From: Grigor Stoichkov, Nikolai Tsekov
So far we haven’t shared much about the Cryptum mining project. Sure, you got hints through our bounty campaign (and we got no answer that is close to the mark), but we have to tell the future CRTM token owners how do we plan to develop this part of Cryptum. The Mining project will be developed only if we sell 150 000 or more CRTM (outside bonuses, mind you, so any sale bonuses are discarded from the total), which is fairly low number - we already got interest of investors well above that.
Our sale is limited to 1.1 million of CRTM, incl. bonuses, and the mining project funding is 30% of the total, so the maximum is somewhere around $1.8 million, and this is a number that mostly refers to our capability of managing the project. If we were asking for $100 million instead, the main funding goal - our board games project, would receive way more funds than it needs, and it won’t scale much above the previous expectations, while the mining project would become more than we could handle, or it would have to be reshaped into purely ASIC mining farm that would be easier to manage.
So we are looking in investment somewhere between $250 000 and $1 800 000. Quite a difference for planning, right?
Well, this is the reason we decided to go for mixed GPU and ASIC mining (to fight availability issues) if we hit a higher sale mark or just ASIC mining at the lowest mark. At the current profitability rates we would be able to pull out decent profit even if the mining continues to trend of becoming less viable, for at least 2 years with the latest hardware. Adding GPU mining gives us the ability to jump on opportunities like new coins and such (including our own) with a proper software support, all the while it has returns still gravitating at well over 300+% of the electricity cost, and this is without optimizations. ASICs availability is an issue, but we will deal with it when the problem raises.
There are many ways to approach even GPU mining: different GPUs, management, number of cards per rig, etc. We would not delve into it right now as we would prefer to keep the know-how we have. Our GPU choices are NVIDIA cards from the top range, like GTX 1080 Ti and GTX 1080, depending on the availability. GTX 1070 is still one of the best cards for certain cryptocurrencies, but they fall short in mining GPU-heavy algorithms, and are mostly preferred for ETH (+dual) or Equihash mining, which are not always the top choices in terms of fast profitability. As it all depends on availability and we want the project to run at full speed as fast as possible, we would separate it into few types of uniformed rigs, ming either whatever is most profitable, or supporting the project by mining ETH, which would be needed for the shared profits.
At the same time, all of the NVIDIA top GPUs are extremely easy to manage and nearly 100% of them work well with much reduced core voltage, resulting much lower spending on electricity and consequent heat reduction. While this is unimportant for small scale mining, for a farm that may include over 1500 cards, heat and electricity are huge thing. A simple voltage reduction could lower a single card power requirements with as much as 30-35%, w/o sacrificing nearly anything from the processing power.
GTX 1080 Ti are the least profitable in terms of initial investment, but they are oriented towards long-term mining. They also allow much better response to opportunities that rise with new currencies or unexpected market changes, and are top choice in regards of space conserved - which is also an issue with such a number of rigs.
GTX 1080s are the easiest to manage and in many regards they scale the same way as 1080 Ti in power efficiency, except for the number of slots; yet they would require less of the PSUs, which is another limited availability rig component.
ASIC mining is more straightforward, we only have to choose among the available miners, With electricity cost on the lower side, we would only care for how much more it can push for a month. Currently we are looking at few miners, but this decision would have to wait for the conclusion of the sale because we cannot predict the number of miners we would require, and there is not much point in planning the meal ahead of catching the fish in this case, as scalability is more of an issue regarding the available hardware.
HardwareBG has been into mining long enough to even develop some tools for rig management of their own, and by partnering with other partners they already have tested countless ways to fight chores. All of the rigs will come with both redundant software remote management capability and hardware monitoring and power management.
Crypto Mining Blog involvement means we will always have the latest compiled miners, and also get news of new opportunities even before they gain much attention. They will also advise on ASICs as we will be looking into them at any of the projected funding goals.
All of the hardware will be hosted at managed and secure mining-related datacenters that already have good experience with mining hardware. Bulgaria is famous for its fast internet access and great many data centers are already located here or being developed. We have a few offers, and all of them are based on electricity spent, so we would probably diversify the locations in order to minimize the risk of power outages.
Right now a 6-card rig (the most preferred configuration for miners expected to work with new currencies) will cost us around 6000-6200 EUR based on GTX 1080 Ti GPUs and depending on the GPU and PSU availability. This includes hardware power state management and links for multiple PSUs when needed. Of course, we’d prefer more efficient GPUs, if possible, but it all boils down to the scale at the end of the sale.
As for ASICS, their prices are public and final, so it really is a matter of choice.
We know this isn’t saying much. We didn’t include exact technical specifications of the rigs, for example. Well, anyone that is deep into mining would tell you there is not telling what’s available on the market right now. CPUs, memory and drives are easy, but with mainboards, video cards and power supplies you can’t rely on the regular supply chain to provide what you need, and sometimes you get what is available. We manufacture our own rig stands up to the datacenter specification (steel and aluminum frames), and the management is done through our own tools, unless we are using ASICs.
