[GUIDE] GridSeed 5-Chip USB, Blade & Black Miner Support/Tuning

[podizzle]

what should i set for BAUD?
cgminer.exe --scrypt -u podizzle.x -p x -o stratum+tcp://us.dogepool.net:2222 --gridseed-options=baud=115200,freq=800,chips=5 --hotplug 5


800 is working much better at what point is too many HW errors is your guys opinion is there some ratio A:HW that i should shoot for?

[podizzle]

hmm i dunno still seems way off to me http://imgur.com/tnVBpzk

[suchmoon]

what should i set for BAUD?
cgminer.exe --scrypt -u podizzle.x -p x -o stratum+tcp://us.dogepool.net:2222 --gridseed-options=baud=115200,freq=800,chips=5 --hotplug 5


800 is working much better at what point is too many HW errors is your guys opinion is there some ratio A:HW that i should shoot for?

baud is good at 115200. HW ideally should be zero, but these cupcakes get some random errors once in a while for no apparent reason, so a few HW per hour is probably ok. Just make sure your accepted hashrate goes up when your increase frequency, once it doesn't anymore you need to step back and use the previous highest setting that worked. Let it run for a few hours to allow the hashrate average to settle down before moving on to the next setting.

[podizzle]

i dunno im thinkin still have some serious issue

[suchmoon]

hmm i dunno still seems way off to me http://imgur.com/tnVBpzk

If you are referring to the pool hashrate, give it some time. Better yet, find a pool that shows 15, 30, or 60 minute average, it would be a much more reliable metric than 5 minute average in that chart.

[podizzle]

well that and cgminer showing 64k accepted shares and my pool showing 2000

[podizzle]

i installed zagid winusb driver do you think that could be the problem?

[wolfey2014]

what should i set for BAUD?
cgminer.exe --scrypt -u podizzle.x -p x -o stratum+tcp://us.dogepool.net:2222 --gridseed-options=baud=115200,freq=800,chips=5 --hotplug 5


800 is working much better at what point is too many HW errors is your guys opinion is there some ratio A:HW that i should shoot for?

baud is good at 115200. HW ideally should be zero, but these cupcakes get some random errors once in a while for no apparent reason, so a few HW per hour is probably ok. Just make sure your accepted hashrate goes up when your increase frequency, once it doesn't anymore you need to step back and use the previous highest setting that worked. Let it run for a few hours to allow the hashrate average to settle down before moving on to the next setting.

Actually there is a 'valid reason' for the 'red nonce' / HW error. It's a failed authentication hence, the dreaded red nonce, which is normal.
See http://en.wikipedia.org/wiki/Cryptographic_nonce
Good read....
Don't forget to try adjusting the FIFO buffers if the problem persists at higher clock rates. When I ran my pods at 850MHz, I had to throttle them back a bit. Receive buffer = 8 - Transmit buffer = 14.
Wolfey2014

[miner765]

Anybody interested in buying some Raspberry PI+AC Adapter+SD Card with Hashra Controlla loaded? PM me

[suchmoon]

i installed zagid winusb driver do you think that could be the problem?

No. You wouldn't be able to run cgminer without it. However you will need to uninstall it if you decide to try bfgminer or cpuminer.

[wolfey2014]

well that and cgminer showing 64k accepted shares and my pool showing 2000

Gotta fill that vacuum with correct data...

I read up on it a little bit, the diff' between local and pool side rates...
It has to do with how the pool calculates hash rates, shares etc.
It's sort of complex.
Here is one explanation given by litecoinpool.org
And I quote:

The pool keeps reporting a much loser hash rate than cgminer. What's going on?

If the estimate displayed on the site stays too low for more than 10 minutes, you are most certainly experiencing hardware errors.
cgminer checks every solution generated by your gpu (ASIC miner in our case) and if one turns out to be incorrect because of hardware errors, it (rightfully) doesn't even submit it. Since pools can only estimate your hash rate based on how many solutions you submit in a given amount of time, their estimate will be lower than your real hash rate.
This raw hash rate, which is correctly displayed by cgminer, is, however, not meaningful if the solutions your card or pod generates are wrong.
You need to make sure that all (not just the top one) of the HW figures in cgminer are zero or next-to-zero. If that's not the case, you need to review your settings.
Make sure you carefully read the documentation about Scrypt mining provided with cgminer.
Remember that Litecoin mining puts a lot more stress on GPU's than Bitcoin mining does. So you don't want to push your hardware too much.

Copied from https://www.litecoinpool.org.help#faq
Edited by Wolfey2014
Hopefully the vacuum has now been filled
Peace!
Wolfey2014

[suchmoon]

well that and cgminer showing 64k accepted shares and my pool showing 2000

These are not the same numbers. cgminer reports shares it submitted to the server, pool reports (depending on difficulty and payout system) what it will pay you for. I don't know if there is an easy way to reconcile these numbers.

[podizzle]

UPDATE: so im running better........ but not optimal. any tips?

http://imgur.com/9BWhsJ8

[podizzle]

thanks wolfey for the help. i have a question

if im mining solo how do i get any feedback other than blocks found? i get no accepts rejects or HWs.... strange

[TrollboxChamp]

thanks wolfey for the help. i have a question

if im mining solo how do i get any feedback other than blocks found? i get no accepts rejects or HWs.... strange

I suffered from similar issues.

So now I run individual instances of CPUminer because CGminer doesnt give work to all my units equally.

Notice how some of your gridseeds have only submitted one share when others have submitted 50.

If you go the CPUminer route with 1 worker/1cpuminer instance per gridseed you will get a higher hashrate

It sucks. i know.

Set the diff to 32 or 64 on scryptguild. I like 32.

I get like 25-27k out of 80 gridseeds poolside clocked @ 850

[wolfey2014]

thanks wolfey for the help. i have a question

if im mining solo how do i get any feedback other than blocks found? i get no accepts rejects or HWs.... strange

I suffered from similar issues.

So now I run individual instances of CPUminer because CGminer doesnt give work to all my units equally.

Notice how some of your gridseeds have only submitted one share when others have submitted 50.

If you go the CPUminer route with 1 worker/1cpuminer instance per gridseed you will get a higher hashrate

It sucks. i know.

Set the diff to 32 or 64 on scryptguild. I like 32.

I get like 25-27k out of 80 gridseeds poolside clocked @ 850

Thanks Trollbox. I didn't know the answer to his question.
Have you tried the latest version of cgminer yet? 4.2.2 I think...
I'm curious to find out what cgminer users think of it.
Maybe you can go back to using it now. Looks like they've handled a lot of bugs and concerns etc.
Wolfey2014

[primer10]

Do you guys have this problem where the cgminer will segfault, complained of Gridseed failure after running for a few days?

[podizzle]

wow that is brutal maybe i can make a bat file to create all the instances?

anyway here is my gridseed eye candy   http://imgur.com/tYtBAh6

[jamieb81]

wow that is brutal maybe i can make a bat file to create all the instances?

anyway here is my gridseed eye candy   8)http://imgur.com/tYtBAh6

sweet!

[organofcorti]

wow that is brutal maybe i can make a bat file to create all the instances?

anyway here is my gridseed eye candy   http://imgur.com/tYtBAh6

Why are they pulsing? Are they some sort of cyborg gridseeds?