Government to provide $1 billion to every chip maker who sets up manufacturing unit in India.

[Death Professor]

In Pakistan?

On mars.

Only a retard would ever compare developing countries, without taking their population into consideration. Your direct competition with any thing related to RnD or manufacturing or production is China not us. We compete you in only one thing and that's military and that is just for OUR survival and balance of power. Nothing more, nothing less.

Don't bother replying, I get it, if this is the first response how retarded the other replies would be.

 

[TigerOfMysore]

What members of PDF need to understand that especially in the last 2+ decades the West has made and continues to make significant investments into India. Many of these investments are in Pharma and HiTech. Every country that is part of the western bloc considers India a critical part of their strategy. You can argue that there is nothing significant around India but her effective policies around Education is paying dividends. Furthermore her market size (real potential is questionable) also makes her an attractive location. Now you add in the mix of the China tensions we are likely to see an expansion and an acceleration of these Western bloc countries into India. This has paid huge dividends for the Indian economy flushing her with massive FDI and export oriented revenue especially in Hi-Tech. This growth is likely to accelerate. Pakistan factors in as a spoiler nation for these western countries. Our Muslim identity does not earn us any points either (given the current and ongoing anti-muslim climate), and at this point we have an openly adversarial relationship with the Western bloc. The Western Bloc is actively pursuing policies that hurt our strategic and short terms goals. Though in the past strategic alignment was limited, short term expediency allowed for some working relationship. That has changed as well. The Western bloc is actively working against the sovereign interest of Pakistan and have been for some time. Pakistan needs to wake up and smell the coffee. If we don't solve the governance issue, we are likely to a die a slow death with a thousand cuts such as but not limited to:

1) Lack of national cohesion and narrative
2) Severe water crisis spurred on by India and poor conservation strategies
3) Active subversion by ethno/religious and secular factions - people don't realize at the extent of external penetration into this group. It is shocking at the amount of foreign money in the political process
4) Military priorities and expenditure that is over sized to our capability & resources
5) Poor educational foundation and limited R&D/Research base and higher Education
6) Hot borders driving and consuming limited resources
7) Development and growth of mafias in many sectors of the country that no government has managed to take on that destroys competition and tax base
8) From a net exporter to now a net importer of Agri products and its impact of FX reserves
9) International pressure through FATF / WB / IMF that drives policies antithetical to PK interests while increasing the fiscal imbalance
10) Poor fiscal performance & tax base given to larger and larger debt to rev ratios
11) Over sized imprint of Military on industries allowing for stifling of organic and private sector development (reminds me of the old PTCL model - that stifled telecom for generations)
12) Continuous brain drain of some of the best minds in the country
13) Highly corrupted judiciary that makes legal balance and accountability very difficult
14) NGOs and foreign entity ingress creating confusion and mistrust in the youth, while building an imbalanced social fabric
15) And last but not least a Bureaucratic system that makes change almost impossible, while aligning and enabling a trifecta of corruption between the three counter parties of politics, judiciary and the bureaucracy.

We have to in parallel hit all of the above issues, short of which we'll keep sliding into this abyss.

 

[Type59]

Just propaganda. Seeing how corrupt bjp is, investors will be forced to pay hundreds of millions in bribes to get project running. 10 years ago, Indians were saying a number of plants were gonna be set up. So won't hold my breath.

 

[Hephaestus]

On mars.

Only a retard would ever compare developing countries, without taking their population into consideration. Your direct competition with any thing related to RnD or manufacturing or production is China not us. We compete you in only one thing and that's military and that is just for OUR survival and balance of power. Nothing more, nothing less.

Don't bother replying, I get it, if this is the first response how retarded the other replies would be.

Your rant aside. I can understand your frustration. I will reply nevertheless. You are so hell-bent on proving that India is a failure in this field and waving the it's hilarious. Chinese don't need you to do their shouting. They are pretty adept at doing it themselves.

You don't compete with us in anything. Hell, you can't even compete with Bangladesh. 212 million people & your literacy rate after 70 yrs of independence is still 59% and you come here & spout inane nonsense.

We will compete & we will get better and better at this as we have in all other fields. We are ready to dole out sops for manufacturers. Hell, we are ready to dole out a few $B. We are not looking and rejoicing at getting handouts for $500M.

Talking about Mars - just as a rejoinder - We have a Mars Orbiter Mission which is still going strong after 6 years.
Which was called "The Pride of Asia"... you can find out by whom?

 

[xyxmt]

That is old tech from 2001 period.

you cannot get free technology unless its 20 years old

 

[Hephaestus]

Just propaganda. Seeing how corrupt bjp is, investors will be forced to pay hundreds of millions in bribes to get project running. 10 years ago, Indians were saying a number of plants were gonna be set up. So won't hold my breath.

ham ko maalum hai......

dil ke khush rakhne ko 'ġhalib' ye khayāl achchha hai

 

[Mirzali Khan]

I will start making chips in India XD

 

[khansaheeb]

jealous neighbors need burnol .

Pakistan defense companies make their own chips/processors on a small scale and not on a global distribution level like AMD, Intel, NEC etc

 

[Arulmozhi Varman]

Sorry but this group at IIT-Madras is not being truthful about the project being "fully indigenous". This series of processors called Shakti is simply the group's implementation of the international, open source project RISC-V.

Many other people in the world are doing it. No great shakes.

It would be more respectable if this IIT-Madras group actually designed a processor from scratch.



Firstly, thank you for tagging me. This field - processor design - is my current interest.

Secondly, about what you said, I don't think Ambani and Adani will invest in this field. They prefer safe investments. Or maybe they will, I don't know, because there is much money to be made by selling processors and memory to local manufacturers.




Well yes, though I am not an electronics engineer I too agree that 130-200 nm is much bigger than say the Leon processor from the ESA ( European Space Agency ) that uses a 32 nm process.



Like which giants ?

It is a from scratch design. 99% of the semis are based on existing architecture. And that architecture is really important for programming. You see when you take a chip to manufacturing to a third party chip manufacturers, they already have fabs on existing architecture. Just like you develop an app for Android or Apple. But you don't need to develop an smartphone ecosystem to develop an app. It's as simple as that. But to develop that app or chip in this case, it is still an capital intensive R&D, dev and testing process. World's top chip companies including Qualcomm or Broadcomm use ARM architecture which is a UK company. RISC V is an upcoming architecture which is new and cost effective. Intel has their own architecture since they make their own chips.

It would be great if many people have an idea before they comment on indigenous stuff. I can personally say Indian R&D in chip design and testing is world class with major companies based in Bangalore and Chennai. But we don't have manufacturing capabilities yet.

 

[Paro_Peagus]

Pakistan defense companies make their own chips/processors on a small scale and not on a global distribution level like AMD, Intel, NEC etc

Liar liar pants on fire.

please name the small scale industry that manufacturers the chips for the military .

I can tell you where they come from but ill leave it to your Chinese frnds to break your bubble. The factory name is a mandarin tongue twister for us

 

[jamahir]

Why do they need to reinvent the wheel? you understand how spec standard work right?
Why new instruction sets when millions of manhours have been spend refining those we have?

That is quite an Indian corporate argument.

One should "reinvent the wheel" in this case mainly because it is intellectually rewarding but also because very few people have achieved it and also one has something to contribute to the world.

In my own processor design project it took me a few years to simplify the processor instructions and the I/O hardware specification. I still have to design the GPU which I think will create need in the general-purpose cores for more instructions. I am learning as I am proceeding.

What I won't be doing is porting C language to the assembly instructions because the assembly will be simple enough.

What I also won't be doing is porting Linux to the processor. I will be writing a microkernel-based OS, for which I have already specified some system calls.

As for app ecosystem my plan is to write a virtual machine for x86.

The next leap is quantum, and the challenge is to being it to the same spec as modern widely adopted interfaces, as well as high robustness, that we enjoy with modern computations on non quantum machines.

From what little I know of quantum computin, the challenge seems to be miniaturizing the hardware and getting it to run at room temperature.

So I think for the next ten years quantum computing will be available as a cloud service.

It is a from scratch design.

RISC-V ? It has been designed from scratch outside India. IIT-Madras is just implementing here. No great shakes. There are others in the world also doing that too.

99% of the semis are based on existing architecture. And that architecture is really important for programming. You see when you take a chip to manufacturing to a third party chip manufacturers, they already have fabs on existing architecture.

Why should that stop someone from designing a new thing, in this case a processor ?

But to develop that app or chip in this case, it is still an capital intensive R&D, dev and testing process.

Designing a processor needs some place to jot the ideas, like a printout or a computer file. Prototyping that processor requires initially a FPGA board which costs about two lakhs or so. I am sure such things are available in India's colleges and companies.

 

[Splurgenxs]

That is quite an Indian corporate argument.

One should "reinvent the wheel" in this case mainly because it is intellectually rewarding but also because very few people have achieved it and also one has something to contribute to the world.

Well corporate interests drive industry , companies need to be secure in there cash flow and the do that by holding back until absolutely necessary. You think they cant add more cores to existing designs ?

Does adding more cores make stuff faster? , not if the software hasten leveraged it in there development.
A large part of software even now run on the main thread.
Innovation for the sake of innovation is not organic or cost-effective.

In my own processor design project it took me a few years to simplify the processor instructions and the I/O hardware specification. I still have to design the GPU which I think will create need in the general-purpose cores for more instructions. I am learning as I am proceeding.

whatever you are doing on your own , is a non silicon, theoretical PD scenario or a large electrical bread board demonstrator, I am assuming.

What I won't be doing is porting C language to the assembly instructions because the assembly will be simple enough.

What I also won't be doing is porting Linux to the processor. I will be writing a microkernel-based OS, for which I have already specified some system calls.

As for app ecosystem my plan is to write a virtual machine for x86.

Well you are just skipping having to write a C compiler , Im assuming , which tokenizes the tasks , converts to instruction sets/assembly and delegates to the OS drivers.

Now since your machine will be unique, so the drivers and OS interface will have to be unique too since you are not going with the standards.
I think is what you are trying to say.

Is this funded ? or your own R&D?

From what little I know of quantum computing, the challenge seems to be miniaturizing the hardware and getting it to run at room temperature.

So I think for the next ten years quantum computing will be available as a cloud service.

Well miniaturization is much later, the current challenge is computation robustness.

if you are dealing with quantum states , the states are finicky and will have issues with holding the info correctly and prone to absorbing entropy from external factors. This makes the storage and processing not robust , to reduces overall entropy they are put in extreme cold states.

Simple analogy
Current arch :
You want is 2+2 =4(since all the bits will hold charge robustly ,especial when dealing with attenuation and noise, local and ambient.)

Quantum Arch.
Not 2+2 =5,(when external factors incite a Qbit to change states especially when trying to read it. Invoke mister Heisenberg)

The Robustness, you enjoy with the current arch , with respect to the issues with cache misses , signal latency , and context isolation to name a very few , was made possible over decades by thousands of people across domains.

 

[VkdIndian]

What members of PDF need to understand that especially in the last 2+ decades the West has made and continues to make significant

Quite an apt analysis of the situation. A lot of issues that you have pointed out plague India too. High population and few other factors make the progress a challenge. There have been few radical moves by the Indian Government to give an impetus to growth.

Introduction of GST was one such step. It is showing signs of change.
Challenge is big though.

 

[jamahir]

Well miniaturization is much later, the current challenge is computation robustness.

if you are dealing with quantum states , the states are finicky and will have issues with holding the info correctly and prone to absorbing entropy from external factors. This makes the storage and processing not robust , to reduces overall entropy they are put in extreme cold states.

Simple analogy
Current arch :
You want is 2+2 =4(since all the bits will hold charge robustly ,especial when dealing with attenuation and noise, local and ambient.)

Quantum Arch.
Not 2+2 =5,(when external factors incite a qubit to change states.)

The Robustness, you enjoy with the current arch , since all the issues with cache misses , signal latency , and context isolation to name a very few was made possible over decades by thousands of people across domains.

Much thanks for the explanation.

Now since your machine will be unique, so the drivers and OS interface will have to be unique too since you are not going with the standards.
I think is what you are trying to say.

Firstly, the processor will be clockless so right there I am going away from current commercial standards.

Next, the hardware will include very few, fixed I/O interfaces ( USB, GPIO, WiFi etc. No PCI interface ). Since the OS will be microkernel-based the drivers will be user-mode programs which will use message-passing IPC to allow operations on the hardware on behalf of other programs.

The idea is to make the setup simple. And the surrounding system modular.

Is this funded ? or your own R&D?

What I know is that FPGA boards cost about 1.5 lakhs and then there will be the engineers. To obtain that kind of money I have a plan which I can't put out on the forum.

whatever you are doing on your own , is a non silicon, theoretical PD scenario or a large electrical bread board demonstrator, I am assuming.

ATM it is theoretical.

Does adding more cores make stuff faster? , not if the software hasten leveraged it in there development.
A large part of software even now run on the main thread.
Innovation for the sake of innovation is not organic or cost-effective.

Multi-core systems have their own issues, both beneficial and not beneficial. It all depends on the structuring of the software. For example, will a multi-threaded driver process run better if it has a single thread or will it run slower if it is multi-threaded since there will have to be mutexes guarding common resources ?

---

Thanks for the vid. Will watch later.

 

[Arulmozhi Varman]

That is quite an Indian corporate argument.

One should "reinvent the wheel" in this case mainly because it is intellectually rewarding but also because very few people have achieved it and also one has something to contribute to the world.

In my own processor design project it took me a few years to simplify the processor instructions and the I/O hardware specification. I still have to design the GPU which I think will create need in the general-purpose cores for more instructions. I am learning as I am proceeding.

What I won't be doing is porting C language to the assembly instructions because the assembly will be simple enough.

What I also won't be doing is porting Linux to the processor. I will be writing a microkernel-based OS, for which I have already specified some system calls.

As for app ecosystem my plan is to write a virtual machine for x86.



From what little I know of quantum computin, the challenge seems to be miniaturizing the hardware and getting it to run at room temperature.

So I think for the next ten years quantum computing will be available as a cloud service.



RISC-V ? It has been designed from scratch outside India. IIT-Madras is just implementing here. No great shakes. There are others in the world also doing that too.



Why should that stop someone from designing a new thing, in this case a processor ?



Designing a processor needs some place to jot the ideas, like a printout or a computer file. Prototyping that processor requires initially a FPGA board which costs about two lakhs or so. I am sure such things are available in India's colleges and companies.

You just totally ignored on technical posts where I had given you idea on what is going on in dev process. I do not expect learn to know all. But when someone gives you an idea on how the tech field works with humility one should accept except of course if one has more experience in that said field than me.

As for new one, as why shouldn't one create an totally different architecture which is new, who will take that product to manufacturing? FYI India doesn't have a single big fab for manufacturing. Contrary to popular belief India has many small fabs doing small scale chip production of timers, oscillators, logical ICs etc but not to make processors etc.

So generally if you or me make an IC, and take it to third party vendor like TSMC, there is a process something like DFM and NPI. They will make sure you do the design based on already existing architecture like ARM or RISC V. Even RISC V hasn't been implemented in all manufacturing locations. So the cost of manufacturing is cost effective. Or else they will ask you to fund manufacturing tech development for new architecture as well and it will be damn costly as hell.