Cryptocoin Mining?

[Magic Carpet]

Only if the power outage corrupted your hard drive.

Nope, that is the whole point- it is a distributed non-centralized system. A repository of some sort would kill those advantages.

You can always back up your wallet. You don't even need to run the PC with the wallet- you can create it on a VM, suspend the VM, copy it to a flash drive, put it in a vault. Send BTC to it's wallet every week, and then 10 years later you get the flash drive out of the vault and run the VM and you see all the coins you earned in the meantime.

Thanks for clearing up the skies for me. So, I would need to back-up the wallet on a flash drive if I want to cache in on a different computer... say in i-Cafe?

So much fuss

 

[IGemini]

Hi all, hoping for a bit of help here. I tried to give bitcoin mining a whirl last night, but I ran into a snag. I loaded up the wallet app, got my wallet code, registered at bitclockers and got a worker name. Then I downloaded the SDK APP for my Radeon drivers, installed them, and downloaded and setup Phoenix Miner (recommended by BitClockers here: http://bitclockers.com/gettingstarted for GPU mining over GUI Miner).

I created the shortcut with the script to start Phoenix miner, but when I ran it, it gave me a "Fatal Error: Could not load OpenCL kernel!", or something similar.

I double checked that everything was as it supposed to be in the guides, but no joy.

Any suggestions regarding what I did wrong?

There's not enough here to tell what exactly the problem is, but I'm guessing you aren't using the right flags for phoenix. Your hardware, driver version, and what syntax you're using to run the miner would help. I'm assuming you're running Windows, and if you installed everything from the default AMD Catalyst package (from 11.6 onward, IIRC), it should already include the APP runtime, which is all you would need. Installing the developer SDK would be redundant. If you have a sandbox/firewall program, it might also be blocking the miner.

It might be easier to just use GUIminer's OpenCL poclbm anyway to get you started. Phoenix is pitched as "more efficient" by using command line but it's a moot point TBH. It's not like the GUI is straining the hardware significantly. I've used both but phoenix takes a lot more trial and error to get it running as efficiently. I'd only use phoenix if your machine was running Linux. GUIminer includes an option to run a phoenix miner anyway, so you can configure it at your leisure much more easily than command-line.

 

[slayernine]

Well I just realized what I forgot to backup before my stupid OCZ Vertex 60GB died, my bitcoin wallet.

Lucky for me it only had like 1 or 2 bit coins sitting in it as I cashed out most of what I had last time I was doing some mining.

 

[Zargon]

not legit at all, people have been fired for it before

 

[Dark Shroud]

Well I just realized what I forgot to backup before my stupid OCZ Vertex 60GB died, my bitcoin wallet.

Lucky for me it only had like 1 or 2 bit coins sitting in it as I cashed out most of what I had last time I was doing some mining.

This is why I send my coins from my mining pool straight into my Mt Gox account.

 

[suklee]

This is why I send my coins from my mining pool straight into my Mt Gox account.

So Mt Gox would be more like a 'traditional' paypal account where you log in and check your balance?

Do you also send it to your Mt Gox account when someone sends you BTC at your Bitcoin address in your sig?

 

[Dark Shroud]

So Mt Gox would be more like a 'traditional' paypal account where you log in and check your balance?

Sort of, unless I plan to buy something from a vendor that uses Bitcoins I just sell them on Mt Gox. Every Sunday night I transfer the balance to my bank via Dwolla.

Do you also send it to your Mt Gox account when someone sends you BTC at your Bitcoin address in your sig?

No, that goes to my Bitcoin wallet. For actual transactions I use my Bitcoin wallet. It's just cheaper for me to send Bitcoins from my mining pool account directly to Mt Gox is I'm going to sell my coins. Which is what I do with most of the coins I mine.

 

[suklee]

This is why I send my coins from my mining pool straight into my Mt Gox account.

Another question, does this depend on the mining pool providing this feature, or do you get a bitcoin address once you sign up for Mt Gox?

 

[tommo123]

i have some funds in mtgox, is there an easy/quick way yet to get it to the UK? without losing half of it on the way?

 

[Dark Shroud]

Another question, does this depend on the mining pool providing this feature, or do you get a bitcoin address once you sign up for Mt Gox?

Mt Gox has to give you an address to send coins to just to be able to transfer coins into Mt Gox.

 

[slayernine]

I've been using the Canadian Virtual Exchange for bitcoin buying/selling.

https://www.cavirtex.com/home

 

[MrTeal]

I'm noticed on the comparison of hardware page that a significant number of people over 300Mhash/s with my card (6870) have their memory clocked extremely low, like 300MHz. Is there a particular performance benefit to doing so, or is it just to reduce power consumption? I'm completely unable to get my memory clocks much below 900 without artifacts, so I'm assuming that most of the people with extremely low mem speeds aren't actually using the cards as a display adapter.

 

[Zargon]

reduce power and heat

 

[Chiropteran]

reduce power and heat

Somehow it increases mhash occasionally. I'm not sure of the technical reason, but I've seen myself sometimes changing only memory from default down to 300 mhz and mhash goes up.

Was doing some mining on my llano laptop and I've noticed something. I'll get about 28mhash with the APU (A6-3400m), but if I run the fritz chess CPU benchmark and push all 4 CPU cores to 100%, the mhash will plummet until the benchmark finishes. I guess when the CPU side is pushed hard the APU takes some resources from the GPU side. Does anyone else mine with a llano? I'm just curious if they all show the same result.

 

[ensign_lee]

I'm noticed on the comparison of hardware page that a significant number of people over 300Mhash/s with my card (6870) have their memory clocked extremely low, like 300MHz. Is there a particular performance benefit to doing so, or is it just to reduce power consumption? I'm completely unable to get my memory clocks much below 900 without artifacts, so I'm assuming that most of the people with extremely low mem speeds aren't actually using the cards as a display adapter.

It's just to reduce power consumption and to reduce temps on the card. Careful though. I downclocked my memory too much one time and I could not stop blue-screening anytime i tried to access MSI Afterburner.

Problem is...the only way to change my memory clock was to fix it was to...access MSI Afterburner...lol

 

[MrTeal]

Somehow it increases mhash occasionally. I'm not sure of the technical reason, but I've seen myself sometimes changing only memory from default down to 300 mhz and mhash goes up.

Was doing some mining on my llano laptop and I've noticed something. I'll get about 28mhash with the APU (A6-3400m), but if I run the fritz chess CPU benchmark and push all 4 CPU cores to 100%, the mhash will plummet until the benchmark finishes. I guess when the CPU side is pushed hard the APU takes some resources from the GPU side. Does anyone else mine with a llano? I'm just curious if they all show the same result.

Fritz is extreme memory intensive, isn't it? Since the APU uses the same DDR3 bus for both CPU and GPU, it would make sense that your performance would suffer even if hashing isn't particularly memory intensive.
If you use Fritz but force it to just 3 cores, do your hashing results drop?


Thanks for the replies. I kind of figured it was just for power reasons. I won't worry about pushing it too low then. I installed an Arctic Cooling Twin Turbo II on my card last night, and even running at a core of 1050MHz thermal imaging shows all the ram sinks staying basically ambient while hashing. I can't imagine the memory is using too much power.

 

[Vesku]

Some CPU time is needed to keep the bitcoin miner software fed, so if you are doing something CPU intensive you'd see a mhash/s drop.

 

[Smartazz]

Any idea whether one 2500k core should be enough for two mining 6870s?

 

[IGemini]

^^Plenty. The CPU time is overstated, I only see my processor used for bitcoin when I tell the card to start hashing, only for a couple seconds. It's barely enough to bring my CPU out of EIST idle.

 

[Vesku]

Should be plenty (although there is a CPU bug involving AMD drivers that will peg a core per GPU https://bitcointalk.org/index.php?PHPSESSID=37de443b4a06dad287a0a6c3000ed2b5&topic=56740.0 ), some miners have had Sempron 140s powering a 2-3 card machine. Just that with something like a Llano A4 if you run a CPU intensive app then I could easily believe a drop in Mhash/s.

 

[Chiropteran]

Some CPU time is needed to keep the bitcoin miner software fed, so if you are doing something CPU intensive you'd see a mhash/s drop.

Ah, that was it exactly. I repeated the test on a non-llano system and observed the same result.

 

[Dark Shroud]

A lot of us use Semprons that we core unlock into Anthlon x2s. They work fine for a few cards.

 

[Sunny129]

reduce power and heat

to elaborate on this response, here's why:

bitcoin mining requires very little VRAM bandwidth. i don't mine for bitcoins myself, so i'll give you an analogy that i'm more familiar with. consider both the Milkyway@Home and Collatz Conjecture DC projects. now, as i'm sure you GPU crunchers already know, the GPU is dependent on the CPU when a GPU task either begins or ends. data for a GPU task must be loaded into VRAM from the system's main memory (via the CPU) before it starts crunching on the GPU. when the task is done, that data must be offloaded from VRAM (again via the CPU). for Milkyway@Home, the amount of data per task that must be transferred onto and off of the GPU is very little (just like bitcoin mining). to be a bit more precise, i think MW@H requires the transfer of roughly 30KB of data, and so the GPU's memory can be clocked extremely low. i downclock my HD 5870's VRAM from 1250mhz to 600mhz while running MW@H, and the tasks complete in as little time as they would had i left my memory at the stock clock of 1250mhz. the benefit is substantially lower power consumption, and thus significantly lower GPU temps. 600mhz seems to be the lowest MSI Afterburner will allow me to go with VRAM, although i've heard of folks using other AMD GPU utilities to go as low as 100-200mhz and still crunch MW@H successfully without errors or invalids.

Collatz Conjecture differs in that it reserves 30% of the GPU's VRAM (with AMD HD 5xxx series cards anyways), which equates to 666MB of my HD 5870's 2GB of VRAM. now whether it is used in its entirety or not is another story...but suffice it to say that the amount of data transferred back and forth between CPU and GPU with this project can easily get into the hundreds of megabytes. so imagine the actual amount of data that must be transferred for a Collatz task is ~300MB - that's 30,000 times more data than MW@H requires in a typical CPU-GPU data transfer. as such, a much higher VRAM bandwidth is required to move all that data in a reasonable amount of time. if it can't move the entire chunk of data at once, it'll be split up into 2 or more groups/transfers, requiring even more time for the entire data transfer. so while nothing is stopping you from downclocking your VRAM to lower power consumption and GPU temps in general, doing so will increase run times for Collatz tasks (unlike tasks for projects whose memory bandwidth requirements are low like MW@H and bitcoin mining).

 

[Smartazz]

Here's a scenerio I was wondering about. When I have two 6870s, will I be able to mine with one and game on the other without slowdown? One 6870 is still enough to power a lot of games at 60fps so I want to mine with the other one. Also, will it be a hassle to disable Crossfire when I want to play a less demanding game?

 

[Zargon]

if you are mining on one you should disable XFIRE anyways to game i beleive