[ANN] US/North American Bitfury sales NEW STOCK ***NOW SHIPPING***

[Beans]

Is anyone else still waiting for a refund on the October batch?

Please shoot both me and Yvonne an email so we can get to the bottom of this.  First time I'd heard of it. 
sales@
support@

Thanks,
Jason

I have still not received my refund from you guys or even heard from you in a couple weeks.

I have sent you back several emails explaining that your case lies with Dave and Dave only.

You emailed me two weeks ago saying that Dave would be sending my refund, that's the last one I got. My case? It should be a simple refund, what do you mean Dave only? He has not made any attempts to contact me. He's not even picking up his cell phone.. What exactly is the problem here?

[salfter]

http://eligius.st/~wizkid057/newstats/userstats.php/1mineL5pWepg1WGKfDWDAFstQw4smUoFe

With six v3 H-boards, an AntMiner S1, and a couple of BFL Jalapeños, I'm getting about 375 GH/s. 

you could be doing a bit better. 6 H-boards with some small pencil modding can give you ~34*6= 204GH + 197GH for an antminer running at 387.5Mhz = 401GH

its worth squeezing a little more out while the difficulty is low

I started with two H-boards that delivered 35 GH/s between them with a 0.2-0.3% hardware error rate.  The four newer boards are doing maybe 30 GH/s.  Collectively, the six boards are running a hardware error rate ranging from 0.8% to 2.6%. 

Pencil-modding is for older boards, right?  These have a (ridiculously small) potentiometer on them for voltage adjustment.  I've not tweaked any of them from how they shipped.  I suspect the newer boards might be set lower than the older ones, but I've not tried taking any measurements.

I have tried briefly overclocking the AntMiner as well, but the power-supply wiring gets even warmer than it already is.  The power supply I'm using with it has both PCIe connectors at the end of a single cable.  I should probably use some Molex-to-PCIe adapters to spread the load across more output lines.  It might even be worthwhile to tap 12V off of the main ATX cable.

[allinvain]

http://eligius.st/~wizkid057/newstats/userstats.php/1mineL5pWepg1WGKfDWDAFstQw4smUoFe

With six v3 H-boards, an AntMiner S1, and a couple of BFL Jalapeños, I'm getting about 375 GH/s. 

you could be doing a bit better. 6 H-boards with some small pencil modding can give you ~34*6= 204GH + 197GH for an antminer running at 387.5Mhz = 401GH

its worth squeezing a little more out while the difficulty is low

I started with two H-boards that delivered 35 GH/s between them with a 0.2-0.3% hardware error rate.  The four newer boards are doing maybe 30 GH/s.  Collectively, the six boards are running a hardware error rate ranging from 0.8% to 2.6%. 

Pencil-modding is for older boards, right?  These have a (ridiculously small) potentiometer on them for voltage adjustment.  I've not tweaked any of them from how they shipped.  I suspect the newer boards might be set lower than the older ones, but I've not tried taking any measurements.

I have tried briefly overclocking the AntMiner as well, but the power-supply wiring gets even warmer than it already is.  The power supply I'm using with it has both PCIe connectors at the end of a single cable.  I should probably use some Molex-to-PCIe adapters to spread the load across more output lines.  It might even be worthwhile to tap 12V off of the main ATX cable.

Safety warning: DO NOT attempt to pencil mode your v2.3 h-cards as you will damage them. The pencil mod  is for v1.2 h-cards only.

Do NOT power your antminer from the same PCIe cable from the PSU. If your power supply is modular run another PCIE 6 pin (or the 6+2 pin connector) from your PSU modular end. If your PSU is not modular then once again find another _separate_ pci-e 6 pin cable and use that!

If you neglect to do this you could burn out the PSU's cables at either end (miner end or power supply end), or worst case start an electrical fire.

[Taugeran]

On a different note has anyone found that bitfury gear did not work well for them on the Eligius mining pool? In my case it seems that they are far more stable on btcguild. Could it be due to network connectivity. Maybe the stratum proxy submit queue is getting overloaded and hashrate is getting destroyed.

I couldn't get my V1 to work correctly on eligius. I had to put it on btcguild for it to work. I am guessing it has something to do with the min difficulty and it overloading, because it starts up, goes to ~400gh/s (it runs at 500gh/s on btcguild), then starts to slowly drop and shoot out a shit load of errors. I tried leaving it going for 30 minutes and no luck. Would keep shooting out errors and sit around 50-100gh/s. I also have been too afraid of using bfgminer with the V1:

v1 M-boards can be fried if the software doesn't compensate.
I don't make any effort to compensate in BFGMiner.
Nobody has tested it, but I cannot recommend this combination.
Do so at your own risk...

Side note: I've never had problems with chainminer on Eligius... (via stratum proxy of course) O.o

That's the exact same thing I've noticed! Damn, so it must be something related to the min diff which supposedly Eligius automatically sets to the "optimal value" - this clearly is not correct. Moving them back to btcguild solved all my stability issues. This is quite sad as I would've like to move all my miners over to Eligius to save on the 3% fee - hey every bit matters these days especially for low hashing power miners like me.

It seems chainminer can't handle the share diff changing once started. Does BTCGuild let you set a fixed or min diff? You don't want the pool changing it on the fly, so you want to set a high enough min that the pool doesn't think it's necessary to change it. I think Eligius is fully auto so is no good for chainminer.

Yeah I got my worker set at 512 min diff in btcguild and it works perfect. It also worked back in Giga's private pool where I manually set the worker min difficulty. Eligius doesn't have the option to manually set a difficulty and chainminer doesn't like that

solution: use bfgminer as the stratum proxy instead. it doesnt adjust downstream difficulty (YET). only side effect is bfgminer will show a lot of clientside rejects of the sort (H-not-zero) or similar. this is basically chainimine submitting to bfgminer a metric carpton of too low shares

I'm curious. @allinvain. Would you be willing to try out my suggestion of using bfgminer as a replacement to stratum-proxy to see if that helps. Only issue I can foresee is if bfg can't keep up with 400Ghs worth of diff 2 shares.

[klondike_bar]

http://eligius.st/~wizkid057/newstats/userstats.php/1mineL5pWepg1WGKfDWDAFstQw4smUoFe

With six v3 H-boards, an AntMiner S1, and a couple of BFL Jalapeños, I'm getting about 375 GH/s. 

you could be doing a bit better. 6 H-boards with some small pencil modding can give you ~34*6= 204GH + 197GH for an antminer running at 387.5Mhz = 401GH

its worth squeezing a little more out while the difficulty is low

I started with two H-boards that delivered 35 GH/s between them with a 0.2-0.3% hardware error rate.  The four newer boards are doing maybe 30 GH/s.  Collectively, the six boards are running a hardware error rate ranging from 0.8% to 2.6%. 

Pencil-modding is for older boards, right?  These have a (ridiculously small) potentiometer on them for voltage adjustment.  I've not tweaked any of them from how they shipped.  I suspect the newer boards might be set lower than the older ones, but I've not tried taking any measurements.

I have tried briefly overclocking the AntMiner as well, but the power-supply wiring gets even warmer than it already is.  The power supply I'm using with it has both PCIe connectors at the end of a single cable.  I should probably use some Molex-to-PCIe adapters to spread the load across more output lines.  It might even be worthwhile to tap 12V off of the main ATX cable.

I would definitely snip the head off the ATX cable and use it for a blade on the antminer. you shouldnt run more than 300W on a PCI cable or you can have a lot of problems when it fails on you.

if your boards are 2.3, dont pencil mod. but a small tweak on the potentiometer will help you get those 30GH cards into the 33-37GH range if you do a tiny adjustment at a time

[cypherdoc]

Has anyone tried bfgminer on a v3 m-board. I was under the impression that it does not work with v3 m-boards and using it could even damage the cards/system. Any info on this?

Runs like a champ for me.  See here for instructions, or you can download a prebuilt image here.  (For the image, default username and password are unchanged from stock; edit /home/pi/bfgminer.conf to set your pool credentials unless you want to mine for me. )

On a different note has anyone found that bitfury gear did not work well for them on the Eligius mining pool? In my case it seems that they are far more stable on btcguild.

I switched to Eligius right before their recent troubles.  I've not had any issues, and I think it's paying a little bit more than BTC Guild did.

http://eligius.st/~wizkid057/newstats/userstats.php/1mineL5pWepg1WGKfDWDAFstQw4smUoFe

With six v3 H-boards, an AntMiner S1, and a couple of BFL Jalapeños, I'm getting about 375 GH/s.  I'm waiting for more heatsinks to arrive before I order more H-boards; I came up one short on the boards I have right now.  I've been buying these...they arrive here pretty quickly, and it only takes a few minutes to stick 17 of them on a board...one for each ASIC and one for the voltage regulator.

salfter, does BSB 0-3 represent the 4 M board banks?

[allinvain]

On a different note has anyone found that bitfury gear did not work well for them on the Eligius mining pool? In my case it seems that they are far more stable on btcguild. Could it be due to network connectivity. Maybe the stratum proxy submit queue is getting overloaded and hashrate is getting destroyed.

I couldn't get my V1 to work correctly on eligius. I had to put it on btcguild for it to work. I am guessing it has something to do with the min difficulty and it overloading, because it starts up, goes to ~400gh/s (it runs at 500gh/s on btcguild), then starts to slowly drop and shoot out a shit load of errors. I tried leaving it going for 30 minutes and no luck. Would keep shooting out errors and sit around 50-100gh/s. I also have been too afraid of using bfgminer with the V1:

v1 M-boards can be fried if the software doesn't compensate.
I don't make any effort to compensate in BFGMiner.
Nobody has tested it, but I cannot recommend this combination.
Do so at your own risk...

Side note: I've never had problems with chainminer on Eligius... (via stratum proxy of course) O.o

That's the exact same thing I've noticed! Damn, so it must be something related to the min diff which supposedly Eligius automatically sets to the "optimal value" - this clearly is not correct. Moving them back to btcguild solved all my stability issues. This is quite sad as I would've like to move all my miners over to Eligius to save on the 3% fee - hey every bit matters these days especially for low hashing power miners like me.

It seems chainminer can't handle the share diff changing once started. Does BTCGuild let you set a fixed or min diff? You don't want the pool changing it on the fly, so you want to set a high enough min that the pool doesn't think it's necessary to change it. I think Eligius is fully auto so is no good for chainminer.

Yeah I got my worker set at 512 min diff in btcguild and it works perfect. It also worked back in Giga's private pool where I manually set the worker min difficulty. Eligius doesn't have the option to manually set a difficulty and chainminer doesn't like that

solution: use bfgminer as the stratum proxy instead. it doesnt adjust downstream difficulty (YET). only side effect is bfgminer will show a lot of clientside rejects of the sort (H-not-zero) or similar. this is basically chainimine submitting to bfgminer a metric carpton of too low shares

I'm curious. @allinvain. Would you be willing to try out my suggestion of using bfgminer as a replacement to stratum-proxy to see if that helps. Only issue I can foresee is if bfg can't keep up with 400Ghs worth of diff 2 shares.

Well, I  used bfgminer directly with mixed results. The main issue with bfgminer is it's autotune feature is not as effective as manually tuning the chips yourself, and you can't get it to save the results of the autotune nor edit a table to get it to forcefully use a specific setting. You could theoretically tell it to use a specific clock for each chip as a command line argument when you start it. That would result in one hell of a long command line argument - although you could script this with a bash shell script. I prefer the way chainminer does things. It just makes sense to have a table with clock values and parameters you can adjust. On the flip side chainminer is crap at other things.

I doubt that bfg will be able to keep up with 400gh of diff 2 shares, but there is only one way to find out.

[allinvain]

http://eligius.st/~wizkid057/newstats/userstats.php/1mineL5pWepg1WGKfDWDAFstQw4smUoFe

With six v3 H-boards, an AntMiner S1, and a couple of BFL Jalapeños, I'm getting about 375 GH/s. 

you could be doing a bit better. 6 H-boards with some small pencil modding can give you ~34*6= 204GH + 197GH for an antminer running at 387.5Mhz = 401GH

its worth squeezing a little more out while the difficulty is low

I started with two H-boards that delivered 35 GH/s between them with a 0.2-0.3% hardware error rate.  The four newer boards are doing maybe 30 GH/s.  Collectively, the six boards are running a hardware error rate ranging from 0.8% to 2.6%. 

Pencil-modding is for older boards, right?  These have a (ridiculously small) potentiometer on them for voltage adjustment.  I've not tweaked any of them from how they shipped.  I suspect the newer boards might be set lower than the older ones, but I've not tried taking any measurements.

I have tried briefly overclocking the AntMiner as well, but the power-supply wiring gets even warmer than it already is.  The power supply I'm using with it has both PCIe connectors at the end of a single cable.  I should probably use some Molex-to-PCIe adapters to spread the load across more output lines.  It might even be worthwhile to tap 12V off of the main ATX cable.

I would definitely snip the head off the ATX cable and use it for a blade on the antminer. you shouldnt run more than 300W on a PCI cable or you can have a lot of problems when it fails on you.

if your boards are 2.3, dont pencil mod. but a small tweak on the potentiometer will help you get those 30GH cards into the 33-37GH range if you do a tiny adjustment at a time

I wish they made the potentiometer larger.

[Keefe]

On a different note has anyone found that bitfury gear did not work well for them on the Eligius mining pool? In my case it seems that they are far more stable on btcguild. Could it be due to network connectivity. Maybe the stratum proxy submit queue is getting overloaded and hashrate is getting destroyed.

I couldn't get my V1 to work correctly on eligius. I had to put it on btcguild for it to work. I am guessing it has something to do with the min difficulty and it overloading, because it starts up, goes to ~400gh/s (it runs at 500gh/s on btcguild), then starts to slowly drop and shoot out a shit load of errors. I tried leaving it going for 30 minutes and no luck. Would keep shooting out errors and sit around 50-100gh/s. I also have been too afraid of using bfgminer with the V1:

v1 M-boards can be fried if the software doesn't compensate.
I don't make any effort to compensate in BFGMiner.
Nobody has tested it, but I cannot recommend this combination.
Do so at your own risk...

Side note: I've never had problems with chainminer on Eligius... (via stratum proxy of course) O.o

That's the exact same thing I've noticed! Damn, so it must be something related to the min diff which supposedly Eligius automatically sets to the "optimal value" - this clearly is not correct. Moving them back to btcguild solved all my stability issues. This is quite sad as I would've like to move all my miners over to Eligius to save on the 3% fee - hey every bit matters these days especially for low hashing power miners like me.

It seems chainminer can't handle the share diff changing once started. Does BTCGuild let you set a fixed or min diff? You don't want the pool changing it on the fly, so you want to set a high enough min that the pool doesn't think it's necessary to change it. I think Eligius is fully auto so is no good for chainminer.

Yeah I got my worker set at 512 min diff in btcguild and it works perfect. It also worked back in Giga's private pool where I manually set the worker min difficulty. Eligius doesn't have the option to manually set a difficulty and chainminer doesn't like that

solution: use bfgminer as the stratum proxy instead. it doesnt adjust downstream difficulty (YET). only side effect is bfgminer will show a lot of clientside rejects of the sort (H-not-zero) or similar. this is basically chainimine submitting to bfgminer a metric carpton of too low shares

I'm curious. @allinvain. Would you be willing to try out my suggestion of using bfgminer as a replacement to stratum-proxy to see if that helps. Only issue I can foresee is if bfg can't keep up with 400Ghs worth of diff 2 shares.

Well, I  used bfgminer directly with mixed results. The main issue with bfgminer is it's autotune feature is not as effective as manually tuning the chips yourself, and you can't get it to save the results of the autotune nor edit a table to get it to forcefully use a specific setting. You could theoretically tell it to use a specific clock for each chip as a command line argument when you start it. That would result in one hell of a long command line argument - although you could script this with a bash shell script. I prefer the way chainminer does things. It just makes sense to have a table with clock values and parameters you can adjust. On the flip side chainminer is crap at other things.

I doubt that bfg will be able to keep up with 400gh of diff 2 shares, but there is only one way to find out.

Does BFG at least display or log per-chip stats in some format? As long as the data is provided somehow, I can find a way to use it the way I use chainminer's .stat.log.

[allinvain]

No, bfgminer does not display nor log chip status. It just tries to continually auto-tune the chips.

[goxed]

http://www.coindesk.com/coinseed-5m-investment-bitfury-mining/

The board design looks interesting and likely is based on the strings design.

[cypherdoc]

No, bfgminer does not display nor log chip status. It just tries to continually auto-tune the chips.

What's your assessment as to how well it auto tunes versus chain miner?

Mine is that it is more stable but at about 10% less hashing speed, although this is just eyeballing it and excludes a few factors.

[RoadStress]

http://www.coindesk.com/coinseed-5m-investment-bitfury-mining/

The board design looks interesting and likely is based on the strings design.

1550 GH/s for 5 mil$? WTF?

[allinvain]

http://www.coindesk.com/coinseed-5m-investment-bitfury-mining/

The board design looks interesting and likely is based on the strings design.

1550 GH/s for 5 mil$? WTF?

That can't be their total, no way!

[Beans]

I don't care who sends my refund, but someone needs to get off their ass. 17k dollars is way to much money to be pulling this crap. The last time I heard from Dave was mid November when I told him to refund my orders. So I don't want to hear some take it up with Dave BS.


This was the last email you sent me regarding my refund. Dated Dec 30th.

I pulled up the orders and sent them to Dave to get the ball rolling. Since you’re nearby, I’m sure you’ll hear from him pretty soon.


Thanks!

-Jason

[superduh]

PSA regarding latest institutional purchase from MBP.

to be honest - this is spells the end of megabigpower's retail arm.
1) besides self mining!
2) and besides selling millions of dollars of equipment at "wholesale/discounted" pricing!
3) the costs for any retail purchase are too high, and what's the point of competing with 1) and 2) when it's the same company doing 1,2,3

hence. today is the day when people should realize that any RETAIL purchase from MBP is like throwing your money into the toilet (or sending bitcoins off into oblivion). please, leave the sympathy to a min unless you have some valid points.

[allinvain]

PSA regarding latest institutional purchase from MBP.

to be honest - this is spells the end of megabigpower's retail arm.
1) besides self mining!
2) and besides selling millions of dollars of equipment at "wholesale/discounted" pricing!
3) the costs for any retail purchase are too high, and what's the point of competing with 1) and 2) when it's the same company doing 1,2,3

hence. today is the day when people should realize that any RETAIL purchase from MBP is like throwing your money into the toilet (or sending bitcoins off into oblivion). please, leave the sympathy to a min unless you have some valid points.

No sympathy from me. It is totally true.

[allinvain]

No, bfgminer does not display nor log chip status. It just tries to continually auto-tune the chips.

What's your assessment as to how well it auto tunes versus chain miner?

Mine is that it is more stable but at about 10% less hashing speed, although this is just eyeballing it and excludes a few factors.

I don't think its better than a human tuning the chips. For example it seems that bfgminer does not know when it should turn some chips off completely. Luke - it's creator - answered me with "why would you want to turn the chips off anyways". I think he does not really know how these bitfury miners really work at the hardware level. I suggested he'd talk with punin, cscape, or the mysterious Russian/Ukranian guy who designed the chips.

[Unacceptable]

http://www.coindesk.com/coinseed-5m-investment-bitfury-mining/

The board design looks interesting and likely is based on the strings design.

Yep,money talks    Centralization is coming & it's NOT about faster transaction times (BTC is "throttled" @ 6 blocks per hour no matter what).......it's about PROFIT.......oh well,Thanks MBP 

[Biomech]

Also, Antminers are not currently for sale from stock.  And wasn't the 2.65BTC price just an introductory offer from the new BitMain website and limited to only the first 30 units sold?

No, first 30 units were priced at 2.40.

At current price and BTC/USD ratio, a full rig will not break-even.

spiccioli

Why do people always insist on offering that same piece of "advice" over and over again?

If I had heeded that advice I wouldn't currently have 2THs of mining gear with another 4THs+ preordered mining gear.  All of this has been paid for (with the exception of the first 200GHs or so) from the BTC I mined.  And I still have approximately $20,000 worth of BTC left in my wallet.

And no, I wouldn't have even more if I had bought and held BTC instead.  I tried doing that and I failed miserably.  Like most people, I panicked at the first dip and lost a good chunk of my modest investment, lol.  "Buy and hold" isn't in my nature.  I'm a miner and proud!

Give this man a cigar! Plus many.

I'm a miner by nature. In the 80's it was real gold. Made a bit of money at it, had a blast. Probably spent more on gas, explosives, and equipment than the gold justified. Wouldn't change it for the world, unless that change was hitting the mother lode.

Digital mining? Same deal. It's what I enjoy. My monetary profit is secondary to the emotional profit. It's worth it. I AM making money. Not a great deal at the moment, but enough to justify it from an accountant's perspective. I'm not an accountant and don't give a tinker's damn about their bean counting perspective. Prudent men do nothing, ever. It's another word for "coward".