So if you base your beliefs on how much money is spent, if you believe in ASICMiner, clearly you should believe in BFL as well.
I am not basing my beliefs on how much money is spent, but what was promised. I simply dont get the maths behind producing 60ghs for ~1000 USD. But this is exactly what ASICMiner has produced - and you believe them! 12 Thash/s divided by $200-300k IPO = 60-40 Ghash/s per $1k You can say that ASICMiner could do it because their 130nm process is cheaper than BFL's 65nm, but BFL is financing themselves with prior FPGA profits, and by manufacturing a lot more chips to recoup the higher NRE - by all estimates they have 100-200 Thash/s of preorders... Anyway we are getting off-topic here. |
|
|
|
I am sure someone like king_pin still does not believe in ASICs even though we have now 5 delivered Avalon units (jgarzik, the Foundation, techwtf, xianfu, and another chinese)  I just cannot understand why you guys at some point did not believe in ASICs. It is a financially and technologically natural and inevitable evolution of Bitcoin, like pretty much all the other computing industries on earth! Let me correct myself. Being in that kind of bussines, I know how much it costs to make a proper asic machine - what avalon made is more of a fpga/asic clone. Avalon delivered 5 items, there is no kind of guarantee that they will deliver all. Just wait a little more time for the entire batch 1 of Avalon to be delivered. You will see. In ASICminer i believed, since he was planing to spend much more money on development.
You are the only person who seems to believe that ASICminer spent more money [than other vendors]. We have proof that BFL have even more capital than ASICminer. ASICMiner's IPO raised only $200-300k while BFL has presold a few million dollars of devices (we have indirect proof through bitpay who inferred that $250k of ASICs were sold the first 24h alone after the opening of pre-orders) and BFL also claims to have investors (not proven, I will give you that). Plus BFL was already established business having sold thousand of FPGAs in 2012. So if you base your beliefs on how much money is spent, if you believe in ASICMiner, clearly you should believe in BFL as well. |
|
|
|
I am sure someone like king_pin still does not believe in ASICs even though we have now 5 delivered Avalon units (jgarzik, the Foundation, techwtf, xianfu, and another chinese) I just cannot understand why you guys at some point did not believe in ASICs. It is a financially and technologically natural and inevitable evolution of Bitcoin, like pretty much all the other computing industries on earth! |
|
|
|
|
r4ez9EWmJ9tAdJ4Rrjmri7u14bVpowQYnS
|
|
|
|
I watch the culture of BFL-influenced mods censoring anti-BFL postings I was very surprised to experience this myself. One of the mods is really a disgrace for this forum! becoin, can you send me the escrow code that casascius generated for you: https://bitcointalk.org/index.php?topic=135914.msg1488827#msg1488827 so that I can reclaim my 0.1 BTC bet that you failed to match? |
|
|
|
iHeartSmartArt, this technique is snake oil. I hope you realize this. A 2-year old can position and resize the spiral to match any part of the price curve going up or down. Or even sideways  |
|
|
|
Bump. My PSUs are perfect for potentially upgrading the PSU of the future 6-module Avalon unit which will need about 1kW |
|
|
|
|
Bitcoins For Gentlemen Miner.
Blocks For Grabs Miner.
Bitcoin Finding Gizmo Miner.
Bigger, Faster, stronGer Miner.
Basically a Freaking Good Miner.
Brazen, First, Genuine Miner.
Boldly Flexible and Generic Miner.
|
|
|
|
Jeff said his 3 module unit pulls 620W. Take away the 20W for the fans and controller, and you're looking at 200W for each module, not 166W. That means 6 modules is (6x200 + 20) 1220W. You could run that on a 1250W, but you'd be pushing the limits. You're either looking at a $300-400 PSU, or 2 750W PSUs.
No. DeathAndTaxes is right. He took the 166W from my computation: https://bitcointalk.org/index.php?topic=144648.msg1537338#msg1537338Basically you fail to take into account the PSU efficiency. For example a 1000W PSU could power 6 modules (100% load), but if it is 82% efficient it would pull 1220W from the wall. All this without exceeding its ratings. |
|
|
|
A picture is worth a thousand words   It would be great if you gave the source of this image. |
|
|
|
From another thread... the units are 620W max, I mean the power supply shipped it's a 650 Antec, so buy UPS accordingly.
Based on the information provided above, the Avalon units are not equipped to handle a 4th hashing module out-of-the-box. The Avalon PSU would be running only 3.5% over its rated capacity with 4 modules. Not great, but it should be able to run 24/7 at least 1 or 2 years at this workload, assuming it is a normal quality PSU rated to run at its full power. The 1- to 2-year lifetime can be derived from the typical 10k hour lifetime of a PSU's electronic components such as capacitors. nghzang reported a 82% efficiency of the PSU and a 5-20 Watt power consumption for the fans. Therefore: (620 (Watt at the wall) * .82 (efficiency) - ~10 (Watt for the fans)) / 3 (modules) = 166 Watt at 12V per module A 4-module Avalon would consume: 166 (Watt per module) * 4 (modules) + 10 (Watt for the fans) = 674 Watt at 12V which is 3.5% higher than 650W 674 Watt at 12V (output) translates, given a .82 efficiency, to 822 Watt at the wall (input), but do not forget that a PSU wattage rating refers to its output power, not input power. |
|
|
|
I did not know about that amazings eletricity rates, my first tough was: "if it's real, google must have a datacenter there". And yes, the have one. Amazing. How do you have your gpu mining farm hosted here, may I ask? I hadn't found any private datacenter with a fast search on datacentermap.
Actually Google does not have one in Douglas County, WA. You are probably confusing with the one they have in Douglas County, GA. However, Douglas County, WA is the location of multiple other datacenters: Sabey's Intergate.columbia (where colo is possible), VMware, and T-Mobile. And Google does have a datacenter 130 miles south in The Dalles, where power is still relatively inexpensive at $0.05/kWh from the local provider Northern Wasco PUD. There is more than just the price of electricity that determines where to host a datacenter: local job market, distance to large cities, etc. I am sure Google had valid reasons for deciding against hosting in Douglas County. Personally, I just rented office space for my GPU farm. Colocation always adds too much overhead costs over the raw cost of electricity. Some colos may say they "sell electricity at cost", but they always force you to pay additional fixed monthly recurring charges (eg. a fixed price per rack cabinet), so at the bottom line, you are paying a lot more than "at-cost" electricity. |
|
|
|
I might be wrong but 2c$/kWh is way too cheap to be real.
It is real. That is the rate in Douglas County in Washington State. I moved my GPU farm over there back in 2011. My last bill was $497 for 21667 kWh (effective rate = $0.023/kWh). You can check their rates here: http://www.douglaspud.org/Service/2013RatesJanuary12013.aspx The reason the rates are so low is because electricity in Washington State is mostly under the control of Public Utility Districts, which are (generally) non-profits and sell electricity at-cost. And most of the power stations are low-key hydro installations, so quite inexpensive. |
|
|
|
There is a simple way ASICminer could gain a 51% position, but demonstrate to the network that they are not attacking. All the have to do is spread their hash power out over several of the larger pools. If they did this then they would need the cooperation of a majority of the pool proprietors to attack the network (already a possibility and not perceived to be a realistic issue). If it were me I and I had 51% of the network then I would put 1/3rd of that power each on deepbit, slush, and btcguild. This way none of those operators would go over 50 (they would each be closer to 33%), and I would not actually be over 50 in practice because I would only be computing the hashes they gave me.
You would be over 50% in practice, even if you compute the hashes they give you. Anyway, an even simpler way to demonstrate a non-malicious majority mining position is to mine solo, and to operate normally, ie. (1) do not cause excessive chain reorgs (monitored by the bitcoin client and by, for example, http://blockexplorer.com/q/reorglog) and (2) do not mess around with transactions (not dropping them, etc). |
|
|
|
Where oh where did cgminer come from, originally? hmmm. ... and the only numerical proof you have of the fact that you are mining is the code I wrote. If you want to inflate your ego with your aggrandizement about the origins of cgminer [...] ...says the guy who is constantly arguing with Luke-jr about the origins of the FPGA mining code in bfgminer/cgminer, who forked who, who rewrote what, blah blah blah A case of the pot calling the kettle black. |
|
|
|
I am selling a $25 Hawken Game Code that newegg gave me as a gift (see picture of the card below). The code is reedemable at http://playhawken.comPrice = 0.5 BTC or best offer.  |
|
|
|
You are right there are many abandoned sites from back them. But that usually speaks for itself.
The keyword is "usually". IOW "not always". Chronicle Tech is 17 years old and has completed hundreds of tapeouts. But what do I know. Perhaps I should discard this evidence, and base my opinion on a marketing tool such as a website. We all know that companies with rich websites all sell amazing products and services, while those with no or crappy websites do not. Right? |
|
|
|
Set up eloipool with the difficulty set to some high value (eg. 64 with ShareTarget = 0x0000000003ffffffffffffffffffffffffffffffffffffffffffffffffffffff). Mine with getwork, not getblocktemplate.
Dynamic targetting was used, which should in theory achieve same... maybe high+static is better. Yes, maybe. I have never tested dynamic targeting. |
|
|
|
Anyway, talking about who shipped or not is irrelevant to prove my point: Avalon is first to mine. This would not have been possible had they chosen a more complex 90nm or 65nm design, therefore 110nm was the right choice in order to beat BFL on time having devices mining as we speak (prototype or not, shipped or not).
Nemesis: watch your language - this is your 2nd warning. Next time you will find yourself in my ignore list.
And i said your point is moot if those AVALONs arent not in their customers hands. Having prototypes with unknown bugs means.... they beat BFL? Its also amusing that you said " shipped or not" ... Sure, ignore me. Or grown up and admit your point is nonsense. (Finally you write a post without insults. Stay polite and I will not ignore you. It is that simple.) I said "shipped or not" because this is an irrelevant semantic disagreement: I consider the 2 units in the wild mean Avalon started shipping. You disagree. Big deal. I also believe them when they say that 53 units are in transit (ie. "shipped"). You will be convinced very soon, probably after Chinese New Year. Just wait and see. |
|
|
|
|
Show Posts
