Modular Python Bitcoin Miner - Official Thread - page 16.

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: allinvain on February 17, 2012, 03:25:05 AM

wtf..how come efficiency is showing as being negative!! looool. I'm pushing an average of -3716 MHs.

I bet this is a race condition of some kind in the Icarus worker module. However I fail to spot/reproduce it. Tongue

Feel free to have a look at the code yourself, it isn't all that complex. An additional pair of eyes can't hurt

Anyway, I'm preparing for a new major version which will overhaul a lot of things (so it might take a while until it's finished), but as this module will probably completely rewritten during the process, I hope that this bug will just be gone afterwards...

allinvain

legendary

Activity: 3080

Merit: 1080

wtf..how come efficiency is showing as being negative!! looool. I'm pushing an average of -3716 MHs.

allinvain

legendary

Activity: 3080

Merit: 1080

Hmm, maybe this can be addressed in the next version of the miner, but it seems after connection/get work issues the miner reports ridiculous hash rates and low efficiency levels. Right now it says my average is 4244 MHs with 9.0% efficiency - I wish!! (I'm running a single Icarus board)

Maybe it should reset it's average/stats after a series of connection problems?

coblee

donator

Activity: 1654

Merit: 1351

Creator of Litecoin. Cryptocurrency enthusiast.

Quote from: TheSeven on February 12, 2012, 10:27:07 AM

Quote from: coblee on February 12, 2012, 06:26:40 AM

I have 2 Icarus boards and after mining for a while, the stats start to look weird:

Code:

               Worker               │Accepted│Accepted│Stale shares│Invalid shares│Current│ Average │Temperature│Effi- │   Current                                             
                name                │  jobs  │ shares │ (rejected) │ (K not zero) │MHash/s│ MHash/s │(degrees C)│ciency│     pool   
Icarus board on /dev/tty.usbserial  │    4389│    2195│   23 (1.0%)│      2 (0.1%)│ 379.70│   373.31│  Unknown  │ 94.7%│   Eclipse                                             
Icarus board on /dev/tty.usbserial20│    4459│    2272│   23 (1.0%)│      7 (0.3%)│ 379.36│129609.80│  Unknown  │  0.3%│   Eclipse

The accepted shares are about the same, but the 2nd board is shwoing a wrong average mhash/s and efficiency. In the logs I do notice a that the 2nd board had timed out a few times. TheSeven, any ideas?

Looks like something goes wrong with the total MHash counter for that worker. No idea what though, I've had a quick look at that code and can't spot any obvious reason for this. I also can't reproduce it by deliberately causing communication problems with the board.

This keeps happening to me. Let me know how I can help you fix this. I am running the latest code.

nbtcminer

sr. member

Activity: 249

Merit: 250

Interesting note:

Found a method for installing Libusb on Windows but haven't had a change to test it yet. It seems like someone make this work around for iPhones but it may work for MPBM purposes:

http://www.limesn0w.org/how-to-install-libusb-on-windows-7-vista-64-bit/

I'll give it a try tomorrow and let y'all know how it goes.

Cheers,
nbtcminer

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: coblee on February 12, 2012, 06:26:40 AM

I have 2 Icarus boards and after mining for a while, the stats start to look weird:

Code:

               Worker               │Accepted│Accepted│Stale shares│Invalid shares│Current│ Average │Temperature│Effi- │   Current                                             
                name                │  jobs  │ shares │ (rejected) │ (K not zero) │MHash/s│ MHash/s │(degrees C)│ciency│     pool   
Icarus board on /dev/tty.usbserial  │    4389│    2195│   23 (1.0%)│      2 (0.1%)│ 379.70│   373.31│  Unknown  │ 94.7%│   Eclipse                                             
Icarus board on /dev/tty.usbserial20│    4459│    2272│   23 (1.0%)│      7 (0.3%)│ 379.36│129609.80│  Unknown  │  0.3%│   Eclipse

The accepted shares are about the same, but the 2nd board is shwoing a wrong average mhash/s and efficiency. In the logs I do notice a that the 2nd board had timed out a few times. TheSeven, any ideas?

Looks like something goes wrong with the total MHash counter for that worker. No idea what though, I've had a quick look at that code and can't spot any obvious reason for this. I also can't reproduce it by deliberately causing communication problems with the board.

Quote from: O_Shovah on February 12, 2012, 09:08:05 AM

Has anybody suceded in running a Icarus board and a x6500 in parallel ?

It seems im missing something in the config.
It only recognices and starts the x6500.
The Icarus remains unseen.

This is a bit tricky because the Icarus board requires the ftdi_sio kernel module, which conflicts with the x6500.
The following might be worth a try:

Configure both workers in the MPBM config file
rmmod ftdi_sio
Start MPBM (Icarus will fail)
modprobe ftdi_sio (Icarus should be detected now)

EDIT: It seems like the Icarus board isn't based on FTDI, but on some other USB to serial converter, which means that it doesn't need the ftdi_sio kernel module either (but possibly some other one, which doesn't hurt x6500 though).

allinvain

legendary

Activity: 3080

Merit: 1080

Quote from: O_Shovah on February 12, 2012, 09:08:05 AM

Has anybody suceded in running a Icarus board and a x6500 in parallel ?

It seems im missing something in the config.
It only recognices and starts the x6500.
The Icarus remains unseen.

Do you have this in the list of modules to be loaded section:

import worker.theseven.icarus

and then in the list of workers section, this:

# Icarus worker
{ \
# Worker module
"type": worker.theseven.icarus.IcarusWorker, \
# Worker module parameters, in this case serial port name (default: /dev/ttyS0)
"port": "COM3", \
}, \

Just remember to change the port variable to your particular one (ie COMx, or /dev/ttyUSBx)

O_Shovah

sr. member

Activity: 410

Merit: 252

Watercooling the world of mining

Has anybody suceded in running a Icarus board and a x6500 in parallel ?

It seems im missing something in the config.
It only recognices and starts the x6500.
The Icarus remains unseen.

coblee

donator

Activity: 1654

Merit: 1351

Creator of Litecoin. Cryptocurrency enthusiast.

I have 2 Icarus boards and after mining for a while, the stats start to look weird:

Code:

               Worker               │Accepted│Accepted│Stale shares│Invalid shares│Current│ Average │Temperature│Effi- │   Current                                             
                name                │  jobs  │ shares │ (rejected) │ (K not zero) │MHash/s│ MHash/s │(degrees C)│ciency│     pool   
Icarus board on /dev/tty.usbserial  │    4389│    2195│   23 (1.0%)│      2 (0.1%)│ 379.70│   373.31│  Unknown  │ 94.7%│   Eclipse                                             
Icarus board on /dev/tty.usbserial20│    4459│    2272│   23 (1.0%)│      7 (0.3%)│ 379.36│129609.80│  Unknown  │  0.3%│   Eclipse

The accepted shares are about the same, but the 2nd board is shwoing a wrong average mhash/s and efficiency. In the logs I do notice a that the 2nd board had timed out a few times. TheSeven, any ideas?

allinvain

legendary

Activity: 3080

Merit: 1080

Quote from: TheSeven on February 11, 2012, 12:45:37 PM

Quote from: allinvain on February 11, 2012, 11:16:25 AM

TheSeven, I'm wondering if you can tell me if the "failed job requests" , "Share upload retries" and "Efficiency" stats are within normal ranges. I know the failed job requests and share upload retries varies from person to person, but shouldn't the efficiency be higher? For a while it was at 98%. What does the efficiency level refer to?

The failed job request, upload retries and stale share values are looking like there's connectivity trouble. Either a very slow (or rather high-latency) internet connection, a DDoS affecting the pool, or something similar.
The efficiency value is basically your luck minus invalids/stales, it should usually be around 85-110% after some time (it needs some minutes to stabilize). However in your case something seems to have gone wrong with hashrate measurement at some point (look at your average hashrate), which will of course screw up the efficiency calculations. If you account for that, your effective efficiency is around 98%, which is pretty much normal.

I think it was the pool as it has now stabilized. The efficiency is slowly climbing back up too. Average hashrate is also dropping to a more realistic figure.

As for latency it's kind of funny cause I chose to use BTC Guild specifically due to the low ping and low number of hops from me to their server.

I'll keep an eye on it and see how it goes.

Props to you for coding a neat miner.

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: freshzive on February 11, 2012, 11:10:24 AM

Sent you a BTC for development of this great software

Thanks a lot, I really appreciate that

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: cypherdoc on February 11, 2012, 10:06:18 AM

To op,

will there ever be a version of this great software for non coders?

Well, you don't really require coding skills to use this software. Currently it isn't really documented well yet (this project was launched just days ago), and some general programming experience will of course make the installation and configuration easier, because programmers will usually grasp the general concepts of how this kind of thing works faster.
However, if you look at the configuration file with some common sense, you'll probably recognize some patterns, how the definitions in there look like, which should enable you to adapt them to fit your needs. It really isn't all that complicated. Just ask for help if you don't understand something.

I'm planning to remove the MPBM configuration file one day, and add a web interface instead through which it can be configured. This will require some major changes and thus take quite some time. Let's see how much spare time I'll have during the next months...

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: allinvain on February 11, 2012, 11:16:25 AM

TheSeven, I'm wondering if you can tell me if the "failed job requests" , "Share upload retries" and "Efficiency" stats are within normal ranges. I know the failed job requests and share upload retries varies from person to person, but shouldn't the efficiency be higher? For a while it was at 98%. What does the efficiency level refer to?

The failed job request, upload retries and stale share values are looking like there's connectivity trouble. Either a very slow (or rather high-latency) internet connection, a DDoS affecting the pool, or something similar.
The efficiency value is basically your luck minus invalids/stales, it should usually be around 85-110% after some time (it needs some minutes to stabilize). However in your case something seems to have gone wrong with hashrate measurement at some point (look at your average hashrate), which will of course screw up the efficiency calculations. If you account for that, your effective efficiency is around 98%, which is pretty much normal.

allinvain

legendary

Activity: 3080

Merit: 1080

It seems it's sporadic. It sometimes goes pretty bad..checkout all the red lol:

Uploaded with ImageShack.us

nbtcminer

sr. member

Activity: 249

Merit: 250

@Theseven:

It's BTCGuild, since they swapped servers yesterday things have been a bit weird.

Cheers,
nbtcminer

Quote from: TheSeven on February 10, 2012, 11:29:38 PM

Sounds like your pool is having issues (or you didn't configure it correctly)?

allinvain

legendary

Activity: 3080

Merit: 1080

Uploaded with ImageShack.us

TheSeven, I'm wondering if you can tell me if the "failed job requests" , "Share upload retries" and "Efficiency" stats are within normal ranges. I know the failed job requests and share upload retries varies from person to person, but shouldn't the efficiency be higher? For a while it was at 98%. What does the efficiency level refer to?

freshzive

sr. member

Activity: 447

Merit: 250

Sent you a BTC for development of this great software

cypherdoc

legendary

Activity: 1764

Merit: 1002

To op,

will there ever be a version of this great software for non coders?

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: freshzive on February 10, 2012, 10:22:02 PM

So I tried this, and it actually seemed to be sending MORE shares to my backup pool than it had previously (this was with priority set to 20 for main and 1 for backup).

With the default settings and priorities of 1000 (main) and 1 (backup), it was sending about 100Mh/s of 2000Mh/s total to my backup pool over the course of ~24 hours.

Any other ideas? Thanks.

Interesting. It shouldn't be doing that unless your main pool is having severe trouble. Can you send me a screenshot of the full statistics and your config file (with passwords removed if neccessary)?

TheSeven

hero member

Activity: 504

Merit: 500

FPGA Mining LLC

Quote from: nbtcminer on February 10, 2012, 06:36:30 PM

@Theseven:

Awesome! I've got it up and running with the D2XX drivers and things look ok so far! However I'm an error that i used to get with the old X6500 as well:

Code:

Error while requesting job from "pool" timed out

Any idea on that one?

Sounds like your pool is having issues (or you didn't configure it correctly)?

Topic: Modular Python Bitcoin Miner - Official Thread - page 16. (Read 74189 times)