Building Cheap Miners : My "Secret"

[spinx]

The Dell R815s that I got are running rock solid.

Hard to justify the price at this point, though, IMHO -- unless you get a sweet deal.

Well... I have to take that back a bit now.

My power flickered last night thus turning off all my hardware. Came in, turned all the Dells back on... started XMR-STAK on them all & walked off.

Noticed my pool hash-rate never returned to how it was... every one of my R815s is now ~400 H/S slower than they were before the power flickered (1600 H/S vs. 2000 H/S).

I've tried re-booting and even re-compiling XMR-STAK... seems to be permanent ?!

Any ideas ?

Im really shocked you have not invested in UPS for your systems man.. i ordered 25 x of the CyberPower CP1500AVRLCD, i got them refurbished off ebay, the guy only had 25 left, so i contacted him to ask if i could buy all 25 at a better rate and he sold them to me for $88 each, but at that time was spring time sales, so i was able to get 20% using the emailed spring code i had sitting in reserves. I have one of these 1500 units on every 6 and less GPU rig i have running right now...

I was about to pull the trigger before all the coins fell off a cliff in value, hah...

I am pricing out a full set of UPS units, though, through some distributors... since I'll need 100+ of them. Gonna just have to suck it up I suppose!

Try to source them refurbished or on ebay with bad batteries. I bought 3 from ebay that were server based UPS that just needed new batteries and i picked them up off amazon for a few dollars each, i even went to my local thrift stores and picked up a handful of them for a few dollars and half of them even worked, just needed fresh batteries. The ones i got from local thrift stores i built custom lithium batteries from 18650 cells, i have been collecting thousands of them over time because i was planning to build a DIY tesla powerwall with them along with build battery packs for my Catrike bicycle that i converted to electric assist, i get them free from thrift stores, hardware stores from the recycle bin just ask the store managers, and colleges. The most i got from thrift stores because when someone donates a laptio they cant resell it, so it gets scrapped and disassembled and they pay for someone to take the batteries away.

This is some of the batteries i got last month from going around and asking politely from many locations and picking up some from locations that told me they would place them to the side for me...


Here are some cells im still slowly testing out, i bought 4 tester/chargers units a while back that allows me to test 16 cells at a time, so its only time to test them, i put them on the test and forget about them really and come back when i remember about them, write the capacity on the side of the cell and move on, i store them in bins based on the capacity so my wall cells i assemble have the highest capacity cells. I have probably close to 40,000 cells assembled in my diy powerwall to date that ive never purchased...

"I have probably close to 40,000 cells assembled in my diy powerwall to date that ive never purchased... "

Thats seriously impressive man!

[1715422443]

1715422443

[1715422443]

1715422443

[1715422443]

1715422443

[1715422443]

1715422443

[1715422443]

1715422443

[swogerino]

Under Linux, I have 6 Nvidia cards working.  Not sure what is up with the Proliant G7 series.  I have tried a DL360, DL380 and the DL580 and none of these will boot with more than 4 GPUs installed.  Windows will lock up at loading screen.

Slap a USB stick with HiveOS on it and 6 get recognized easily.

I have a ML350P G8 that has taken 6 GPUs in windows easily.  

Edit* Make that 9 1060s in one 580 in HiveOS.

I hit a wall at 9 cards on the 580 also, using SMOS. Spent countless hours getting 9 cards to work. In one way i love the HP servers, H/s on cryptonight is a bonus but as a gpu-mining platform they look good on paper but is in my experience not fun maintaining. Did anyone try to run Windows 10 on the 580 with success?

That is a problem with the HP server series like the G7 which are a bit old so HiveOS maybe does not have the best support for them. I will also blame HP or any other old server like Dell as I have tested HiveOS with 12 RX 580 cards in ASUS B250 Pro Mining and they worked out of the box perfectly without problems with a great hashrate of 360 Mhs.

The strong Xeon processors are good to mine Bytecoin or Monero.

[PharmEcis]

That is a problem with the HP server series like the G7 which are a bit old so HiveOS maybe does not have the best support for them. I will also blame HP or any other old server like Dell as I have tested HiveOS with 12 RX 580 cards in ASUS B250 Pro Mining and they worked out of the box perfectly without problems with a great hashrate of 360 Mhs.

The strong Xeon processors are good to mine Bytecoin or Monero.

Actually HiveOS and the HP's work GREAT.  Currently have 2 running 9 cards each.

It's HiveOS and the Dell R815's that DOESN'T and that's just a NIC driver issue.

[spinx]

That is a problem with the HP server series like the G7 which are a bit old so HiveOS maybe does not have the best support for them. I will also blame HP or any other old server like Dell as I have tested HiveOS with 12 RX 580 cards in ASUS B250 Pro Mining and they worked out of the box perfectly without problems with a great hashrate of 360 Mhs.

The strong Xeon processors are good to mine Bytecoin or Monero.

Actually HiveOS and the HP's work GREAT.  Currently have 2 running 9 cards each.

It's HiveOS and the Dell R815's that DOESN'T and that's just a NIC driver issue.

Guess i have to give HiveOS a second chance then

[jakkwb]

So with Monero back up again I decided to fire up my DL385 and give it another try. Put new XMR-Stak on it, configged and fired up. Still getting 660hs. This is stock with no GPUs in it.

Has anyone out there achieved a better rate than this?

[PharmEcis]

So with Monero back up again I decided to fire up my DL385 and give it another try. Put new XMR-Stak on it, configged and fired up. Still getting 660hs. This is stock with no GPUs in it.

Has anyone out there achieved a better rate than this?

Cranking out 1050 on a pair of 6380s on a 385...

[jakkwb]

Mine has the 6220s. Sounds like I need a processor upgrade...

[PharmEcis]

Mine has the 6220s. Sounds like I need a processor upgrade...

I have a set of spare 6276s I will sell.  They will do ~950H/s.

[sundownz]

The Dell R815s that I got are running rock solid.

Hard to justify the price at this point, though, IMHO -- unless you get a sweet deal.

Well... I have to take that back a bit now.

My power flickered last night thus turning off all my hardware. Came in, turned all the Dells back on... started XMR-STAK on them all & walked off.

Noticed my pool hash-rate never returned to how it was... every one of my R815s is now ~400 H/S slower than they were before the power flickered (1600 H/S vs. 2000 H/S).

I've tried re-booting and even re-compiling XMR-STAK... seems to be permanent ?!

Any ideas ?

Some servers have a bios setting for throttling in the event of a power failure. It's possible it got stuck in that state? Have you tried a full proper power shut down and restart?

I have not done a complete "hard" power off -- I will give that a shot.

Okay tried on two systems. One just a "shut down" then power back on. One with a "shut down" and unplug / replug the system. No effect on either system. Both still doing only 1600 H/S

Apparently this IS, in fact, a thing :

https://community.spiceworks.com/topic/1447342-dell-r820-slow-after-power-outages

So... that being said I'll have to work on them tomorrow when I have some time.

One of the responses in that thread... Seems like it'd be the fastest solution in your case.

I did discover that if I shut the server all the way down, pull the plugs, wait 10-15 seconds, then plug it back in. After rebooting, it seems to be acting normal again. So what is it about pulling the power OR changing the performance profile that would fix whatever issue is happening?

So... I have tried on 5 systems so far to power off, unplug AC power, AND unplug the CMOS battery -- leave them unplugged long enough to clean out the system with some air. Re-assemble... and start mining. They are all STILL doing only 1600 H/S vs. 2000 H/S before the power flicker.

[sundownz]

The Dell R815s that I got are running rock solid.

Hard to justify the price at this point, though, IMHO -- unless you get a sweet deal.

Well... I have to take that back a bit now.

My power flickered last night thus turning off all my hardware. Came in, turned all the Dells back on... started XMR-STAK on them all & walked off.

Noticed my pool hash-rate never returned to how it was... every one of my R815s is now ~400 H/S slower than they were before the power flickered (1600 H/S vs. 2000 H/S).

I've tried re-booting and even re-compiling XMR-STAK... seems to be permanent ?!

Any ideas ?

Some servers have a bios setting for throttling in the event of a power failure. It's possible it got stuck in that state? Have you tried a full proper power shut down and restart?

I have not done a complete "hard" power off -- I will give that a shot.

Okay tried on two systems. One just a "shut down" then power back on. One with a "shut down" and unplug / replug the system. No effect on either system. Both still doing only 1600 H/S

Apparently this IS, in fact, a thing :

https://community.spiceworks.com/topic/1447342-dell-r820-slow-after-power-outages

So... that being said I'll have to work on them tomorrow when I have some time.

One of the responses in that thread... Seems like it'd be the fastest solution in your case.

I did discover that if I shut the server all the way down, pull the plugs, wait 10-15 seconds, then plug it back in. After rebooting, it seems to be acting normal again. So what is it about pulling the power OR changing the performance profile that would fix whatever issue is happening?

So... I have tried on 5 systems so far to power off, unplug AC power, AND unplug the CMOS battery -- leave them unplugged long enough to clean out the system with some air. Re-assemble... and start mining. They are all STILL doing only 1600 H/S vs. 2000 H/S before the power flicker.

I have now also tried this from the link :

" I have to go into the BIOS and switch the performance settings from Max Performance to something else, then back. After that, the server performs at lightning speed again. "

Tried on two systems so far -- zero effect.

So... I'm kind of stumped.

[PharmEcis]

Try clearing the NVRAM by changing jumper position on the board. 

* Locate the 3-pin CMOS jumper (CLR CMOS) on the motherboard
* Remove the jumper plug from pins 1 and 2
* Place the jumper plug on pins 2 and 3 and wait approximately 5 seconds
* Replace the jumper plug on pins 1 and 2

Are you running both PSUs or did you remove one like I did?  Maybe try swapping PSU to the other bay or even have both PSUs installed and plugged in.

Another thing to try is resetting BIOS to defaults in the BIOS itself.

I'll do what I can to help you trouble shoot even though you won't take my phone calls!

<joke>

[MinersRus]

For those here who are using HiveOS I recommend that you do a reinstall with the 0.5-45 Image and not just do an update.

https://hiveos.farm/install

0.5-45 IMAGE RELEASE 2018-04-19

    Logs are OFF by default
    Linux 4.13 kernel
    Installed ROCm with AMD driver to support Vega
    Configs partition is now NTFS
    Intel's E1000E updated network driver by default

0.5-45 2018-04-17

    NEW: cpuminer-opt
    ethminer 0.14 and 0.15 in the box, version can be selected in the wallet
    amdmeminfo updated

The image uses Linux 4.13 kernel and if you just do the update you will still be on the 4.10 kernel.

On thing I noticed is that on the Linux 4.13 "mpstat -P ALL" now reports the 32 cores on my 4x 8837's correctly whereas Linux 4.10 reported 40 cores with 2-dummy per processor. With 4.10 I couldn't do a custom XMR-Stak CPU configuration now with the Linux 4.13 kernel I can.

[sundownz]

So I have been on ZergPool for a while now since they launched X16R on there.

But... the earnings seem a bit low lately so I am experimenting with my friends farm first (I manage it for him)... as he has only 10 machines.

He has 60x GTX 1080Ti but has only averaged ~$1.90 USD / Day per card for the last week or so -- so I am pushing him back onto ZPOOL to see how it stacks up (when we both swapped to ZERG it was doing much higher).

[honestinvestor]

IN GENERAL DO NOT SEE MEET TO TAKE ALL THE EARLY CHEAP EQUIPMENT, LOOK BETTER BETWEEN THE PRICE AND THE QUALITY THEN WILL GO ON TOLY

[sundownz]

IN GENERAL DO NOT SEE MEET TO TAKE ALL THE EARLY CHEAP EQUIPMENT, LOOK BETTER BETWEEN THE PRICE AND THE QUALITY THEN WILL GO ON TOLY

I'm not sure I understand your statement.

All of my Z400s have been bullet-proof... close to a year now on the first ones I bought running 24/7.

[PharmEcis]

IN GENERAL DO NOT SEE MEET TO TAKE ALL THE EARLY CHEAP EQUIPMENT, LOOK BETTER BETWEEN THE PRICE AND THE QUALITY THEN WILL GO ON TOLY

I'm not sure I understand your statement.

 

[m.vina]

This is some of the batteries i got last month from going around and asking politely from many locations and picking up some from locations that told me they would place them to the side for me...

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.

[sundownz]

Try clearing the NVRAM by changing jumper position on the board.  

* Locate the 3-pin CMOS jumper (CLR CMOS) on the motherboard
* Remove the jumper plug from pins 1 and 2
* Place the jumper plug on pins 2 and 3 and wait approximately 5 seconds
* Replace the jumper plug on pins 1 and 2

Are you running both PSUs or did you remove one like I did?  Maybe try swapping PSU to the other bay or even have both PSUs installed and plugged in.

Another thing to try is resetting BIOS to defaults in the BIOS itself.

I'll do what I can to help you trouble shoot even though you won't take my phone calls!

<joke>

So I have tried:

1) Reset NVRAM with Jumper.
2) Completely re-install BIOS with Dell Utility.

Neither works... still ~1600 H/S.

I am running both supplies... I'll try swapping them with one another.

----

I just set up two more R815s that I had not previously powered on or set up at all... they both run at 2100 H/S just like they should / my other ones used to. I cannot find a single difference in any setting.

[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 maintanance.

[sundownz]

Try clearing the NVRAM by changing jumper position on the board.  

* Locate the 3-pin CMOS jumper (CLR CMOS) on the motherboard
* Remove the jumper plug from pins 1 and 2
* Place the jumper plug on pins 2 and 3 and wait approximately 5 seconds
* Replace the jumper plug on pins 1 and 2

Are you running both PSUs or did you remove one like I did?  Maybe try swapping PSU to the other bay or even have both PSUs installed and plugged in.

Another thing to try is resetting BIOS to defaults in the BIOS itself.

I'll do what I can to help you trouble shoot even though you won't take my phone calls!

<joke>

So I have tried:

1) Reset NVRAM with Jumper.
2) Completely re-install BIOS with Dell Utility.

Neither works... still ~1600 H/S.

I am running both supplies... I'll try swapping them with one another.

----

I just set up two more R815s that I had not previously powered on or set up at all... they both run at 2100 H/S just like they should / my other ones used to. I cannot find a single difference in any setting.

SOLVED... But leaves me with more questions than answers...

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... the weird part is I did NOT have to do this before the power flicker. I also did NOT have to do this on the two new systems I setup today to get the full speed.

I have no idea how a power flicker could make changing these settings a necessity.