Building Cheap Miners : My "Secret"

[PharmEcis]

Have you been using hugepages already?  I've been using it since day one.

The only thing I can think of is that because of the power flicker something happened kernel wise maybe?  Something changed that required the need for the band-aid you had to apply.

For shits and giggles as I know you have a lot of boxes that are going to need your remedy, image one of the new boxes if using HDD, just swap it into a broken box and see if it magically starts to work right because it appears the fix was software based and had nothing to do with hardware.

I'm really leaning to something on the reboot screwed up in the OS and has caused all your grief which is why I'm assuming a fresh install worked just fine on the new boxes.

I'm gearing up to start removing all my drives and do network boot here soon.

I found something that is working awesome on these dells...  but ... do I share?  You have enough firepower to drastically increase diff and still won't return my calls...

[PharmEcis]

The lithium batteries for your power wall i understand man. Good job on saving money on those.. But these old laptop batteries?

If you don't mind may i ask what you use them for? Thats a lot of junk that could be flammable if i understand correctly.

I hope you understand that the batteries for his power wall come from those laptop batteries right?  He tears the laptop battery casings open to extract the Lithium cells then proceeds to test and bin accordingly.

[MinersRus]

ı also build some cheap rigs. it is ok if you have max 5-6 rigs but if you plan 10-20 or more rigs best is 8 gpu rigs with new motherboard and gpus. reason is maintenance.

I beg to differ.

I will put my Oldish HP DL580 G7 from 2010 upgraded to 4x E7-8837's with 9-11 GPUs up against those newfangled rigs you mentioned. Being a Server it has robustness designed in. Redundant High Efficiency Gold/Platinum Power Supplies, ECC Error Corrected Memory, Superior Cooling. The parts in it are made to last for a very long time and if maintenance is needed it is easy to do.

The cost savings on these older systems vs new systems far out-way what little maintenance costs they might have sometime in the future.

If you use the same systems to build your farm it shouldn't matter if they are oldish or newish when it comes to maintenance as long as you go with quality hardware.

[Sevarchik]

What you think about using one simple ATX supply per card and connect their black and green wires for sync
Simple atx supply cost around 8$ and it easy cat drive one GPU

[sundownz]

What you think about using one simple ATX supply per card and connect their black and green wires for sync
Simple atx supply cost around 8$ and it easy cat drive one GPU

It's something I've thought of but haven't tried... I have 80 or so of the stock Z400 supplies at this point, haha... each one can power one fairly robust card.

I built a gaming PC for my son with a Z400 and stock PSU and when he isn't using it I do mining... with a GTX 1080. Just have to turn power down a bit in afterburner =)

[sundownz]

ı also build some cheap rigs. it is ok if you have max 5-6 rigs but if you plan 10-20 or more rigs best is 8 gpu rigs with new motherboard and gpus. reason is maintenance.

I beg to differ.

I will put my Oldish HP DL580 G7 from 2010 upgraded to 4x E7-8837's with 9-11 GPUs up against those newfangled rigs you mentioned. Being a Server it has robustness designed in. Redundant High Efficiency Gold/Platinum Power Supplies, ECC Error Corrected Memory, Superior Cooling. The parts in it are made to last for a very long time and if maintenance is needed it is easy to do.

The cost savings on these older systems vs new systems far out-way what little maintenance costs they might have sometime in the future.

If you use the same systems to build your farm it shouldn't matter if they are oldish or newish when it comes to maintenance as long as you go with quality hardware.

Indeed.

I have a few "brand new" hardware miners... and I've had several motherboard failures (total or single PCI-E slot).

Z400s... zero system hardware has failed (I had one hard drive go out.. but that is unrelated to the Z400 itself).

[sundownz]

Have you been using hugepages already?  I've been using it since day one.

The only thing I can think of is that because of the power flicker something happened kernel wise maybe?  Something changed that required the need for the band-aid you had to apply.

For shits and giggles as I know you have a lot of boxes that are going to need your remedy, image one of the new boxes if using HDD, just swap it into a broken box and see if it magically starts to work right because it appears the fix was software based and had nothing to do with hardware.

I'm really leaning to something on the reboot screwed up in the OS and has caused all your grief which is why I'm assuming a fresh install worked just fine on the new boxes.

I'm gearing up to start removing all my drives and do network boot here soon.

I found something that is working awesome on these dells...  but ... do I share?  You have enough firepower to drastically increase diff and still won't return my calls...

Sorry... I am terrible at returning PMs so I tend not to look at them on forums (I'm on a lot of forums for my main business as well) 

-----

I never setup huge pages on the Dell systems as I had already achieved what I had read to be max hash-rate of ~2100 H/S without doing so -- and being new to Linux saved myself the hassle. I am going to just go ahead and do it on the new ones I setup just in case.

On my Windows systems I always enable it as I see a direct difference.

[sundownz]

I had to do the following:

1) Type "sudo sysctl -w vm.nr_hugepages=128"
2) Edit "/etc/security/limits.conf"
3) Add "* soft memlock 262144" and "* hard memlock 262144"
4) Change in XMR-STAK "config.txt" to "use_slow_memory : never"

This resolved the issue.

So... it seems I have to do #1 every time I re-boot as well now.

The other steps are saved but it seems to forget #1 upon re-boot.

[timbereagle]

ı also build some cheap rigs. it is ok if you have max 5-6 rigs but if you plan 10-20 or more rigs best is 8 gpu rigs with new motherboard and gpus. reason is maintenance.

I beg to differ.

I will put my Oldish HP DL580 G7 from 2010 upgraded to 4x E7-8837's with 9-11 GPUs up against those newfangled rigs you mentioned. Being a Server it has robustness designed in. Redundant High Efficiency Gold/Platinum Power Supplies, ECC Error Corrected Memory, Superior Cooling. The parts in it are made to last for a very long time and if maintenance is needed it is easy to do.

The cost savings on these older systems vs new systems far out-way what little maintenance costs they might have sometime in the future.

If you use the same systems to build your farm it shouldn't matter if they are oldish or newish when it comes to maintenance as long as you go with quality hardware.

If somebody has those uniqe server MB, yes you are %100 right.

However I am talking about old generation MB which has maximum 4- PCIE slots.
and with power supplies of 500w.

If you can find 8-10 PCIE MB that is %100 fine.

What I mean is number of rigs is important as more rigs takes more time usually. Also CPU and system consumes energy therefore minimum 8 GPU is better.

[sundownz]

I had to do the following:

1) Type "sudo sysctl -w vm.nr_hugepages=128"
2) Edit "/etc/security/limits.conf"
3) Add "* soft memlock 262144" and "* hard memlock 262144"
4) Change in XMR-STAK "config.txt" to "use_slow_memory : never"

This resolved the issue.

So... it seems I have to do #1 every time I re-boot as well now.

The other steps are saved but it seems to forget #1 upon re-boot.

SOLVED:

Replace Step #1 as follows:

1) In console go to "/etc" -- type "sudo gedit sysctl.conf" -- add "vm.nr_hugepages=128"

This saves it permanently.

So... not a huge fan of Linux over here, haha... such a PITA.

[jmigdlc99]

What you think about using one simple ATX supply per card and connect their black and green wires for sync
Simple atx supply cost around 8$ and it easy cat drive one GPU

That would probably be highly inefficient. Each ATX PSU would have to run some sort of fan to cool the PSU components. So imagine if you had 8 GPUs and 8 PSUs. You would have an additional 7 PSU fans that would be running all consuming more electricity that would eat up your mining profits. It would be so bulky and space consuming too.

Not cool.

[HLB11]

Is cheap mining a cheap hard drive?

[PharmEcis]

If somebody has those uniqe server MB, yes you are %100 right.

However I am talking about old generation MB which has maximum 4- PCIE slots.
and with power supplies of 500w.

If you can find 8-10 PCIE MB that is %100 fine.

What I mean is number of rigs is important as more rigs takes more time usually. Also CPU and system consumes energy therefore minimum 8 GPU is better.

Your reading comprehension must be slow.  The z400's came with 4 PCIE slots and like a 450W PSU.  He's been using like 40 of them perfectly fine.  In fact, he's even got more than 4 cards running in them thanks to PCIE 4 in 1 splitters.

The system energy is minimal and all CPUs are mining as well.  Another thing to point out is that when an 8 GPU rig goes down, you have much more hashpower out of the game.

Also, with HIVEOS, system updates are effortless.  Literally, even for a ton of boxes.

And honestly, Sundownz should be running a proxy because then literally all your boxes point at one, and then you make changes on the ONE, and EVERYTHING points to whatever you changed it to...

Please, this thread now has multiple combined years of testing these theories with the various people doing it.

[PharmEcis]

So... not a huge fan of Linux over here, haha... such a PITA.

It's the quirks, once you figure them out, it's actually WAY easier and WAY more stable.  I don't really touch them anymore.  Now my Windows Vega boxes... GRRRR

[PharmEcis]

Is cheap mining a cheap hard drive?

What?  There is mining that can be done with hard disk drives but you need TERABYTES and it don't pay.  You might even be storing goat porn.



Jacob, you better check your PM this time. 

[sundownz]

The system energy is minimal and all CPUs are mining as well.  Another thing to point out is that when an 8 GPU rig goes down, you have much more hashpower out of the game.

Also, with HIVEOS, system updates are effortless.  Literally, even for a ton of boxes.

And honestly, Sundownz should be running a proxy because then literally all your boxes point at one, and then you make changes on the ONE, and EVERYTHING points to whatever you changed it to..

I still have a ways to go in refining it all -- Linux at all is completely new to me.

But yes... the proxy idea sounds solid & I have been meaning to try out HIVEOS as well.

Thanks for the suggestions =)

[MinersRus]

Is cheap mining a cheap hard drive?

I take it that you mean to use cheap hard drives in mining rigs.

I'm using cheap 60GB SSD's that I pick up for $14-$17. Much more reliable than spinning hard drives and use less power.

[MinersRus]

The system energy is minimal and all CPUs are mining as well.  Another thing to point out is that when an 8 GPU rig goes down, you have much more hashpower out of the game.

Also, with HIVEOS, system updates are effortless.  Literally, even for a ton of boxes.

And honestly, Sundownz should be running a proxy because then literally all your boxes point at one, and then you make changes on the ONE, and EVERYTHING points to whatever you changed it to..

I still have a ways to go in refining it all -- Linux at all is completely new to me.

But yes... the proxy idea sounds solid & I have been meaning to try out HIVEOS as well.

Thanks for the suggestions =)

Not to toot my own horn, well okay I am  , but I got started on HiveOS March 28th. I got the idea to check out HiveOS from spinx who should get the original credit.
https://bitcointalk.org/index.php?topic=1955358.msg33341701#msg33341701

Doing some PMs with PharmEcis regarding the DL580's I mentioned how well HiveOS was working with them. He moved to using HiveOS on the DL580's because Windows was being a bitch on them.

Glad to share info with fellow miners so that we all may prosper.

[MinersRus]

I'm gearing up to start removing all my drives and do network boot here soon.

I found something that is working awesome on these dells...  but ... do I share?  You have enough firepower to drastically increase diff and still won't return my calls...

Why not share with the class and help out fellow miners.

[PharmEcis]

I'm gearing up to start removing all my drives and do network boot here soon.

I found something that is working awesome on these dells...  but ... do I share?  You have enough firepower to drastically increase diff and still won't return my calls...

Why not share with the class and help out fellow miners.

Because I want a MN before everyone else gets it...   You have PM.