourself in many traditional software development jobs, even for full-stack developers. Front-end developers? Forget about it. Pick something that you know many, if not most people, hate and are weak/poor at but that you love. You will be much better off working in that sweet spot, which will enable you to out-innovate the large amount of software developers in Latin America, Europe, South Asia (including India), and East Asia (e.g., China, South Korea, Japan, and Taiwan).Also, the hot emerging technological trends in big data, cloud computing, and cyber-physical systems (facilitating the Internet of Things) allow you to exploit ECE skills better than CS graduates. Most CS graduates are weak in numerical analysis, physics, and engineering. They cannot handle the engineering, statistical analysis, stochastic modeling (ever heard of CS students volunteering to take graduate ECE classes in random processes at MIT, Berkeley, Stanford, and USC? Probably not.), and continuous-time domain aspects of cyber-physical systems (CPS). Exploit this. Now, you are collecting so much information in your CPS devices anyway. What should you do with them? Big data analysis! Domain-specific analytics, whether it be for basketball, financial analysis, or medicine will be the sweet spot for you to tap into your wisdom and insight regarding your passion (whatever it is).Bottom line: An ECE degree may help you separate yourself from the pack more easily. However, it is not for everybody.P/S: Whatever you do, don't listen to bigots like Razvan. He is not just anti-American, anti-CS, but also foolish and not-so-smart.

IEEE and its societies, as well as ACM and it SIGs (Special Interest Groups) hold frequent seminars.See the following:Distinguished Speaker Series from IEEE CEDA.Distinguished Lecturer Program (DLP) from IEEE Circuits and Systems (CAS): IEEE Circuits and Systems Society and Online Lectures | IEEE Circuits and Systems SocietyDistinguished Lecturer Program from IEEE Solid-State Circuits Society: Distinguished Lecturer ProgramACM Distinguished Speaker Program (DSP): ACM Distinguished Speakers ProgramDepartments in computer science (CS), electrical engineering (EE), computer engineering, or combinations thereof (e.g., electrical and computer engineering, ECE, and EECS -- EE + CS), at research universities may hold departmental seminars as well as seminars for research areas. Check them out. Some of them are online: e.g., the University of Washington's Department of Computer Science runs it Computer Science and Engineering Colloquia regularly, and posts videos of the talks at UW CSE Colloquia Search and UW CSE Colloquium and Televised Talk Information.Many research universities offer seminar-based classes in CS, EE, ECE, and EECS. They may include seminar classes on advanced topics in embedded systems and VLSI design, and related research areas. E.g,, some topics that may be covered as a separate class each could be: network-on-chips, hardware/software co-verification, and nanoscale design-for-manufacturability.Other professional societies may offer seminars on embedded systems and VLSI design, too. However, they are usually of an inferior quality. See The Institution of Engineering and Technology and Engineers Australia as examples of what I mean.