Butterfly Labs - Bitforce Single and Mini Rig Box - page 70.

Inspector 2211

sr. member

Activity: 448

Merit: 250

Quote from: platorin on March 10, 2012, 01:08:35 PM

Guys, how many of you received their equipment? Is this company reliable or is this another scam? I would like to order few basic platforms, but I do not know if they can be trusted.

I think we are now beyond the question of whether this is a scam or not.
Just from counting the guys who have posted about it on this forum, more than 10 systems have been delivered so far.

However, that doesn't mean they cannot suddenly go under, for instance from cash flow problems or yield issues.

Lots of small companies succumb to cash flow issues (less of a risk here because orders have been pre-paid for 4 months, giving them an ample cash cushion) or yield issues.

Yield issues: What if they find out during QA that they have to [expensively] rework or even dump 70% of their circuit boards and can ship only 30%?

Yield issues like that can doom a start-up company, when you spend more money per shipped product than you take in.

Thus, I'd say, no, it's not a scam, but that doesn't necessarily mean it's 100% certain that you will receive product for the money you pre-pay.

platorin

sr. member

Activity: 392

Merit: 250

Guys, how many of you received their equipment? Is this company reliable or is this another scam? I would like to order few basic platforms, but I do not know if they can be trusted.

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

Quote from: cablepair on March 09, 2012, 09:56:04 PM

...
kano: i been on these forums awhile man and you gotta be one of the smartest dudes on here , if you dont mind me askin what do you do for a living?

LOL - thanks I'll presume that was a compliment

(I can never be 100% sure when people are typing text - I sometimes misunderstand and give back rather unfriendly replies Tongue

)
Anyway, just an IT person like many? around here.
I currently setup and run the IT for half a dozen very small companies.
I'm definitely not a Satoshi level (coz I absolutely HATE statistics as is probably obvious from some of the comments I've made) but I like to know a lot about what I'm dealing with (though I have a rather bad memory)
I'm sure that's more than most people would want to know Cheesy

fred0

sr. member

Activity: 349

Merit: 250

Quote from: jddebug on March 09, 2012, 09:29:42 PM

I have 3 singles running now. 258 watts including the laptop. Averaging 2426 Mh/s after 10 minutes of mining using cgminer.

I've not really mined before so I don't know if this is common but my hash rate takes a hit when stale shares happen. I get a few rejects either just before or just after the stales and it saying something about a Long Poll detecting a new block or a new block detected before long poll.

Is this due to the way the Bitforce Single does its thing or ?

Can I change settings to reduce the stale shares and rejects?

Thanks

I have the same. 10 Minutes is too short a period to measure.

The long polling/new block, there is nothing you can do about.

If Pool 0 is not providing work fast enough, I have increased my queue to 4 (1 for each single +1).

[Edit] updated benchmarks
us.ozco.in:8332
rejects 179
getworks 2072
shares 15301
time(s) 26901
time 07:28:21
mh/s 2442.9
mh/s/BFS 814.3
reject % 1.16%
efficiency 738%

You should run benchmarks on pools that you like, if for nothing other than finding a good backup pool. CGMiner is really outstanding for its failover capabilities.

Matthew N. Wright

hero member

Activity: 588

Merit: 500

Hero VIP ultra official trusted super staff puppet

Quote from: cablepair on March 09, 2012, 09:56:04 PM

kano: i been on these forums awhile man and you gotta be one of the smartest dudes on here

Agreed.

cablepair

hero member

Activity: 896

Merit: 1000

Buy this account on March-2019. New Owner here!!

Quote from: kano on March 09, 2012, 09:51:49 PM

Quote from: jddebug on March 09, 2012, 09:29:42 PM

I have 3 singles running now. 258 watts including the laptop. Averaging 2426 Mh/s after 10 minutes of mining using cgminer.

I've not really mined before so I don't know if this is common but my hash rate takes a hit when stale shares happen. I get a few rejects either just before or just after the stales and it saying something about a Long Poll detecting a new block or a new block detected before long poll.

Is this due to the way the Bitforce Single does its thing or ?

Can I change settings to reduce the stale shares and rejects?

Thanks

It is due to the way the Bitforce currently works.

The Bitforce replies with a list of shares after it has processed the full 2^32 nonce range.

Since it hashes at about 830MH/s then it will take approx 5 seconds to do a nonce range.
If you get an LP during that 5 seconds, then the work returned at the end of the 5 seconds for each Bitforce you have will of course be rejected if sent to the pool
(which also means this can happen: if it found the share before the LP but the 5 second reply is after the LP, then the share that could have been valid will of course be rejected)

5 seconds is 0.8% of an average 10 minute block on non P2Pool

IF however you mine P2Pool it will be 50% of an average 10 second block ... so don't mine P2Pool with a Bitforce.

kano: i been on these forums awhile man and you gotta be one of the smartest dudes on here , if you dont mind me askin what do you do for a living?

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

Quote from: jddebug on March 09, 2012, 09:29:42 PM

I have 3 singles running now. 258 watts including the laptop. Averaging 2426 Mh/s after 10 minutes of mining using cgminer.

I've not really mined before so I don't know if this is common but my hash rate takes a hit when stale shares happen. I get a few rejects either just before or just after the stales and it saying something about a Long Poll detecting a new block or a new block detected before long poll.

Is this due to the way the Bitforce Single does its thing or ?

Can I change settings to reduce the stale shares and rejects?

Thanks

It is due to the way the Bitforce currently works.

The Bitforce replies with a list of shares after it has processed the full 2^32 nonce range.

Since it hashes at about 830MH/s then it will take approx 5 seconds to do a nonce range.
If you get an LP during that 5 seconds, then the work returned at the end of the 5 seconds for each Bitforce you have will of course be rejected if sent to the pool
(which also means this can happen: if it found the share before the LP but the 5 second reply is after the LP, then the share that could have been valid will of course be rejected)

5 seconds is 0.8% of an average 10 minute block on non P2Pool

IF however you mine P2Pool it will be 50% of an average 10 second block ... so don't mine P2Pool with a Bitforce.

rjk

sr. member

Activity: 448

Merit: 250

1ngldh

I've seen a few of the same, a few rejects around the time of a longpoll. Not sure what's up.

jddebug

sr. member

Activity: 446

Merit: 250

Quote from: cablepair on March 09, 2012, 09:31:21 PM

Quote from: jddebug on March 09, 2012, 09:29:42 PM

I have 3 singles running now. 258 watts including the laptop. Averaging 2426 Mh/s after 10 minutes of mining using cgminer.

I've not really mined before so I don't know if this is common but my hash rate takes a hit when stale shares happen. I get a few rejects either just before or just after the stales and it saying something about a Long Poll detecting a new block or a new block detected before long poll.

Is this due to the way the Bitforce Single does its thing or ?

Can I change settings to reduce the stale shares and rejects?

Thanks

what pool you mining on?

bitcoin.cz

cablepair

hero member

Activity: 896

Merit: 1000

Buy this account on March-2019. New Owner here!!

Quote from: jddebug on March 09, 2012, 09:29:42 PM

I have 3 singles running now. 258 watts including the laptop. Averaging 2426 Mh/s after 10 minutes of mining using cgminer.

I've not really mined before so I don't know if this is common but my hash rate takes a hit when stale shares happen. I get a few rejects either just before or just after the stales and it saying something about a Long Poll detecting a new block or a new block detected before long poll.

Is this due to the way the Bitforce Single does its thing or ?

Can I change settings to reduce the stale shares and rejects?

Thanks

what pool you mining on?

jddebug

sr. member

Activity: 446

Merit: 250

I have 3 singles running now. 258 watts including the laptop. Averaging 2426 Mh/s after 10 minutes of mining using cgminer.

I've not really mined before so I don't know if this is common but my hash rate takes a hit when stale shares happen. I get a few rejects either just before or just after the stales and it saying something about a Long Poll detecting a new block or a new block detected before long poll.

Is this due to the way the Bitforce Single does its thing or ?

Can I change settings to reduce the stale shares and rejects?

Thanks

fred0

sr. member

Activity: 349

Merit: 250

Quote from: P_Shep on March 09, 2012, 05:48:42 PM

See if you can put a multimeter on the power input to the single and get the actual current draw. That'll be very useful.

The last time I used a multimeter was like 20 years ago. I don't even know where it is...it probably has the same batteries!

SgtSpike

legendary

Activity: 1400

Merit: 1005

Quote from: P_Shep on March 09, 2012, 05:48:42 PM

See if you can put a multimeter on the power input to the single and get the actual current draw. That'll be very useful.

You need an ampmeter specifically (not usually included on most multimeters), and it only works if you can separate the wires. Otherwise, the positive/negative power draws cancel each other out, and you'd see 0 amps on the ampmeter.

P_Shep

legendary

Activity: 1795

Merit: 1208

This is not OK.

See if you can put a multimeter on the power input to the single and get the actual current draw. That'll be very useful.

fred0

sr. member

Activity: 349

Merit: 250

On another note, I measured about 83.3W for the Wal-Wart power supplies. Does anyone have any idea about what the power would drop to if I make a number for ATX to coax power adapters? I've a couple of bronze PS units idle. I also am using a couple of Gold units which may free up in the future.

Edit: I may wind up doing it in any case to avoid having a bunch of wall-wart psus cluttering up the place.

rjk

sr. member

Activity: 448

Merit: 250

1ngldh

Quote from: DeathAndTaxes on March 09, 2012, 05:11:58 PM

Quote from: coblee on March 09, 2012, 05:01:28 PM

That's awesome. Inaba, you should as the people in the know if they've thought about splitting the rigbox into 2 devices then? Sell them for $15,000 each, mining 25.2 Ghash/s for 1250W. That would make more sense, right? Send 2 of these devices to the people that ordered a big rig already. Or would they not be able to achieve those specs then?

That's a good idea. Honestly since they need to use 2 PSU and multiple boards it isn't like 2x 25 GH units would be significantly more expensive than a single unit. With dual PSU I hope they have though about what happens if one trips offline and the other remains powered. Having a single PSU unit eliminates that risk/complexity.

When we first heard about it on the website so many months ago, I remember specific reference to pluggable modules, up to 32 such per rig. I haven't looked to see that is still specified on the site, but you could possibly buy them with fewer modules for lower power consumption and cost.

fred0

sr. member

Activity: 349

Merit: 250

Quote from: DeathAndTaxes on March 09, 2012, 05:11:58 PM

Quote from: coblee on March 09, 2012, 05:01:28 PM

That's awesome. Inaba, you should as the people in the know if they've thought about splitting the rigbox into 2 devices then? Sell them for $15,000 each, mining 25.2 Ghash/s for 1250W. That would make more sense, right? Send 2 of these devices to the people that ordered a big rig already. Or would they not be able to achieve those specs then?

That's a good idea. Honestly since they need to use 2 PSU and multiple boards it isn't like 2x 25 GH units would be significantly more expensive than a single unit. With dual PSU I hope they have though about what happens if one trips offline and the other remains powered. Having a single PSU unit eliminates that risk/complexity.

It definitely makes it look more attractive to me.

DeathAndTaxes

donator

Activity: 1218

Merit: 1079

Gerald Davis

Quote from: coblee on March 09, 2012, 05:01:28 PM

That's awesome. Inaba, you should as the people in the know if they've thought about splitting the rigbox into 2 devices then? Sell them for $15,000 each, mining 25.2 Ghash/s for 1250W. That would make more sense, right? Send 2 of these devices to the people that ordered a big rig already. Or would they not be able to achieve those specs then?

That's a good idea. Honestly since they need to use 2 PSU and multiple boards it isn't like 2x 25 GH units would be significantly more expensive than a single unit. With dual PSU I hope they have though about what happens if one trips offline and the other remains powered. Having a single PSU unit eliminates that risk/complexity.

DeathAndTaxes

donator

Activity: 1218

Merit: 1079

Gerald Davis

Quote from: Inaba on March 09, 2012, 04:14:00 PM

Ok folks, I have some good news with regards to the power requirements for the RigBox. According to "People Familiar with the Matter" (TM) - the RigBox will use dual 1500w PSUs, meaning they will be 120v/220v switchable units... also meaning you can plug, (presumably) each PSU into a separate circuit to prevent overloading one circuit with impossible current draw.

This also means, I presume, that each PSU will be field replaceable by the end user without too much hassle and downtime.

This pleases me a great deal.

Thanks for the details. Hopefully anyone with the $30K also is aware that two different outlets isn't necessarily two different circuits.

kano

legendary

Activity: 4592

Merit: 1851

Linux since 1997 RedHat 4

Quote from: BFL-Engineer on March 09, 2012, 04:56:13 PM

Quote from: kano on March 09, 2012, 04:48:50 PM

Well ... actually ... the original faulty estimate was 20W per 1GH/s
That, as everyone knows, blew up to 80W for 832Mh/s (if you ignore Inaba's results except the very best one he got)

The web site now says 50.4GH/s @ 2,500w

Doing a simple 50.4/0.832 x 80W gives 4846W so we are talking almost a halving of the power requirements ... and an 18% reduction in price.
That's gonna be interesting.

The RigBox power spec was calculated after what we learned from singles. So the increased power has
already been taken into consideration...

Good Luck,

Hmm interesting - so I guess that means there must be some relationship between the single and the board used in the rig box

Edit: so looking at this in a different way ... it seems to me that the power/temperature is the driving force with these.
Anyone got any more guesses to make about the chip - taking that into consideration?

Topic: Butterfly Labs - Bitforce Single and Mini Rig Box - page 70. (Read 186946 times)