[CLOSED] Bitmine CoinCraft A1 28nm chip distribution / DIY support

[silver71]

Why 16, when you already have ref.design for 8, and 8 is easier to cool...

Even when you think about immersion cooling, if enything goes wrong, your whole chain (16) goes offline...

Wouldn't it then be smarter to stick with 8 ?

16 was just a number I pulled from thin air

As for the technical challenge: with the experience collected so far, I'd say once you have an 8-chip chain working, it is not a huge step to move to 16 chips. At least from the A1 side, I don't understand the DCDC part of it to state it would be easy.

The immersion cooling idea follows DaT's approach here, where due to the high costs of the fluid it is essential to stuff as much hashing power into as little volume as possible. I think one could get a 4x4 A1 matrix onto a 10x10cm^2 PCB and stack them with 1cm distance. Resulting in a 6kW burner in a 1 liter cube.

That would be more of a fun than a serious project and I proposed this to be added as a challenge for Bitmine's planned design contest - which for obvious reasons was put at the back of the priority queue and might never leave the announcement phase


Takers? I'd supply the chips and the fluid.

Let's first solve present rig delay (non-delivery), and go further with improved cooling ...

About the fluid...you meant oil or 3M-like liquid ?

It would be less expensive to just hook the rig to central heating (as a source), and use household radiators as coolers

[Cheshyr]

Why 16, when you already have ref.design for 8, and 8 is easier to cool...

Even when you think about immersion cooling, if enything goes wrong, your whole chain (16) goes offline...

Wouldn't it then be smarter to stick with 8 ?

16 was just a number I pulled from thin air

As for the technical challenge: with the experience collected so far, I'd say once you have an 8-chip chain working, it is not a huge step to move to 16 chips. At least from the A1 side, I don't understand the DCDC part of it to state it would be easy.

The immersion cooling idea follows DaT's approach here, where due to the high costs of the fluid it is essential to stuff as much hashing power into as little volume as possible. I think one could get a 4x4 A1 matrix onto a 10x10cm^2 PCB and stack them with 1cm distance. Resulting in a 6kW burner in a 1 liter cube.

That would be more of a fun than a serious project and I proposed this to be added as a challenge for Bitmine's planned design contest - which for obvious reasons was put at the back of the priority queue and might never leave the announcement phase


Takers? I'd supply the chips and the fluid.

I'd be glad to poke at that after we finish these first two A1 designs.  I've got a gallon of Novec 7000 in the corner, and was planning on playing with it using A1s anyway.

DCDC will be the real problem, and it really depends how fast you plan to clock the chips. 

[goodney]

I haven't looked at power at the board level at all yet - hopefully that's the issue. I think this board is running a little low on the voltage front so hopefully I can clean things up with that.

totalslacker:

Have you been able to measure the current or power consumed by one chip at 800Mhz?

Thanks!

-a[g

[zefir]

Notification: A1 CoinCraft Desk cgminer driver now upstreamed

Since some of you asked for it: as of today, Con pulled the current A1 driver (as variant for CoinCraft Desk) into cgminer. I will follow soon with the variant for the Rig, and after we get the products into delivery state, I'll also work on a bfgminer integration.

[totalslacker]

Have you been able to measure the current or power consumed by one chip at 800Mhz?

Not a super accurate measurement, but it looks to be about 20A. This is when the chip is around 40C as it gets hotter the current goes up (I was seeing around 22-24A at 60C). I do have the board instrumented so that I should be able to get more accurate results, I just haven't written code to look at that yet

Oh, this was with a core voltage of 0.84V. So looks to be pretty spot-on at 0.67W/GH at 25GH/s (assuming you have decent cooling).

[drinkmorecoffee]

Code merged into master cgminer, thanks.

zefir previously mentioned that cgminer has support for MCP2210-based boards (with USB comms to a PC rather than a RasPi).  Does this merge now mean that a current-version build of cgminer will run the A1 chips direct from a PC, if the MCP2210 is included?

That would be sweet...

[zefir]

Code merged into master cgminer, thanks.

zefir previously mentioned that cgminer has support for MCP2210-based boards (with USB comms to a PC rather than a RasPi).  Does this merge now mean that a current-version build of cgminer will run the A1 chips direct from a PC, if the MCP2210 is included?

That would be sweet...

No, the current version runs exactly the CoinCraft Desk which is driven by a RasPi over its SPI interface. To make it working over MCP2210, one would need to wrap the SPI interface (which is abstracted over spi-context.c already) over MCP2210. Should be easy to implement, alas I currently am too busy so look at.

[drinkmorecoffee]

Code merged into master cgminer, thanks.

zefir previously mentioned that cgminer has support for MCP2210-based boards (with USB comms to a PC rather than a RasPi).  Does this merge now mean that a current-version build of cgminer will run the A1 chips direct from a PC, if the MCP2210 is included?

That would be sweet...

No, the current version runs exactly the CoinCraft Desk which is driven by a RasPi over its SPI interface. To make it working over MCP2210, one would need to wrap the SPI interface (which is abstracted over spi-context.c already) over MCP2210. Should be easy to implement, alas I currently am too busy so look at.

Very good, thanks for the quick reply.  I've heard a couple comments that the RasPi is not suitable for this sort of thing, but hearing that it's used in the Desk product (a professional, commercial product) is reassuring.

Back to work!

Thanks again!

[zefir]

Very good, thanks for the quick reply.  I've heard a couple comments that the RasPi is not suitable for this sort of thing, but hearing that it's used in the Desk product (a professional, commercial product) is reassuring.

I guess what you refer to is the known problem of the RasPi wearing out the SD-card if you run it over standard Linux system over long time - which I personally did not notice working ~3 months mostly compiling with continuous write access to the card.

To be on the safe side, the FW provided with the Desk is mostly read-only, i.e. it writes to the SD-card only on re-configuration, while everything else goes into memory mapped filesystems.

[jbcheng]

Code merged into master cgminer, thanks.

zefir previously mentioned that cgminer has support for MCP2210-based boards (with USB comms to a PC rather than a RasPi).  Does this merge now mean that a current-version build of cgminer will run the A1 chips direct from a PC, if the MCP2210 is included?

That would be sweet...

No, the current version runs exactly the CoinCraft Desk which is driven by a RasPi over its SPI interface. To make it working over MCP2210, one would need to wrap the SPI interface (which is abstracted over spi-context.c already) over MCP2210. Should be easy to implement, alas I currently am too busy so look at.

hi Zefir
Our team also stuck on this part at the moment, Known you are quite busy, while would you please take some time and help to develop a version over a PIC? as this is a important step for multi modules integration. Thanks

Jbcheng

[zefir]

hi Zefir
Our team also stuck on this part at the moment, Known you are quite busy, while would you please take some time and help to develop a version over a PIC? as this is a important step for multi modules integration. Thanks

Jbcheng

Hi,

'quite busy' is quite some understatement - I am out of this world for at least the next 4 weeks to finalize the SW for the CoinCraft products and clean up the mess I will have been left behind me until then.

I'd support the community as best as I can (like I have been doing so far), but working on another driver / FW now is completely out of reach for me.

But this is exactly the idea behind the open source DIY approach: stick together and share your potential. There are at least 20 projects working on something, and while at some point we all are competitors, we at the same time profit from cooperation. I know marto74 did his firmware for a PIC, as well as the WASP team is working on some framework for mining endpoints. There are also at least 2 teams working on a USB->SPI bridge approach. I'm pretty sure that if one party starts to publish their work on FW as open source, everybody else will join to be the second. It only takes one team to step up...


Cheers,
zefir

[silver71]

Very good, thanks for the quick reply.  I've heard a couple comments that the RasPi is not suitable for this sort of thing, but hearing that it's used in the Desk product (a professional, commercial product) is reassuring.

I guess what you refer to is the known problem of the RasPi wearing out the SD-card if you run it over standard Linux system over long time - which I personally did not notice working ~3 months mostly compiling with continuous write access to the card.

To be on the safe side, the FW provided with the Desk is mostly read-only, i.e. it writes to the SD-card only on re-configuration, while everything else goes into memory mapped filesystems.

With current SD-card prices this would not be an issue, if reasonable user, clones SD-cards prior of using them, and swaps them if they fail...on the other hand I would rather choose Arduino over Pi, but...that's me...now it's all - whatever works first...

...and when talking about sharing mutual knowledge and learning from it...

maybe something usefull comes from Scrypt-ASIC teams...and their open-source PCB's :
https://github.com/losh11/aura-asic

[drinkmorecoffee]

Why is Bitmine no longer allowing payment in BTC?  I was just about to order a couple samples and the BTC payment option is just... gone.  Is this a temporary thing, or what? 

I wrote their support people but haven't heard back.  Anyone know anything?

[jbcheng]

That's really bad news for Bitmine do not accept BTC payment. We can use the realtime exchange rate and this shall not be too complicated?

[Bicknellski]

Why 16, when you already have ref.design for 8, and 8 is easier to cool...

Even when you think about immersion cooling, if enything goes wrong, your whole chain (16) goes offline...

Wouldn't it then be smarter to stick with 8 ?

16 was just a number I pulled from thin air

As for the technical challenge: with the experience collected so far, I'd say once you have an 8-chip chain working, it is not a huge step to move to 16 chips. At least from the A1 side, I don't understand the DCDC part of it to state it would be easy.

The immersion cooling idea follows DaT's approach here, where due to the high costs of the fluid it is essential to stuff as much hashing power into as little volume as possible. I think one could get a 4x4 A1 matrix onto a 10x10cm^2 PCB and stack them with 1cm distance. Resulting in a 6kW burner in a 1 liter cube.

That would be more of a fun than a serious project and I proposed this to be added as a challenge for Bitmine's planned design contest - which for obvious reasons was put at the back of the priority queue and might never leave the announcement phase


Takers? I'd supply the chips and the fluid.

Ice Wasp is something we are keen on doing... just need to get our Wasps finished first.

[zefir]

Update: Chip Distribution closed

That's it, I assigned all remaining chips for assembly and with that this DIY chip distribution is closed.

Chips will remain available through Bitmine's web-shop, even in sample quantities.


Thank you all for your trust and support. Good luck with your projects, especially since times became more stormy in bitcoin land recently.

[marto74]

Thank you for the opportunity to start our design.
As usual it was pleasure doing business with you.
Looking forward for future projects.
We wish you and the Bitmine team all the best.

[vs3]

Zefir - thanks again for all of your work and helping the community with this project!

[loshia]

Zefir - thanks again for all of your work and helping the community with this project!

+1

[giorgiomassa]

Thank you Zefir for handling this process, as for all the DIY people we wish you a great success with your designs and thanks for all the feedback given, we are working on fixing the reference design files and committing them to the git repository soon.

The base files for building the OpenEmbedded/Yocto-based rpi image will also be committed soon so that everybody willing to develop forks based on it is very welcome to do so (and share back)