Avalon ASIC users thread

[jgarzik]

Good news.  Mining on BTC Guild with vardiff seems to have dramatically reduced the restarts.

Which cgminer version are you using?  
For stratum Im using the latest xiangfu openwrt solution with cgminer 2.10.4 running 27 icarus and my rig is constantly restarting,

before update was using cgminer 2.7.5(non stratum) and the 703n was more stable.

Avalon's cgminer 2.10.4, as described in the review linked in the OP: http://garzikrants.blogspot.com/2013/01/avalon-asic-miner-review.html

Make sure you are not mining at difficulty 1.0.  diff-1 is the original difficulty for pools, and it generates too much network traffic for this ASIC miner, possibly leading to many restarts.

Find a pool with "variable difficulty" or "vardiff."  p2pool supports a higher-than-1.0 difficulty by appending "+N" to the worker's name, e.g. "1BrufViLKnSWtuWGkryPsKsxonV2NQ7Tcj+8" for difficulty 8.0.

13 hours uptime and counting, since switching back to BTC Guild/stratum/vardiff.

[1714967676]

1714967676

[1714967676]

1714967676

[1714967676]

1714967676

[1714967676]

1714967676

[1714967676]

1714967676

[tacotime]

This is encouraging.  Hopefully more units will roll in this week.

[spiccioli]

hi jgarzik,

did you try http://hhtt.1209k.com/ ?

it is a PPS variable diff pool with very low fees for high diff shares, but no stratum.

best regards

spiccioli

[tacotime]

hi jgarzik,

did you try http://hhtt.1209k.com/ ?

it is a variable diff pool with very low fees for high diff shares, but no stratum.

best regards

spiccioli

That pool is fixed difficulty share (32), vardiff adjusts the difficulty every x minutes for incoming shares based upon your hash rate.  Share difficulty of 32 is still very low for 60 gh/s, ideally it should be in the hundreds or thousands.

[spiccioli]

hi jgarzik,

did you try http://hhtt.1209k.com/ ?

it is a variable diff pool with very low fees for high diff shares, but no stratum.

best regards

spiccioli

That pool is fixed difficulty share (32), vardiff adjusts the difficulty every x minutes for incoming shares based upon your hash rate.

hi tacotime,

no, 32 is default difficulty if you don't ask for a different one.

you use a bitcoin address as username and suffix it with desired difficulty, for example an address like 1....._512  asks for 512 difficulty shares.

_NNN is the difficulty you're asking, with 66 GH/s you can use 512 or 1024 without too much variance.

spiccioli

[tacotime]

ah, okay, that makes sense.  i am used to ltcmine.ru that constantly adjusts based on your hash rate (not sure if pool.itzod.ru also does this).

[Integ]

Make sure you are not mining at difficulty 1.0.  diff-1 is the original difficulty for pools, and it generates too much network traffic for this ASIC miner, possibly leading to many restarts.

Find a pool with "variable difficulty" or "vardiff."  p2pool supports a higher-than-1.0 difficulty by appending "+N" to the worker's name, e.g. "1BrufViLKnSWtuWGkryPsKsxonV2NQ7Tcj+8" for difficulty 8.0.

13 hours uptime and counting, since switching back to BTC Guild/stratum/vardiff.

Remains try it in a solo. This power is enough for the production of a block of time in 3 days, but that's not the case, and in the stability.

[jgarzik]

Remains try it in a solo. This power is enough for the production of a block of time in 3 days, but that's not the case, and in the stability.

Already tried straight solo mining to bitcoind, as noted elsewhere.  Saw at least 10% "get failure" rate, and it was falling back to the 2nd and 3rd pools.

It probably makes sense to set up a 1-person pool server, for solo mining, than go straight to bitcoind.

[Inaba]

You need to be using GBT if you are solo mining on bitcoind.  You can't mine via getwork on bitcoind at those speeds.

[jjiimm_64]

You need to be using GBT if you are solo mining on bitcoind.  You can't mine via getwork on bitcoind at those speeds.

I do getwork solo with 60G, but it is 24 seperate miners, not one unit hitting all at once if that makes a difference?

[kano]

You need to be using GBT if you are solo mining on bitcoind.  You can't mine via getwork on bitcoind at those speeds.

I do getwork solo with 60G, but it is 24 seperate miners, not one unit hitting all at once if that makes a difference?

The avalon effectively hits getwork 24? times at once, each ~1.6s
No idea how the work scheduler in cgminer handles that or if it needs adjustment ... then of course there is that commit mentioned ...

But of course ckolivas doesn't have an Avalon or even the source code so your stuck with what you got.
Damn shame about that hey ...

[-ck]

Good news.  Mining on BTC Guild with vardiff seems to have dramatically reduced the restarts.

That would almost certainly then be due to the memory leak associated with stratum share submission since you're submitting less of them you are getting a slower leak. The leak has been fixed in git, but unfortunately, you only have a binary... and we have no modified source code from them yet. Where is that anyway?

[jborkl]

I thought this was opensource?  So why are the source files not available if a product is shipped with it?

[darkip]

I thought this was opensource?  So why are the source files not available if a product is shipped with it?

China.

[Inaba]

But, but!  PuertoLibre says there's no IP problem in China!  He is SURE of it!

[Vicus]

But, but!  PuertoLibre says there's no IP problem in China!  He is SURE of it!

But, but! Josh said BF will ship at Oct! He is SURE of it!

[wknight]

But, but!  PuertoLibre says there's no IP problem in China!  He is SURE of it!

But, but! Josh said BF will ship at Oct! He is SURE of it!

No No.. he said November.. I mean December.. No.. no.. I mean January.. No No.. I mean February. No No. ohhh fuck it..

[hotdealdave]

Children, children...

I want to say thanks to the Avalon customers for beta-testing ASIC for everyone. Looks like you'll have plenty of time to get all the software bugs fixed before my singles arrive.

[jgarzik]

More good news.  Machine uptime and cgminer uptime are both now 24 hours.  Pool was BTC Guild, and stratum seems to most commonly use a share difficulty of 32.00000.

Now going to play with p2pool some more, with higher difficulty rates.

It sounds like a theory of a memleak is certainly plausible, at least.

[eleuthria]

More good news.  Machine uptime and cgminer uptime are both now 24 hours.  Pool was BTC Guild, and stratum seems to most commonly use a share difficulty of 32.00000.

Now going to play with p2pool some more, with higher difficulty rates.

It sounds like a theory of a memleak is certainly plausible, at least.

Glad to hear that you stopped at the 24 hour mark and not as the result of a restart.  Good luck on future testing!  Glad to see the unit is going to good use and not just being stuck on a single pool and ignored.

Regarding earlier comments about share difficulty, 60 GH/s actually hits ~32 on most pool vardiff implementations.  Most vardiff ipools are set between 12 and 30 SPM.  These settings leave 24 hour variance relatively low (+/- 1%).  While an argument could be made for more aggressive vardiff (1-6 SPM), the 24-hour variance does increase a bit when you go much under 12 SPM [unless I'm not remembering organofcorti's post about it].