Author Topic: [LTC-GLOBAL] CIPHERMINE Industrial Cryptocoin Mining & High Perf. Computing  (Read 316235 times)

Offline Ethera

  • Full Member
  • ***
  • Posts: 488
    • Litecoininvest - litecoin investment platform
Indeed klondike K16's from BKKcoin design
PCB assembly is carried out to IPC A 610 CLASS 2 where the board design permits, and assembly plant is on stand by.
Any floating (ie < 16) chips will be put to either klondike nano's (K1's) or the arduino miner shield, which I personally have sample chips sent
to.
There maybe other boards that appear, and I will be keeping a close eye on all boards (inc scrypt prototypes) for speed/cost/energy etc.
Board repairs will be done in house, as I can reflow the odd defective part and test it.
As will any revision changes, ie K16 v1 to K16 v2, we will have to look at the changes and if they can be retro fitted (and or worth it)
I currently have some floating spare sample chips that maybe used for spares/repairs should we need them

architecture does not really fit in this context does it.....or should I just state QFN48 ?

cheers for the answer. architecture i meant how will be stacking boards. sorry, english is not my first language :)

Offline evilscoop

  • Jr. Member
  • **
  • Posts: 170
Other than available now.....
At 1 btc they do represent better value short term than fpga's
I dont think Kate is about to sink a lot into them tbh :D


Ethera, sorry :D
Simon and Ross will deal with casing, Id think back to back ala terrahash model looks promising....but I just play with chips :D

Offline Benny

  • Sr. Member
  • ****
  • Posts: 667
    • www.buyahash.com
We are currently evaluating how best to spend the rest. Since the USB Erruptors are now only 0.9 BTC each they are looking like an interesting purchase,

They may be interesting but they aren't profitable.  With any realistic model of future difficulty they'll barely (if at all) ever mine back their costs.  They're great as novelties and a decent way of transferring cash from the innumerate to ASICMINER shareholders but in no way a serious investment for profit.

The most profitable thing about USB Erupters?

Selling them to others.

That is how you're going to make money on them... Not by mining with them. Really, there's very little of what ASICMiner sells that is going to be cost-effective for a PMB or any sort of mining company.

Offline Sahtor

  • Jr. Member
  • **
  • Posts: 77
...
Further, with LTC's value trending strongly upward and a drop in Sapphire 7950 cards there is an improving case for spending more on GPU rigs.
I'm in favor of more scrypt machines right now. I'd even support a motion to raise GPUs to 10% if asked.

Anyways thanks for an exciting operation. It's been fun reading all the updates.

Offline WoodTech

  • VIP Donator
  • Full Member
  • ****
  • Posts: 278
  • Kate Craig-Wood
    • Wood Technology LLP
Hardware orders / tracking
« Reply #49 on: June 27, 2013, 07:59:03 PM »
Dear All,

Rather than posting information on individual hardware orders for evidence purposes and such I have decided to streamline the process by simply uploading screenshots of the order confirmation documents to docs.ciphermine.com/orders/. Please review them at your leisure. Note that I'm not bothering to include the small stuff (fans, cable ties, tools, CPUs, PCI-E risers, DVI-VGA adaptors, etc).

We have spent an initial 172 BTC out of 423 raised at IPO (15,667 LTC x 0.027 BTC/LTC) on Avalon ASICs (see order CM-SB-05).

We've decided to spend 15% of the fund (25 BTC) on GPU hardware rather than the planned 5%. This is mainly because of the fervour around LTC at the moment with its expected launch on Mt. Gox in July, but also because the prospects of our diversification into renting out big GPU clusters is looking increasingly promising as is their resale value (we could make 10% just by flogging them on eBay!). This is also likely a good hedge; spending the full 95% on SHA256 mining hardware was probably a bit overkill, especially with new, improved devices on the horizon.

We will be spending the rest of the fund (224 BTC) on SHA256 mining hardware in the next few days. Giles and I are figuring out what the best option is going to be; we're tempted by the ASICminer USB erruptors which are now 0.9 BTC each, but with Steamboat expecting to be shipping his ASICs (which are in 300% more efficient in terms of ROI) in mid August we might just stick with those! We do also need to purchase the parts for the Klondike boards so some of the funds are reserved for that.

Back to GPUs, we now have hardware for an additional three quad-Sapphire 7950 rigs on order due to arrive in the next week. We've determined that these cards are definitely the sweet spot in terms of cost and power consumption (those rigs only use 1kW but generate 2.43 MH/s scrypt - way better than 7970s for instance). We are also looking at squeezing a fifth card into the rigs.

For your interest I have also made a folder for pictures of the mining rigs and associated hardware. It is at docs.ciphermine.com/hardware/. See below for what our preferred rig configuration looks like (that one's back on the build workbench to try to squeeze in the extra GPU):



Kate.

Offline WoodTech

  • VIP Donator
  • Full Member
  • ****
  • Posts: 278
  • Kate Craig-Wood
    • Wood Technology LLP
...
Further, with LTC's value trending strongly upward and a drop in Sapphire 7950 cards there is an improving case for spending more on GPU rigs.
I'm in favor of more scrypt machines right now. I'd even support a motion to raise GPUs to 10% if asked.

Anyways thanks for an exciting operation. It's been fun reading all the updates.

Thanks for the input. On reflection, we agree. I've increased it to 15% (see last post). We put in the plan that we can fiddle with the ratios based on current conditions so I didn't feel it necessary to raise a motion, especially as the sooner we buy stuff the sooner its mining! :)

Kate.

Offline Benny

  • Sr. Member
  • ****
  • Posts: 667
    • www.buyahash.com
Just curious, who is your PCI-E riser vendor?

If you need parts, I'd love to work out an arrangement with you, since I know you'll be (likely) creating a few GPU rigs in the future. I know they're quite hard to get in the UK, and we have a lot of customers there. We also have a few other smaller items that may be hard for you to source. PM me or let me know, and I can work on giving you a discount or something.

Offline WoodTech

  • VIP Donator
  • Full Member
  • ****
  • Posts: 278
  • Kate Craig-Wood
    • Wood Technology LLP
Just curious, who is your PCI-E riser vendor?

If you need parts, I'd love to work out an arrangement with you, since I know you'll be (likely) creating a few GPU rigs in the future. I know they're quite hard to get in the UK, and we have a lot of customers there. We also have a few other smaller items that may be hard for you to source. PM me or let me know, and I can work on giving you a discount or something.

eBay.co.uk. ;) For the 16x-16x risers I just get the from the cheapest, but for the powered 1x-16x I used http://myworld.ebay.co.uk/litecoins

I'd be open to a discussion about sourcing from you though certainly - assuming that's not you?

Kate.
« Last Edit: June 27, 2013, 10:29:57 PM by WoodTech »

Offline Benny

  • Sr. Member
  • ****
  • Posts: 667
    • www.buyahash.com
That's not me. We're about 1/2 the price of them, but would have to ship from the US.

Offline WoodTech

  • VIP Donator
  • Full Member
  • ****
  • Posts: 278
  • Kate Craig-Wood
    • Wood Technology LLP
A mini update for you all: Jan has been beavering away on our private pools with a bit of testing help from me. It is going really well and we've now got about 50% of them running. :)

As part of the process we've also been experimenting with new coins like Cryptogenic Bullion (cgb). While these are not yet traded anywhere, new coins with at least a little merit do tend to get quite good value in the first few days, and since not many people are mining them right now we're finding zillions of blocks! We are therefore considering putting some of our scrypt hashing towards the emergent coin du jour.

I would welcome feedback / thoughts on this activity. It is not something we have done before and is obviously not without risks.

I am also re-working the model to take into account some new factors, including a much improved deal we have secured for Sapphire 7950s, the fact that we can get 5 of them in a rig now, and that we are consistently outperforming Litecoin by using CipherMiner's automagic coin selection. My aim is to determine whether there is merit in investing even more of the IPO funds in GPU hardware or whether we should stick to spending the rest on ASICs/FPGAs. While I felt it was reasonable to take an executive decision on increasing the GPU investment element from 5% to 15% I will want to ask for a motion if the numbers suggest increasing it further makes sense.

Kate.

Offline xavier999

  • Newbie
  • *
  • Posts: 8
A mini update for you all: Jan has been beavering away on our private pools with a bit of testing help from me. It is going really well and we've now got about 50% of them running. :)

As part of the process we've also been experimenting with new coins like Cryptogenic Bullion (cgb). While these are not yet traded anywhere, new coins with at least a little merit do tend to get quite good value in the first few days, and since not many people are mining them right now we're finding zillions of blocks! We are therefore considering putting some of our scrypt hashing towards the emergent coin du jour.

I would welcome feedback / thoughts on this activity. It is not something we have done before and is obviously not without risks.

For my input, I think it is OK to do a little speculative mining of untraded coins, though only on a limited basis. There is the potential to make some good returns...and the potential to make nothing and waste time and resources that could be devoted towards established coins. I would say no more than 10% of the script mining resouces should go to this, but that's just my two cents. And thanks for all the hard work so far!

Offline shacky

  • Jr. Member
  • **
  • Posts: 155
Woodtech, I had read the new Motion.

Im with you. Buy ASICMiner's USB Erruptors won't make a good ROI in the long term. Im calculate it, and maybe buy it now, will be a loss.

About ASIC orders. I think that when the company ship the hardware, will be near impossible to get the money back. Check the AVALON 3 Batch, The owners of the first batch got a good ROI, but the batch # 2 and #3 won't run with the same luck.

When BFL ship all the hardware pre-ordered the difficulty will grow a lot.

My opinion:

The life of ASIC is starting, maybe in Q1 2014 we will see a more powerful ASIC, and the old ones won't work. GPU is the way to go. the BEST re-sell value, and NO ASIC for a long time.

Offline WoodTech

  • VIP Donator
  • Full Member
  • ****
  • Posts: 278
  • Kate Craig-Wood
    • Wood Technology LLP
Please see our latest news and new motion on our security's page.

Please note that I intend to abstain from the motion with my votes. I want this to be truly down to the shareholders; if I believe something to be the right course of action then I should be able to convince you of that. If I cannot, it is probably a bad idea.

Kate.

Offline WoodTech

  • VIP Donator
  • Full Member
  • ****
  • Posts: 278
  • Kate Craig-Wood
    • Wood Technology LLP
Quote from: woodtech
As part of the process we've also been experimenting with new coins like Cryptogenic Bullion (cgb). While these are not yet traded anywhere, new coins with at least a little merit do tend to get quite good value in the first few days, and since not many people are mining them right now we're finding zillions of blocks! We are therefore considering putting some of our scrypt hashing towards the emergent coin du jour.

For my input, I think it is OK to do a little speculative mining of untraded coins, though only on a limited basis. There is the potential to make some good returns...and the potential to make nothing and waste time and resources that could be devoted towards established coins. I would say no more than 10% of the script mining resouces should go to this, but that's just my two cents. And thanks for all the hard work so far!

I am inclined to agree, certainly until we have proven the scheme. We have some earlier models of our rigs which are not terribly high hashrate (the new quituple-7950s are 3MHs each, representing 1/6th of total power roughly since the first four were just quads) so we'll probably just leave one of them working on our emerging coin of choice. CGB does seem to be a good one at the moment.

To mitigate the risk I am happy to cover the potential lost earnings from that rig as well. If it proves to be a complete failure I'll make up the difference from my own pocket so that our dividend growth is not harmed. If it works then we'll pay out the bonus in full! :)

Kate.

Offline WoodTech

  • VIP Donator
  • Full Member
  • ****
  • Posts: 278
  • Kate Craig-Wood
    • Wood Technology LLP
Im with you. Buy ASICMiner's USB Erruptors won't make a good ROI in the long term. Im calculate it, and maybe buy it now, will be a loss.

About ASIC orders. I think that when the company ship the hardware, will be near impossible to get the money back. Check the AVALON 3 Batch, The owners of the first batch got a good ROI, but the batch # 2 and #3 won't run with the same luck.

We have some in those batches, and the extra money will go to Steamboat's batch 5 which is already ordered. Expected delivery is August - they have advanced their timelines.

Quote
When BFL ship all the hardware pre-ordered the difficulty will grow a lot.

"When"? I thought consensus was that no-one will ever see any hardware from that crook. ;) Also, compared to ASICs the capex required is very high to acquire hashrate (comparable to GPUs/FPGAs).

Quote
The life of ASIC is starting, maybe in Q1 2014 we will see a more powerful ASIC, and the old ones won't work. GPU is the way to go. the BEST re-sell value, and NO ASIC for a long time.

I disagree that ASICs don't have a future: difficulty has to go to >500m before the current batch stop turning a profit, ASICminer themselves are going to bump into limits due to 51%, and the capital required if buying their hardware is significant - in other words only those of us with the skills/patience to integrate ASICs will be significant players. However I do tend to agree that there will likely be improved models coming soon. It looks like the motion will be passed, but before taking action the team and I will consider potential future ASIC scenarios.

I must confess to liking GPUs. They are fun and interesting to build, tune and manage, and the opportunities are the plethora of alt-coins are very interesting. However, in the grander scheme of things it does not look like they will make us a fortune any time soon. ;) Still, increasing the GPU spend to 15% is a good start and we will continue to review that ratio.

Kate.

PS. We are expecting scrypt FPGAs to become available late this year, and if not we intend to commission some!

Kate.