OFFICIAL CGMINER mining software thread for linux/win/osx/mips/arm/r-pi 4.11.1

[jmc1517]

Downloaded and started running cgminer-3.7.2

After about 40 minutes, 2 AMU LEDs (27/28) came on solid, but by the time I had checked the display only 1 was showing as zombie, then both LEDs went out and no zombies remained in the display. Yay!  Both AMUs had re-initialised as AMU 34/35.  This looks very promising

[1713585922]

1713585922

[1713585922]

1713585922

[Jeffrey]

Hi,

Maybe a weird question, but is there a limit on the amount of Block Erupters cgminer is able to run? I'm asking this because I'm trying to set 98 up, (2x49port hubs) and it doesn't work.

When I connect hub 1 (with 49 BE's) it's fine. When I disconnect 1 and connect hub 2, it's fine. (so it's not a power issue, both are powered by an ATX supply with 30a on 5v) When I connect both of them at the same time, cgminer tells me it can't find any device and it's waiting for a hotplug.

When connect 1 full and a couple on the 2nd hub, it works. cgminer crashes on hotplug, so I have to close it, add more, and restart it. This works until I have 59. 59 is fine, 60 gives this error. I've tried putting the 60th in different usb ports (also on the laptop host itself), also other sticks, but it just won't accept  60 or more.

This is the behavior on cgminer 3.6.4, 3.7.0 and 3.7.2 (and often, 3.7.2 crashes after 1 or 2 minutes regardless of how many I've attached)

Using windows 7 64bit

Any advise?

[SpAcEDeViL]

CGMiner 3.6.6. Crash Raspberry Pi with Raspbian after 20 min, or 2 h or 6h... random system freeze.

Only restart with power off.

[chadtn]

CGMiner 3.6.6. Crash Raspberry Pi with Raspbian after 20 min, or 2 h or 6h... random system freeze.

Only restart with power off.

I did an update & upgrade yesterday in anticipation of my K16's showing up.  As soon as I popped them in my Raspberry Pi running Raspbian Wheezy started crashing constantly any time I tried to hash.  zipiju posted the following on www.projectklondike.com and it fixed my problem instantly:

Raspberry Pi is locking up when running CGMiner

If your RPi is freezing randomly after some time (could be minutes, could be hours or days) with error like Oops: PREEMPT ARM, CPU0 Not tainted printed out over HDMI, and with locked up ethernet and USB interfaces, you have to add following option to the /boot/cmdline.txt file on SD card:

slub_debug=FP
Make sure you do not add that option to the new line if some other options are already present in that file. Always paste it behind all other options delimited with space. Also, before editing, make sure your /boot partition is mounted by issuing this command:

mount | grep boot
It should print out something like this when it is mounted:

/dev/mmcblk0p1 on /boot type vfat (rw,noatime)
If it is not mounted, that command returns nothing. To mount /boot partition, just do:

mount /boot
Freezing is caused by some bug probably in Kernel itself, as you can read on http://www.raspberrypi.org/phpBB3/viewtopic.php?t=55370&p=421883:

"We recently changed the kernel memory allocator to SLUB for increased performance (incidentally also allowing this sort of debugging) which meant that this bug, which has probably always been there, has crawled into the open." by jdb.

To this date, this bug is still present in Master and Next branches of RPi firmware/kernel (https://github.com/raspberrypi/firmware) - up to a Kernel version 3.10.15 and maybe even latter.

Chad

[-ck]

Yay finally a night's sleep after a release without showstopper bugs.

[SpAcEDeViL]

Hy,

i have

Code:

Linux raspberrypi 3.10.18+ #577 Tue Nov 5 12:33:36 GMT 2013 armv6l GNU/Linux

Code:

Oct 18 2013 16:07:43
Copyright (c) 2012 Broadcom
version dbda126981820ea5681263042435388dddf65131 (tainted) (release)
root@raspberrypi:~#

Last kern.log before crash:

Code:

Nov  5 21:19:00 raspberrypi kernel: [   10.741945] bcm2708_i2c bcm2708_i2c.1: BSC1 Controller at 0x20804000 (irq 79) (baudrate 100k)
Nov  5 21:19:00 raspberrypi kernel: [   10.836278] bcm2708_spi bcm2708_spi.0: master is unqueued, this is deprecated
Nov  5 21:19:00 raspberrypi kernel: [   10.846583] bcm2708_spi bcm2708_spi.0: SPI Controller at 0x20204000 (irq 80)
Nov  5 21:19:00 raspberrypi kernel: [   12.340492] usbcore: registered new interface driver rtl8192cu
Nov  5 21:19:00 raspberrypi kernel: [   18.815515] EXT4-fs (mmcblk0p2): re-mounted. Opts: (null)
Nov  5 21:19:00 raspberrypi kernel: [   19.255702] EXT4-fs (mmcblk0p2): re-mounted. Opts: (null)
Nov  5 21:19:00 raspberrypi kernel: [   20.328267] i2c /dev entries driver
Nov  5 21:19:00 raspberrypi kernel: [   25.039911] FAT-fs (mmcblk0p1): Volume was not properly unmounted. Some data may be corrupt. Please run fsck.
Nov  5 21:19:00 raspberrypi kernel: [   28.192313] smsc95xx 1-1.1:1.0 eth0: hardware isn't capable of remote wakeup
Nov  5 21:19:02 raspberrypi kernel: [   50.173568] Adding 102396k swap on /var/swap.  Priority:-1 extents:2 across:507900k SS
Nov  5 21:25:37 raspberrypi kernel: imklog 5.8.11, log source = /proc/kmsg started.
Nov  5 21:25:37 raspberrypi kernel: [    0.000000] Booting Linux on physical CPU 0x0

Now i have install cgminer 3.7.2
And reboot it with slub_debug=FP in my cmdline.txt

Code:

dwc_otg.lpm_enable=0 console=ttyAMA0,115200 kgdboc=ttyAMA0,115200 console=tty1 root=/dev/mmcblk0p2 rootfstype=ext4 elevator=deadline rootwait slub_debug=FP

i hope the system runs now...

[jmc1517]

Downloaded and started running cgminer-3.7.2

After about 40 minutes, 2 AMU LEDs (27/28) came on solid, but by the time I had checked the display only 1 was showing as zombie, then both LEDs went out and no zombies remained in the display. Yay!  Both AMUs had re-initialised as AMU 34/35.  This looks very promising

I hate to tempt fate, and I also hate quoting myself, but this baby's been running for 5 hours now with only the two temporary glitches reported above.  Nothing untoward in the logfile. Fingers and toes crossed.

[-ck]

Downloaded and started running cgminer-3.7.2

After about 40 minutes, 2 AMU LEDs (27/28) came on solid, but by the time I had checked the display only 1 was showing as zombie, then both LEDs went out and no zombies remained in the display. Yay!  Both AMUs had re-initialised as AMU 34/35.  This looks very promising

I hate to tempt fate, and I also hate quoting myself, but this baby's been running for 5 hours now with only the two temporary glitches reported above.  Nothing untoward in the logfile. Fingers and toes crossed.

Well, even if it isn't 100% and we still need to make changes, you can take part of the credit for your extensive testing.

Your failures were always still interesting because you routinely have one device disappear by itself followed by write errors on another device. I suspect ultimately there is a communication limitation/issue power issue or something after a fixed amount of time in your setup and only if things are running for an extended period does it reach a new steady state. The changes I made were to both increase reliability of communications on windows and to more gracefully handle failures.

[SpAcEDeViL]

Yeah.... Zombie Farm... only reboot helps to give the usb ports free....

[jmc1517]

Downloaded and started running cgminer-3.7.2

After about 40 minutes, 2 AMU LEDs (27/28) came on solid, but by the time I had checked the display only 1 was showing as zombie, then both LEDs went out and no zombies remained in the display. Yay!  Both AMUs had re-initialised as AMU 34/35.  This looks very promising

I hate to tempt fate, and I also hate quoting myself, but this baby's been running for 5 hours now with only the two temporary glitches reported above.  Nothing untoward in the logfile. Fingers and toes crossed.

Well, even if it isn't 100% and we still need to make changes, you can take part of the credit for your extensive testing.

Your failures were always still interesting because you routinely have one device disappear by itself followed by write errors on another device. I suspect ultimately there is a communication limitation/issue power issue or something after a fixed amount of time in your setup and only if things are running for an extended period does it reach a new steady state. The changes I made were to both increase reliability of communications on windows and to more gracefully handle failures.

Well I guess that did it. First permanent zombie AMU 20 after about 6 hours, and timeouts from AMU 0
Re-plugged AMU 20 (twice) but no hotplug message, and it's not being recognised, so AMU 20 remains a zombie.
Guess I'll kill it now and revert to 3.5.1 - which just to clarify does seem absolutely stable and will run for days without problems - but I'll stick a log on it anyway just to see if any temporary errors pop up.
Logfile for 3.7.2 here (without --debug):
 https://dl.dropboxusercontent.com/u/44240170/logfile-3.7.2.txt

[aneutronic]

3.7.2 ran a few minutes on the avalon at much lower hashrate then quit with a usb error.

http://pastebin.com/DQaP9S6Y

 

[SpAcEDeViL]

mh i become freeze crash, or USB crash... only a power off reboot helps... back to 3.1.1...

[-ck]

Well I guess that did it. First permanent zombie AMU 20 after about 6 hours, and timeouts from AMU 0
Re-plugged AMU 20 (twice) but no hotplug message, and it's not being recognised, so AMU 20 remains a zombie.
Guess I'll kill it now and revert to 3.5.1 - which just to clarify does seem absolutely stable and will run for days without problems - but I'll stick a log on it anyway just to see if any temporary errors pop up.
Logfile for 3.7.2 here (without --debug):
 https://dl.dropboxusercontent.com/u/44240170/logfile-3.7.2.txt

Yep same pattern. The libusb write errors are it failing to hotplug every single time. Weird fuckage but so be it.

Next test:
http://ck.kolivas.org/apps/cgminer/temp/cgminer-allsync.exe

[techman05]

Sorry to add a new twist to cgminer but have a possible unrelated bug to play with when all is done. I tried load balancing between bitcoin and unobtainium pool and found even though I had set the unobtainium pool higher and was hashing higher, it got no shares back even though it was accepting all of them. Lesson learned for now "don't mix coin pools" even if there both compatible coins for my hardware.

Also I think the mixed difficulties fudged this up more as difficulty listed higher when trying to do both and dropped when I did just the Unobtainium(even bitcoin was the one that reported properly, the difficulty of the Unobtainium was the one shown as affected.

I'm hoping this is a bug and you can find a way for them to play together. Bitcoins fine ,but  I just want to have backup option to play with.

I'm currently running 3.6.6 on windows 7 (didn't see 3.7.2 yet but if you tripped over this and fixed it) send me a "fixed it" pm and I'll delete the post so you don't have others asking.

[Krak]

Sorry to add a new twist to cgminer but have a possible unrelated bug to play with when all is done. I tried load balancing between bitcoin and unobtainium pool and found even though I had set the unobtainium pool higher and was hashing higher, it got no shares back even though it was accepting all of them. Lesson learned for now "don't mix coin pools" even if there both compatible coins for my hardware.

Also I think the mixed difficulties fudged this up more as difficulty listed higher when trying to do both and dropped when I did just the Unobtainium(even bitcoin was the one that reported properly, the difficulty of the Unobtainium was the one shown as affected.

I'm hoping this is a bug and you can find a way for them to play together. Bitcoins fine ,but  I just want to have backup option to play with.

I'm currently running 3.6.6 on windows 7 (didn't see 3.7.2 yet but if you tripped over this and fixed it) send me a "fixed it" pm and I'll delete the post so you don't have others asking.

It's not a bug. It's covered right in the first few sentences of the readme.

[techman05]

Which read me? I didn't see it.

[Krak]

Which read me? I didn't see it.

The main one. The one that says "README." Second sentence on the third line.

Do not use on multiple blockchains at the same time!

[crazyates]

3.7.2 uploading as we speak/type. Besides, even numbers always sound more stable.

EDIT: And it's up.

I like even numbers. A lot.

Great work on this! 3.7.2 seems pretty stable.

[-ck]

Next test:
http://ck.kolivas.org/apps/cgminer/temp/cgminer-allsync.exe

And here's the mother of all usb changes for testing:
http://ck.kolivas.org/apps/cgminer/temp/cgminer-ymmv.exe

[-ck]

is this planed to add?

Not unless someone sends us developers one or more of them.