Featured Project Azm: Pakistan's Ambitious Quest to Develop 5th Generation Military Technologies.

[Goenitz]

I do not agree with your assumption that you need more PhDs. Unless you are doing pure research and the intention is churn out few research papers then sure PhDs would be ideal.

Even PhDs are not enough. You need a researcher with sufficient industry or post-doc experience. But you need PHDs. The research work broaden your mind. Post-doc researchers can monitor the output, optimise and can give new solutions.

But to produce them you need engineers (i dont care much about the degrees) who have experience in the industry and know best practices

Your are right that once the production/prototyping/manufacturing starts, you don't need researchers.. Apollo rocket computer was made by women bcz they were good at stitching/knitting.

 

[Bilal.]

Money talks and if you are willing to pay good dime you can always go to ASML. That is where China got its latest fab as well, although they are now doing their own research as well.

But don't discount the gulf. If you knew what I know, you wouldn't Unfortunately, revealing what I know leads to revealing names. And that is something I do not want to do under any circumstances.

For our defence needs, we don’t need the cutting edge ASML equipment. A 90nm fab should be quite sufficient. Having said that Chinese are close to qualifying ArF lithography machines which will enable 28nm and eventually even 14, 10 and 7nm in 3-4 years. Additionally, EUV research may reap results within a decade.

 

[JamD]

To be honest I don't really care whether they are PhDs or Masters or Bachelors or nothing. You can hire 80 PhDs with no experience in the industry or product development and the output will be nothing. It does not mean you do not need PhDs, you do need them because if you want a very specific knowledge of a particular domain the most eligible candidate would probably have a PhD. But usually you do not need many.

Point well taken. However, I will maintain that you can and should hire more PhDs, especially since they are becoming increasingly numerous - more and more people are getting PhDs.

You gave example of Raytheon and Lockhead Martin and it is true that they have a lot PhDs but they are not hired because they have PhDs, they are hired because they have vast practical knowledge of a specific domain.

Actually, in the R&D departments they are hired primarily if they have PhDs. Their high-tech research side is almost all PhDs. I know this because I have friends at Raytheon, LockMart, GE, and Boeing. These companies have dedicated research centers which are primarily staffed by PhDs. They have masters and undergrads for grunt work in research (like doing lots of CFD). I also know this because I have gone through their hiring process and know their requirements and work environment intimately.

In Pakistan the PhDs do theoretical research in academia and do not have any practical experience of any industry what so ever. As you have to hire them locally they are not the ideal candidates to begin with. Not many PhDs or engineers with the requisite experience will be willing to come to Pakistan to get hired in AvRid.

Don't even get me started on the quality of Pakistani PhDs. That deserves its own thread/rant lol. I agree that our PhDs are marginally better than our undergrads and suffer from the same issues of experience. I should've said foreign PhDs, which is what I really meant. There are many foreign PhDs that would come to AvRID and contribute if the right environment exists. I am one, and I know a couple at least.'

That is where private sector shine as they can offer a package depending upon the candidate rather than painting everyone with the same brush.

Funny you should say that. There have been hires at AvRID like that with negotiated packages. So that's one positive thing I've seen there.

Unexperienced PhDS are not going to guide what so ever other unexperienced fresh graduates. The only way you solve this issue is hiring experienced people and actually listening to them on technical as well as management level. The degrees don't help here what so ever.

Agreed mostly, but I will not go entirely azad chaiwala and claim degrees are no help whatsoever.

You are absolutely right. I don't think AvRID with its government scale structure and baboo mentality can build something complex on limited time scale. You need to adhere to modern way of working and hiring principles to achieve that. Providing funding and even creating semi-private institutions through AvRID funding which don't have governmental red tape can achieve it.

Right. We need policy level rethinks, not just fancy names and doing more of the same.

 

[sal29]

For our defence needs, we don’t need the cutting edge ASML equipment. A 90nm fab should be quite sufficient. Having said that Chinese are close to qualifying ArF lithography machines which will enable 28nm and eventually even 14, 10 and 7nm in 3-4 years. Additionally, EUV research may reap results within a decade.

Sure we can invest in fab just for military if we have 100 Billion dollar budget. But with our budget I would rely on Chinese supply of electronic components and build capacity of developing our own systems around it. The Chinese already have capacity to build up to 14nm I think and they are investing heavily in EUV (although you dont need EUV for military applications as the EUV fab produces cutting edge application processors which are used in latest gen laptop and mobiles application processor not the one used in military (at least not yet). The military uses older technology which has become much more reliable over time.

We should invest in fabs with Chinese cooperation but not specifically for military but for mass market. With that we can develop capacity over time for military applications.

 

[Bilal.]

Sure we can invest in fab just for military if we have 100 Billion dollar budget. But with our budget I would rely on Chinese supply of electronic components and build capacity of developing our own systems around it. The Chinese already have capacity to build up to 14nm I think and they are investing heavily in EUV (although you dont need EUV for military applications as the EUV fab produces cutting edge application processors which are used in latest gen laptop and mobiles application processor not the one used in military (at least not yet). The military uses older technology which has become much more reliable over time.

We should invest in fabs with Chinese cooperation but not specifically for military but for mass market. With that we can develop capacity over time for military applications.

Yes, if you see my previous post I said we can leverage the Chinese fabs (which most likely we already are). My suggestion was not for a cutting edge node fab. A 45 or 90nm fab with limited production capacity would cost a couple of hundred million dollars. Same as the price tag of a single frigate.

 

[sal29]

Point well taken. However, I will maintain that you can and should hire more PhDs, especially since they are becoming increasingly numerous - more and more people are getting PhDs.

Sure if they are there then why not. But I think you can also find very competent Bachelor and Master degree holder with industry experience provided the packages are good.

Actually, in the R&D departments they are hired primarily if they have PhDs. Their high-tech research side is almost all PhDs. I know this because I have friends at Raytheon, LockMart, GE, and Boeing. These companies have dedicated research centers which are primarily staffed by PhDs. They have masters and undergrads for grunt work in research (like doing lots of CFD). I also know this because I have gone through their hiring process and know their requirements and work environment intimately.

As I mentioned its because of there knowledge and experience not because they have PhD but I understand your point. I think that is mostly because they invest heavily in next generation weapons research. We have to yet to develop base to be able to design and develop this generation products

Don't even get me started on the quality of Pakistani PhDs. That deserves its own thread/rant lol. I agree that our PhDs are marginally better than our undergrads and suffer from the same issues of experience. I should've said foreign PhDs, which is what I really meant. There are many foreign PhDs that would come to AvRID and contribute if the right environment exists. I am one, and I know a couple at least.'

You are right creating the right environment is key. This old school notion of government scale does not bode well in this new way of working in industry. Also I think this secrecy NOC and all that stuff also creates a hurdle. It should made easy if someone need to leave or something there shouldn't be a river you have to jump across.

Funny you should say that. There have been hires at AvRID like that with negotiated packages. So that's one positive thing I've seen there.

That is good news. I hope they do this more often to get the necessary people.

Agreed mostly, but I will not go entirely azad chaiwala and claim degrees are no help whatsoever.

I never said they are no help. I said I don't care for them much. I care about experience, knowledge and competency. If you have those it is most likely you have some form of formal education, hence a degree.

Right. We need policy level rethinks, not just fancy names and doing more of the same.

Agreed.

 

[CriticalThought]

For our defence needs, we don’t need the cutting edge ASML equipment. A 90nm fab should be quite sufficient. Having said that Chinese are close to qualifying ArF lithography machines which will enable 28nm and eventually even 14, 10 and 7nm in 3-4 years. Additionally, EUV research may reap results within a decade.

I am not sure where you get that 90nm number and why you are so focused on solely Chinese options. Don't get me wrong, the Chinese are awesome, but I don't know what you have against the Gulf nations.

 

[sal29]

Yes, if you see my previous post I said we can leverage the Chinese fabs (which most likely we already are). My suggestion was not for a cutting edge node fab. A 45 or 90nm fab with limited production capacity would cost a couple of hundred million dollars. Same as the price tag of a single frigate.

I understand but what exactly are we going to do with that fab. Having a fab does not mean you can produce your own processors. For that you have to invest also in different domain which is processor architecture development. You also have to invest in translating that architecture into silicon. You have invest in packaging of that silicon. You have to invest in producing tooling (compilers, debuggers, linkers etc) for that architecture. You can always license an architecture of course but you can also just also buy it then. If you license it then you still have an issue as you do not control the entire supply chain. I think investing in fab is very good for Pakistan but we dont need it for military purpose at least not now. We do need it so that we can develop a commercial and leverage it for military use like every other country does. It is not as simple as investing couple of hundred million dollars. You have to invest it in workforce development which does not come in an year or even 5 years unlike a frigate.

 

[CriticalThought]

I understand but what exactly are we going to do with that fab. Having a fab does not mean you can produce your own processors. For that you have to invest totally in different domain which is processor architecture development. You also have to invest in translating that architecture into silicon. You have invest in packaging of that silicon. You have to invest in producing tooling (compilers, debuggers, linkers etc) for that architecture. You can always license an architecture of course but you can also just also buy it then. If you license it then you still have an issue as you do not control the entire supply chain. I think investing in fab is very good for Pakistan but we dont need it for military purpose at least not now. We do need it so that we can develop a commercial and leverage it for military use like every other country does. It is not as simple as investing couple of hundred million dollars. You have to invest it in workforce development which does not come in an year or even 5 years unlike a frigate.

The RISC-V work happening in Pakistan is the answer to a lot of your concerns.

 

[Bilal.]

I am not sure where you get that 90nm number and why you are so focused on solely Chinese options. Don't get me wrong, the Chinese are awesome, but I don't know what you have against the Gulf nations.

The number I got is from inference. India is utilizing 180nm for production of its indigenous defence and space application chips. So a bit of future proofing would land us at 45-90nm fab requirement.

As for the unhinged comment on whether I have or not have something against any country or region, well I don’t and I don’t know how you reached the conclusion, let’s have positive discussion like we are having instead of devolving down that path.

 

[sal29]

The RISC-V work happening in Pakistan is the answer to a lot of your concerns.

Yes I know about RISC-V but the main company behind it SiFive is not Pakistani. I think they might have R&D center in Pakistan (not sure). Even with RISC-V in Pakistan it would only satisfy just a very small part of a complex system. I would invest heavily in commercial fabs for mass market to get the work force trained and have revenue.

 

[Bilal.]

Yes I know about RISC-V but the main company behind it SiFive is not Pakistani. I think they might have R&D center in Pakistan (not sure). Even with RISC-V in Pakistan it would not satisfy just a very small part of a complex system. I would invest heavily in commercial fabs for mass market to get the work force trained and have revenue.

No some good work being done by MERL, Lampro Mellon and probably others.

Micro Electronics Research Lab (MERL)

Accelerating Engineering Innovation

merledupk.org

Lampro Mellon | Design Tomorrow, Today

lampromellon.com

 

[sal29]

No some good work being done by MERL, Lampro Mellon and probably others.

Micro Electronics Research Lab (MERL)

Accelerating Engineering Innovation

merledupk.org

Lampro Mellon | Design Tomorrow, Today

lampromellon.com

Are they Pakistani .? .. if so then that is very good news. As per my understanding Lampro Mellon is not a design firm but rather a design verification and compliance firm. They deal with architecture issues but I don't think they deal with architecture to silicon and fabrication process.

 

[Bilal.]

Are they Pakistani .? .. if so then that is very good news. As per my understanding Lampro Mellon is not a design firm but rather a design verification and compliance firm. They deal with architecture issues but I don't think they deal with architecture to silicon and fabrication process.

Yes they are Pakistani. MERL has designed a couple of chips as you can see in their website. They are fabless entities. MERL is now involved in OSFPGA initiative too.

 

[sal29]

Yes they are Pakistani. MERL has designed a couple of chips as you can see in their website. They are fabless entities.

At this point I am happy if any work is going on. I hope government invests in this sector. This is something crucial for this century and we don't have the knowledge base for it.