Why I Only 'Mine' Gridcoin, and My Unusual ~200 TFLOPS Mining SetupsteemCreated with Sketch.

in #gridcoin7 years ago (edited)

I have only been mining cryptocurrencies for 2 months, but from the get go I have thrown myself 100% behind mining only a single currency - GRC. This may be non-conventional, as most serious miners switch between coins to maximise their profits, but mining Gridcoin has benefits above and beyond the bottom line:

  • Gridcoin pays out in proportion to research work done through the BOINC platform. This replaces POW of the classic coins, and means that while 'mining' I am actually generating tangible research results. I have personally contributed to mapping the Milky Way, several cryptographic applications, solving Riesel and Sierpinsky bases, and folding protein structures - among other things.

  • The community looks out for each other. Everyone is endlessly helpful in working together and supporting the network. There is even a tradition that every new member of the network is given some token amount of GRC to get them started on their journey!

  • At the end of the day, even if all of crypto falls to pieces, my time and compute has been well spent. As a researcher myself, I am stoked to see compute being spent wisely instead of hashing into the empty void.

My Mining Set-up

My mining set-up is a little out of the ordinary both in terms of size and hardware, comprising a mix of CPUs and GPUs across almost the entire spectrum from the last 5 years. Without going into full detail, here are some of the key statistics (click the picture for my Gridcoin stats page):

Computer.png

This currently yields a mint just shy of 800 GRC per day. What is missing from this picture is that this is the amount of hardware I have running on average. Some of this is often diverted to research jobs outside the BOINC infrastructure, and it takes a significant amount of time on my part to manage this. Of course, I would much rather spend time optimising the use of hardware than having it sit idle on its way to becoming obsolete (as all hardware is within 5 years to a decade).

You're Doing Something Right When People Think You Are Cheating!

Ever since joining the crypto mining scene, I have been mining so much Gridcoin that some people suspected foul play was involved. I'll take that as a compliment - must be doing something right to get that kind of praise!

To be fair, the concern of the user was kind of understandable. We have in the past seen a user who distributed the BOINC research software as a virus and collected the earnings on the research done by the PCs he infected. Needless to say, the days of that user were numbered. Both the BOINC project admins and the Gridcoin devs rapidly put a stop to the operation after running the virus in a sandbox to see what it was doing.

The Bottom Line

Gridcoin exists to support people already contributing to BOINC, and to encourage more people to get involved. It is not the most expensive coin, and may not be the most efficient coin to mine - but that is OK. Everyone contributing to the 'mining' effort is helping to make tangible progress in the scientific community. As the Gridcoin community grows, this will result in simultaneous natural growth in the value of the coin - especially once research institutes use it to buy compute. This infrastructure already exists and is 100% functional!

With the community having grown significantly in the past month alone, and GRC being worth over 5 times as much as it was this time last year, it will be exciting to see where BOINC and Gridcoin go next!

If you would like to get involved, I published a full guide on how to get set up with Gridcoin from a cold start on Steemit a while ago. You could also get started with pool mining, as it is quicker and easier to get going. Finally, there is a wealth of information available on Gridcoin.us.


Content credit:
Banner, @joshoeah
Infographic, @me-shell
Footer, @me-shell

Dutch.gif

Sort:  

What an incredible setup!
I just started mining Gridcoin one week ago and I got really into it thanks to you. Your posts on steemit and reddit were about 90% of all the important information that I needed during this week. Your work is not just helpful, but your writing style is also fantastic.

Since your are running that many GPUs and CPUs, I wanted to ask you for some advice on my setup and the project I am crunching at the moment. I have an i5-7600k and a gtx 1060 running at the moment and I use grcpool.com. The projects I mainly crunch are SRBase (mainly), NFS@home (backup) and Amicable Numbers (GPU). Do you think that makes sense?

I'm very glad to hear that you are able to get something useful out of my writing, so thank you a lot for letting me know! =)

Nice CPU, and the graphics card is nothing to look past either. I used to run SRBase about a month or two ago when the project was still very quiet at less than 700k RAC. It now has over 3 million RAC so it has become a lot more competitive! This is the trend with most projects, where they are optimal choices one day and not the next - best not to min-max it too hard.

That being said, looking on gridcoinstats we can see that for CPU the VGTU, Cosmology and LHC projects are good choices. Sztaki is about to be removed from the whitelist, so not a good one to crunch for. That being said, SRBase is not bad as it hands out credit (which builds RAC) more generously than most other projects. I have never looked at NFS, sorry. =/

For GPU, seeing as you have a single precision focused gaming card, maybe have a look at something like SETI.

Remember that the beauty of the GRC minting model is that it encourages evenly distributed compute from Team Gridcoin - most projects get a turn at being optimum.

One follow up on GPU: You mentioned SETI. However, in the other guides I thought to have read to go for low number of users. Amicable Numbers is one of the lowest and SETI the second highest. Is Amicable Numbers not made for single precision cards and I should look at SETI because of that or did I get something else wrong? Thanks for your help!

All GPU projects are fine for single precision cards, except for MilkyWay@Home where the double precision cards will shred all the competition.

You are right that SETI has lots of users, but many are inactive or running only their CPUs on the project. I have not actually run Amicable Numbers myself, so I cannot compare the rate of credit handouts though... I cannot say which is more efficient.

The nice thing about SETI is it can use your integrated graphics chips, which most other projects can't. A little extra power that otherwise goes unused. Regardless of what your GPU does, it's worth pointing your integrated chip to SETI.

In general, low users and low RAC is good - but low RAC could be due to the project handing out credit slowly, and high users could be largely inactive... You never know.

Okay, very helpful again :) I therefore have a closer look at the Team RAC instead of the number of users!

as well as the number of user

Both are important. :D

I just wanted to try out SETI now and faced a problem: If I run SETI, it uses 0.3CPUs compared to the 0.0899 Amicable Numbers needs. This means, that I can either let 3 CPU workloads + 1 SETI workload rund, or 4 CPU Workloads and SETI at 60%CPU usage, since the gpu is bottlenecked by the lack of CPU I guess.
Both solutions do not seem very satisfactory to me. Can I somehow use the other .7 of this CPU core or would I just have to let do not much?

Yes - just run your CPUs at 100% load. The GPU will take priority over the CPU work, and only be very slightly throttled at worst. I tested this and noted very limited throttling.

The only time when you need to keep a whole core or two free for your GPU is when your GPU is effectively doing all your mining (it is the star of your rig), such as an FP64 GPU on MilkyWay@Home would be.

Effectively, setting your CPUs to 100% will let your system run one GPU job at practically 100%, three CPU jobs at 100%, and one CPU job at 70% load. Check task manager and you can verify. =)

I am running Ubuntu, and watching the GPU usage, I see only 40-60% being used when I run SETI and 4 cores. Amicable Numbers used the GPU 100%. Could it be, that my CPU is just not fast enough with its 30% for the GPU?

I doubt it. What does the GPU usage on SETI look like with one CPU left idle for it?

That is 18kW/hr that would not be powered on or that would be at a much lower power state using much less draw. That seems to go against team Gridcoin's moto and claim to fame that Boinc/Gridcoin uses the IDLE left over unused power of machines that normally would be on but not doing anything but sitting running system background processes. Unless you are a University ( no better than kikipope and that brings up more questions and gray areas and legalities so I wont go there ) you are actually causing more power usage over all because not only are you running all of that hardware but the rest of the team has to run more hardware just to maintain earnings. I think it's time someone re-writes " instead of generating unnecessary heat and wasted power for the proof of work algorithms required to keep the coin network running. It is estimated that bitcoin is using approximately 23,312 megawatt hours per day " in our team description and puts down truths and facts vs knocking another. Thank you for pointing out the massive power sucker Gridcoin is growing to be , maybe we can fix that before it gets too out of hand and the quest to actually help the world by science is actually harm defeating the intended outcome. No I am not saying hug a tree , that tree needs to die for more to grow and stand in its place but maybe its time for a revamp and Gridcoin v3.0.

Mining Gridcoin since 20 Sep 1999 here! Of course, cryptos didn't exist then, but SETI@home did and BOINC arrived soon thereafter.

Yes, the idea of volunteer computing is much older than crypto and I missed the Bitcoin train because of it (didn't want to abandon my scientific computations because of some dubious hashing). However, I made sure I don't miss the Gridcoin train :)

Exactly, I've missed allot but gridcoin i won't.

I remember back in the day when i was hashing with a poor 9800 nvidia gard. Sold 12 bitcoins for 12$. Now it seems such a bad decision, but that was a high price at that time.

I have been only mining Gridcoin, and doing that since late 2014, I have been BOINCing for over 10 years but I lost my original account so dont have an exact start date.

Gridcoin helped me to focus on researching dilligently. As I mine on solar power its also very green :)

I would kill (well, maybe not kill, but you know...) for an 18kW solar set-up.

Would probably need to be at least double that to cover the night time. Do you only mine during the day, or do you have a battery?

I only mine in the day, its just personal choice, I wouldnt sleep for fear of my PC burning the house down anyway :)

I have a 4kW solar array onmy roof.

Im in the UK (51 degrees North) so our winters are not great for solarPV, because they are very cloudy and the sun is very weak, sometime I make less than 1kWh (on 10th Dec 2016 I made 0.21kWh) on those days I do still run BOINC and pay for the electric from the grid, but I tend to reduce output. Cold weather is good for solarPV though, so even in winter if its a clear cold day then my array makes a fair bit of power (a few days earlier on 4th Dec 2016 I made 11.3kWh!!), but its only producing anything decent for about 4 hours in the middle of the day. Summer of course is the opposite, with 16 hours a day of daylight and much less cloud, my 4kW system can produce over 25kWh a day, and even a really cloudy and rainy day will still do over 5kWh.

So to get 18kW for 24 hours a day, you would need a solar PV setup that produces 432kWh every day, even in winter, that is a significant sized system, plus batteries which aint cheap. I see you are in NZ though so at around 36 degrees South your winter/summer production would be a lot less variable than mine.

However, you dont have to do it all in one go, you can make a system that covers your summer requirements in daylight and then just contributes in the winter, this could still significanty reduce your energy bills. Self consuming 100% of your solarPV power is usually the most efficient in terms of ROI.

Haha! That reddit post is hilarious. "Apparently a killer with the ladies too..."

I'm pretty new to cryptocurrency and somehow stumbled upon Gridcoin while trying to get a Monero node running. I have an old i7 with an antiquated 1GB GeForce card but I've been contributing to SETI@home and LHC@home for almost a week now (internet connection willing). My rig earns a measly 0.4 GRC per day at the moment but it feels good being able to contribute. My eyes start to water when I think what a system capable of 800 GRC per day looks like...

Your mint rate will increase a lot, as it it based off your project RAC, which in turn takes a month to peak. =)

This means that when you start mining you will earn very slowly at first, but also that when you stop mining you will continue to receive diminishing payouts for the month or so after.

Thanks or clearing that up! I have seen it go up from 0.1 GRC a day to 0.4 over the past week but thought it was because I changed my GPU processing settings from "based on preferences" to "run always".

I'll be following to see what more I can learn about Gridcoin!

So much to learn and everyday comes with something interesting to learn and enjoy.
By the way, joining any of the community will be of great value and blessing, most especially the IRC #gridcoin and #gridcoin-help channels.
Good-luck

Great post as usual ;) I'm just a little bit confused about the electricity cost of your setup - at a rough estimate it seems to exceed the yield of the GRC earned that way?

Yup - you are 100% correct. Power costs about NZD$0.22 per kWh when weight averaging the day and night rates. This means I use:

18KW * 24h/day * NZD$0.22 = NZD$95/day

Meanwhile, I could sell the Gridcoin for about NZD$0.04 per coin, so I earn about:

800 GRC/day * NZD$0.04/GRC = NZD$32/day

That being said, recently GRC traded for almost NZD$0.15 at one point, which would have yielded NZD$120/day.

In reality, there are still several reason I continue doing this:

  • I need the hardware for my research anyway, and this is running on idle cycles. The power bill of using those cycles is only a fraction of the total compute powerbill.

  • The price of GRC will go up, so I will see a good return in future if I were to sell

  • I have plans to use the GRC I mine on idle cycles to commission compute in the future through the BOINC and GRC network, if I can overcome the technical hurdles of the models I work with.

There's always a blessing for selfless services but the selfish can never see or understand it.
Good luck in your endeavors.

GRC is not profitable like other coins, its a reward system for volunteering computing since 1999

It was very profitable a month ago. =)

It also will be again in the near future. Best to mine those coins before the flocks of miners arrive for the next GRC gold rush!

It's entirely possible that the NN participants is going to inflate a couple orders of magnitude when the team req is removed.. the BOINC community is huge!

Nailed it - and that would result in hugely increased mining difficulty and a rise in GRC price to match.

Difficulty - Yes, however there is always room for additional BOINC projects.

Price - Potentially, with greater scarcity and demand the likely outcome is an increase in market cap.

Thanks for sharing! I thought you were in this for a lot longer, because of your high quality posts ;)
Now you told us about your setup, but could you describe it a little bit more? Do you have several servers running for your research anyway? I am really interested :)

Haha no I am still a spring chicken! I have been running BOINC for a very long time though, on a different account. Now, all my compute is under Team Gridcoin.

My set-up is composed of the workstations in our lab and the research cluster when it is not running in-house research. I may write something about it down the line if people are interested.

I for one am defiantly interested.

On a serious note though - I will do a writeup with some photos in future. =)

Thanks for letting me know you're interested!

Touché. I don't claim spelling to be one of my greatest skills.

Thanks for the info - and yes, I would be really interested! :)

Wow - Very nice! There is no way I can compete with your setup! My small back office rig runs 24/7 overclocked to the bloody edge producing 330,000 RAC w/ 33 MAG (45 with 35% boost.) My wallet has a meager 3,800 GRC balance of which over 3,500 was purchased at an exchange in hopes of jump starting my staking chances. After 50 days I've only staked twice. First stake was the "newbie" / "gimme" stake of 1 GRC on 6/17, then over a month later my true non-newbie stake on 7/27.

To say I am envious of your results is an understatement. My results are quite disappointing. Certainly is deflating and depressing to know despite all my best efforts I could never obtain a tenth of your results.

Let me make an effort to inflate you then, and explain how you too are benefiting from my set-up.

Currently, competition on Gridcoin is still relatively low. It is increasing rapidly, but we are very much in the early days. As more and more people join the team, we will be increasing the compute power of the network. This means competition for the GRC minted every day goes up.

As a result of it being harder to mint Gridcoin, the price of Gridcoin will rise. It will further rise when research teams start using it to incentivise crunchers to compute for their specific project.

Compute added to network --> GRC price goes up

Every Gridcoin you possess or mine now, you will be able to sell for easily ten times that in future. This is thanks to competition from set-ups like mine. To add to that, there are set-ups running BOINC outside Team Gridcoin that utterly dwarf mine - soon enough they will join as well.

Your results are far from disappointing - they are just 'normal'. Almost everyone on Team Gridcoin will have a set-up similar to yours. You're doing great! Also, remember that research is not paid out from when you join, but from when you staked the noobie block. So, whatever you earned was earned between 6/17 and 7/27.

So all the for about $24/day? Do you have strong beliefs it will go up a lot? I'm running on one old PC, and getting about 2, using the GRCPool, and getting less then .3 GRC per day.

Yes, although the price has actually gone up 10% since yesterday. =)

I have strong beliefs:

  • In the vision of Gridcoin, and what is it achieving for the scientific community
  • That the coin will increase significantly in price in the near future
  • That I would still be doing this even if there was no GRC on offer

How long have you been with the pool? It takes just over a month for your magnitude, RAC, and thus daily GRC mint to peak.

Looks like 19 days where everything was working. Had one issue where it was down for a week or so, and I didn't notice. I'll keep watching with interest. I don't really mind running it anyway, except this old computer is my bedroom and make a wee bit of noise.

People in the BOINC community are running far more powerful setups, completely for free (i.e. they haven't joined Gridcoin yet). There are 500k active BOINC crunchers and over 4.3 million registered BOINC users. So yes, there is a lot of room for growth. Not to mention that, for the general population, the idea of computing for science is far more understandable and fascinating than repetitive Proof-of-Work hashing (which has zero benefits except securing the blockchain and is therefore difficult to understand for non-crypto people).

The aspect of doing something positive through science for humanity will hoodwink the kindhearted, the reward aspect is a catchy phrase which gives you a "hard-on" for everything Gridcoin | ; )

staking computer GRC.png

Hey guys! Just wanted to share this. You can Earn while you Trade or Hodl GRC at Btcpop.co!

If there are any lazy stakers out their Btcpop stakes deposits for you and gives you proportionate rewards