Cryptocoin Mining?

[Ghiddy]

I wonder what will happen when difficulty goes way up and every one with less than 1 or 2 GHps will call it quits and drops out...

A couple of things:

1. Electricity & cooling costs will overtake the revenue generated. Those with cheaper costs will continue to mine. They will also be able to buy up cheap hardware from those with more expensive electricity, so the drawdown in profitability will be gradual. (hash rate lost by some will be gained by others). Overall it will take place pretty quickly.
2. Oscillation for a while, while people figure out if it is profitable or not to mine. Drop-off will decrease network hash rate, which will decrease difficulty, which will draw some back in. Some dweebs will even try to time these cycles, but only those who already have hardware and are borderline between profit and loss will do that.
3. Every time new hardware releases that improves on hashrate economics people will storm that and stuff that was borderline profitable will devalue and people will sell it off.

So I don't know. I wouldn't be surprised if there wasn't a dump of cheap 5830's, and it happened more gradually.

What's more surprising to me is that there is a market for bitcoins. I saw a post online today in the bitcoin marketplace where a guy wanted to buy a "very large amount" of bitcoins using silver and cash, and he wanted to meet in person to make the exchange. Shady. Pretty sure he wanted to buy illegal stuff with bitcoins, and wanted to keep his transfer private.

 

[notty22]

Here is a good explanation of how the 'bitcoin' system allows for 'mining'. Basically the more people mining spread out whatever possible profit that can be had. Verifying the bitcoin system, its checks and balances.
http://arstechnica.com/tech-policy/...nside-the-encrypted-peer-to-peer-currency.ars

Mining and make-work

As a digital currency, Bitcoin suffers from a tangibility problem. Unlike other currencies traded online, you can't go to a bank and withdraw physical coins, so what are they? More importantly, where do they come from? Coins are essentially agreements between all the Bitcoin nodes to accept a particular coin as currency. They are created gradually according to a precise protocol in order to reward those who contribute and maintain the network, control the rate of creation of the currency, and maintain the integrity of the transaction list.
In a process known as mining, individual Bitcoin users attempt to generate new coins by checking the integrity of the transactions list. They confirm the previous transactions and attempt to solve a difficult proof-of-work problem which involves exhaustively trying different solutions. There are a very large number of such potential solutions, so the likelihood of finding the solution depends how many other people are looking for it and how much computing power you devote to the problem. The first client to find the solution announces its good fortune to the whole network and earns a little reward for itself in the form of some shiny new Bitcoins.
By finding the newest solution to the proof-of-work problem, a Bitcoin client confirms the history of previous transactions and moved the transaction register forward, allowing new debits and credits to form part of the next block that can be mined to earn more coins. Future coins can't be mined in advance, because the computation to find the new block (and hence create new Bitcoins) relies on the the chain of previous blocks and the history of transactions since the most recent block.
The number of new coins generated per block gradually decreases over time. It started out at 50 BTC, but will dwindle to zero sometime in future when all 21 million coins have been generated. Fortunately, coins can be divided down to the eighth decimal place, which may prove increasingly useful if their value grows.

 

[pcm81]

I see that people are making money with this, but does anybody not see this is a giant pyramid scheme?

Yes!!! is the Ultimate Pyramid scheme, so get out now while you still can and send me all of your bitcoins for proper cryptographically secure disposition.

 

[YoungGun21]

BTW deepbit is too big atm, please choose another solid pool if you are just starting out. When a pool server controls too large a percent of the processing power it gains the ability to cheat (not that I'm accusing deepbit of anything nefarious, just the ability).

Agreed here. I'm running Slush's pool, but I've heard great things about BTCGuild.

 

[Muyoso]

I have a question. Why would you use multiple workers with deepbit instead of just using the same login on multiple computers? Is it so that you can more closely monitor what is going on, or is there some other reason?

 

[OVerLoRDI]

Agreed here. I'm running Slush's pool, but I've heard great things about BTCGuild.

Slush had too many issues for me. The opportunity cost of his pool being flaky ended up being too great.

 

[Muyoso]

Why didn't anyone say that the 5830 can overclock like crazy? I have had it running at 1030mhz up from the stock 800mhz for over an hour now. I had it running at 1050 for 15 minutes without a problem. I literally just got it and didn't want to mess around an entire day trying to find the max, but its obviously a GIGANTIC overclocker. I could probably hit 1100mhz. Of course I am not running graphical benches on these clocks, but for mining it seems to be doing fine.

 

[YoungGun21]

I have a question. Why would you use multiple workers with deepbit instead of just using the same login on multiple computers? Is it so that you can more closely monitor what is going on, or is there some other reason?

You have one main account, but each card actually counts as a worker. This is how it works with Slush anyway. So we have one main account, but three workers (2 5830s, 1 5750) labeled "accountname.rig1" "accountname.rig2" "accountname.rig3".

Also, Slush has had no problems for us. Website has been down before, but mining has only even been down for maybe 2 hours. Deepbit on the other hand I've heard has been getting downtime like crazy.

 

[YoungGun21]

Why didn't anyone say that the 5830 can overclock like crazy? I have had it running at 1030mhz up from the stock 800mhz for over an hour now. I had it running at 1050 for 15 minutes without a problem. I literally just got it and didn't want to mess around an entire day trying to find the max, but its obviously a GIGANTIC overclocker. I could probably hit 1100mhz. Of course I am not running graphical benches on these clocks, but for mining it seems to be doing fine.

How many Mhash/s is that OC giving you?

 

[NoStateofMind]

How many Mhash/s is that OC giving you?

Yes this please.

 

[Chiropteran]

What pool are you using?
I'm trying BTCGuild which seems even slower.

slush's, seems to be working well for me so I don't have much reason to switch. I did try it an one of my miners, but BTC Guild went down or something the other day and I lost some mining time, so I've fully switched back to slush's.

 

[dust]

How many Mhash/s is that OC giving you?

I don't know about his, but in my case, I simply moved the slider from 800Mhz to 840 and got an increment of 28-30 Mhash/s, from 298 to 328ish.

Is there any version of the Afterburner that will allow me to downclock the memory to the 2D level? I'm using 1.5.1 ATM and the lowest I can do is 940 as opposed to 1250, the standard clock(6950).

 

[Muyoso]

How many Mhash/s is that OC giving you?

I was using guiminer and getting only like 310Mhash/s from that. Also since I don't know how stable it is in the long term I backed off a little to 1000mhz and then used Phoenix instead of guiminer as gives an extra 10Mhash/s for me. At 305 Mhash/s now at 1000mhz using phoenix.

Also, this is from a 5830. Tomorrow when I have more time I will find out exactly what is stable. Sure hope 1000mhz is stable and it doesn't crash overnight.

 

[v8envy]

Hmm. Only at 249.9 Mhash/sec on one card, 244.5 on the second (identical cards, same machine, tested independently and together). Maybe I'll give this a shot under Windows, 249ish appears average for stock clocks on a 5830. I'd be very excited about 300.

I'm a bit leery of cranking the clocks. Sure, you get 20% more Mhash/sec. But is it worth it to increase power consumption and heat by say 40%?

 

[Muyoso]

Hmm. Only at 249.9 Mhash/sec on one card, 244.5 on the second (identical cards, same machine, tested independently and together). Maybe I'll give this a shot under Windows, 249ish appears average for stock clocks on a 5830. I'd be very excited about 300.

I'm a bit leery of cranking the clocks. Sure, you get 20% more Mhash/sec. But is it worth it to increase power consumption and heat by say 40%?

Absolutely. How expensive is your electricity really? Nothing compared to a 30 dollar BTC, not that it will stay there forever. Also, I set the fans to a fixed speed of 75-80%.

 

[dakU7]

You have one main account, but each card actually counts as a worker. This is how it works with Slush anyway. So we have one main account, but three workers (2 5830s, 1 5750) labeled "accountname.rig1" "accountname.rig2" "accountname.rig3".

Also, Slush has had no problems for us. Website has been down before, but mining has only even been down for maybe 2 hours. Deepbit on the other hand I've heard has been getting downtime like crazy.

Yeah but you can log in with the same worker on multiple GPUs on slush's, so what's the difference between that and creating dedicated workers for each GPU?

 

[Muyoso]

Yeah but you can log in with the same worker on multiple GPUs on slush's, so what's the difference between that and creating dedicated workers for each GPU?

This is exactly my question.

 

[v8envy]

It's interesting to see the various workers compete. That's why I keep them separate.

Interesting discovery. Different bios versions on 5830s behave differently. One (the 244 Mhash) 5830 can have memory downclocked to 300 mhz. The 249 one can only go down to 500. The second runs a bit hotter. The first one runs best with SDK 2.4 and a worker size of 256, the second runs best with SDK 2.1 and worker size of 128. Both are now at 260.5 and 261.7, I gave in and bumped the clock up 5% to 840 on both.

 

[Vesku]

v8envy, what temps do you get at load on your 5830s?

 

[v8envy]

At the moment 66.5C on the hotter but 1 Mhash/sec slower one at 81% fan speed, 64.5C with 60% fan speed on the faster one. At default fan profile they hit nearly 90C. At a bit over a mile of altitude there's 17.5% fewer air molecules to cool with, so heatsinks designed for sea level don't do so well.

 

[Vesku]

Both the same model? My Sapphire 5830 matches pretty closely with the first card you mentioned about (70F ambient). Compared it to reviews of the same model and it seems I lost the gpu lottery and got a hot running unit.

 

[v8envy]

While I bought them at the same time the box for one says XFX HD-583X-ZAFC version 1.2 and the other version 1.3! At the very least that translates to a different bios, max and min core and memory clocks, and a 5% performance difference at stock settings with SDK 2.1 and work unit size of 128. I didn't keep track of which is which unfortunately.

 

[Muyoso]

Mine is running at around 70c overclocked to 1000mhz at 75% fan speed.

 

[Vesku]

Mine is running at around 70c overclocked to 1000mhz at 75% fan speed.

I think I'd smell burnt pcb if I tried for those settings, that one of those extra nice cooling models?

 

[YoungGun21]

Setting the fan to 80% and overclocking core to 875, memory to 700 we get 264MHash/s with our 5830s. They run about 58C.