-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
USB FPGA #3
Comments
I hacked it to at least start. the bus id that gets returned is 0. I have the system key off of that to return a false if the bus id is 0. to skip that miner in the main loop. I will put together a pull request to show it, but my python skills are no where near as good as yours. |
@snoby - thanks for the JSON files. I’ll take a look at this either tonight or tomorrow. I don’t have an FPGA card, so this is a case I hadn’t written code for yet |
@snoby - I'm digging around in the various HiveOS bash scripts to figure out how an FPGA is reported to Hive. A few requests for you:
I have yet to find anything that looks FPGA specific in the various HiveOS detection scripts that look for hardware. Everything I'm finding is GPU specific. What I'm aiming for is some way to properly detect the FPGA from the monitoring script and provide all the labels that are also provided for the GPUs. |
It's not listed nor detected in Hive. Only teamredminer sees it. It is a USB device that teamredminer knows about. What's interesting is that HiveOS is more than happy to let teamredminer start and run and report hashrate. Output of usb-devices
It's an odd / complete corner case. This is the card: Teamredminer config info for the fpga. I will add that the 0 that is the bus id, seems to be a default, as the usb bus id from the above output is actually "01"
|
@snoby - I put in a minimal change that is likely similar to what you mentioned you had tried previously. In this case, the script still reports hashrate for the FPGA, but many of the labels are filled with "unknown". The card number is also just reported as the bus number from the last_stat.json (which in this case is "0"). I don't know for sure that is the correct approach, but can you test out #4 for me and validate that it works for you? If it does, I'll get it merged and then you can at least get monitoring you need without crashes. As for proper FPGA support, I likely would need to have that type of hardware to really dive into how to detect it. I don't foresee getting that kind of hardware any time soon. |
I've been digging further and could use one last bit of information from you. On your rig with the FPGA, can you give me the output of I'm suspicious that HiveOS doesn't even properly pull out the stats from the miner for an FPGA device. |
|
Everything works great with the exporter, except I found 1 corner case. On a mining rig that has some gpu's, like AMD, and one of those USB FPGA cards ( the C1100) the initialization gets all messed up. Because there is no entry for the FPGA in /run/hive/gpu-detect.json, but there is a miner entry for it ( teamredminer)
I've put some debug in and some try except blocks, but not sure how to work around this.
in Class Miner:
log.debug(' return is: %s ',self.stats['bus_numbers'][0])
will say 0. not None, maybe because it's initialized based on the AMD cards that are already there...
I think the call to Miner.is_gpu_miner (), might have to change, but I'm not sure how... maybe do a cross reference of the busid ...
Not sure how i would fix this. - any suggestions?
json_files.zip
The text was updated successfully, but these errors were encountered: