HASHRA CONTROLA on Raspberry Pi for Gridseed

[HASHRA]

Please see hashra.com/support to download version 1.4.0

Now called "HASHRA MINI CONTROLA" as we will have a version for BLADE and another for ATLANTIS.

"MINI" is what we are now calling the Gridseed "5 Chip Devices".

[maxzilla]

Quick question for you guys. I have 20 gridseeds running on controlla. would it be worth it to dual mine? also if it is worth it i have a spare seasonic 1050watt psu, would that be enough power?

Personally I don't think it's worth it.

I've tested it with 10 miners. 20 makes controla unresponsive in dual mode. I set the btc cores to 5 in system settings. It get's like 3.5 gh for sha and 300k for scrypt per miner. So for a 10 unit, 35ghs / 3mhs

[bexwhitt]

Quick question for you guys. I have 20 gridseeds running on controlla. would it be worth it to dual mine? also if it is worth it i have a spare seasonic 1050watt psu, would that be enough power?

Personally I don't think it's worth it.

I've tested it with 10 miners. 20 makes controla unresponsive in dual mode. I set the btc cores to 5 in system settings. It get's like 3.5 gh for sha and 300k for scrypt per miner. So for a 10 unit, 35ghs / 3mhs

BTC mining on the gridseed  would probably  not even pay for the electric used these days, it's a feature that when designed was at a difficulty that would have been making money but that ship has long since sailed.

[bexwhitt]

what's the root password for 1.4?

 it's not root, hashra or pi

[HASHRA]

what's the root password for 1.4?

 it's not root, hashra or pi

Sorry Bexwhitt but the user and password will not be public for this version.

We are working on some features for the future where having the user and password public is just not going to work.

Sorry about that.

[mazedk]

Hey Hashra

Any idea if/when you guys are gonna be working on individual clock speeds and/or auto-tuning as there is in the lightningasic controller?

Ditto.. this would be a huge benefit, esp for those of us that are running multiple units.  Right now I have to downclock to the weakest unit, I have 14 others that can run 50-100 Mhz faster, it's a huge disadvantage to using your product, but I really like the rest of it.  Another feature I'd like is in addition for each unit showing the shares/reject like it does now, is to show the HW errors for each unit, it would make it a lot easier to figure out which units are having issues when running so many of them.

And set a failover pool..or long shot add load balance between pools.

what is the algorithm for auto tuning? the slower the unit the lower the clock speed? measured by pool hashrate and hw errors?

if you really want to set different clock speed per miner, this also possible with CONTROLA btw, it's a hack, but it works. I don't know if it will have any effect in overall mining performance though.

PM me if you want to know how to do this.

failover is already available in version 1.4.0

LA coder explained a bit here:

To elaborate a bit on how the auto overclocking works, if the chip has hashed for over 600s it will attempt to change the frequency. If there is more than 1 HW error or if the current hashrate is less than the previous hashrate (after that 600s period), it will decrease the frequency, otherwise increase it. It will stop adjusting if the frequency has been decreased atleast once, so make sure to pick a low enough frequency to start with. The frequency adjusts in steps of 25MHz, I have found this to give the most stable results while overclocking, and it also makes sense if you look at how the frequency is calculated from the PLL variables.

[sandor111]

Hey Hashra

Any idea if/when you guys are gonna be working on individual clock speeds and/or auto-tuning as there is in the lightningasic controller?

Ditto.. this would be a huge benefit, esp for those of us that are running multiple units.  Right now I have to downclock to the weakest unit, I have 14 others that can run 50-100 Mhz faster, it's a huge disadvantage to using your product, but I really like the rest of it.  Another feature I'd like is in addition for each unit showing the shares/reject like it does now, is to show the HW errors for each unit, it would make it a lot easier to figure out which units are having issues when running so many of them.

And set a failover pool..or long shot add load balance between pools.

what is the algorithm for auto tuning? the slower the unit the lower the clock speed? measured by pool hashrate and hw errors?

if you really want to set different clock speed per miner, this also possible with CONTROLA btw, it's a hack, but it works. I don't know if it will have any effect in overall mining performance though.

PM me if you want to know how to do this.

failover is already available in version 1.4.0

LA coder explained a bit here:

To elaborate a bit on how the auto overclocking works, if the chip has hashed for over 600s it will attempt to change the frequency. If there is more than 1 HW error or if the current hashrate is less than the previous hashrate (after that 600s period), it will decrease the frequency, otherwise increase it. It will stop adjusting if the frequency has been decreased atleast once, so make sure to pick a low enough frequency to start with. The frequency adjusts in steps of 25MHz, I have found this to give the most stable results while overclocking, and it also makes sense if you look at how the frequency is calculated from the PLL variables.

I have changed the algo meanwhile, https://github.com/siklon/usb-miner/blob/master/software/cpuminer/gc3355.h
Getting 3700 KH/s poolside consistently with 10 miners, unmodded.

Feel free to compile it for Rpi.

[nogdog21]

Just wanted to say thank you to the Hashra guys and everyone else who helped me narrow down the possible issues with my machines.  As it turns out a bad ethernet hub was creating a data collision domain with all the machines.  Once removed all was well.  Thanks again!!

-Andrew

[HASHRA]

Just wanted to say thank you to the Hashra guys and everyone else who helped me narrow down the possible issues with my machines.  As it turns out a bad ethernet hub was creating a data collision domain with all the machines.  Once removed all was well.  Thanks again!!

-Andrew

Good to hear!

[lajz99]

Please see hashra.com/support to download version 1.4.0

Now called "HASHRA MINI CONTROLA" as we will have a version for BLADE and another for ATLANTIS.

"MINI" is what we are now calling the Gridseed "5 Chip Devices".

When will the BLADE version be released?

[bexwhitt]

what's the root password for 1.4?

 it's not root, hashra or pi

Sorry Bexwhitt but the user and password will not be public for this version.

We are working on some features for the future where having the user and password public is just not going to work.

Sorry about that.

so there in no safe way to shut down the pi if I need to make some adjustments, clever

[compumine]

Please see hashra.com/support to download version 1.4.0

Now called "HASHRA MINI CONTROLA" as we will have a version for BLADE and another for ATLANTIS.

"MINI" is what we are now calling the Gridseed "5 Chip Devices".

Under the current Hashra controla, you can't do an update?

So I just want to confirm, I have to load the image to the sd cards manually?

[mazedk]

Please see hashra.com/support to download version 1.4.0

Now called "HASHRA MINI CONTROLA" as we will have a version for BLADE and another for ATLANTIS.

"MINI" is what we are now calling the Gridseed "5 Chip Devices".

Under the current Hashra controla, you can't do an update?

So I just want to confirm, I have to load the image to the sd cards manually?

You do, it was mentioned earlier that the 1.4 will need a new image.

[compumine]

The SD card I got with my PI is a 16GB.  I read that some of you able to use 8GB!  Will the new 1.4 fit on an 8GB?  What class of cards better to use with PI?

[HASHRA]

The SD card I got with my PI is a 16GB.  I read that some of you able to use 8GB!  Will the new 1.4 fit on an 8GB?  What class of cards better to use with PI?

New IMG file is 3.88GB, I would suggest to use an 8GB card.

Please see this list for details on SD cards best for the Pi.
http://elinux.org/RPi_SD_cards

[bexwhitt]

The SD card I got with my PI is a 16GB.  I read that some of you able to use 8GB!  Will the new 1.4 fit on an 8GB?  What class of cards better to use with PI?

I have used a class-4 8GB card with no problems


but hashra controla will work with a 32GB card as it is Up my stairs ATM

[ZBC3]

First I want to thank Hashra for making the image available for download. Now I need some help setting up the Controla. Looking at the output below, it appears that only one gridseed mini is working. Does anyone know what I may be doing wrong?

1.036 Mh/s current hashrate 1.019 Mh/s avg hashrate 69 / 20 Acc./Rej.
28.99 % rejection rate
3 Hardware errors

SCRYPT Miners hashrate 1017 Kh/s

Scrypt Miner 1
Mining...5/5 (0%)339 Kh/s
Scrypt Miner 2
Mining...64/67 (4.48%)339 Kh/s
Scrypt Miner 3
Mining...0/0 (0%)

Total hashing listed on clever mining is 363.

[maxzilla]

First I want to thank Hashra for making the image available for download. Now I need some help setting up the Controla. Looking at the output below, it appears that only one gridseed mini is working. Does anyone know what I may be doing wrong?

1.036 Mh/s current hashrate 1.019 Mh/s avg hashrate 69 / 20 Acc./Rej.
28.99 % rejection rate
3 Hardware errors

SCRYPT Miners hashrate 1017 Kh/s

Scrypt Miner 1
Mining...5/5 (0%)339 Kh/s
Scrypt Miner 2
Mining...64/67 (4.48%)339 Kh/s
Scrypt Miner 3
Mining...0/0 (0%)

Total hashing listed on clever mining is 363.

I think your clock speed might have been set too high ?

[ZBC3]

I have it set at 800, but I've tried going as low as 700.

[maxzilla]

I have it set at 800, but I've tried going as low as 700.

hmm strange. What kind of hub are you using?