new miner - is it worth it?

[pandemic]

I'm still looking into mining to see if it's really worth the costs.

I've been messing with kradminer.com and bitcoinplus.com. I've found that the GPU dominates the CPU for this job.

Here's my rig:
2.66ghz core 2 duo overclocked to 3.2ghz
Radeon HD 4870 512mb
4gb ram
150gb 10k rpm hdd

I know disk space, ram, and CPU don't mean much.

I do have a second PCI-e slot and I was looking around for another Radeon HD 4870 to set up via crossfire.


I'm just not sure if it's worth mining. I only started playing around with and reading about bitcoin the other day. I'm still not exactly sure how they even work or why they are worth anything.

I read about someone who averaged it cost about $0.65 per BTC generated. I also heard that the going rate right now is $7/BTC. That's a pretty large profit.

I still have lots of reading to do. Read somewhere that you can strike gold and get 50BTC somehow but I'm not even sure how that works, lol.

[1713947651]

1713947651

[1713947651]

1713947651

[1713947651]

1713947651

[1713947651]

1713947651

[1713947651]

1713947651

[kjj]

You are unlikely to recover the cost of even a single 4870 even if you run a pair of them until they crumble into dust.

Opinions are mixed on this.  I'm pretty sure the profit ship sailed a couple of weeks ago.  Others think you can still do it.

If you don't mind gambling, or if you want the second card for better gaming, go for it.  If you need that money for something more important than entertainment, I wouldn't.

[pandemic]

Yeah, that's the thing I have to look into. I only pay about $0.12 CAD per KWH but I'm still not sure if it'd justify building a dedicated rig. I have a 4u server at home in the US, 4 xeon procs with HT. I'm thinking about setting that up next time I'm home. I also have a HD 2XXX radeon that I'll put into another machine and run that.

I was thinking about building a budget miner. Budget CPU with 2,3 or 4 GPU's. I'd have to see what the cost would be and then go from there.

[Gameover]

not worth it IMO.  the biggest gotcha that people don't realize is the difficulty rating, basically in less than 2 days bitcoins will be almost half as profitable to mine, and 10 days after that half again, this keeps continuing given current trends.  this is based on the assumption bitcoins sell for the same amount, if you believe that bitcoins value will increase (a big assumption) then building a rig will be worth it, of course you could just buy bitcoins directly and most likely profit more.  its a gamble either way.

[pandemic]

So when was it worth it? I see it's meant to run out with current trends in 2140 after 21 million coins are mined. That's still a lot of coins to be found.

[Jaime Frontero]

So when was it worth it? I see it's meant to run out with current trends in 2140 after 21 million coins are mined. That's still a lot of coins to be found.

it's still worth it, and always will be.  you're right - "That's still a lot of coins to be found."

the thing is, you need to calculate your investment pretty carefully.  there's quite a few threads on this forum about that, with many formulae for doing so.

but, in the final analysis, it all depends on your own, personal assessment of what the future value/exchange rate of Bitcoin will be.

if you think - as vladimir (one of our better-known Hero Members) does - that Bitcoin will exchange for USD at 100 to 1 by the end of the year and 1,000 to 1 by the end of 2012, then you'll be willing to mine at a loss.

at this point though, the conventional wisdom is that for a newcomer to buy Bitcoin outright for cash is more profitable.  that's probably true.

but nowhere near as much fun...

[darbsllim]

A lot of people have a mental block about exchanging money for bitcoins rather than exchanging money for hardware, which can find bitcoins.

If bitcoin fails, at least you still have the hardware to sell to gamers.

If bitcoin takes off into the $100/btc level relatively quickly, you would have been better to buy.

[Gameover]

So when was it worth it? I see it's meant to run out with current trends in 2140 after 21 million coins are mined. That's still a lot of coins to be found.

But the coins are issued reverse exponentially, so the first year will release a huge portion of the coins and the last year hardly any, the people mining at last 1% will take millions of times more computation power to mine than the first 1%.

it's still worth it, and always will be.  you're right - "That's still a lot of coins to be found."

Not true at all, bitcoins could go to $0 today, it then wouldn't be worth it.  More likely they will increase slightly and the exponentially increasing difficulty (its actually increasing faster than exponentially right now!) will merge with price and raw power cost to mine them in about 45 days, at which point it won't even be cost effective to mine on a ATI GPU if you pay for your own power and all your hardware is already paid for.

Tomorrow we will be at the tipping point where buying a dedicated rig is no longer a return but a loss, 45 days from then even people who already own their hardware won't make money if they pay for power.  This is assuming big increases in bitcoins value of $25-$50.  If it increases like others hope to $100 - $1000 then you will still make money.

Like all commodity rushes the first that got in make the most while the rest fight for 5% of nothing and most lose money.  The difference with bitcoin is its virtual, so even the people holding massive amounts of bitcoins can be burned in an instant.

[Chucksta]

hmm, I'd only start mining now if I already had the equipment. No way I'd invest in new hardware, unless it was very very very cheap.

If you've got zero overheads and equipment at hand, then go for it. Each card can still make a few dollars per day, but there's a high chance that will not last for much longer.

I reckon the best thing to do here is to monitor the price and just buy low and sell high. And only with money you don't need.

[rograz]

I just bought another Sapphire 5850 extreme myself, but that was only cause I got it for 115 euro and I had a slot left in one of my rigs. Building a completely new rig at this point I would say is risky in the best of cases(finding sub 140$ 5850s or something) and just silly when it comes to building the 6990 rigs with high initial investments that some ppl seem so fond of. At this point it's all about maximizing $/Mhash and Mhash/W to stay in the game as long as possible unless we see another crazy rally that could change everything again, but we might just as well see a drop down to the 2-3$ range or even lower :<

Then again it's not the first time mining has been proclaimed dead and unprofitable and still we are here with more hashing power than ever :p

[bcpokey]

So when was it worth it? I see it's meant to run out with current trends in 2140 after 21 million coins are mined. That's still a lot of coins to be found.

But the coins are issued reverse exponentially, so the first year will release a huge portion of the coins and the last year hardly any, the people mining at last 1% will take millions of times more computation power to mine than the first 1%.

it's still worth it, and always will be.  you're right - "That's still a lot of coins to be found."

Not true at all, bitcoins could go to $0 today, it then wouldn't be worth it.  More likely they will increase slightly and the exponentially increasing difficulty (its actually increasing faster than exponentially right now!) will merge with price and raw power cost to mine them in about 45 days, at which point it won't even be cost effective to mine on a ATI GPU if you pay for your own power and all your hardware is already paid for.

Tomorrow we will be at the tipping point where buying a dedicated rig is no longer a return but a loss, 45 days from then even people who already own their hardware won't make money if they pay for power.  This is assuming big increases in bitcoins value of $25-$50.  If it increases like others hope to $100 - $1000 then you will still make money.

Like all commodity rushes the first that got in make the most while the rest fight for 5% of nothing and most lose money.  The difference with bitcoin is its virtual, so even the people holding massive amounts of bitcoins can be burned in an instant.

Misuse of the term exponential aside, I'm not sure where you are getting your numbers from. I have a limited data set to work from but the previous difficulty changes have been: 79000 -> 101500 -> 157000 -> 244000 -> 435000 (estimated). Somewhere in there has been a huge flood of media attention, a large computer enthusiast forum hopping on, etc. and it's held to about 50-60% increase for difficulty/hashrate. Even assuming that this trend continues at 60% increase in difficulty every 10 days in 40 days it will yield a difficulty of about 2850816, and a hash rate of 23THash/sec (which I'd find a surprise). Even a meager 1 GPU rig of 400MHash/sec will still generate an estimated $25/wk @ $25/coin, far far far far above the cost of electricity. Even at the next 60% difficulty increase to 4561305 this same machine would generate roughly $16/wk.
Now if rates stayed stagnant @ $8.5/coin while difficulty increased > 10fold then yes, you have a point that existing mining machinery would no longer be profitable, but if that were the case I can hardly imagine such a continued pressure to add computing power to the pool.

I'm not saying that someone is still ok buying a dedicated rig, it's become much more of a PITA and profits are hard to be found, just saying that you are making bold statements that I'm not sure the source of.

For those wondering about whether or not to build a mining rig at this moment, I'd advise extreme caution. All number crunching aside there is currently a large collection of people with $ signs in their eyes overbuying that you will be competing with entering the market, while we don't necessarily see the same headlong rush of investors into the bitcoin market. People with a real need to sell (to recoup costs) without an equivalent need to buy doesn't forecast well for offsetting the diminished returns that are already in the pipeline. This is why you will see so many people promoting the investment in bitcoins rather than mining hardware, both because it will be more profitable for them personally, as well as that it is the only way to keep the system running.

[Jaime Frontero]

i agree generally with the posts above - that within 45 days or so it will be unprofitable (at current BTC exchange rates, if they matter to you) to mine unless your equipment is paid for and/or you have free electricity.

my equipment is paid off, so i'll be mining at least until the cost of electricity exceeds the Bitcoin i mine.  but who knows where the exchange rate will be in 45 days?  not i - but uppish, i would think.  it could easily take 90 days before electricity costs overrun a more valuable bitcoin, at bitcoin's current rate of growth.

that said, i'll note that - historically - the primary driver of hardware advances in the PC business has always been gaming.  bitcoin will likely change that, if it takes off into the mainstream in a big way.  like if google payments takes note, and offers some sort of bitcoin exchanger.

in any case, the thing i'm watching right now is the FPGA stuff.  i'm pretty sure that's where the future of individual and pooled mining lies.

and 45 - 90 days is quite a while in geektime.

[Gameover]

I just posted a more in depth analysis:

http://forum.bitcoin.org/index.php?topic=10015.0

[cdb000]

Probably not.

I recently considered expanding my mining cluster and decided it would be a mistake.  I am still uncertain about whether the expansion from 9 5870s to 12 a small number of weeks ago is ever going to pay for itself. Maybe it will if I manage to sell the used cards on Biddingpond or eBay when I no longer want to mine... but there will probably be a lot of used 5870s on eBay/Biddingpond when a future difficulty rise makes 5870 mining unprofitable.

If you think a better GPU is worth buying for improved performance playing games, then buy it. Mine some coins and they could help pay for the GPU, but I think it is very unlikely that a GPU bought today is going to pay for itself through mining.

Difficulty is currently projected to rise by somewhere over 75% in the next few hours. A 5870 (if you can find one - they are getting scarce) is going to be mining less than 1 Bitcoin per day.

For those of us who have 5870s, already paid for, there is still a small amount of profit to be gained after power is paid for as long as the value of Bitcoins stays in the current $8 - $9 region.

Of course, you might take the view that the value of the Bitcoin will rise (it does seem to be forever rising) - but in this case your best course of action might be to simply buy some Bitcoins.

[cbeast]

I'm not sure if this has been discussed, but right now I see this as a game similar to an mmo. I see 'insiders' that believe in the concept and are willing to speculate are mining and investing to support the miners. There is some outside interest and trading volume is slowly creeping up. At some point folks like Lucas and Spielberg may find it profitable to turn their rendering farms into bitmining enterprises, but that day is not yet. When they do, maybe the trading volume will accelerate.

OTOH, if it doesn't, is it possible to turn these bit mining operations into animation rendering cooperatives to work with the entertainment industry?

[fasti]

If you are about to buy new graphics card for gaming upgrade, do it fast :) You can get at least some money back from mining.

I would strongly suggest not to use 2x cards in one mobo, it builds up so much heat for the inner card that it's insane...

[pandemic]

I think I'll just get another 4870 and just mine with that. I need to start looking into pool mining. Soooo much reading involved with this, lol.

Anyone think a quad xeon server would be good at mining? I'm pretty sure my 2x4870's will blow it out of the water. Just wanted to see if it could be put to use.

cdb000, how much for two 5870's?

Fasti, thanks for the heads up. I'm still thinking about putting them both in. I've got an Antec 900 case and it has quite a number of fans. When all else fails, I can just take the side off

[allinvain]

A lot of people have a mental block about exchanging money for bitcoins rather than exchanging money for hardware, which can find bitcoins.

If bitcoin fails, at least you still have the hardware to sell to gamers.

If bitcoin takes off into the $100/btc level relatively quickly, you would have been better to buy.

True but you know there is a problem with this idea. Most games are going to want to get the latest generation hardware. Many miners (including me) bought cheaper used last gen cards. We're all going to have a much harder time selling them. Plus all this supply being dumped on the market at once is going to make competition for clients really tough. So we'd have to sell the cards at a much lower price than we could get for them if we sold them right now. If a lot of people stop mining and decide to sell their hardware that will create a tremendous supply of video cards which I'm not sure will be met with enough demand from buyers.

I look at it this way, if you payed for your cards already then anything from here on is all gravy. If you feel like taking a gamble sure go ahead and buy that 6990 and mine away. I said 6990 because you're going to probably have a much easier time selling that than a 5870 or 5830 etc.

[allinvain]

I think I'll just get another 4870 and just mine with that. I need to start looking into pool mining. Soooo much reading involved with this, lol.

Anyone think a quad xeon server would be good at mining? I'm pretty sure my 2x4870's will blow it out of the water. Just wanted to see if it could be put to use.

cdb000, how much for two 5870's?

Fasti, thanks for the heads up. I'm still thinking about putting them both in. I've got an Antec 900 case and it has quite a number of fans. When all else fails, I can just take the side off

Don't bother with CPUs - even Xeons. 4870s are okay but pretty power hungry so if you can get your hands on a 6000 series or even 5000 series card (or cards) then that's what I'd recommend. In fact I'd recommend the 6000 series simply because you're going to have a much easier time selling them after you got bored of mining - they have better tesselation performance in gaming with DX 11 games.

5870s go for $200 USD and up. I found a few for 150 a piece but those are much harder to find now. Basically hunt around.

[pandemic]

Huh, so I'd be better off getting a better card rather than just putting two 4870's in it? That wasn't what I was expecting. I'm not sure what the 512mb 4870 uses, power wise.

[um0rion]

I found that the best way to go is with low power amd builds. Get a cheap (CHEAP) processor, a motherboard (also cheap) that can handle 1-2 powerful cards, and calculate what its power requirements are carefully for the power supply. Buy as little ram for it as you can, same with the hard drive. Use an old dvd player laying around as an optical drive, or again, but a cheap one. the hard drive Im still looking into, as some people can get mining working from just a flash drive.

The 5850 currently is a massively powerful card for mining, 270 mhash/s stock, and Ive got one running overclocked pushing 392 mhash/s. My main miner cost $1000 to build, and puts out 1.2ghash/s. There's a potential to put out as much as 1.6 ghash/s with this one, but Id have to spend $500 on cooling.

[pandemic]

I found that the best way to go is with low power amd builds. Get a cheap (CHEAP) processor, a motherboard (also cheap) that can handle 1-2 powerful cards, and calculate what its power requirements are carefully for the power supply. Buy as little ram for it as you can, same with the hard drive. Use an old dvd player laying around as an optical drive, or again, but a cheap one. the hard drive Im still looking into, as some people can get mining working from just a flash drive.

The 5850 currently is a massively powerful card for mining, 270 mhash/s stock, and Ive got one running overclocked pushing 392 mhash/s. My main miner cost $1000 to build, and puts out 1.2ghash/s. There's a potential to put out as much as 1.6 ghash/s with this one, but Id have to spend $500 on cooling.

What's your build?