ASIC Development SHA 256

[aeronautical]

I'm an ex Cadence design systems engineer, just thought i would post this here. I don't understand what the fuss is about
developing a SHA 256 ASIC and why many more have not already done it.


ASIC design IP for SHA 256

Good start is to download the chip estimate software developed by Cadence.
http://www.chipestimate.com/  lookup SHA 256

http://www.chipestimate.com/ip.php?256-bit+SHA+Cryptoprocessor+Core&id=14006
This looked good the TSMC core is also available that many are working with.
We use Cadence and Altium electronic CAD, here ready for hire!!!

Thanks
Aeronautical Pty Ltd

[DarkPunk]

I'm an ex Cadence design systems engineer, just thought i would post this here. I don't understand what the fuss is about
developing a SHA 256 ASIC and why many more have not already done it.


ASIC design IP for SHA 256

Good start is to download the chip estimate software developed by Cadence.
http://www.chipestimate.com/  lookup SHA 256

http://www.chipestimate.com/ip.php?256-bit+SHA+Cryptoprocessor+Core&id=14006
This looked good the TSMC core is also available that many are working with.
We use Cadence and Altium electronic CAD, here ready for hire!!!

Thanks
Aeronautical Pty Ltd

The problem isn't the cores, almost any electrical design hobbiest can design a SHA256(SHA256(x)) core with the help of google and some opensource IPs, its the massive overhead that is the hurdle.  Not many here have the 6-figure bank account to make a mass order if ASICs, nor the man power to put all of them to work.

Hell, if you want to find the capitol, and are willing to sell singles, I'll buy a few, but anything beyond that takes more than most have.

[aeronautical]

I been looking into this for about a week and it's been fun.
I think considering the money or bits or shit whatever it's worth a very close look funny we have our extra PC's here grinding away think they call it mining.
Another spot is www.opencores.org they also have SHA 256 .
I would go for an IP core not sure what one maybe the TSMC looked like the best speed, Multiplex < Spartan > (USB) or Ethernet.

We are also a Microchip design partner but they don't have anything with enough power.

[DarkPunk]

I been looking into this for about a week and it's been fun.
I think considering the money or bits or shit whatever it's worth a very close look funny we have our extra PC's here grinding away think they call it mining.
Another spot is www.opencores.org they also have SHA 256 .
I would go for an IP core not sure what one maybe the TSMC looked like the best speed, Multiplex < Spartan > (USB) or Ethernet.

We are also a Microchip design partner but they don't have anything with enough power.

The crypto core on opencores.org is FPGA based, your not going to break any banks there, but it's nothing a passing hobbiest can't manage either, more over, it's designed for a single SHA256 hash, you would have to write a script to pass every hash twice before moving on...

[aeronautical]

I'm an ex Cadence design systems engineer, just thought i would post this here. I don't understand what the fuss is about
developing a SHA 256 ASIC and why many more have not already done it.


ASIC design IP for SHA 256

Good start is to download the chip estimate software developed by Cadence.
http://www.chipestimate.com/  lookup SHA 256

http://www.chipestimate.com/ip.php?256-bit+SHA+Cryptoprocessor+Core&id=14006
This looked good the TSMC core is also available that many are working with.
We use Cadence and Altium electronic CAD, here ready for hire!!!

Thanks
Aeronautical Pty Ltd

The problem isn't the cores, almost any electrical design hobbiest can design a SHA256(SHA256(x)) core with the help of google and some opensource IPs, its the massive overhead that is the hurdle.  Not many here have the 6-figure bank account to make a mass order if ASICs, nor the man power to put all of them to work.

Hell, if you want to find the capitol, and are willing to sell singles, I'll buy a few, but anything beyond that takes more than most have.

Sorry missed that comment typing at the same time, agree it is expensive designing new electronics. I think that there is also a lot of BS in it and scammers it's bad for Bitcoin and the mums and dads that want a simple solution.
But the underlying code, idea looks great or fantasy!! If you have soem Hardware ideas please post I'II have a look, right no just need a block diagram. I don't think that is so hard though.

[Gabi]

I don't understand what the fuss is about
developing a SHA 256 ASIC and why many more have not already done it.

It takes millions to start making an ASIC. No matter if you design the chip, then you have to get the masks etcetc, and that means money, a lot of it.

[aeronautical]

I been looking into this for about a week and it's been fun.
I think considering the money or bits or shit whatever it's worth a very close look funny we have our extra PC's here grinding away think they call it mining.
Another spot is www.opencores.org they also have SHA 256 .
I would go for an IP core not sure what one maybe the TSMC looked like the best speed, Multiplex < Spartan > (USB) or Ethernet.

We are also a Microchip design partner but they don't have anything with enough power.

The crypto core on opencores.org is FPGA based, your not going to break any banks there, but it's nothing a passing hobbiest can't manage either, more over, it's designed for a single SHA256 hash, you would have to write a script to pass every hash twice before moving on...

Your right FPGA is not going to do it, looked at that too slow and cost is high. I think the CAST or TSMC that runs at 500mbps?

[2x0ninja]

I bet a transparently developed AISC would go nuts on Kickstarter if BFL doesn't deliver soon...

[aeronautical]

I don't understand what the fuss is about
developing a SHA 256 ASIC and why many more have not already done it.

It takes millions to start making an ASIC. No matter if you design the chip, then you have to get the masks etcetc, and that means money, a lot of it.

Disagree it's not millions, but it is ~200k to get the foundry warmed up plus design charges. TSMC Shuttle service..
http://www.tsmc.com/english/dedicatedFoundry/services/cyberShuttle.htm

 

[DarkPunk]

I bet a transparently developed AISC would go nuts on Kickstarter if BFL doesn't deliver soon...

I don't mind developing an ASIC mining chip, but there is no way I have the funds to get them made.  I'm sticking to my Spartan 6's for now.

[silzero]

The processing of one 512-bit block is performed in 66 clock cycles and the bit-rate achieved is 7.75Mbps / MHz on the input of the SHA256 core.

Its important to note performance/SHA-256 IP is not the issue. The issue is getting everything else together in a cost effective manner.

-SZ

[aeronautical]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

[DarkPunk]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

Agreed,
https://bitcointalk.org/index.php?topic=170977.0

[aeronautical]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

Agreed,
https://bitcointalk.org/index.php?topic=170977.0

It good, but looks like most of that thread are talking FPGA on PCI and i don't think that is going to do the job, too slow. Using an FPGA is a good idea if power is a problem but our office graphics cards are doing more work. So moving forward i think an ASIC is the only way to go.

[DarkPunk]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

Agreed,
https://bitcointalk.org/index.php?topic=170977.0

It good, but looks like most of that thread are talking FPGA on PCI and i don't think that is going to do the job, too slow. Using an FPGA is a good idea if power is a problem but our office graphics cards are doing more work. So moving forward i think an ASIC is the only way to go.

Agreed, but were still talking about magical ASICs here.  They don't fall from the sky.  Perhaps BitSyncom will start selling singles when Avalon's 3rd batch ships?  We can dream, huh?

[aeronautical]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

Agreed,
https://bitcointalk.org/index.php?topic=170977.0

It good, but looks like most of that thread are talking FPGA on PCI and i don't think that is going to do the job, too slow. Using an FPGA is a good idea if power is a problem but our office graphics cards are doing more work. So moving forward i think an ASIC is the only way to go.

Agreed, but were still talking about magical ASICs here.  They don't fall from the sky.  Perhaps BitSyncom will start selling singles when Avalon's 3rd batch ships?  We can dream, huh?

I don't think that there stuff works or they are sitting back testing all there boxes and not shipping them, reading the forum many have not shipped, again i don't get it how did these guys get so much money and not ship anything?? Is everyone nuts, give us the job the last chip we designed sold 8,000,000 units it was an LED driver.

[DarkPunk]

Avalon has given way more than most.  Their software is opensource, hosted on github, and updated simi-regularly.  They produced chip specs, even a die, and have plenty of pictures, and has several confirmed shipments.  Honestly, they are the only ASIC mining rig company I would trust with my coins.

[pheaonix]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

Agreed,
https://bitcointalk.org/index.php?topic=170977.0

It good, but looks like most of that thread are talking FPGA on PCI and i don't think that is going to do the job, too slow. Using an FPGA is a good idea if power is a problem but our office graphics cards are doing more work. So moving forward i think an ASIC is the only way to go.

Agreed, but were still talking about magical ASICs here.  They don't fall from the sky.  Perhaps BitSyncom will start selling singles when Avalon's 3rd batch ships?  We can dream, huh?

I don't think that there stuff works or they are sitting back testing all there boxes and not shipping them, reading the forum many have not shipped, again i don't get it how did these guys get so much money and not ship anything?? Is everyone nuts, give us the job the last chip we designed sold 8,000,000 units it was an LED driver.

wanna make some coin?

develop a board that simply requires the avalon chips to be soldered into them.

do it any way you want, but do it fast.

avalon stated they will be selling off mass orders of the chips.

[DarkPunk]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

Agreed,
https://bitcointalk.org/index.php?topic=170977.0

It good, but looks like most of that thread are talking FPGA on PCI and i don't think that is going to do the job, too slow. Using an FPGA is a good idea if power is a problem but our office graphics cards are doing more work. So moving forward i think an ASIC is the only way to go.

Agreed, but were still talking about magical ASICs here.  They don't fall from the sky.  Perhaps BitSyncom will start selling singles when Avalon's 3rd batch ships?  We can dream, huh?

I don't think that there stuff works or they are sitting back testing all there boxes and not shipping them, reading the forum many have not shipped, again i don't get it how did these guys get so much money and not ship anything?? Is everyone nuts, give us the job the last chip we designed sold 8,000,000 units it was an LED driver.

wanna make some coin?

develop a board that simply requires the avalon chips to be soldered into them.

do it any way you want, but do it fast.

avalon stated they will be selling off mass orders of the chips.

It can be done now, with the specs give by BitSyncom, with little guess work.

[pheaonix]

I'am simple person, and need few block diagrams but think that a PCI card packed would be a better idea than all of the Boxes out there.
It would also transport better fit's in a post pack.

Agreed,
https://bitcointalk.org/index.php?topic=170977.0

It good, but looks like most of that thread are talking FPGA on PCI and i don't think that is going to do the job, too slow. Using an FPGA is a good idea if power is a problem but our office graphics cards are doing more work. So moving forward i think an ASIC is the only way to go.

Agreed, but were still talking about magical ASICs here.  They don't fall from the sky.  Perhaps BitSyncom will start selling singles when Avalon's 3rd batch ships?  We can dream, huh?

I don't think that there stuff works or they are sitting back testing all there boxes and not shipping them, reading the forum many have not shipped, again i don't get it how did these guys get so much money and not ship anything?? Is everyone nuts, give us the job the last chip we designed sold 8,000,000 units it was an LED driver.

wanna make some coin?

develop a board that simply requires the avalon chips to be soldered into them.

do it any way you want, but do it fast.

avalon stated they will be selling off mass orders of the chips.

It can be done now, with the specs give by BitSyncom, with little guess work.

next step is getting the chips out of the shipping hell that is china. we need at least two 10k orders in the US and another 10k EU order.

get the boards manufactured and sent out to individuals and bingo everyone is happy