Topic: BFGMiner 5.5.0: CPU/GPU/FPGA/ASIC mining software, GBT+Stratum, RPC, Linux/Win64 - page 54. (Read 834507 times)

You mean you never underclocked your Monarch?

It's hardware, you can't change it after it ships.

Edit: Ok, I get it the 1.4.2 firmware has the possibility to set frequence and voltage and you built your part. But BFL has to compile the special bfgminer?

Because the 5.0.0 doesn't work at all with the Monarch. I get 500GH vs 700GH with the old special bfgminer on the BFL site.

All I can say is ask BFL... Probably you should make a debug log and put it on Google Drive or something to send them (feel free to open an issue on GitHub if you want me to confirm from that log everything is okay in BFG).

Correct it's basically 2 ft232h smacked together on some Si. Oh have I got s fun project ahead. 😁

Time to get my dev board and start coding my baby

Not sure, but I think the ft2232h is just a dual-channel version of the ft232h, which lowl-ftdi already supports (to some extent) for the BFx2.

@Luke-jr: how much work would be needed to modify the lowl-ftdi code to support the ft2232h as well?

I ask before I go usb to SPI&I2C interface shopping

Hm, I tried this command, but the Monarch still goes full blast:


Also I get loads of these:

ok I will try that. more concerned that I was doing it right and nothing was awry than anything.

Try adding #skipcbcheck to the pool URL, e.g. stratum+tcp://stratum.whatever.com/#skipcbcheck

Fired up 5.0 pointed it at a pool running peercoin and this error basically kept showing up.

Coinbase check: multiple inputs in coinbase: 0x26
It appeared it was mining ok, shares seemed to be accepted and credited. But that coinbase check wouldn't go away.
I'm curious to know what it is and why its doing it.
I wasn't mining to any coinbase address or to a wallet address. Just to normal username.worker.

Reverting back to 4.* works with no output like pasted above.
I'm guessing it's checking bitcoin related and because I'm not mining bitcoin but another sha it isn't handling it well?
only a guess

Anyone want to throw out their lowest stable underclock command line?

Well it's not like they've been known for making great haste at anything :/

Ok, BFL is a bit slow then: https://forums.butterflylabs.com/the-monarch-discussion/8936-underclocking-psu.html

They go back all the way to 4.0.0

Are those in the 5.0.0 release?

Hello again,

I've setup a Tomato development environment for Shibby's latest build using GitHub. I built the 2.6.22.19 kernel with the Silicon Labs cp2101 driver option enabled as a module in the .config file. I then extracted the cp2101.ko.

If you want to use this kernel object, you can find it here: https://drive.google.com/file/d/0B9zGF2pK75nCUmEzZE5obTlCVFk/view?usp=sharing

Simply put it in your /opt directory if you mount Optware or Entware FS at boot time. In the web GUI, set Tomato to load the usbserial (included in firmware image at /lib/modules/2.6.22.19/kernel/drivers/usb/) and cp2101 kernel objects after boot with insmod. Use lsmod to make sure the kernel objects loaded. The device mounts to /dev/ttyUSB0. Make sure it is using the driver with "cat /proc/bus/usb/devices." The result should be "Driver=cp2101."

Alternatively you can use a firmware extract tool, insert the cp2101.ko into /lib/modules/2.6.22.19/kernel/drivers/usb/, rebuild, and call on usbserial and cp2101 kernel objects with insmod in the web GUI.

You can run bfgminer command line or JSON config file, just reference the /dev/ttyUSB0 device.

I'm currently hashing 2 GH/s or 1uBTC/hr!

One more feather in the cap for embedded Linux!

I agree it would be useful. Perhaps some non-coder can contribute these kind of changes?

Luke-Jr, it would be very useful if you would include a section in README explaining all #xxx pool parameters and the correct usage of them (syntax). Currently these options/features are only mentioned in NEWS...

Yes

Hi,
I'm optimizing a Titan for Nicehash.

I read I should add #xnsub for extranonce but I'm already using #skipcbcheck (got the hint reading KnC forums).

So, a couple of questions:

Is #skipcbcheck of any use for Nicehash?
What would be the right grammar to add both, "stratum+tcp://stratum.nicehash.com/#xnsub#skipcbcheck"?

Thanks

Hi Luke-Jr.

Just a tip: the README.ASIC.txt, Gridseed section, reports the possibility to set individual clock per device, but not to force the use of individual devices.
Indeed, it reports only the parameter "-S gridseed:all", so it seems to be impossible to use a single device.
After reading the README.txt (lines 370-380) and some search (also this thread, post #67 and #68) I find that BFGMiner allows this possibility, simply specifying a "-S gridseed:\\.\COMx" for every device needed (also the simplest "-S \\.\COMx" seems to works correctly).
It may be useful to update the README.ASIC.txt with these information, as i.e. reported in the Zeus section

Now, the implementations request

1) Can be added the possibility to show in the main interface (at least) the specified clock for every Gridseed device, like CGMiner?

2) Can be added the possibility to specify a single (or a set of) device(s) to a specific pool?
Explanation: I have 2 Gridseed Mini (5 chip) and 1 Gridseed Blade (2x40 chip), and I need to use them separately (every single device dedicated to a specific pool).
The reasons of this choice are multiples, last but not least the use of pools with rapid difficulty changes, that seem to work better (higher speed on dashboard) with a single dedicated device, instead to use all devices and set the load-balancing with quota for every pool.
To make this, currently I run 4 instances of BFGMiner, each with its config file: it works well, but it's a little complicated to monitor them via RPC.
A single BFGMiner instance, with a single RPC to monitor, will be more simple to monitor

Greetings in advance