My switch from scrypt to X11

Carbon_Rod

Gawd
Joined
Apr 2, 2012
Messages
1,022
I just made the switch to X11 on 1 machine to give it a try. I pointed it at Nicehash for the time being. But being that I'm not yet familiar with it, what I'm getting doesn't seem normal.

It gets like 1 Accepted share message every so often, but then like 20 or so "Stratum from Pool 0 requested work restart" messages in between. It still says my hash rate is at about 2MH per card (280X) so it still looks like its hashing away nicely, though I see "Reconnect requested" messages from the server as well and my hash rate drops for a few seconds. Is this a quirk of Nicehash? X11 itself? Is there something else I should be setting?

EDIT: A few more minutes later, the work restart messages have lessened in number but I'm still getting a few of them... accepts are few and far between, but no rejects from what I see. Hmmm.
 

mkrohn

2[H]4U
Joined
Apr 30, 2012
Messages
2,345
I just made the switch to X11 on 1 machine to give it a try. I pointed it at Nicehash for the time being. But being that I'm not yet familiar with it, what I'm getting doesn't seem normal.

It gets like 1 Accepted share message every so often, but then like 20 or so "Stratum from Pool 0 requested work restart" messages in between. It still says my hash rate is at about 2MH per card (280X) so it still looks like its hashing away nicely, though I see "Reconnect requested" messages from the server as well and my hash rate drops for a few seconds. Is this a quirk of Nicehash? X11 itself? Is there something else I should be setting?

EDIT: A few more minutes later, the work restart messages have lessened in number but I'm still getting a few of them... accepts are few and far between, but no rejects from what I see. Hmmm.

Lower the diff. WIth them you can force it to be lower through the password field. I don't know the syntax off the top of my head but they have an FAQ page. Inch the diff lower and wait and see. I'm only using pimp so never actually sit there watching the miners
 

mkrohn

2[H]4U
Joined
Apr 30, 2012
Messages
2,345
OK, I started monitoring my rigs and I'm seeing the same thing as you Carbon. I was on the lowest diff already so I'm cranking it up now
 

Carbon_Rod

Gawd
Joined
Apr 2, 2012
Messages
1,022
I just started watching it again right now and it appears that things have sort of settled down a bit. I'm getting a lot more accepts than restarts now. I'll let it go for now and continue to monitor it.

But ya, I'll try adjusting the difficulty if things go all weird again.

EDIT: Just switched my other rigs over and noticed the behavior happening again... lots of restart requests, few accepts.
 
Last edited:

Mega666

[H]ard|Gawd
Joined
Apr 19, 2008
Messages
1,277
"Stratum from Pool 0 requested work restart" happens every time NB increases or a new block is found. The rate of NB increasing or how fast the network finds new blocks is dependent on your difficulty and hashrate. "Diff"

time = difficulty * 2**32 / hashrate
 
Last edited:

Carbon_Rod

Gawd
Joined
Apr 2, 2012
Messages
1,022
Things seem to have settled themselves down again. I'm thinking much of this issue is due to Nicehash itself.

I also just finished trying to set the diff manually via the command line and it doesn't seem to work. I followed the instructions on Nicehash and even used their own version of sph-sgminer to no avail. In fact their version doesn't report the difficulty for accepted blocks either. It keeps saying the diff is 0.000. This is even after giving it a manual setting.

So, I'm sticking with the sph-sgminer linked in one of the threads here (the one found at the Bitcointalk forums). It works and I'm fine with it.

EDIT: The 0.000 difficulty display issue is a known bug with the nicehash version of sph-sgminer... I may revisit that version tomorrow to see if it has any additional quirks that could cause it to perform better.
 
Last edited:

Spidey329

[H]F Junkie
Joined
Dec 15, 2003
Messages
8,683
I have noticed NiceHash giving about 50% of MH because of the switching.

I put one of my rigs on DrkPool for X11 to test out. The SGMiner 4.1.0-103-g00015 shows no accepted shares, despite getting consistent "accepted" messages in the screen. The pool itself shows my full MH/s.

Display error? Anyone know if there's a known bug in SGminer?
 

Carbon_Rod

Gawd
Joined
Apr 2, 2012
Messages
1,022
What the X11 community needs is a multipool that mines X11 coins but pays out in BTC or LTC. Or one of the major multipools like Waffle, or Coinswitch should start supporting X11.

I agree that Nicehash is convenient and pays relatively well but the constant work restarts and the reconnect requests every so often just kill things.
 

Nexillus

[H]ard|Gawd
Joined
Nov 6, 2012
Messages
1,116
Good to know about the work restarts and reconnects have dealth with that before and didn't end well for my machines >>
 

mkrohn

2[H]4U
Joined
Apr 30, 2012
Messages
2,345
What the X11 community needs is a multipool that mines X11 coins but pays out in BTC or LTC. Or one of the major multipools like Waffle, or Coinswitch should start supporting X11.

I agree that Nicehash is convenient and pays relatively well but the constant work restarts and the reconnect requests every so often just kill things.

yeah the problem seems to be having to use a different kernel. Look at the questions around the various forums. The majority of miners are completely STUPID. X11 is actually by far the easiest setup I've come across. I'd like to see pimp just default to X11 so that people can just plug in, punch in the pool info and be mining since tweaking doesn't really do anything for X11.
 

pheonix991

Limp Gawd
Joined
Jun 24, 2009
Messages
338
What the X11 community needs is a multipool that mines X11 coins but pays out in BTC or LTC. Or one of the major multipools like Waffle, or Coinswitch should start supporting X11.

I agree that Nicehash is convenient and pays relatively well but the constant work restarts and the reconnect requests every so often just kill things.

I've been on trademybit for a while, been working out great.
 

Spidey329

[H]F Junkie
Joined
Dec 15, 2003
Messages
8,683
I have noticed NiceHash giving about 50% of MH because of the switching.

I put one of my rigs on DrkPool for X11 to test out. The SGMiner 4.1.0-103-g00015 shows no accepted shares, despite getting consistent "accepted" messages in the screen. The pool itself shows my full MH/s.

Display error? Anyone know if there's a known bug in SGminer?

So I figured this out. The reason no "accepted" shares are there despite getting consistent "share accepted" messages when viewing the console is because X11 does share fractions and SGMiner rounds up or down.

When I connected with CGRemote to see what the rig was doing, I saw this:
Accepted: 2.08984375 <--- As soon as it reaches 2.5, the console reports A:3
Rejected: 0.046875
Stale: 1.43359375

I assume this is just how X11 works? My pool dashboard still says it sees the full hashrate.

But I have noticed once the rig is running for awhile (I left it overnight), I did notice accepted dropped to 60% and stale was around 30%. I'm assuming there was an internet outage or something. That or the pool sucks.

With that in mind, temperatures of X11 seem to be sporadic. Some cards sit around 73c, while others sit around 39c. They were so low, that my autorebooter RaspPi was rebooting them because it thought they died. I had to adjust the config on that.
 
Last edited:

Mega666

[H]ard|Gawd
Joined
Apr 19, 2008
Messages
1,277
So I figured this out. The reason no "accepted" shares are there despite getting consistent "share accepted" messages when viewing the console is because X11 does share fractions and SGMiner rounds up or down.

When I connected with CGRemote to see what the rig was doing, I saw this:
Accepted: 2.08984375 <--- As soon as it reaches 2.5, the console reports A:3
Rejected: 0.046875
Stale: 1.43359375

I assume this is just how X11 works? My pool dashboard still says it sees the full hashrate.

But I have noticed once the rig is running for awhile (I left it overnight), I did notice accepted dropped to 60% and stale was around 30%. I'm assuming there was an internet outage or something. That or the pool sucks.

With that in mind, temperatures of X11 seem to be sporadic. Some cards sit around 73c, while others sit around 39c. They were so low, that my autorebooter RaspPi was rebooting them because it thought they died. I had to adjust the config on that.

Auto-rebooter Pi? Good god. I just dialed in the temps with intensity, Temps are stable in X11 and all other algorithms. It's the environment that causes temperature fluctuations. Missing hash is pretty much gone where you guessed it did. If it's not in the miner stats, question the network and the pool.
 

Spidey329

[H]F Junkie
Joined
Dec 15, 2003
Messages
8,683
Auto-rebooter Pi? Good god. I just dialed in the temps with intensity, Temps are stable in X11 and all other algorithms. It's the environment that causes temperature fluctuations. Missing hash is pretty much gone where you guessed it did. If it's not in the miner stats, question the network and the pool.

Yeah. I've had two switches get cranky, so they sometimes get intermittent issues. I assume it might be heat related, since everyone seems to be switching to the "fanless green" standard. With temperatures, I was referring to cards in the same rig. So there at one point was a wide separation of temps with cards sitting next to each other. Ambient doesn't play a huge role in that situation.

As for the autorebooter. My rigs are off site - I don't want to have to monitor them 24/7. Although the majority are 100% stable 24/7, I have a few that are made up of the "misfits" (lemons) - all these cards were swapped from 24/7 rigs to lemon rigs. Cards that work for X time and then just die. I could under clock them to get them stable, but I did calculations that showed if they got a reboot after 6hours at higher hashrates, they'd produce more work then if they were underclocked to he 24/7 stable for mining. What the Pi does is monitor that. As soon as one card dies, it reboots the rig. Sometimes they stay up for 10 hours, others it's 4. I have them tuned to get the best benefit.

It monitors the other "stable" rigs just to he sure. Nothing more annoying than coming into the office after a long weekend and find some rigs offline because of some random SGMiner crash or super-low hashrates. Regardless of how stable the rig might be, sometimes things get cranky. Whether it was an internet outtage that screwed things up or a random kernel crash.
 
Top