Topic: OLD: BFGMiner 3.10.0: modular ASIC+FPGA, GBT+Strtm, RPC, Mac/Lnx/W64, AntU1, DRB - page 162. (Read 1193364 times)

Can bfgminer run on a windows 8 tablet? Also, should any of nicer android tablets work with the BF units?

I posted the meat of this in a post on reddit, but I figured it could be helpful here, since there's no real "Quick-start" guide for BFGMiner. This will get someone mining, and fairly optimized within about 15 min.

This post will be using commands and settings based on BFGMiner so go grab a copy and play along. I'm also going to assume you're in windows, but it really isn't much different in *nix.

Run the client with the options like this either in a command prompt or from a shortcut(or you can just run the .exe and follow the onscreen instructions to set up a pool):

bfgminer.exe -o stratum+tcp://mmpool.bitparking.com:3333 -u YourWorkername -p WorkerPassword

Once you're in the program, and you see your cards spitting out shares, pres "S" then "W" and hit enter twice to get back to the main screen. This creates a config file, saving all your settings. Now to run the program in the future all you need to do is run the .exe (no need for the extra options now)

To change anything, now all you have to do is edit the bfgminer.conf Make sure you don't mess with the formatting, however. If you want to learn how to properly format it by yourself, look up JSON formatting. The easiest way to add more pools (for backup when your primary pool goes down, etc) is from bfgminer itself by hitting "P" then "A" and follow the onscreen prompts. once your new pool is added, make sure you save your config file like we did before.

For how to squeeze every last MH/s out of your chips, you'll want to fiddle with some configuration settings.
In your bfgminer.cfg at the top should be your primary and backup pools, followed by something that looks like this:

"intensity" : "9",
"vectors" : "2",
"worksize" : "128",
"kernel" : "diablo",

with a bunch of other options at the bottom that you don't need to change(if you have 2 or more gpus, it will look similar, but with comma separated values like "d,d" for intensity). You can mess with the intensity, 9 is the max you'll want to go at all, but if you still want to use the desktop that is being rendered, you'll want to go down to "d" which is a dynamic intensity. Try modifying the vectors try out 1, 2, 4, and 8 to see what gives better results. Worksize is the same, 64,128,256 are what you should try. Kernel is a big one I think the default is "poclbm" but i find diablo has given me better results. If you're running a 7000-series try "diakgcn" that is designed for it.

Play with these settings, edit them, save the .conf and then hit "S" "B" and "Y" to quickly restart bfgminer without having to open the .exe again. You'll have to do this lots of times until you find the right combination of kernel, vectors, and worksize that works for your GPU.

If anyone has corrections or criticism, feel free to let me know, I'd like to improve this guide.

Thanks for that info.
So i will just keep ignoring this message


Ps:

X6500 and Icarus happily mining side by side gathering some 750 Mh/s.

It's a largely-useless command to get the transaction list that the pool is planning to include in a block. In addition to not being necessary for mining, it uses significant additional bandwidth, so most pools (and most miners) don't bother to implement it.

What if it doesn't? What if it keeps going unabated a full 30 minutes after internet disconnection, and p2pool's window simply reports that bfgminer is still doing the usual 200-ish MH/s work with 100% DOA?

It should usually stop within 5 minutes when it can't generate any more work.

Is there anything I can do to make BFGminer automatically stop mining when the network connection is lost? Frankly I don't even understand what it's mining when there's no more tasks coming in from the network.

no sorry :/ i cant reproduce this easily it happens after 1h or after 10h
if needed im sitting at freenode #bitminter with same nick

Hello,

afther a bit of roaming the README and try an d error i have compiled bfgminer 3.00 on my ubuntu serversystem.
I am running a Icarus board against the ozcoin pool.

Its all running fine so far except one message that nontheless seems not to be critical.

" Unknown stratum msg: {"error": [-3, "Method 'get_transactions' not found for service 'mining'", null], "id": "txlist84f8", "result": null} "


Just for your notice.

Might be interesting to know what this "get transactions" originally was supposed to do.

Regards

Jens

It sounds like you might be able to reproduce the crash easily?
If so, are you willing to go through a series of test builds (preferrably on IRC) to track down the cause?

Thanks,

Luke

same - crashes on bfgminer 3.0 and 2.10.6 with ztex and 6950 im on bitminter too
2.10.5 was stable
cgminer 2.11.4 is stable too
im using backup pools and i think problem is in this;)

I can add my anecdotal experience here as well. Win7/64, 3 separate machines, 3 separate instances of bfgminer (2.99.1 on all 3, not 3.0.0, but I've had this issue on 3.0.0 as well which is why I downgraded to 2.99.1). All BFL Single FPGAs in my case; no GPUs. All 3 instances crashed simultaneously coincident with bitminter going down temporarily last night.

I had two instances of BFGMiner 3.0.0 crash simultaneously today; each was running on its own computer.  Both video cards are MSI Radeon 7770 running AMD's Catalyst 13.1 on Windows 7 64 bit OS.

There was an interruption (my ISP connection went down) followed by a new block detected message and both instances of BFGMiner 3.0.0 trapped.  Due to the trap, I wasn't able to gather any further info.

Machine #1

Faulting application name: bfgminer.exe, version: 0.0.0.0, time stamp: 0x515f07a4 (2013-04-11 17:22:45.000000000Z)
Faulting module name: bfgminer.exe, version: 0.0.0.0, time stamp: 0x515f07a4
Exception code: 0xc0000005
Fault offset: 0x000000000000bf96
Faulting process id: 0x1b28
Faulting application start time: 0x01ce3647d3d57e57
Faulting application path: (removed)\bfgminer.exe
Faulting module path: (removed)\bfgminer.exe

Machine #2

Faulting application name: bfgminer.exe, version: 0.0.0.0, time stamp: 0x515f07a4 (2013-04-11 17:22:45.000000000Z
Faulting module name: bfgminer.exe, version: 0.0.0.0, time stamp: 0x515f07a4
Exception code: 0xc0000005
Fault offset: 0x000000000000bf96
Faulting process id: 0x264
Faulting application start time: 0x01ce35ec278316ab
Faulting application path: (removed)\bfgminer.exe
Faulting module path: (removed)\bfgminer.exe

Model path names contain identifying information so I removed those details.

Actually, this is what I'd consider completely vanilla - "make install" isn't really supported right now.
I'd welcome a patch to fix this, but it's not really a high priority.

For some reason, I can't seem to access the archlinux.org URI.

Almost vanilla, I only had to do a slightly patch to permit both cgminer and bfgminer installed, using /usr/share/bfgminer for kernels instead of /usr/bin. Maybe you should do that upstream too (having .cl files in /usr/bin is against FHS hier)

bfgminer in community now

I think I spoke to soon about my pi setup not crashing ...

Last night I had a network glitch when my network was down for about 10-15 minutes ... obviously in this time my mining was stopped and all my pools where not responding but the bfgminer 3.0.0 behaviour was not as expected ...

First let me say that I use only stratum protocol with both GBT and getwork disable via command line (--no-gbt and --no-getwork).
My setup use only 2 pools, both setup using "stratum+tcp://".

I think for my network glitch it was expected to retry pool connection alternating available pools for a while until one of them was available but ... after all stratum pools where not responding bfgminer tried to switch to getblocktemplate and even getwork and get stuck there for about 6 hours -  I guess because both where disable.

Here are the tail of the log as I found it in the morning before restarting it:

For my case BFL = Butterfly Labs's FPGA BitForce Single

BFL?

Running bfgminer 3.0.0 on a pi with BFL with no crashes for almost 4 days (3d 19h 51m 55s more exactly).