Bitcoin Forum
April 18, 2014, 07:00:21 AM *
News: Due to the OpenSSL heartbleed bug, changing your forum password is recommended.
 
  Home Help Search Donate Login Register  
  Show Posts
Pages: 1 2 3 4 [5] 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
81  Bitcoin / Bitcoin Discussion / Re: What are the risks of having blockchain wallet backup on: March 10, 2014, 04:32:23 PM
If you don't know the exact details of how blockchain.info encrypts the wallet you shouldn't assume it is done properly.   Have they made the encryption/decryption process open source?

I think it is multibit compatible. (have not actually tested though)
82  Bitcoin / Bitcoin Discussion / Re: Co-workers talking about bitcoin.... so much misinformation still today on: March 10, 2014, 04:28:15 PM
When i visited my parents last month, I was asked if I lost all my Bitcoins. Apparently they were referring to the Mt.Gox melt-down: With the exchange reporting a large number of missing Bitcoins.

Heard about the apparent suicide at work. Had to filter through all the Mt.Gox news to find reference to it here.
83  Bitcoin / Bitcoin Discussion / Re: What are the risks of having blockchain wallet backup on: March 10, 2014, 04:20:37 PM
I have trouble understanding why people think wallet back-ups are secure if they are encrypted: you need to store that encryption key somewhere! That is all wallets are: sets of encryption keys.

I still recommend paper back-ups in at least two geographically separate locations. Physical theft is a concern though. Using n-of-m transactions and a Pay to script hash address (read: 2 of 3 locations) would be better. Blockchain.info does not support that as far as I know.
84  Bitcoin / Bitcoin Discussion / Re: MtGox database leak: why you should always mix your coins. on: March 10, 2014, 04:09:44 PM
Never recommend noobs to use Tor, it's a honeypot where they are worse off than not using Tor at all.
Noobs should use a trustworthy VPN instead.
The optimal solution is VPN + Tor.

Not if you stay in-network. Unfortunately, my services (bitcoin node) are not tor-enabled yet. Namecoin has the potential to facilitate this with human-readable addresses as well.
85  Bitcoin / Hardware / Re: [Guide] Dogie's Comprehensive ASICMiner Cube Setup on: March 10, 2014, 06:27:37 AM
I have not read the entire thread, but I have my blade working.

I had a hard time getting it to take settings. My guess is that if the HTTP Accept: header pushes the POST submission over two Ethernet frames, the second half of it is ignored. I found the w3m browser works to reliably program the machine.

Also, I am using molex connectors. This is not a problem if you make sure each connector has it's own dedicated pair of wires going back to the power supply (they don't melt at 5A).

Also learned:
  • Uses getwork protocol
  • Each hashing chip asks for work mostly independently (12.8 second round time)
  • Does not support long-polling
  • Draws a lot of power idle (and will reset every 150 seconds while idle).
  • 10Mbps interface (fast enough)
  • No mounting holes (unless you count heatsink)
86  Bitcoin / Bitcoin Discussion / Re: opening a Great Firewall on: March 10, 2014, 06:07:11 AM
Suppose a Great Firewall splits the world in two, and transactions that are broadcast on one side of the wall aren't heard on the other, and vice versa.  Blockchains are built in parallel in each subworld.

Then suppose a revolution happens and the Firewall opens up.  Doesn't the unlucky (slower) side of the wall lose all their transactions?

All it would take is about 5 modem links to leak data through the wall. Radio hams could even broadcast and receive that data. If the outage lasts more than a week, you can start smuggling transactions and block-chain updates by mail (Eg: Micro SD cards).

The Internet was designed to survive a nuclear exchange between the super-powers.
87  Bitcoin / Hardware / Re: Pictures of your mining rigs! on: March 09, 2014, 06:59:23 AM
Here's my setup. From top to bottom:

CCD-600 (Not Visible)
...
CCD-1000


My Search-engine-fu failed me. "CCD" commonly stands for "Charge Coupled Device". Adding "bitcoin -camera" to the search terms brought up Sony console game .torrents (apparently .ccd is a common extension).

A search of this forum for "CCD-600" only brought up the above post.
88  Bitcoin / Bitcoin Discussion / Re: POLL: How likely do you think it is that Dorian is Satoshi, founder of Bitcoin? on: March 08, 2014, 06:39:35 AM

Means nothing.  Dorian could've just logged in and posted that.

Except by doing that, he risks confirming that he is Satoshi.
89  Bitcoin / Bitcoin Discussion / Re: POLL: How likely do you think it is that Dorian is Satoshi, founder of Bitcoin? on: March 08, 2014, 06:30:54 AM
It is telling that satoshi broke his silence to say he was not Dorian.

P2Pfoundation link

The real Satoshi risked exposure to clear Dorian's name.
90  Other / CPU/GPU Bitcoin mining hardware / Re: Project: Stealth Mining Rig on: March 07, 2014, 07:17:55 AM
It only took me 10 months, but I built a 10Ghash/s stealth miner. It draws less than 300W.
91  Bitcoin / Hardware / Re: Pictures of your mining rigs! on: March 07, 2014, 07:12:26 AM
I would like to second the complaint about people quoting images. My "good" computer (below) runs bitcoind, so the machine I am using to browsing the web is 10 years old.

This is my response to to Project: Stealth Mining Rig thread. Only took about 10 months Tongue

Specs
  • Antec Sonata Case (not sure of specific model)
  • Antec EarthWatts 380W "green" PSU (see? it is painted green...)
  • 3Ghz dual core Pentium D processor (power hungry)
  • 4GB RAM (3306MB available)
  • 60GB SSD (expect to need to replace it by the end of the year)
  • 80GB SATA disk (used for unrelated encrypted jail)
  • 100Mbps RTL8139 NIC
  • 9600 baud serial console
  • FreeBSD 9.2 (apparently I have to figure out how to migrate to 10.0 now)
  • ASICminer new-style blade (10.7 GHash/s)



The grey cable is a serial cable (serial console over a null-modem cable). Removing the 2MB video card freed 124MB of addressable memory. I had to install a RTL8139 NIC because the on-board LAN would lock up when I tried to transfer bootstrap.dat over using SSH. There is no performance loss, since it is only plugged into a 100Mbps router/switch anyway.

Besides being discrete, the metal case blocks Electromagnetic Interference (EMI). I actually had to move the machine almost directly below the TV antenna (out of picture) because the front of the case does not block EMI as well as it should. The machine is also relatively silent (case is designed that way). With the fan on high, I can easily hear it, but the controlling machine (Pentium 166) is still louder.



As you can see, I crammed an ASICminer new-style blade into the case. The only fans are the PSU, the CPU, and the one to the upper left (120mm IIRC). I tested that the air-intake on the CPU stayed below 38C with the fan on low; with no miner in the case. Did not re-test with the fan on high and the blade in the case.

Cable management is key. It helps that the drives are partitioned from the rest of the case. I wanted to use a PCI-E 6 Pin -> dual molex adapter , but was not able to find one locally. Ended up using a 2 molex -> PCI-E 6 Pin adapter I got from a local computer shop free-of-charge. Apparently when they are building systems they always get an extra one. They come with video cards, but sane builds will use a PSU that includes the 6 pin connector. I made sure that each connector uses its' own wire pair going back to the PSU. This meant I had to run a different cable to the drives: one of the SATA power chains had a molex connector at the end, while another did not. I do share blade power with the vent fan (about 3W, 0.25A) and front LED though.

I was disappointed by the lack of mounting holes. The blade sits on a boxy plastic tube with a slot for the fingers cut into it (possibly not needed, but did not want to risk it). I have a zip-tie wrapped around the fan connector (left), fastened to the case with a series of half-hitches. I have insulated wire running through the two small holes on the right side of the board. That is looped around internal case structure.



I did not want any add-on cards at all. However, I think the NIC is low-profile enough to not cause problems. Because the spinning disk is SATA, and only has data used by its' jail on it: it is actually hot-pluggable. I learned the hard way that the ASIC miner blade is not hot-pluggable. It worked no problem once (the one wire gets hot until you plug in the second one). However, the second time I tried it, I got a big blue spark, and the PSU shut-down. Toggling the hard power switch got things working again.



I was surprised how much extra power P2Pool uses. With Bitcoind and Slush's stratum proxy running (pointed at Eligius), the CPU idled between 700-350Mhz (0.07 load average). Running P2Pool, the CPU stays between 2-3 Ghz (0.30 load average). This results in a power draw of an extra 20 Watts or so. I saw one report that pypy (a just-in time python compiler) works with P2Pool, but have not been able to get it working yet.

Update: Got pypy working. Results are inconclusive so far, but looks like about a 0.20 load average (33% CPU usage reduction) the same CPU usage at the cost of more than double triple the memory usage (from about just under 0.5GB 400MB to over 1.0 GB 1200MB). CPU still stays at high clock speeds.
Edit: It appears the initial drop in CPU usage was simply reflecting the reduced network traffic the node was seeing after restart.

Power Usage
  • Pentium D Machine = 94 Watts (idle, with fan on full)
  • ASICminer blade = 80W (28 W idle)
  • Pentium 166 = 36W
  • VDSL modem = 4W
  • Netgear Router = 2W
  • Subtotal = 216W

PS: I am aware that newer boards are more power efficient. The Pentium D was probably the most power-hungry design ever made by Intel. They went back to Pentium II based designs for their next chip (the Core 2). At 92Watts of idle power draw (fan low), I am convinced something is wrong with the machine. I don't think the board supports a calibrated power supply output resistance of 1.6 milliohms required by the processor. Stopped short of thermal imaging because I could not find an appropriate camera. I disabled everything not in use on the board, with no effect. Most my old computers draw about 40W with non-essential components stripped.

92  Bitcoin / Bitcoin Discussion / Re: Curious- Does Satoshi really own over 1 milion of BTC? on: February 27, 2014, 04:49:53 PM
The "satoshi premine" theory bugs me. When Bitcoin first started, the difficulty stayed at 1 for 11 months because nobody was interested in mining it.

I think I recall reading about it on Slashdot in 2009 2010. I thought it was a neat idea, but that it would never take off. One year later, Bitcoin is worth actual money. Suddenly, I can't CPU mine anymore.
93  Bitcoin / Pools / Re: [185 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool on: February 27, 2014, 04:14:44 PM
Hi guys, do you have any idea what might be causing this hashrate drops on my p2pool ?

http://freebtc.eu:8336/static/

I was going to guess no internet until you posted the error log.

"0197 > Failure: twisted.internet.error.ConnectBindError: Couldn't bind: 24: Too many open files."
 suggest you should look into raising the number of file descriptors. Have not run into that problem myself yet.
94  Other / Politics & Society / Re: Is a Madmax outcome coming before 2020? Thus do we need anonymity? on: February 27, 2014, 02:57:00 PM

My point was that you can't just assume the cause-and-effect you want emotionally. You would need to prove it. And the climate science is no where near even predicting temperature (they were so far off it is embarrassing), so don't even mention this more complex interactions. There are many possible scenarios that can create climate changes. The climate is always changing. This is normal. Been going on for millions of years btw. Wink

Everyone can have their emotional bias, but this isn't science.

The Arctic Ocean is now ice-free in the summer. My government has literally burnt books in an apparent attempt to cover up historical records that contradict their message that climate change is not a problem. That suggests that even utltra-right-wing Conservatives believe in the science.

Science is all about observing cause an effect. If you destroy the observations, I am going to have to rely on emotions.


Quote
Thermal efficiency. But other holistic measures of efficiency ICE wins, which is why even here in the Philippines where I am where we can get electric motorcycles imported from China, they are not competitive with gasoline motorcycles.
The ones I have seen in Canada are limited to 32 km/h, 500W and use lead-acid batteries. Unless you are talking about actual electric motorcycles and not vehicles classified as "electric bicycles". I can beat those things up a (relatively short) hill by brute-force.

Quote
Battery patents also appear to be why we do not have Electric SUVs pulling generator trailers on road-trips

You are using an internal combustion engine (ICE) in that design Wink

Yes. My point is that electric vehicles are only non-competitive if you assume that you need to quickly replace the energy expended. Most cars spend about 95% of their time parked, and are used for commuting less than 100km.

Pure Electric vehicles are not appropriate for continuous commercial use (like a taxi service) or long road-trips. That said, I have seen electric vehicles used all day in an industrial setting (intermittent use, read: forklifts and electric pallet jacks).

If I was going to spend a lot of time on the highway, I would consider a diesel vehicle.

95  Bitcoin / Pools / Re: [185 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool on: February 27, 2014, 01:30:35 PM
Hi! I have the following problem. Sometimes, if the restart p2pool (quark-algoritm), when the power pool > ~ 150 MH / s, process starts to load the processor to 100% and all shares the dead. All logentries "DEAD ON ARRIVAL". Why this happens and how to solve this problem?

If it takes your node more than 30 seconds to verify a block, you will get 100% stales. Even without 6 different hashing algorithms, I have seen up to 12 seconds of CPU lag on my machine processing P2Pool.

Part of the problem may be that python is an interpreted, not compiled language. A Just-In-Time compiler like PyPy may help. (I think I will test that on my own machine)

Edit: I think actual block processing is handed off to Bitcoind. P2Pool still has its own chain though.
Edit: pypy Does not work:
Code:
Traceback (most recent call last):
  File "app_main.py", line 72, in run_toplevel
  File "./p2pool/run_p2pool.py", line 3, in <module>
    from p2pool import main
  File "/usr/home/P2Pool/p2pool/p2pool/main.py", line 17, in <module>
    from twisted.internet import defer, reactor, protocol, tcp
ImportError: No module named twisted

Update:
Got it working.
96  Other / Politics & Society / Re: Is a Madmax outcome coming before 2020? Thus do we need anonymity? on: February 26, 2014, 09:44:05 PM

More cold and ice means less snowfall. You are obviously not a scientist.

I live in Edmonton. We get most of our moisture from the Pacific Ocean, which does not freeze over.

Quote
Ironically, I believe the conspiracy that electric cars are being held back by battery patents on NiMH batteries. As a result, they remain expensive and in limited supply.

And where is your science to support that?

If you knew enough about energy density and chemistry, you would realize batteries can not compete with hydrocarbon fuels for transportation.
ICE motors only get about 30% efficiency at best. They also have more moving parts and require more maintenance. Lithium-ion batteries may be lighter, but in practice you are only allowed to use about 60% of the capacity (30-90% charge). This negates the weight advantages over NiMH.

Patents don't require science to prove they exist: Patent encumbrance of large automotive NiMH batteries.

I ran into those limitations trying to build a lighting system for my bicycle. You don't see rechargeable D cells in stores: even though "real" NiMH cells have more capacity than the alkaline cells you can buy. The NiMH C and D cells you may see in stores are really AA cells: which is why advertise the same capacity as AA cell.

I got to play with some real 10Amp-hour cells imported from China by a friend. They cost about $3 each (packages of 16 totaling about $50). The "fake" D cells cost $5 each, and only have 2 amp-hour capacity. I asked about buying high-capacity cells locally: and was quoted $50 per cell. Patent fees discouraging automotive use is the only rational explanation.

Battery patents also appear to be why we do not have Electric SUVs pulling generator trailers on road-trips
Quote from: EVnut
A 500cc motorcycle engine is used, housed in a small, aerodynamic package. ~20kW DC output is sufficient for extended high-speed travel. The micro trailer incorporates intelligent "BackTracker" steering which automatically maintains trailer-to-vehicle alignment during backing to avoid jack-knifing. There is little question that this 350 pound trailer functioned as planned - sustaining freeways speeds for as long as the 9.5 gallon tank had gasoline. Amazingly, even with all the conversion losses added up, the gas mileage of this combo is comparable or BETTER than the pure gasoline version of the same vehicle.
(Bold is mine)

Quote
This leads to devices designed to betray the users.

You have not articulated a connection. Were you thinking IP == closed source, and open source is more security vetted?

No, that since the 1996 WIPO "Copyright" and "Performances and Phonograms" Treaties, devices and software have been designed to betray the user in order to enforce Effective Technological Measures.

As an example, HDMI was invented to encrypt the video signal going from video players to displays. The standard sucks: it's only purpose is to require device manufacturers to agree to a patent license. The patent license in turn, requires gives a $1-2/port discount for the use of HDCP, which does the actual encryption. The real "single cable digital audio-visual" standard is DVI, which does not use HDCP. The cabling is actually cheaper and higher quality (and can send signals much longer distances).


97  Other / Politics & Society / Re: Is a Madmax outcome coming before 2020? Thus do we need anonymity? on: February 26, 2014, 11:09:29 AM
You go into a lot of detail, but most of it strikes me as slightly wrong.

Not everything is part of some global conspiracy to control the masses. As you point out: delaying problems inevitably leads to collapse. Any secret cabal in power must realize this.

Global warning is real, and appears to be happening in my life-time. We are seeing more and more snow-free days. When I was a kid, snow-free Decembers were rare in Edmonton. In the last 15 years, we have come close to something like 5 "brown Christmases".

Ironically, I believe the conspiracy that electric cars are being held back by battery patents on NiMH batteries. As a result, they remain expensive and in limited supply. This conflicts with the theory that smart meters are control measures. (Electric cars can make very interesting use of smart-meters).

You also seem to be under the delusion that "Intellectual Property" rights should trump physical property rights. This leads to devices designed to betray the users. The rabbit hole goes so deep that the thought of setting up a VPN service users may rely on for their safety sickens me. Because modern computers are inherently insecure, sometimes actively made that way: there is no easy way to prove your infrastructure is not back-doored.

The think is, we need to risk setting up privacy-enhancing infrastructure. We need to start auditing our routers, NICs, keyboard and storage firmware. We need to start proving software (especially embedded software) is written correctly.

98  Bitcoin / Pools / Re: [185 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool on: February 25, 2014, 09:10:06 PM
Well I think I found out where the extra 6 seconds of lag came from:
Code:
2014-02-25 13:53:22.835376 Punishing share for 'Block-stale detected! height(6f2f28d90af36171b9bc41e7986c7bb24943c37bec4ef054) < height(151b94b7ca0744b491b3fcbf950435984bc36c5a97c9a08e9) or
19015f53 != 19015f53'! Jumping from 1ce9c6ad to 212471a6!
...
2014-02-25 13:53:24.455802 New work for worker! Difficulty: 1.000000 Share difficulty: 529462.142225 Total block value: 25.027840 BTC including 106 transactions
.
.
.
2014-02-25 14:05:10.552019 Transaction db65517b5b4b7314e86aa52a1f9e0ea3be37044eab1a73e408ea07f74ec5e383 rescued from peer latency cache!
...
2014-02-25 14:05:17.463766 New work for worker! Difficulty: 1.000000 Share difficulty: 553972.015873 Total block value: 25.105139 BTC including 529 transactions
.
.
.
2014-02-25 14:14:57.135562 P2Pool: 17337 shares in chain (9424 verified/17341 total) Peers: 6 (0 incoming)
...
2014-02-25 14:15:00.183437 P2Pool: 17338 shares in chain (9425 verified/17342 total) Peers: 6 (0 incoming)
...
2014-02-25 14:15:05.433806 New work for worker! Difficulty: 1.000000 Share difficulty: 565095.416800 Total block value: 25.057434 BTC including 304 transactions

I may be misinterpreting things, but I appears to take my machine 1-7 seconds to put work together.
99  Bitcoin / Pools / Re: [185 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool on: February 25, 2014, 05:52:01 PM
Edit: Since AFAIK, the getwork protocol does not allow workers to be interrupted with new work, we should be able to estimate the expected stales given a specific block frequency. If we assume a 13s worst case latency, that works out to at most 43% stales with a 30 second target. If we assume a 6.5s average, that works out to 21.7% stale. -- that does seem high.

Edit: Apparently Longpolling works around HTTP limitations by having the miner request new work immediately. The server then does not respond until new work is ready. Testing time.


See this:

https://bitcointalk.org/index.php?topic=18313.msg4556966#msg4556966

I don't know of anyone who has got slush's proxy to work with p2pool, I've tried countless times myself - it just doesn't work, in fact I have a feeling that that's what is now causing my p2pool start up error/warning. Your only choice with a blade is to use +1 at the end of your user name/addy which, tbh, is a bit of a waste..... Wink

OK, I finally got P2Pool running and learned the truth. One thing that confused me is that the +1 option does not appear to be documented anywhere. What it does is tell the blade to report diff-1 shares. Apparently, the blades do not support higher difficulty.

Test results are in: no longpoll support. 21% stale for me, I guess. (confirmed with netstat -- only transient connections show up on port 9332) The reported DOA rate varies between 20-40% for some reason.

PS: Slush's proxy does not work because P2Pool does not appear to support the stratum protocol. I may try a different proxy if I think it will result in less CPU usage (for example, using compiled, rather than interpreted code). Better power savings would result from replacing the CPU though.

Edit: P2Pool does support Stratum. Slush's proxy not working is more complex. P2Pool tries to automatically detect whether the worker is using stratum or not. While at the same time, the slush proxy tries to automatically detect if the pool is using Stratum or not (using the getwork protocol). I tried removing that detection from the slush code, and it still didn't work.
100  Bitcoin / Pools / Re: [185 TH] p2pool: Decentralized, DoS-resistant, Hop-Proof pool on: February 23, 2014, 11:18:59 PM
Hey guys, what bad things come from having a high getwork latency? I'm under the impression that stratum makes getwork irrelevant, but I'm a bit worried that mine is quite high (sometimes a few seconds, you can see it here http://www.blisterpool.com/stats in the graphs bit). Is this a symptom of anything in particular? The cpu doesn't seem to be under any enormous stress, but it does peak every now and then...is this the cause?

Also, I helped a miner get configured to run on my p2pool node, and he has some ASICMiner blades (~10.7GH each). Google search results (most seem to be from a year ago) have all told me that they naturally have high DOA rate with p2pool, without any real way to fix it. Is this still the case? His dead rate is around 40-50%. I also noticed the server was getting absolutely hammered with hash > target spam, and I suggested to the miner to use his bitcoin address+1 for his username. It reduced the server spam drastically (from 100/sec to several/sec), and reduced his DOA a little bit, but it also reduced his mean hashing power by about 10%. Could anyone explain to me what's going on here? I'd like to help him get better results.

Is this miner running a stratum proxy locally? each of the 32 chips seem to mostly ask for work independently.
My blade gets work 150 times per minute, meaning that the latency should be at most 12.8 seconds (150/32*60).

I don't actually have P2Pool working yet, but moved the midstate calculations to my Stratum proxy on the assumption it can calculate mid-state faster than the blade.

Edit: Since AFAIK, the getwork protocol does not allow workers to be interrupted with new work, we should be able to estimate the expected stales given a specific block frequency. If we assume a 13s worst case latency, that works out to at most 43% stales with a 30 second target. If we assume a 6.5s average, that works out to 21.7% stale. -- that does seem high.

Edit: Apparently Longpolling works around HTTP limitations by having the miner request new work immediately. The server then does not respond until new work is ready. Testing time.
Pages: 1 2 3 4 [5] 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44
Sponsored by , a Bitcoin-accepting VPN.
Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!