about future of ASIC design engineer

[haosg]

any future for asic design

Hi, all

I know most of you are hardware designer, from architecute to synthesised netlist. maybe some through backend to gdsii.
But I'm pure implementation engineer. that from rtl to synthesized netlist.
i.e. synthesis and DFT(internal scan, BIST, BSD). and help backend for timing closure. Though I have finish some SOC chip, but I'm also fear my future.
you know the tools will be more and more powerful. the stage will don't need persons.

About my job , any senior give some proposal.

Thanks very much.

 

[arunragavan]

you seriously stir me dude..
am aiming to become a design engineer.. i agree with tools gettin more and more complex.. it becomes seriously very difficult to learn and understand each and every thing n the market as one seem to be better than the other.. as always to capture the market or customers we need to get the best in best hand in hand..

lets hope we get a new glimpse of hope each time some new tool comeup in the market.

 

[JesseKing]

i am front end designer. after finishing some projects, i began to think that everyone can do my job.
to my honesty, i don't think my job is a challengable one, but simple manul work.

 

[khorram]

I don't agree with you that the ASIC designer will have some problems in the future. As you mentioned, the tools will be better to overcome to the complexity problems. On the other hand, the complexity makes designers to do some activities manually. For a real example consider the software developement tools like Visual C++. A few years ago, Turbo C was used to develop a program and most of works should be done manually. After introducing object-oriented programming and Visual capabilities, now all of programmers work on different projects which need new features. We can expand it to the hardware design. After improving to the EDA tools, a designer will be able to do bigger projects.

Regards,
KH

 

[niuniu]

yes, i agree. don't worry about future, machine is always stupid, and there are always job for you to do. just remember to keep on learning. because no one can do everything, we need cooperate, everyone only need to be good at some specific field.

 

[andromeda]

Today, everybody make SoCs because chips are to big and functionality is to complex, nobody could develop them from beginning in reasonable amount of time with reasonable team size.
Average chip today includes several IPs, several embedded processors (with different complexity) and on-chip memories occupie 50+% of area. Number of IPs and embedded processors and memory % is expected to increase in following years.
This is the reason why companies are being restructured in a way that they organize IP teams for developing of IPs (in general to be used in company only, but eventually for sell to others also) and separate teams for chip integration.
IP development is in general more interesting than chip integration (for front-end engineers), but also easier for outsourcing (especially true for software IPs).

In following years ASICs it is expected that more things will go into software and SW/HW partitioning and co-simulation will be hot area. Because memory system organization for chips with embedded processors is critical it will be also one of the most interesting area. Architecture development for this kind of systems (several embedded CPUs) is interesting job. Usage of instruction-level simulators is expected to increase.
Usage of configurable CPUs (Tensilica, ARC) is expected to increase also (this approach change a way how things are designed).
Low-power system design is and will be extremely hot.

Whole methodology of ASIC design could be changed with wide usage of PSL (for properties specification), languages for executable specification, assertions …

From my point of view ASIC design is more challenging then ever and requires continual education. There is problem with outsourcing, of course, but the best way to have job security and high salary is to be the best.

 

[tukken]

It's a good explaintion, But I think the architecture and algorithm will do a core role in future digital design.

 

[sudhirtj]

Hi haosg,

Your concern is true about the future of design engineer. As the upturn and downturn are cyclic, it is same with respect to the designers. Some years back, there was huge requirement for the verification engineers (frontend). Now the trend is Physical Designers since the backend design in the VDSM and UDSM use to take considerable amout of time in design cycle. now since the tools are good enough to cut the design time, the industry is looking back for the verification side to cut the design time. This could be with respect to the resources.

The other thing is the ASIC design itself. Because of large number of IPs available, now we are going for the SoCs. Integration will be the main concern. Here you need to know from synthesis to gds. Other than this, R&Ds are going on in the field of Structural ASICs. That is one more future other than SoCs.

The next thing is the combining of physical domain and the manufacturing domain with PSMs at the physical domain itself. This is one more future.

One main thing what i have seen in the industry is you should never be in one domain atleast in your early career. you should have worked in architecture design, frontend, backend, physical verification and some knowledge of packaging and manufacturing. I was backend designer and worked in fab for creating IPs and now Integration engineer doing synthesis to gds.

-Sudhir

 

[mic_huhu]

To my honesty, I am worry about it too. I do not see the future of design engineer. I also found my job is easy so that everybody can do. The life is "yesterday once more".

 

[diablo1222]

The digital design isn't difficult to study, so I worry the furture. But for RF and analog , it is better than digital.

 

[kelvin_sg]

The digital design isn't difficult to study, so I worry the furture. But for RF and analog , it is better than digital.

It is not true. I do agree that digital design is easy to learn by itself. However, there is no digital designer that does *digital design* alone. Majority of them thoroghly understand the knowledge of the area of their working, for example wireless communication, often an engineer can both code, synthesis and write C/Matlab codes to simulate a wireless standard, say Bluetooth, 11A/B/G, etc.
It is the understanding of some field of study that appeals to the employer.

We learn Mathematics in order to work in engineering, and we learn digital
design in order to work in wireless communication. The former is a tool, and
later is a specialty.

People have to constantly upgrade themselves in order to stay competitive,
that applies to both digital, analog and RF.

 

[gerade]

what a hot discussion here! and very happy to learn from you!

In some degree we will all worried, and as when we worried, we will look for new solctions. and this is how the IC evolve itself today and has been since the first day.

As a IC engineer, We are facing a situation that like a ICs, we can develop some specialities, but the same time to lose some flexiblity, or reversely, and we are not going to be allmighty any more like the guys 20 or 30 years. But for sure good products will have good markets, in both directions, As the today's market will always require verisity and qualility, we should notice ourself of the new requirement from the market, and know what we want to do and achieve in long-term period.

And as an implemention designer, the future is for sure, as andromeda and sudhirtj(really nice guys right? ) mentioned, SOCs need you guys. the same time, you as implemention engineer need to know what is the new issues you will face, low power, mixed signal environment and so on. and same time know the interface to other engineer, know how to communicate with them.

in such way our worry today will become our fortune in future!

 

[ram]

yeah i too agree ... It needs a lot of effort to update with technology. Also now a days verification is hot. So this also needs to be learned. Also systemc is one who is initiating study..

 

[baso4hao]

i agree with gerade, for a IC designer we must think more rather than only the design.

 

[tony_taoyh]

After woking digital IC for three years, I have touched both front-end and back-end design.
My concept is, Digital IC is tools-related, heavilly.
What we need is:
(1) The concept is logic design;
(2) Some tips for special design: multiple clock domain and reset, scan.
(3) Know some basic rule in DIC.

Now I am beginning to touch Mixed-signal design, It is more experience-based...

For the future of the DIC enginner, am not sure.
What I think is, the change of all of us will constittute the change of the world.

 

[siawklai]

For me, i think the micro-architecture and algorithm (especially in DSP and security) will do a core role in future digital design.

 

[Zeng Jie-jun]

If you enter this field more, you will understand further, digital engineer should have rich experience on logical design, tool is easy to learn, but circuit knowledge should be accumulated for many years and personal skill.

 

[jjww110]

i agree with Zeng Jie-jun, tool is easy to learn, but circuit knowledge and personal skill should be accumulated for many years

 

[salma ali bakr]

why worry about future guys....if we remain on worrying, then we won't do our best in the present and worrying will just keep us back...we get updated always with the future hot topics but without notice maybe, and engineers in general know where to go and how to go there...no need to worry...!!!

 

[test_out]

yeah, too many thing to learn to become an designer. Help me!!!!

Where should I start!