Located in the Netherlands, also pretty expensive here:

During the day: $ 0.2831
During night: $ 0.2563

But last year the electricity meter was broken for 6 months, and they didn't charge me anything.
I suspected they would just take some average... but no :-D

Krypta:

I've seen that before as well, I didn't change it, but it sure looks like it shouldn't even compile!
I haven't tried the alterations yet, but will do so when I come home!

UPDATE: It works great, increase on both 6870's from 287 to 296 (> 3%!)

This is my "simple" miner. Nothing fancy, 2x 6870 with room for a third card. I need to get a SATA flat-cable before it fits probably, I've already ordered one.



I've tried to keep it as clean as possible, not blocking any airflow. Running the GPU fans on half speed gets it down to 74 degrees for both cards, mostly thanks to the bit of cardboard jammed between the GPU's to allow just a bit more air in between.

Cardboard! Somebody told me that his whole rig is running because of stategically placed cardboard. And I'm now doing the same.

My motherboard has a small problem, it has 3x double PCIe x16 slots, but I can't use the bottom one because of a SATA connector that is in the way. So I've got 2x 6870 in there at the moment. They are placed next to eachother and they reach temperatures of 99 and 82 degrees (the top one has serieus cooling problems, it is blocked). Jamming a piece of cardboard between the cards, forcing a bit of room between them has greatly reduced this problem, they are currently running at 87 and 81. I'm now waiting for 2 case fans I've ordered to fix the cooling problem for good. And I've also ordered a flat SATA cable to allow we to put in another card.

Oh, and yes, I'm currently also using a desk fan instead of a good case fan...

+1 as well!

I'm now using phatk (with Phoenix) for my double 6870's, it is working like a charm. But the tought that it might do better with an optimized kernel is killing me ;-)
Are there indications a better/optimized kernel for the 6xxx series can be created?

I'm currently using my laptop (Quadro FX 1800M, NVidia) on Windows 7, and it works like a charm. Averaging around the ~10 MHash/s, which is pretty good. DiabloMiner gives me ~13 MHash/s on this laptop.

Hmm, it seems that adding a third 5830 will force me into a lot more expensive motherboards... and more expensive PSU. I'd rather buy another case with two cards in the future if it turns out to be worth while.

So I guess I'm going for the cheapest option:

- Rasurbo BC-10 mini tower
- ASRock M3A770DE
- AMD Sempron 140
- 2x HD5830 Extreme
- OCZ ModXStream Pro 700W (should be enough for 2x 5830)
- 1 GB mem and a small HDD

(resulting in +/- 430 euro)

Just a feeling I have... No particular reason.

Any more advice on the choice of cards (5830/5870), the amount (2 seems easier to accomplish then 3), or building a PC from scratch?

Just looked it up, getting myself two 5870 in this setup will costs me 550 euro, and getting 2x 5830 will cost me 417 euro.

Which is actually a good point, as I'm not likely to ever see a ROI :-) Why spend that extra 133 euro?

Or maybe I should put 3x 5830 in the machine? And obviously upgrade the power supply!

Hi everybody,

I don't have a lot of experience building PC's, and I doubt I'll ever get ROI on my first mining rig heh. But I just like the idea of building one!

This is my current plan, any advice? Will it work? Can it be done (a lot) cheaper?

- Rasurbo BC-10 mini tower
- ASRock M3A770DE
- AMD Sempron 140
- 2x XFX HD5870
- OCZ StealthXStream 2 600W (Is this enough? Or would you suggest going 750W+?)
- 1 GB mem and a small HDD

This will set me back 550 euro, if I build it myself (is this hard?). I've already replaced some videocards, HDD's, DVD players etc in my life, just never build a PC from scratch.

Roy

Not that I don't want to dive into the source code to find out, but it can't really be that hard to explain how a miner works can it? Most likely it just takes some data, hashes it with one or two changing variables and tests if it fits the current target? I doubt knowing the data structures intimately is really necessary, but I get the feeling it is more adviced and/or encouraged?

Rather then having to read sources and decypher the algorithm I'd rather see it described somewhere. I know a little C/C++ so I'll probably be able to decypher the code. When/if I find the time I'll dive into the source code to find out how it works. If I do, I'll write a more elaborate description.

Good questions, I'm trying to answer those as well. Some things I've found out:

- The nonce is what a miner itself increases after every try.
- I think you need to hash quite some fields, the nonce, timestamp, and more (merkle root?)
- The generated hash needs to be smaller then the current target (which you can retrieve)

I'm also quite puzzled after reading the wiki-page, I don't think there is enough information currently on that page to create you own miner (which I plan to do).

Who can enlighten us? When you get a block, which steps do you need to do to generate a (possible) new block? Can somebody write it down in pseudo code?