Topic: OFFICIAL CGMINER mining software thread for linux/win/osx/mips/arm/r-pi 4.11.0 - page 95. (Read 5805649 times)

Hmm this is weird...
I restarted CGMiner, the problem no longer shows up and hashrate has increased. Could be a memory leak? I'm taking a look.

Might mean your network connection had a hiccup.
Otherwise, ignore it.
I found that message in my cgminer logs 39,007 times with my one BTB board from 12-Sep-2013 until I permanently switched it off on 16-Apr-2014

Thanks!  I also get duplicates on p2pool when using my Ants.  I think I only saw that with S2s, though.


I appreciate it!

M

Hi, I noticed a strange behaviour in my mining equipment:
Sometimes out of a sudden I receive BTB: Idled 1 miners for no reason at all. No new block detected by stratum.

http://imageshack.com/a/img841/5870/h19c.png

I am running BitBurner stacks with --avalon-auto on GHash.io

What are the reasons for "idling" miners at such low temperatures and no block detected? Is something wrong with the equipment?
However utility remains pretty high at 22,24 shares/min.

Thanks in advance.

You are running bitmain's S2 driver. Don't do that on p2pool. Also don't do that with bitmain's S1 driver.

Bitmain does some terrible things in their driver including ... discarding stale work in the driver before passing it to the work validation code.
This means on p2pool that if you find a stale block that is still a valid network block, it will be discarded. Oh well.

Use my driver for S1 if you don't want to throw such valid blocks away on p2pool.

My S2 driver is still in development - I'm still trying to work out why the device thinks it's a good idea to have >8000 work items queued in it - more than 30s of work - and to work out how to reduce that without letting it get idle - so I'm still waiting on source code from bitmain.

My blackarrow minion RPi driver is able to handle > 2TH/s of work at 1diff verifying every share returned at 1diff.
The S2 driver, on a faster BBB, should be able to handle 1TH/s of work at 1diff also. It doesn't. So I need to work out how to fix that - but I need the SPI kernel driver code bitmain wrote, for me to work that out.

I have two Antminer S2s.  Both misbehave the same way with p2pool.  Within minutes of pointing them at p2pool, my hashrate drops to 930gh/s or lower. 

It shows this way on p2pool, on the S2 LCD, and on the S2 web UI.

It stays this way, or gets worse, as time goes by.  I've run against p2pool for days this way trying different things, and the result is always the same.  Yet running on a conventional pool (like Eligius), things run fine immediately and ongoing.

I've tried adjusting my pseudo share size.

I've tried running on a local node, my public node (other side of the country), and at least one other public p2pool node that was close to me.

Since S2s use cgminer, I was hoping a dev might be able to shed some light on this for me.

BTW, my S1s work fine with p2pool.  At 1/5 the hashrate, but they do work fine.


I apologize if this has been mentioned already.  I stopped following this thread a while ago.

Thanks.

M

Yes I did. I've run it again.
On the miner console I get this:
Debug:

Thanks. There's not much information in that. Did you use the debug cgminer.exe binary from that directory as well?

Hardware is actively cooled and working in air conditioned room - always below 42C.

Here is the debug:

That is a sign of hardware instability, not software. If you're 100% sure you have enough power (and most people don't have as much as they think they have) then the next thing to do is actively cool them better.

EDIT: Do you mean "crashes" when you say "crushes"? In that case you may be able to help by running a debug version and following the instructions in here:
http://ck.kolivas.org/apps/cgminer/debug/

Read the README-debug file.

Im running 3x Bi*Furry on powered USB HUB - power is sufficient.
Randomly (after 10m - 4h) I get:


and miner crushes (WIN 7).
Any idea?

Hmm, works fine for me.

There is no command to maximize performance for my setup, which is USB BE's and a Jalapeno.  I just run CGMiner with the  pool stuff and it detects and sets everything for me.

I mean, I often fail to use it on win 7 x86,
and if you can help command for maximum performance?

Is that a question or a statement?

If your asking if CGMiner works with 32 bit Windoze 7?  The answer is yes.

win 7 x86 if it works stable?
I often have problems: (

@choklivas

THX for the new version, i will test and report

regards

Dups are expected.
I get quite a lot and I wrote the driver
Expiring the work earlier risks rejecting valid work.
The interface to the chip makes it difficult to ensure you aren't rereading the same results sometimes.
You may have noticed the low HW error %?
That's due to using more CPU and sometimes getting dups.

I've been working on the AntS2 driver and wrote a simple dup check that I can implement in any driver with a few lines.
Once I've finished the AntS2 and the updates to the BlackArrowMinion work soon, I'll probably add it to the BaB driver also, since I see the dups also - and avoiding sending them to the pool is nicer to the pool and saves ever so slightly on network usage.

Oh well about your choice of pool ...

FYI: I've got 7 boards on a V2 4 SPI controller always above 270GH/s
The code works well up to 8, but will need some SPI tuning to do more than 2 per SPI reliably.
Mine runs forever without failing - only restarts on power outages, config changes or cgminer updates.
1.25W/GHs at the wall, really is pretty amazing for these boards that have been around for so long!

Are you maybe using load balancing or too low of a difficulty. I only get problems when I switch pools or exit and come back before it knows I 'm gone.

I'm still using 3.4.0 on a BeagleBone to run my BFL Jalapeno 7 GHs.
An oldie but a goodie. 

New release: Version 4.3.4, 25th May 2014


Human readable changelog:

- Driver updates to spondoolies, cointerra, avalon2 and minion.
- Support for nanofury NF2 (good working) and NF6 (flaky, ?design related) USB sticks.
- Extra output in the API for NF* devices and BXM in the API for per chip information.
- Fix for pools going idle if stratum is restarted unsuccessfully.
- Fix for a crash in libusb when devices are frequently plugging/unplugging on linux.
- New benchmarking code with the --benchmark option which uses a deterministic set of work items to find a known number of different diff share nonces at regular intervals. Note that devices that don't check the entire range (mainly bitfury based devices) will not find some of these nonces so the benchmark code is suited to devices like the hashfast,  cointerra etc. Give it an hour before assessing results.
- Fix the reset counter on hashfast devices to be properly inherited.
- New ruby example for API.
- Low level code fixes and cleanups.


Full changelog:

- Add support for 2 nonces per block in spond driver
- Increase timeout on reset in cta driver to 5 seconds
- Increase max diff on spondoolies driver slightly to be well below spi comms
limitations
- Use the active contents lock and safe list iteration within the linux usbfs
code
- Add Ruby Api Example
- Automatic detect the small miners
- Update default modules from 3 to 4
- Fix the temp max. we should use currect max temp
- add avalon2-cutoff options
- Enable the cutofftemp to Avalon2. ignore longer coinbase and longer merkles
stratum
- Fix the diff value used on MM firmware
- Mark pool as idle if stratum restart is failed
- Add hacky workaround for double list removal race in libusb
- Make the work given in benchmark mode deterministic on a per-device basis
- Rework the benchmarking code to use a deterministic set of work items with a
known number of diff share nonces at regular spaced intervals
- minion - restrict nonce read result size to ioctl() limit
- minion - must check temp when overheated
- minion - idle chips that hit >100C until back to 80C
- minion - report the chip/reg when aborting due to an invalid ioctl() size
- minion - all freq in Mhz but only convert when used
- minion - remove unused ioctl debug
- minion - command queue is now larger
- minion - check rolled in stale work cleanup
- Work stats should be based on device_diff not work_difficulty since non-shares
haven't been filtered out yet
- Prevent a segfault when writing a config file containing 'rotate' option
- minion - comment out HW debug message
- minion - roll work to reduce CPU
- minion - report init_freq in stats
- api - howoldsec is only used for USB
- minion - allow setting the frequency
- minion - disable iostats by default since it slows down mining
- minion - define frequency value table
- minion - report temp/cores/freq and handle temp formatting
- minion - item is undefined
- Rationalise diffs stored in the work struct and document them to avoid further
confusion
- Add basic API stats for nfu drivers to see how many submits each chip returns
- Add output direction for the EN0 pin on nfu driver
- Support power management optimisations in newer nf* firmware
- Support variable numbers of chips with NFU and BXM drivers
- Identify number of chips in nanofury devices and change name accordingly
- Rename nf1 driver to nfu in anticipation of support for more chips
- Make hashfast reset counter rise on old instances when inheriting the value on
new ones