FPGAs are basically programmable circuits. Basically, rather than running an encryption algorithm sequentially, it can run the entire thing in parallel (in just a few clocks, rather than hundreds/thousands/millions). The problem with FPGAs is that they aren't that cheap, but I don't know how the rates on the cheaper ones would compare to GPUs. ASICs are custom circuits, basically the FPGA turned into a non-programmable version. These cost hundreds of thousands of dollars to get a tape-out, but the per-part price goes way down, as does power consumption. edit: No one is going to make a bitcoin hashing ASIC. It is financially unfeasible. Someone has probably already written VHDL/Verilog code to do hashing, but it sounds like from the above posters that it's not as fast as GPUs. FPGAs are pretty slow (since all signals are routed through many gates due to the programmable nature) compared to ASICs (a GPU is just a massive ASIC), and they're also expensive, getting into the hundreds of dollars for low quantity on the high-end fast chips.
|
|
# ¿ Jun 27, 2011 03:56 |
|
|
# ¿ Apr 28, 2024 00:05 |
I've been following this thread for a few days now (since I've been on "vacation", i.e. being lazy at home). I figured I'd post a bit. It really surprises me how much money people are sinking into this. Watching the MtGox live chart, it looks like $50-$100k has been sunk into bitcoins since trading resumed. I have a former co-worker who left a few weeks ago (shortly after the article about Silk Road caused a lot of interest in Bitcoins). When I talked to him, he told me he was just going to live off his savings and work on some side projects for a bit (he was surprised that I didn't have enough money saved to retire... at 26, and claims he does). It turns out, his plan was to move back in with his family and one of his side projects is a Bitcoin exchange. He claims he got into Bitcoins when they were pennies, and I have a feeling he has a significant amount of money in it. He also thinks the housing market is going to drop to about 25% of where it is now, and that we're going to go through a time of hyperinflation in the US in the near future. That's my reasoning for following all this. I am so disappointed that the coin didn't crash as everyone cashed in their Bitcoins. The amount of conspiracy theory around every other currency coming from these people makes me believe that there are people crazy enough to sink tens of thousands of dollars in at this point. Wreckus posted:and once they design and build an ASIC ($$$$$) it'll be even faster/cheaper/mass produceable. I don't think they would be fast enough to make it financially viable unless the price of bitcoins goes waaaaaay up. You get an ASIC designed (not a simple feat) and debugged using an FPGA. Now you've got a huge NRE investment to start getting actual chips in stock, and this is going to be weeks of lead time. Then you have to spin boards to populate the chips onto (more $$$$$). Despite all that, I have no doubt that someone with some ASIC design experience could basically do a kickstarter approach and get people to pre-order in BTC to fund all that. By the time product exists, the difficulty would be so much higher.
|
|
# ¿ Jun 27, 2011 04:07 |
homeless snail posted:Tell a Bitcoin user that Calculating a hash in one clock is probably not practical given propagation delays, unless he's ramping that 8 MHz clock up to something higher internally for registering. This guy is delusional. ASIC designers make big bucks for a reason (and they don't accept pay in BTC).
|
|
# ¿ Jun 27, 2011 04:21 |
Police Automaton posted:^^^ like I said earlier, I wouldn't be suprised if most of the exchange websites are basically complete BS ^^^ He's an idiot that doesn't know a drat thing about design. He's designing with schematic capture (rather than VHDL/Verilog), and doesn't know to pipeline the design with registering. There are some people there who know what they're talking about (100-200 MHz designs by using pipelining in an FPGA), and they're all the ones who know that a Bitcoin ASIC is a dead-end.
|
|
# ¿ Jun 27, 2011 04:28 |
The $110k of buy orders at $15.97 boggles my loving mind. Was the market always like this? I guess someone's got big money thinking that it's going to go back up to $30.
|
|
# ¿ Jun 27, 2011 04:40 |
fishmech posted:Do remember that most of that money has been sitting at mt gox for like a week while it was down. I would have thought it would be still sitting there, but coming out of the system, $1,000 at a time.
|
|
# ¿ Jun 27, 2011 05:53 |
Jombo posted:This has been a fun show, but I can't handle the constant drama anymore of following the bitcoin saga. There needs to be some sort of service that txts your phone when the crash finally happens so we can all watch it real-time. But until then I must give this topic a rest, it's consumed more of my time () than Fallout 3 + New Vegas. I don't know what the price difference is, or how the resources are utilized, but I would have thought the Virtex would give you better performance (unless you're just doing it for the engineering, rather than the bitcoin glory).
|
|
# ¿ Jun 27, 2011 06:04 |
CRUISECONTROL posted:Is the FPGA stuff going anywhere either? I would be pretty surprised if they could program an FPGA to keep up anywhere near a full blown GPU. In the thread that was linked earlier, someone claimed to have an FPGA design running at 100-200 MHz. Assuming the pipeline length is negligible (hundreds or thousands of stages isn't an issue since it stays full, but I doubt it has this depth), then depending on the size of the FPGA, you're looking at 100-200 MHash/sec for each pipe you can fit for a lot less power than a video card. I don't know if these claims are true, just culling from the information I read in the thread filled with misinformation.
|
|
# ¿ Jun 27, 2011 06:15 |
CRUISECONTROL posted:The guy in the ASIC thread talks about how he has a BS in physics and is one year in to getting a masters in electrical engineering. It's basically like every goon run game, except instead of being better at game design than Valve they are better at hardware design than Intel, AMD, and NVidia To be fair, if someone actually knew what they were doing, they could make a better Bitcoin mining card than the AMD GPUs. It would take months, lots of money, and lots of skill. Also, it would be nigh worthless outside of Bitcoin mining (I guess it could still be useful if you need to process tons of SHA-256 hashes for some other purpose). I have no doubt that AMD/Intel/NVidia could build something that would get huge throughput if they wanted to invest the money and resources to make something dedicated to the task.
|
|
# ¿ Jun 27, 2011 06:21 |
The $110k at $16 turned into $100k at $15.50. I think whoever said that was probably the MtGox circuit breaker is probably right (or some wise investor has lowered his price goal on a big buy).fishmech posted:The best part of that thread is this tho: I guess I just don't get why that's better than plastic, which can already do that (minus the spontaneous generation of new money).
|
|
# ¿ Jun 27, 2011 06:29 |
Jombo posted:Virtexes would give much greater performance, but I'm not in it for mining - just to practice my VHDL/ISE and see how much of performance increase over the standard SHA256 code from opencores I can achieve (for personal epeen glory). I really don't know the details of what goes in to ramping up performance, but if the limiting factor is the propagation delay of the combinatorix, wouldn't it just be as simple as adding pipeline stages (also, optimizing the routing in the FPGA, but I think that's automatic). I should probably be more familiar with the SHA-256 algorithm, since I'll probably need it at work in the next few years. I'm on the other side of the FPGA fabric, though (embedded uC code to control FPGA functionality).
|
|
# ¿ Jun 27, 2011 06:58 |
So, maybe I am not understanding the amazing cyrptocurrency correctly, but as I understand it, a bitcoin transaction gets submitted to the network, and miners end up verifying it until a consensus is reached that the transaction is valid. Fewer miners means transactions take longer to validate. What's to stop someone from flooding the network with invalid transactions that would still need to get verified (and fail), which will use up more resources and slow down real bitcoin transactions (not exchange trades, though)? Am I misunderstanding something?
|
|
# ¿ Jun 30, 2011 23:10 |
Harik posted:Except you don't need to do billions of hashes to recognize an invalid transaction, there's only a few rules that need to be followed. The from-address must be signed with a matching private key, it must have the funds (determined via transaction history that every daemon has on record) and if below a certain threshold size must offer up a payment to be processed. So, you have a botnet out there. You have them all running modified clients with valid wallets/keys. These keys are signing invalid transactions on coins and sending to the network. Each drone in the botnet is sending these out as fast as possible. Basically, craft something for the drones to send out that comes off as potentially valid to get distributed around the network, but isn't actually valid. In this case, you're not interested in convincing the network that your transaction is valid, but getting the network to waste time on all your invalid transactions and drowning out the real/valid ones. I have no interest in actually trying to breakdown the bitcoin network, but I find avenues of doing so interesting. It seems like it might be the same issue that befell Gnutella: distributed search meant that a modified node could return results for things that you didn't actually search for. The network still worked, but it made it difficult to actually find what you were searching for. In the bitcoin world, it might be possible to interfere with the network such that legitimate transactions become more and more delayed.
|
|
# ¿ Jul 1, 2011 06:52 |
You know, I heard a story on NPR last week about a style of home schooling where kids just learn what they want to when they want it. It's like the ultimate libertarian style of education. They call it unschooling. It sounds retarded, and it kind of is. http://www.npr.org/2011/06/06/137009154/unschooled-how-one-kid-is-grateful-he-stayed-home Kid learned to read at age 10 because he wanted to play Magic the Gathering. Learned to spell "a year ago" (looks like he's about 16). It wouldn't surprise me if this was Atlas, except the article isn't filled with libertarian buzzwords.
|
|
# ¿ Jul 5, 2011 06:13 |
HineyBorelTheorem posted:HOMEWORK He's just trying to effect change in the education system. (Throwing a curve ball here)
|
|
# ¿ Jul 5, 2011 06:53 |
max4me posted:Yes I mean real Teachers, who know their poo poo. I complained about busy work back in high school, and then in college. I was only accepted into one college because I didn't do the busy work. I barely made it out of college because I never did the busy work. To this day, I never finish projects. All that busy work serves a purpose: preparing students for the real world, where you're going to be stuck doing busy work poo poo.
|
|
# ¿ Jul 5, 2011 07:34 |
ManlyWeevil posted:Except MP isn't saying that the busy work is a waste, in fact the contrary that school itself is a rather effect way if preparing the students to finish tasks, even those they'd rather not be doubt because the swingset outside looks so awesome this time if day. Atlas is actually a very good example of what its intending to avoid; his ideas are all over the place and he doesn't stuck to one long enough to even write a mediocre business proposal. Yeah sure we got along relatively fine for centuries with on the job training, but that really doesn't cut it anymore for most professions. Grade school provides an important foundation for the modern world and very rarely can an individual make muster without that basis. Thank you for writing this. I kept writing posts then deleting it because it was way too many words for a subject that I don't actually care strongly about. That, and I never finish posts as a result of me not being forced to do busy work in school. edit: Even now I just started editing this post and added way too many
|
|
# ¿ Jul 5, 2011 08:29 |
There are sizable bid and ask walls just above and below $12.50 right now. I'm curious to see which one falls first.
|
|
# ¿ Jul 5, 2011 17:46 |
Molotov Cock Tale posted:The bid wall will fall, I was watching a huge bid wall at 14 last night, vanished as soon as it got sold into. They're probably set up by people with lots of coins, and a few $k in the pokemon currency exchange, who are trying to keep the market up so they can sell more coins before it all goes to poo poo. They certainly won't be looking to buy at that price. The bid wall fell. There is almost nothing above the $12 cliff. edit: If that is a fake bid, or if that vanishes, that is going to bring the volume of bids wayyyyyy down and maybe we'll see the price start pushing towards the much smaller wall at $11.
|
|
# ¿ Jul 5, 2011 19:25 |
God, I am watching this way too intently. We're almost at the wall.
|
|
# ¿ Jul 5, 2011 19:29 |
Nux posted:Could you explain the walls you're talking about? I assume you men buy orders over market price, but how can you see them before you're there? http://mtgoxlive.com/orders?volumeon;oneaxis Walls would be the "walls" in the bid/ask volume. They represent a large buy or sell order, which would take a large amount of volume to overcome. Generally. this would give the price stability, but apparently the orders have been disappearing as soon as the price hits it (i.e. someone is pulling their order or it was never a real order to begin with). They're buy orders under market price, but they serve as a deterrent for the price dropping. Anyone with real buy orders will want to place them above the large wall so that their order has a chance of getting filled. It basically props up the price. The same can be done on the selling side if you want to keep the price below a certain point, but who would want to keep the value of bitcoins low? Also, with single axis on that chart, you can see that all the sell orders showing on the chart wouldn't be enough to satisfy that buy order.
|
|
# ¿ Jul 5, 2011 19:37 |
12 bux. Also, Molotov, that second wall is at $10, not $11. There's only a small amount of price support above the $10 wall.
|
|
# ¿ Jul 5, 2011 19:45 |
Nux posted:Thanks both of you! I think I understand now, though the http://mtgoxlive.com/orders?volumeon;oneaxis graph is loving with my mind haha. The orange bar is bid orders. Vertically is how many are available at or above that price. The blue bar is ask orders. Basically the same thing as above, but vertically is how many are available at or below that price. These two lines are handy because you can see how much volume is required to move the price in either direction. The green line is trade history. Price is on the bottom, vertically it is showing over time. The darkness is volume.
|
|
# ¿ Jul 5, 2011 19:54 |
Dre2Dee2 posted:Oh poo poo oh poo poo oh poo poo... my kid's college fund Seriously, how many of these Unschooled kids could pass the GED test when they turn 18? edit: Let's go, $11.0!
|
|
# ¿ Jul 5, 2011 19:56 |
spidoman posted:So, what exactly set the drop off? It stayed somewhat stable for a few days following coming back, what changed that? It could be as simple as a speculator cutting his losses and cashing out at a slightly accelerated pace. In other words, this may have been the natural outcome over a longer period of time, but someone wanting to get out has sped this up. I think after it hit $14, we started seeing the panic sells.
|
|
# ¿ Jul 5, 2011 20:01 |
What just happened? Price is all over the place! edit: crazy volatile right now.
|
|
# ¿ Jul 5, 2011 20:33 |
A millionaire in lovely fiat currency money. Why would someone so invested in bitcoin care that they are a millionaire in lovely fiat currency, when the true value is in the bitcoin. Also, good luck cashing out 50k+ BTC without crashing the market.
|
|
# ¿ Jul 6, 2011 07:58 |
So, what's the opposite of trying to catch a falling knife? In this case, I'm guessing people are buying on the rebound bandwagon, but it could just as easily drop back down to $10 today.
|
|
# ¿ Jul 6, 2011 18:27 |
Kavros posted:Why announce this? Now they're all just going to be on the lookout for the new gormless pedagogue who churns out abortive, naive proposals. Because he loves attention. Even negative attention.
|
|
# ¿ Jul 7, 2011 15:38 |
BTC.spengler posted:I am officially shutting my miners down now. .125btc/day isn't worth it. If you go to church you can confess your sins and be absolved of all wrongdoing from selling them! A clean slate!
|
|
# ¿ Jul 9, 2011 20:12 |
hayden. posted:Why doesn't he just go steal poo poo from someone's yard, it's essentially the same thing and he'd make more money. You should be getting tons of donations, because all the internet libertarians post about how the government shouldn't provide for those without, and that they donate to charities that already do it. You should pick up on the cross section of internet libertarian and bitcoin user.
|
|
# ¿ Jul 10, 2011 00:31 |
Beowulfs_Ghost posted:Would it be bad of me to start a little business trading bitcoins for pre-paid Visa gift cards? With a shipping and handling fee of course. Compensate me for going to an evil bank and dealing with that filthy fiat money and all. You should just make a business out of selling all sorts of anonymous goods in BTC. You can sell "burner phones", prepaid cards. Hell, why not sell anonymous cash for BTC? Send me 1 BTC and I will send back four non-sequential one dollar bills. For 4 BTC, will get a well-used twenty. 15 BTC will get you a Benjamin.
|
|
# ¿ Jul 10, 2011 07:30 |
Lodin posted:I think Buttcoins need their own definition, no scrip. Since the Dollar and other government backed currencies are called fiat money, Buttcoins should be lada, trabant or maybe gremlin money. škoda notes
|
|
# ¿ Jul 13, 2011 15:26 |
FCKGW posted:Could you link this thread? I want to read it.
|
|
# ¿ Jul 13, 2011 21:06 |
ChubbyEmoBabe posted:Sorry for trolling again. Mods don't ban me I don't usually have any respect for WHORES, but I have to agree with you.
|
|
# ¿ Jul 13, 2011 21:08 |
Tasty and Delicious posted:I'd prefer he interview Vegetta or both of them together. Jessy doesn't sound that crazy at all if I wanted to talk to a normal person I'd just talk to the woman next to me in the office. She's still a dominatrix that runs hew own successful business. Not crazy, but probably genuinely interesting.
|
|
# ¿ Jul 13, 2011 22:04 |
Lots of downward market pressure at the moment. $200k of bitcoins for sale at $14 or under. A big buy order at $13.5 just fell, but it looks like it may have been filled rather than vanished (basing this off volume).
|
|
# ¿ Jul 17, 2011 06:00 |
Lolie posted:Atlas thinks that we're a social threat to their community. Aren't we?
|
|
# ¿ Jul 18, 2011 04:14 |
homeless snail posted:I'm just glad they don't have the to buy an account, if a Bitcoin forum guy started posting in this thread, all our plans would be ruined! I think they'd be okay with dumping $10 to post here and get banned. I mean, those ten dollars are fiat poo poo that aren't backed by anything, and they'll be valueless within decades. Now, if it were to cost .8 bitcoins, then it would be throwing away something of value.
|
|
# ¿ Jul 18, 2011 05:42 |
|
|
# ¿ Apr 28, 2024 00:05 |
Angela Christine posted:He is an advocate of homeschooling, and apparently goons think that on this one topic he wouldn't give an opinion on something he knows nothing about. I had posted the article because I thought I had read that Atlas was home schooled. Regardless, it had a very libertarian vibe to it and it made me laugh. I'll be kind of sad if we find out that Atlas wasn't actually pulled from public school to get more TIME THAT WILL TO BE ONLY HIS. (Yes, I know I broke English)
|
|
# ¿ Jul 20, 2011 14:05 |