 Show Posts
|
|
Pages: « 1 [2] 3 »
|
Sorry guys, small delay. It'll be.. two more weeks.
</sarcasm>
Months, not weeks...  Eligius != BFL.  |
|
|
|
But I do think it's fair question to wonder why anyone would spend tens of thousands of their own money, and 12 hour days to support something for free.
I just want to briefly address this comment. As for money, the data center we are hosted at I've worked with for years. They're very helpful. Most of the hardware we have is donated, however I have been able to score some upgrades and additional items from donations and such and by sweet talking the data center owner. There are actually a few techs there who have their own mining hardware mining on Eligius from inside the same data center. (How's that for low latency?) So, we have a good deal of support from various aspects of the community, and it is much appreciated. When I contacted the data center this afternoon and was going over the issue they were all for getting it resolved and went to work seeing what they could come up with to help without breaking the bank. Turns out we were able to get a very nice upgrade to the main database server for pennies on the dollar. I put in my personal time and effort because I believe Eligius is successful, I believe it is the best mining pool, and I believe a legitimate and transparent no-fee pool needs to exist for miners to use to lower their variance without giving up huge amounts of their earnings to do so in pool fees. I personally want such a service to exist, and I use Eligius myself with my own hardware. Hopefully this helps. Thank you everyone for your support. -wk Thanks for the update. I think this is a good point to address. A lot of folks here are operating strictly on numbers, and if they can't quantify exactly what the profit margin will be, then they don't understand the motivation. Sometimes the technology, or the community, have their own rewards for participation. And people forget about donations helping to offset costs as well. |
|
|
|
P.S. - My cable connection seems fine now that the OS install (the bandwidth-need part of this process) is done... figures.
So it usually goes, eh? Seriously. I've spent many years as a sysadmin. I know how it goes... so my sympathies.  |
|
|
|
I was talking with a corporate farm owner that had invested 500k into 1st gen mining equipment which only did double digit Gh/s, now they're stuck and can't afford to get the single digit Th/s due in 2014.
This makes absolutely no sense, unless the "500k in 1st gen mining equipment" happens to have been block erupters purchased in the past 2 months @ $100/ea. 500k in Avalon batch 3s or clones, or BFL gear would leave you pretty broke right now. There's always eBay... |
|
|
|
So in simple terms, assuming I understand this correctly (which I admit may not be the case... ) is that the node with 1 GHs versus the node with 100 GHs is like a guy buying one lottery ticket versus a guy buying 100 lottery tickets? ...per second/minute/whatever. yes. It wasn't really a complaint, more of an observation, but your (and others) explanations has helped me in somewhat solidifying my understanding of the whole process! Yep, it can be very confusing at first. But yes, that's the gist. Faster hashing = more lottery tickets. |
|
|
|
That's why. The first two numbers, for Icarus-based devices, are essentially estimates based on an internal # that is different for each device, e.g. Erupter is ABC, Antminer is XYZ, etc. The third number is your effective hashrate based on actual work you are doing.
When you start to override the Icarus timings using arguments on the command line or config file it throws off those first two numbers, but not the real work done.
OK, that's good to know. So I either need to ignore the other averages, or ditch the custom settings. I'm not sure how those settings are affecting performance anyway, so I might remove them. It looks like overall performance is better with the defaults. Thanks guys! |
|
|
|
Actually yes, I copied someone's settings from a couple pages back in this thread: "scan" : [
"erupter:all",
"antminer:all",
"all"
],
"set-device" : [
"antminer:clock=x0981"
],
"icarus-options" : "115200:2:2",
"icarus-timing" : "2.5=90",
"klondike-options" : "362:65"
} That's why. The first two numbers, for Icarus-based devices, are essentially estimates based on an internal # that is different for each device, e.g. Erupter is ABC, Antminer is XYZ, etc. The third number is your effective hashrate based on actual work you are doing. When you start to override the Icarus timings using arguments on the command line or config file it throws off those first two numbers, but not the real work done. OK, that's good to know. So I either need to ignore the other averages, or ditch the custom settings. I'm not sure how those settings are affecting performance anyway, so I might remove them. |
|
|
|
What do the three speeds for each device mean, relative to each other? I just updated to 3.10.0, and plugged in the antminers that arrived today, and I see the third speed reported as 2GH, but the other numbers are half that. Is this normal?
What arguments are you using to launch bfgminer? Are you specifying manual Icarus timings? Actually yes, I copied someone's settings from a couple pages back in this thread: "scan" : [
"erupter:all",
"antminer:all",
"all"
],
"set-device" : [
"antminer:clock=x0981"
],
"icarus-options" : "115200:2:2",
"icarus-timing" : "2.5=90",
"klondike-options" : "362:65"
} |
|
|
|
The info is in the readme.txt The totals line shows the following:
6/32 75.0C | 171.3/170.8/171.2Gh/s | A:729 R:8+0(.01%) HW:0/.81%
Each column is as follows:
The number of devices and processors currently mining
Hottest temperature reported by any processor
5 second exponentially decaying average hash rate
An all time average hash rate
An all time average hash rate based on actual nonces found, adjusted for pool
reject and stale rate
The number of Accepted shares
The number of Rejected shares and stale shares discarded (never submitted),
and the percentage these are of total found.
The number of HardWare errors, and percentage invalid of nonces returned Awesome thanks, I must have missed that. But that doesn't explain to me why there would be such a huge difference. For instance, I see a line like this: AMU 3: | 0.59/ 1.06/ 2.00Gh/s How does the "all time" average end up being about half of the one that is based on nonces found? |
|
|
|
What do the three speeds for each device mean, relative to each other? I just updated to 3.10.0, and plugged in the antminers that arrived today, and I see the third speed reported as 2GH, but the other numbers are half that. Is this normal?  |
|
|
|
My point is, with the puzzle difficulty getting higher and higher, wouldn't we need equipment that actually solves a single hash faster? Or is that not a factor? What matters is the GH/s display in your miner software. Nobody cares if that is achieved by massive parallelisation or massive increase in clock speed (which means solving a single hash faster). Over the long run, both will happen. Right, the time-to-hash for any single hash attempt doesn't matter that much. Everyone is in a race to find a nonce that solves the criteria for the difficulty of the current block, and everyone is attempting different nonces. If everyone counted from "0" and incremented from there, and there was a fixed integer that solved the "puzzle", then the fastest hasher would always win. But that's not how it works. The nonce that will solve the puzzle depends a lot on the header of the block you're solving, which is different for each node, plus I suspect each node starts at a random nonce position in an attempt to not duplicate work with other nodes. |
|
|
|
|
Also, I'd look to see what sort of time-frame the average is calculated over. For instance, on Eligius, the 3-hour average is much more accurate than the 128 second average, since the number of shares submitted is used to calculate the average, and in a very short time-sample, it's not granular enough to give you accurate results.
|
|
|
|
SHA is most profitable at those speeds.
currently unobtanium is the most profitable SHA coin to mine but that changes very quickly,
To work on BTC you would get around 0.084 per day
and Litecoin you would get around 0.184 per day.
you are about 25% more profitable in SHA than Scrypt with quoted speeds.
are you crazy?? 30GH/s 0.084btc / day?  wtf dude? try to add another "0"!!! 0.0084! Yeah, I was going to say maybe 0.01 at most. |
|
|
|
When people are putting that they've preordered "1" Prospero X-1, they mean 10, right? since 10 is the minimum order... folks who have ordered "1" will be paying $3k+, not $300+, correct?
No. The minimum order order requirement was a recent addition. It didn't exist for the first few rounds of the discount. You could order any volume direct from BAS originally. Also, there's no minimum order through MinerSource. |
|
|
|
I ordered a Jalapeno 7 GH/s in August, I received it a week ago, the unit was hashing at 8,5 GH/s at 50 Celsious.
I removed the stock paste and replaced it with Arctic Silver and I got rid of the surrounding aluminium.
The temp droped at 30-35 Celcious but the PSU died 4 days after.
I tried to find the same PSU but I couldn't find a 13V 6A so I bought a 12V 10A and I am waiting for it to see if the Jalapeno works.
The PSU is garbage, I 've read hundreds of posts for PSUs that died in the first week, it's a shame, 425$ down the drain.
Anecdotally, I've heard that anything in the 12V-13V ballpark will work with the Jallies, so you should be good to go with the new PSU. So technically the only thing down the drain is the cost of the second PSU, since the first shouldn't have failed on you like that. Well... that and many months of missed easier difficulties and hashing time. |
|
|
|
When I looked at eBay a week or so ago, some were selling for as much as $2,000, if you want fiat. |
|
|
|
ASICMINER mine themselves, and with some partners. They don't sell "retail". There is no fixed costs to "build rigs", it depends on many factors.
I beg to differ, there is a minimum cost you can manufacture an asic based miner for (in volume), and it's easy enough to calculate if you have commercial design experience or are willing to do a reasonable amount of research work on the net. Granted, some companies have more 'overheads' than others, but that's just greed or poor management. I'm talking about a well run company with experienced people and modern manufacturing methods. We're also seeing a very rapid development cycle on the ASICs that is quickly catching up with current top-end chip fabrication technology. That is, we'll soon be seeing hashing-speed advances slow down to the pace of Moore's Law, at best (short of just packing more and more cores into a box). Combine that with the exponential growth of the entire network's hashing speed we've been seeing lately, plus the difficulty increases that result from it, and it may reach a point where none of the hardware can be sold at a price that is sufficient to justify its manufacture. This can be staved off for a while if the value of BTC continues to climb, but we're facing a situation of diminishing returns. More people are competing to get BTC, and the minting payouts will continue to halve every few years. So I think the question is when, and not if, we'll see mining for the sake of profit become useless. Before that happens though, all hardware has to be bought with the expectation that its worth will diminish rapidly, and the electrical cost may soon overshadow its earning rate. |
|
|
|
|
You can add me:
Username: NecroBones
Order Date: Dec 31, 2013
Payment Date: Jan 02, 2014
Order Number: R6232351xx
Currency: Wire Transfer
Machine Type: Prospero X-1
Quantity: 2
Delivery Country: US
Vendor: minersource.net
|
|
|
|
After 6 months from my order I finally received my Jally, the speed is above 8 GH/s at 40 degree.
6 months? Holy hell... I'd say the difficulty has gone up just a tiny bit since then... |
|
|
|
Are there going to be like nextgen hardware that would be like 10x faster than the current mining hardware now or is this the best it's going to be? I can't decide whether to buy an ASIC block erupter cube or save for nextgen stuff if there is.
An ASIC block erupter cube is very old tech. I know its cheap but ... you get what you pay for. But no point waiting for "something much faster", you'll wait forever because hardware will only get faster and faster. Buy something now if you really want to mine. hey now i just got my block erupter cube today some of us arent as btc rich as you The BE cubes are relatively cost-effective, if your goal is < $1000 on eBay. Compared to the BFLs, small USBs, etc, that are out there. The pre-order gear that's slated for Q1/Q2 2014 from Black Arrow, Cointerra, and even the BFL Monarchs (among others) will be significantly more cost-effective. But they're not out yet, and who knows which ones will deliver on their promises the best. Compare a BE Cube at 38 GH/s (overclocked) for, let's say $600 for the sake of argument, that's $15.79/GH. Much of the unreleased hardware is promising cost rates under $4/GH. And they look even better when you also factor in the lower power consumption per GHash as well. |
|
|
|
|
