The Open Source Block Erupter Project

[friedcat]

Any chance the chips might be open source too?

Yes. But there are no written plans in the near term yet.

Also it's much harder than making a piece of software open source, because the whole flow to generate GDS has tons of manual operations and very fab and technology node specific.

Please don't do this in the near future
On a serious note: I'm curious, though, how would this profit AM?

Whenever open-source or not doesn't make difference in AM profits, both immediate and potential, we will do.

[1714935193]

1714935193

[Luke-Jr]

Any chance the chips might be open source too?

Yes. But there are no written plans in the near term yet.

Also it's much harder than making a piece of software open source, because the whole flow to generate GDS has tons of manual operations and very fab and technology node specific.

Please don't do this in the near future

Why not? Open source mining chips benefit the community the best, I think? Or am I missing something?

[JoTheKhan]

Any chance the chips might be open source too?

Yes. But there are no written plans in the near term yet.

Also it's much harder than making a piece of software open source, because the whole flow to generate GDS has tons of manual operations and very fab and technology node specific.

Please don't do this in the near future

Why not? Open source mining chips benefit the community the best, I think? Or am I missing something?

Miner is an investor in AsicMiner, by open sourcing chips wouldn't that give advantage to the competition? I assume this is why he is saying. "don't do this." you are revealing your hand to your opponents I would assume.

[minerpumpkin]

Any chance the chips might be open source too?

Yes. But there are no written plans in the near term yet.

Also it's much harder than making a piece of software open source, because the whole flow to generate GDS has tons of manual operations and very fab and technology node specific.

Please don't do this in the near future

Why not? Open source mining chips benefit the community the best, I think? Or am I missing something?

Miner is an investor in AsicMiner, by open sourcing chips wouldn't that give advantage to the competition? I assume this is why he is saying. "don't do this." you are revealing your hand to your opponents I would assume.

Yes, if they completely open-sourced the chip, practically everyone (who can afford a $1.3m mask set) would be able to produce their chips. Hence, AM could get serious problems selling their chips - large customers could just produce them themselves. AM and its investors would be screwed.
After selling the chip isn't profitable anymore or only self-mining is being done, and assuming the chip doesn't contain too much trade-secrets for gen 4, I think it's a nice thing to release the blueprints.

[Bicknellski]

Any chance the chips might be open source too?

Yes. But there are no written plans in the near term yet.

Also it's much harder than making a piece of software open source, because the whole flow to generate GDS has tons of manual operations and very fab and technology node specific.

Please don't do this in the near future

Why not? Open source mining chips benefit the community the best, I think? Or am I missing something?

Miner is an investor in AsicMiner, by open sourcing chips wouldn't that give advantage to the competition? I assume this is why he is saying. "don't do this." you are revealing your hand to your opponents I would assume.

Yes, if they completely open-sourced the chip, practically everyone (who can afford a $1.3m mask set) would be able to produce their chips. Hence, AM could get serious problems selling their chips - large customers could just produce them themselves. AM and its investors would be screwed.
After selling the chip isn't profitable anymore or only self-mining is being done, and assuming the chip doesn't contain too much trade-secrets for gen 4, I think it's a nice thing to release the blueprints.

Dump the old on the public keep the new for the AM shareholders. Just have to time that right so you don't get caught with your pants down so to speak. That might have worked in the 1st and 2nd Gen stuff but with other players and chips in the market it the demand is going to go to the cheapest price chip / board combo. Fun times indeed.

[Chris_Sabian]

Anyone interested in designing a Cube Block Erupter retro-fit?

now we're talking!!!  there's a large market for this and the earlier blades too, since the power boards are out there.
if people could buy replacement blades for either platform, they would.  and if they could replace one at a time that would be phenomenal as well.

This would be an amazing thing.  I would like to re-use the cubes since they fit together and stack nicely. 

[Luke-Jr]

Anyone interested in designing a Cube Block Erupter retro-fit?

now we're talking!!!  there's a large market for this and the earlier blades too, since the power boards are out there.
if people could buy replacement blades for either platform, they would.  and if they could replace one at a time that would be phenomenal as well.

Slightly related to this, is there perhaps the possibility of open sourcing some or all of the Cube's information?
It would be nice for someone to try to debug the TCP stack and figure out how to make it work correctly

[minerpumpkin]

Yes, if they completely open-sourced the chip, practically everyone (who can afford a $1.3m mask set) would be able to produce their chips. Hence, AM could get serious problems selling their chips - large customers could just produce them themselves. AM and its investors would be screwed.
After selling the chip isn't profitable anymore or only self-mining is being done, and assuming the chip doesn't contain too much trade-secrets for gen 4, I think it's a nice thing to release the blueprints.

Dump the old on the public keep the new for the AM shareholders. Just have to time that right so you don't get caught with your pants down so to speak. That might have worked in the 1st and 2nd Gen stuff but with other players and chips in the market it the demand is going to go to the cheapest price chip / board combo. Fun times indeed.

Well, AM is a company after all, why should they just now give away their assets for free? Open sourcing the designs at a later point might be considered a mere nice gesture, IMO...

[Bicknellski]

Yes, if they completely open-sourced the chip, practically everyone (who can afford a $1.3m mask set) would be able to produce their chips. Hence, AM could get serious problems selling their chips - large customers could just produce them themselves. AM and its investors would be screwed.
After selling the chip isn't profitable anymore or only self-mining is being done, and assuming the chip doesn't contain too much trade-secrets for gen 4, I think it's a nice thing to release the blueprints.

Dump the old on the public keep the new for the AM shareholders. Just have to time that right so you don't get caught with your pants down so to speak. That might have worked in the 1st and 2nd Gen stuff but with other players and chips in the market it the demand is going to go to the cheapest price chip / board combo. Fun times indeed.

Well, AM is a company after all, why should they just now give away their assets for free? Open sourcing the designs at a later point might be considered a mere nice gesture, IMO...

Agreed. Just pointing out the fact that "open" source doesn't always mean altruism.

[xzempt]

Where is intron and cscape..... they should be all over this

[intron]

Where is intron and cscape..... they should be all over this

Working on other stuff now...

[friedcat]

We had AM_BE200_X24 boards sampled and tested. We also sampled the compatible heatsinks and supporting structures for them. It hashes at 768GH (each board 192GH) with 0.85-0.92W/G (PSU loss excluded). Power draws on higher/lower hashrate with higher/lower voltage is under testing. The design files of heatsinks/structures are going to be released a little later.



It should be emphasized that QUAD and X32 are also sampled and well tested before X24.

[jimmothy]

We had AM_BE200_X24 boards produced and tested. We also sampled the compatible heatsinks and supporting structures for them. It hashes at 192GH with 0.85-0.92W/G without PSU. Power draws on higher/lower hashrate with higher/lower voltage is under testing. The design files of heatsinks/structures are going to be released a little later.



It should be emphasized that QUAD and X32 are also produced and well tested before X24.

My god it's beautiful.

Where do we buy them??

Also how many chips @ 192gh?

[friedcat]

My god it's beautiful.

Where do we buy them??

Also how many chips @ 192gh?

You can purchase chips and make this devices of your own.

You can also buy from our customers who are going to produce according to this open source X24 design.

Each board has 24 chips and hashes at 192gh at this setting. So the total is 192*4=768gh. The hash speed
can be got higher or lower according to your voltage.

[WinterParker]

The power draw is looking good!

[friedcat]

It's not the cube.

When we release source of BE100 based cubes there might be some contributors who can do BE200 based cubes for us.

[loshia]

........

Dump the old on the public keep the new for the AM shareholders. Just have to time that right so you don't get caught with your pants down so to speak. That might have worked in the 1st and 2nd Gen stuff but with other players and chips in the market it the demand is going to go to the cheapest price chip / board combo. Fun times indeed.

Bicknellski,

I am wandering what happened to your plans to include BE200 In your WASP project?

https://bitcointalk.org/index.php?topic=641957.msg7175102#msg7175102

Thank You
PS: I will appreciate very much if we stay on topic and you leave the usual off topic if you ever answer me.

Once again Thank you very Much

[minerpumpkin]

My god it's beautiful.

Where do we buy them??

Also how many chips @ 192gh?

You can purchase chips and make this devices of your own.

You can also buy from our customers who are going to produce according to this open source X24 design.

Each board has 24 chips and hashes at 192gh at this setting. So the total is 192*4=768gh. The hash speed
can be got higher or lower according to your voltage.

Maybe you should point out in the original post that the whole miner hashes at 768 GH/s and the single boards hash at 192 GH/s, each. One may be led to think that each chip only performed at 2 GH/s otherwise - which would be quite low.
Good work apart from that, it is a remarkable strategy to open up the designs and let potential customers/resellers produce their own miners. Also, the increase in communication over the last few days is very much appreciated!

[friedcat]

My god it's beautiful.

Where do we buy them??

Also how many chips @ 192gh?

You can purchase chips and make this devices of your own.

You can also buy from our customers who are going to produce according to this open source X24 design.

Each board has 24 chips and hashes at 192gh at this setting. So the total is 192*4=768gh. The hash speed
can be got higher or lower according to your voltage.

Maybe you should point out in the original post that the whole miner hashes at 768 GH/s and the single boards hash at 192 GH/s, each. One may be led to think that each chip only performed at 2 GH/s otherwise - which would be quite low.
Good work apart from that, it is a remarkable strategy to open up the designs and let potential customers/resellers produce their own miners. Also, the increase in communication over the last few days is very much appreciated!

Fixed. Sorry.

[aahzmundus]

Amazing work, that is a beautiful product.  You should be applauded for how your efforts will help to distribute the overall network hashrate.

I would also like to second the thank you for increased communication from AM.

Next time anyone complains about GHash.io, I will link them to this project.