[ANN] Spondoolies-Tech - carrier grade, data center ready mining rigs

[klondike_bar]

The "Spondoolies hero of the day" medal goes to user Klondike_bar, for discovering a problem of high error rate on his machine. We released a fix (1.4.3) that should bring the HW error rate down from ~8% to 2%.

Please update ASAP and let me know if the fix works.

my HW errors on firmware version 1.4.2 settled down to about 5%

I will try 1.4.3 and report back tomorrow with results.

thanks zvisha and klondike_bar

BTW, what did klondike_bar discover?

His machine had exceptionally high error rate of over 10%. I have error rate of about 3-4%. He let me use his machine to understand the situation, and I found out that the fix for his machine improves HW errors in all my machines too. It is the way I use BIST to find optimal point, the BIST I did was too weak for many ASICs.
 

After ~4 hours running with zvisha's fix my error rate is reading as 0.66% down from a 9-13% range.

I used to get 1.42TH at the miner and ~1.26-1.31TH at the pool
Now i get ~1.38TH at the miner and 1.33-1.37 TH at the pool (eligius)

I am loving the quality of both the hardware and the team behind it!  (my only ongoing issue is that it is freakin' loud - but I bought knowing that it would be)

[1715278883]

1715278883

[1715278883]

1715278883

[1715278883]

1715278883

[RoadStress]

After ~4 hours running with zvisha's fix my error rate is reading as 0.66% down from a 9-13% range.

I used to get 1.42TH at the miner and ~1.26-1.31TH at the pool
Now i get ~1.38TH at the miner and 1.33-1.37 TH at the pool (eligius)

I am loving the quality of both the hardware and the team behind it!  (my only ongoing issue is that it is freakin' loud - but I bought knowing that it would be)

1.4.3 is best firmware up to this date! My miners saw a decrease in the miner speed, but an increase at pool (same eligius, BTCGuild doesn't seem to recover from the bad luck)

Thank you zvi and klondike_bar!

[wpgdeez]

My errors are much lower as well. Ambient is pretty warm right now, sure I can get more hash overnight when the temps drop.

SP10  1311.075 GH\s 1851 accepted 6 rejected 9 errors 18.34 utility

[zvisha]

Gotcha.  Thanks

According to Con the problem is now resolved. I didn't test it since I am not at work.
I created an image 1.4.4 (experimental) with latest cgminer, you can test and see if it works for you (remember to rm /mnt/config/etc/bin/cgminer).

[hardhouseinc]

The "Spondoolies hero of the day" medal goes to user Klondike_bar, for discovering a problem of high error rate on his machine. We released a fix (1.4.3) that should bring the HW error rate down from ~8% to 2%.

Please update ASAP and let me know if the fix works.

my HW errors on firmware version 1.4.2 settled down to about 5%

I will try 1.4.3 and report back tomorrow with results.

thanks zvisha and klondike_bar

BTW, what did klondike_bar discover?

His machine had exceptionally high error rate of over 10%. I have error rate of about 3-4%. He let me use his machine to understand the situation, and I found out that the fix for his machine improves HW errors in all my machines too. It is the way I use BIST to find optimal point, the BIST I did was too weak for many ASICs.
 

Firmware rev 1.4.3 looks like a real winner. I also have a SP10 that was registering around 100 GH/s lower at the pool due to high hardware errors, but firmware rev 1.4.3 seems to have raised the pool hash rate of this unit substantially.

Thanks zvisha!

Yup!!! Me too.  The web GUI is  showing a bit lower than before, but the pool stats are showing a bit higher and both are
staying right around 1.4TH/s each on 208V power with 22 - 25 degree air.  Great work Zvisha and SP-TECH for keeping up the
great customer service!!!

[hardhouseinc]

After ~4 hours running with zvisha's fix my error rate is reading as 0.66% down from a 9-13% range.

I used to get 1.42TH at the miner and ~1.26-1.31TH at the pool
Now i get ~1.38TH at the miner and 1.33-1.37 TH at the pool (eligius)

I am loving the quality of both the hardware and the team behind it!  (my only ongoing issue is that it is freakin' loud - but I bought knowing that it would be)

1.4.3 is best firmware up to this date! My miners saw a decrease in the miner speed, but an increase at pool (same eligius, BTCGuild doesn't seem to recover from the bad luck)

Thank you zvi and klondike_bar!

YES, the bad luck there is really
The last few blocks have been FAR apart and no multiple blocks in the same
hour like before after a 3 or 4 hour run of bad luck.
24 hour rate is down almost .04 BTC per day now from the other day.
I know people bash GHash.IO for their high pool rate but Im about to jump for a
week and see if it makes a difference.

[ManeBjorn]

Having a pool at the near 50% mark is what makes variance so pronounced elsewhere.
I understand the worry though.

BOT
With the new firmware 1.4.3 I get almost perfect sync between the SP10's rate and pool rate.
1.347 th/s Miner  --  1.365 th/s poolside. 
Very cool.
Thanks for a great job Zvi!!

After ~4 hours running with zvisha's fix my error rate is reading as 0.66% down from a 9-13% range.

I used to get 1.42TH at the miner and ~1.26-1.31TH at the pool
Now i get ~1.38TH at the miner and 1.33-1.37 TH at the pool (eligius)

I am loving the quality of both the hardware and the team behind it!  (my only ongoing issue is that it is freakin' loud - but I bought knowing that it would be)

1.4.3 is best firmware up to this date! My miners saw a decrease in the miner speed, but an increase at pool (same eligius, BTCGuild doesn't seem to recover from the bad luck)

Thank you zvi and klondike_bar!

YES, the bad luck there is really
The last few blocks have been FAR apart and no multiple blocks in the same
hour like before after a 3 or 4 hour run of bad luck.
24 hour rate is down almost .04 BTC per day now from the other day.
I know people bash GHash.IO for their high pool rate but Im about to jump for a
week and see if it makes a difference.

[hardhouseinc]

Well if you cant beat them, join them.
Ive read a few reddit articles about them pool jumping and sending blocks
to their pool. 
Ive also never researched it, so I don't understand double spending or how its
even possible, but its something that seems worth knowing more if there is any sort of
quick explanation.

http://www.reddit.com/r/Bitcoin/comments/279sex/ghashio_double_spending/?sort=new

http://www.reddit.com/r/Bitcoin/comments/27js1b/ghashio_24_hours_45_again/?sort=new


Possible, but who knows.  I don't really buy into it.  The bigger the
pool, the more stable the block finding it seems by looking at their charts I can find.

My hash rate is too low to deal with variables that big per day.
The network rate is jumping like 10 FKNG PETAHASH a week or more lately so
I think they are going to be well under 50% with all the big additions to the network
that are going on weekly if not daily.

[opentoe]

None of these automatic firmware upgrades change or reset the IP information, correct? Since these really are data center rigs most people would be upgrading the firmware remotely. If the IP info changes then no way to get back into miner. Safe to update to latest firmware remotely?

Thanks!

[Zelek Uther]

Well if you cant beat them, join them.
Ive read a few reddit articles about them pool jumping and sending blocks
to their pool. 
Ive also never researched it, so I don't understand double spending or how its
even possible, but its something that seems worth knowing more if there is any sort of
quick explanation.

http://www.reddit.com/r/Bitcoin/comments/279sex/ghashio_double_spending/?sort=new

http://www.reddit.com/r/Bitcoin/comments/27js1b/ghashio_24_hours_45_again/?sort=new


Possible, but who knows.  I don't really buy into it.  The bigger the
pool, the more stable the block finding it seems by looking at their charts I can find.

My hash rate is too low to deal with variables that big per day.
The network rate is jumping like 10 FKNG PETAHASH a week or more lately so
I think they are going to be well under 50% with all the big additions to the network
that are going on weekly if not daily.

For the love of Bitcoin don't mine on GHash.IO!

Your variance is lower on larger pools, but over time your payout is no different.

Learn about the >50% attack (aka 51% attack) and why this would be very bad for bitcoin.

[hardhouseinc]

I keep hearing about a 51% attack but no time to read that long about it....
Any cliff note versions?

KNC is shipping a crapload soon. 

LOTS of more mining power Im sure will hit the network and by July be a ridiculious number.
I dont see them hitting 51%, but then again Im probably wrong.

I think BTCGuild, Slush, and Eligius should all merge to compete.
Or bad idea??

Ill stay with BTCGuild, theyve been stable so far, just shit luck lately
Over the long run it seems to get worse....

[hardhouseinc]

None of these automatic firmware upgrades change or reset the IP information, correct? Since these really are data center rigs most people would be upgrading the firmware remotely. If the IP info changes then no way to get back into miner. Safe to update to latest firmware remotely?

Thanks!

My SP10s are hosted and Ive rolled the dice and with the big bug fixes in the firmware, I updated mine
to 1.4.2 and 1.4.3 using static IPs.  NO ISSUES with the updates and they work well, held their IPs after the update.

I mentioned my static IP need to be VERY stable to Zvisha and he said they worked on that to make sure its OK.

So far so good

[JoTheKhan]

I keep hearing about a 51% attack but no time to read that long about it....
Any cliff note versions?

KNC is shipping a crapload soon.  

LOTS of more mining power Im sure will hit the network and by July be a ridiculious number.
I dont see them hitting 51%, but then again Im probably wrong.

I think BTCGuild, Slush, and Eligius should all merge to compete.
Or bad idea??

Ill stay with BTCGuild, theyve been stable so far, just shit luck lately
Over the long run it seems to get worse....

Think Ghash.io should (needs to) put up a fee, something small to the point where it keeps people off it. Also making them money and helping qualm anxiousness about a 51% attack.

Unfortunately some mining pool is probably going to have close to 51% for the for-seeable future, especially with professional mining companies (who make hardware) starting to gain maturity and especially with each additionally halving. Companies are going to start being able to push out Hash Rate that might be equal or even exceed the NHR with every successive generation.

Especially if BTC price keeps going up, once companies are able to stay afloat while they R&D, Manufacture & Set Up the Farm, they will have less of a reason to sell (this specific view might be an ignorant view and I am probably missing some information).

It would be best if some consensus is put in place between the big pools to agree to not allow a certain % on their pool or to raise fees in accordance with their percentage of the network.

Example:

Wanna mine at Ghash.io:

When they are at 50% of the NHR: 2.5% fee
When they are at 40% of the NHR: 2% fee
When they are at 30% of the NHR: 1.5% fee (or even 0% I suppose, just to keep it from getting into the high 40s).

EDIT: Sorry I wouldn't be the most effective at explaining a 51% attack. My apologies.

[RoadStress]

None of these automatic firmware upgrades change or reset the IP information, correct? Since these really are data center rigs most people would be upgrading the firmware remotely. If the IP info changes then no way to get back into miner. Safe to update to latest firmware remotely?

Thanks!

Safe to update to the latest firmware remotely. I have already did it and nothing changes regarding to the IP.

[Zelek Uther]

I keep hearing about a 51% attack but no time to read that long about it....
Any cliff note versions?

It would allow a "double-spend" attack. 

That is, the attacker could buy something using bitcoins, and it would seem to the receiver that they had actually received the bitcoins (they would be confirmed on the blockchain), so they would ship the product.

Then using >50% of the network the attacker could create an alternate blockchain where that spend never happened. This new false blockchain would overtake the original blockchain because the false blockchain has more hashing power on it (i.e. >50% of the network). In this way the attacker can claw their bitcoins back, allowing them to spend the SAME bitcoins AGAIN with someone else. They could keep doing this until the world finds out about it (via news media), at which point everyone would panic-sell their Bitcoins until they had no value.

If this ever happens, no-one would trust Bitcoin or any Proof-of-Work cryptos (SHA256, Scrypt), since all PoW cryptos have this same vulnerability.
Basically it would be the end of Bitcoin. All our Bitcoins would be worthless (ditto for Litecoin, Dogecoin, Namecoin, etc).

Why would someone accept payment in Bitcoin if they knew it was possible that payment could be clawed back?  If we ensure that no-one has more than 50% of the global hash rate we KNOW a double-spend is not possible. If someone has >50% of the hash rate double-spends ARE POSSIBLE and therefore the integrity of Bitcoin is at stake.

See http://codinginmysleep.com/bitcoin-attacks-in-plain-english/
Note: >50% attack possibility is very real from the largest mining pool (currently GHash.IO).

tl;dr
The >50% attack would kill Bitcoin. No-one would trust Bitcoin, therefore no-one would use it.

LOTS of more mining power Im sure will hit the network and by July be a ridiculious number.
I dont see them hitting 51%, but then again Im probably wrong.

Let's hope the people taking delivery of those miners do the right thing for Bitcoin and don't point them at GHash.IO (or whatever the largest pool might be).

I think BTCGuild, Slush, and Eligius should all merge to compete.
Or bad idea??

We need to avoid having one dominant pool. A good selection of pools of roughly the same size would be best.

Ill stay with BTCGuild, theyve been stable so far, just shit luck lately
Over the long run it seems to get worse....

Great stuff!  Thanks for doing your part in helping to secure Bitcoin.

[mrmichaelvau]

any one selling their sp10 in texas?? hit me up..

[murdof]

About 51% the best solution is P2Pool - but the problem is that there is even more variance due to low speed.
If more people jump on P2Pool then we will not have to speak about 51% again.
In the long run you get same and better payouts but some weeks it shit.
So for example you might see that you get no payments for 3 days - but then get 3 payments in one day.

[murdof]

Zvi i have seen difference with the numbers reported by the miner and the numbers reported by cgminer api (i use miner.php)

One miner says 1394.55 GH/s - while cgminer api reports(5m): 1288.18.
Other miner says 1448.1 GH/s - while cgminer api reports(5m): 1339.96.

As far as I remember those numbers were pretty close.

Running latest firmware and mining at BTCGuild. As other mentioned it hasn't had luck last night but I just want to confirm with you if these api numbers are affected by Pool or they are real miner numbers.

[Zelek Uther]

About 51% the best solution is P2Pool - but the problem is that there is even more variance due to low speed.
If more people jump on P2Pool then we will not have to speak about 51% again.
In the long run you get same and better payouts but some weeks it shit.
So for example you might see that you get no payments for 3 days - but then get 3 payments in one day.

I agree.

I read up about P2Pool a while ago, and it has been on my To Do list for a while. Don't you need to run a full Node for that? I tried to run a Node on a RaspPi, then gave up.

[opentoe]

None of these automatic firmware upgrades change or reset the IP information, correct? Since these really are data center rigs most people would be upgrading the firmware remotely. If the IP info changes then no way to get back into miner. Safe to update to latest firmware remotely?

Thanks!

Safe to update to the latest firmware remotely. I have already did it and nothing changes regarding to the IP.

Thanks, and also for the previous response.