Whatever

[jamahir]

Cool. What CAD are you using for designing the processor? Are you defining your own NMOS and PMOS or using pre-existing cells/libraries?

Sounds like an interesting project.

nothing that forward until now... the design is still on paper... be that the processor signal-enable table or be that block diagram... once the fpga comes in, we will begin with other things.

and i am a designer, not engineer... so these heavy technical words like nmos and pmos will take some learning for me...

as for cells/libraries, i presume you mean designs for device blocks... well, everything is new within the processor.

please do reply if my reply is wrong.

 

[ghazi52]

 

[Donatello]

nothing that forward until now... the design is still on paper... be that the processor signal-enable table or be that block diagram... once the fpga comes in, we will begin with other things.

and i am a designer, not engineer... so these heavy technical words like nmos and pmos will take some learning for me...

as for cells/libraries, i presume you mean designs for device blocks... well, everything is new within the processor.

please do reply if my reply is wrong.

Lol okay, you can be a design engineer.....something i am looking forward to.
Cells/Libraries means you already have the standard transistor/diode/component designs ready. You can play with them or literally create your own, depending on where you want to get your chip fabricated. Every FAB has it's own design rules that you must adhere to. So for example, if you are working on the 350nm process, there is a certain limit to which you can modify the transistor before it becomes un-usable. If you go for a smaller process, like the more common 90nm, 65nm etc, it becomes expensive to get a prototype fabricated.

So you are basically implementing an architecture in FPGA?

I thought you were doing custom IC design. No problem, i just misunderstood.

If you can prove your credentials/academic work, you can sometimes get free HDKs and FPGA sample kits from likes of Texas Instruments, NI, etc......so why don't you try contacting them?

@Skull and Bones

Abey oye, this topic deserves a better thread than 'Whatever page number 3496'

 

[jamahir]

So you are basically implementing an architecture in FPGA?

fpga is faster and relatively cheaper way of demonstrating prototypes.

If you can prove your credentials/academic work, you can sometimes get free HDKs and FPGA sample kits from likes of Texas Instruments, NI, etc......so why don't you try contacting them?

actually, i have already contacted xilinx and have met their local authorized distributor ( in bangalore ) who have recommended me their development board model, ml509, which uses their virtex-5 fpga chip.

just that we have to arrange the 120,000+ rupees for the board...

there is the distributor's "university program" which is a subsidized pricing for the "student community"... must find a student then...

Lol okay, you can be a design engineer.....something i am looking forward to.

herr designuer enginuer... sounds good...

Cells/Libraries means you already have the standard transistor/diode/component designs ready. You can play with them or literally create your own, depending on where you want to get your chip fabricated.

well, component-level designs ( existing ) won't be useful in this project because of the clock-less design... once we have the xilinx fpga board and have the xilinx "ise webpack" design tools ( that xilinx offers for free download ) we will know what libraries from there can be used by us... other than that, at the moment i don't know what tools to use...

Every FAB has it's own design rules that you must adhere to. So for example, if you are working on the 350nm process, there is a certain limit to which you can modify the transistor before it becomes un-usable. If you go for a smaller process, like the more common 90nm, 65nm etc, it becomes expensive to get a prototype fabricated.

do you know the pricing arrangements of fabs?? particularly those in china or any in russia?? this will be important info to put in proposal documents.

this topic deserves a better thread than 'Whatever page number 3496'

true.

 

[jamahir]

@Horus @waz

why has my post ( Whatever | Page 3496 ) gone for moderation?? all it contains is discussion about microprocessor design...

@Donatello ^^^

 

[waz]

@Horus @waz

why has my post ( Whatever | Page 3496 ) gone for moderation?? all it contains is discussion about microprocessor design...

@Donatello ^^^

I have cleared it my friend. Carry on. It is a good post.

 

[jamahir]

I have cleared it my friend. Carry on. It is a good post.

thanks for the prompt reply and clearance...

 

[Donatello]

fpga is faster and relatively cheaper way of demonstrating prototypes.



actually, i have already contacted xilinx and have met their local authorized distributor ( in bangalore ) who have recommended me their development board model, ml509, which uses their virtex-5 fpga chip.

just that we have to arrange the 120,000+ rupees for the board...

there is the distributor's "university program" which is a subsidized pricing for the "student community"... must find a student then...



herr designuer enginuer... sounds good...



well, component-level designs ( existing ) won't be useful in this project because of the clock-less design... once we have the xilinx fpga board and have the xilinx "ise webpack" design tools ( that xilinx offers for free download ) we will know what libraries from there can be used by us... other than that, at the moment i don't know what tools to use...



do you know the pricing arrangements of fabs?? particularly those in china or any in russia?? this will be important info to put in proposal documents.



true.

Well, that is a lot of money for a board that will eventually be obsolete like all kits today. What i meant was, try contacting their office in USA. Maybe make a few phone calls? They are generally very good at coming up with some sort of solution. Tell them you don't have that kind of money, but you have a ready design to implement.

A clock less design would be considered as 'custom IC design'.....it's overlapping term.

I don't know about the exact pricing for FABs, but again, contact their representatives. You will need a lot more than a prototype and lots of $$$ to have any hope of them fabricating a single chip for you. They are generally more open towards building large orders, because processing silicon for just one wafer is too damn expensive.
Here are some FABs you can contact:
TSMC
UMC
X-FAB

X-fab is based in Germany, and they are generally the go to place for academic research type prototypes. Other fabs deal in large orders.

One IC that we needed fabricated for a bio implant cost 30,000GBP at X-Fab and that was on 0.35micron (350nm) process with a total lead time of 14 weeks.

But again, it depends on lots of things, from process size to the type of metal you are using for interconnects, to the number of layers you have and above all, the die size. The more they are, the more expensive.

Some foundries that can do IC fabrication:
MOSIS Fabrication Processes

You have to adhere to their strict timetables...a typical project process is like this:
The MOSIS Service: Vendors : GlobalFoundries : GF_TIMELINE

Though i am still confused on how you would go from prototyping on the FPGA to real time fabrication of the IC???
@Skull and Bones is that possible without the DRC and LVS checks?

 

[jamahir]

Well, that is a lot of money for a board that will eventually be obsolete like all kits today. What i meant was, try contacting their office in USA. Maybe make a few phone calls? They are generally very good at coming up with some sort of solution. Tell them you don't have that kind of money, but you have a ready design to implement.

good idea... will remove the red-tape of the local office/distributor.

A clock less design would be considered as 'custom IC design'.....it's overlapping term.

okay.

X-fab is based in Germany, and they are generally the go to place for academic research type prototypes.

that is a good tip.

One IC that we needed fabricated for a bio implant cost 30,000GBP at X-Fab and that was on 0.35micron (350nm) process with a total lead time of 14 weeks.

thanks for that... really helps in clearly knowing the kind of money needed, and to provide numbers to my colleagues/support here... so 30,000 pounds is 28+ lakh rupees ( 2.8+ million rupees ) for that particular process.

but i didn't understand how many chips or rounds of chips did that money buy for your project??

But again, it depends on lots of things, from process size to the type of metal you are using for interconnects, to the number of layers you have and above all, the die size. The more they are, the more expensive.

i suppose our focus on simplification and minimalization will cost us lot less than something that uses arm or sparc architecture.

Though i am still confused on how you would go from prototyping on the FPGA to real time fabrication of the IC???
@Skull and Bones is that possible without the DRC and LVS checks?

i admit to being ignorant about that... maybe skull would explain.

View attachment 161516

what does this picture mean??

 

[Leader]

haha... the moderator is okay with other members being called jerk.. he just corrected this guy.. who this moderator is anyway? @Horus @WebMaster just because I reported so I wanted to see if its okay, dont want to create a fuss though.

 

[Kompromat]

View attachment 161745

haha... the moderator is okay with other members being called jerk.. he just corrected this guy.. who this moderator is anyway? @Horus @WebMaster just because I reported so I wanted to see if its okay, dont want to create a fuss though.

Link?

 

[Parul]

@Prometheus He is duplicate ID of Manvantantri, I ignored his posts when he quoted me. Onw can make out from his ID and way of writing who he is, why you becoming his prey? Just Ignore him.

 

[rmi5]

@Kamil_baku
I don't know why the other thread was closed. here is my 2 cents:

Actually, Azerbaijan will only supply 16 billion m3 gas in the beginning and will increase it to 50 billion by 2023. This number(50) is nothing compare to European needs. its not even 5% of their needs. Also, EU purchase almost 10% of their gas as LNG which is twice the price of gas they can import by pipeline. Turkey will need way more gas pipelines in the future to supply EU, so its funny to compare Russian pipeline with Azerbaijani. Even two of them together is not enough to supply EU. Turkish demand for gas also growing... Turkey wants new pipelines from Qatar,(probably after syrian conflict finish, another reason west want to continue it and dont agree with turkey for no fly zone. they want to push Turkey to purchase it from Israel and South Kibris..

As americans say, too many chiefs, not enough indians
European economy has a slow GDP growth, so it's gas needs is not increasing much. Even for Turkey, their gas consumption has been decreased compared to a couple of years ago, and is not increased. We need to talk based on actual economical facts, not vague baseless claims. Also, Europe already produces a big amount of gas, and imports from North Africa(Algeria), and Russia, and is fully saturated, specially for western european market. The only new markets are south eastern europe, like Croatia, and in some lesser extent South Europe, like Italy. This market is very limited as well. Their other concern is having another pipeline so they can increase their voice against Russia. Russia's strategy was planning another pipeline to South East europe, to fully drive us out of the market. But, the crimea crisis, ... helped us to move Southern Corridor project forward. Then, Russians have hastily proposed this plan, which even the finances and its routes, and share holders are not defined, to just futile Southern Corridor. Anyway, They will either 1. cancel this project, or 2. southern corridor will be canceled or have very long delays, or 3. they will ask for some of our share in Southern corridor to merge the two pipeline projects

Guys, no need to prove our brotherhood. I give you one example.. in 1915, Armenians started huge genocide in Azerbaijan. They wanted to kill everybody. When Ataturk was told about it, he said there is no difference if Turks die there or Turks die here. He sent his army to defend the nation. Thanks to them, there is a country called Azerbaijan now, Azerbaijan also became first democratic republic in the east (1918) and collected tons of gold to send Turkey as a support. This is how it should be.
Only together, we can develop and have huge influence on Russian speaking middle asian countries too.

What Ataturk has to do with Erdogan?

 

[Pboy]

Hello...........

@Akheilos

What happened with the crazy kid alpha1?

 

[Kompromat]

हाँ नहीं आती हिंदी भार जा तू

What does this mean?

@levina