What would you buy if... ?

[ATC777]

Let's suppose you were wanting to do a "trial run" of a small mining operation to demonstrate the viability of Bitcoin mining to shareholders in your company -- you want to show them some tangible results and some projections/estimations for the future to bring them aboard... keeping in mind that Bitcoin is a totally new and strange concept to them. And you want to adapt your plan to the rise of ASIC chips. You have a limit of $5,000.00 and no more (you don't necessarily have to spend all of it though)!

Now, how would you spend your $5k and what would you do to accomplish your goal? ...and where would you buy the things you'd need?

Would you go for a cost-effective GPU-based operation? Or opt for FPGA technology instead? Or do you think it's insane to do anything other than wait on ASICs? And if so, what would you do with the money when ASIC chips hit the shelves and are released into the wild?

Try to be as detailed and thoughtful as possible in your ideas. Do you think you could profit? Would you attempt to re-sell the rig(s) you'd buy, and at how much depreciation/loss? You get my point now...

If you give particularly helpful/useful advice I may want to hire you as a paid freelance/contract consultant in the near future and stay in touch...  

Regards,

--ATC--

[crazyates]

Let's suppose you were wanting to do a "trial run" of a small mining operation to demonstrate the viability of Bitcoin mining to shareholders in your company -- you want to show them some tangible results and some projections/estimations for the future to bring them aboard... keeping in mind that Bitcoin is a totally new and strange concept to them. And you want to adapt your plan to the rise of ASIC chips. You have a limit of $5,000.00 and no more (you don't necessarily have to spend all of it though)!

Now, how would you spend your $5k and what would you do to accomplish your goal? ...and where would you buy the things you'd need?

Would you go for a cost-effective GPU-based operation? Or opt for FPGA technology instead? Or do you think it's insane to do anything other than wait on ASICs? And if so, what would you do with the money when ASIC chips hit the shelves and are released into the wild?

Try to be as detailed and thoughtful as possible in your ideas. Do you think you could profit? Would you attempt to re-sell the rig(s) you'd buy, and at how much depreciation/loss? You get my point now...

If you give particularly helpful/useful advice I may want to hire you as a paid freelance/contract consultant in the near future and stay in touch...  

Regards,

--ATC--

Cost-effective GPU operation? Most GPU based miners are barely making even, depending on power costs. FPGAs are a little more forgiving for higher electric rates, but they are also more expensive.

Either way, I expect both of those to be dwarfed by ASICs in less than 60 days. I expect all 3 ASIC manufacturers to be shipping in the next 90 days or so. By then, you will have a hard time convincing investors it's worth it, as your earnings will be very minimal.

If you wanted earnings now, buy $5,000 worth of BFL Singles. You couple prolly pick up 8-9 for that price, which would put you somewhere around 7.5GH/s for under 1kw. Then you can earn something now, and when ASICs hit, you can use your earnings and investor's cash to buy upgrades to ASICs and still remain viable. Just a thought.

[kwoody]

I would buy a small GPU farm based on 7xxx cards with half the funds, and spend the other half on pre-ordering ASICs, and leave fpga out all together. gpu's can be easily sold to a second hand market whereas fpgas are somewhat limited in what they can be repurposed for and therefore doesn't have a huge resale market.

6x   SAPPHIRE 100351SR Radeon HD 7970 3GB 384-bit GDDR5 PCI Express 3.0 x16 HDCP Ready CrossFireX Support Video Card, $2160

2x   ASRock 970 EXTREME3 AM3+ AMD 970 SATA 6Gb/s USB 3.0 ATX AMD Motherboard,  $129

2x   AMD Sempron 145 Sargas 2.8GHz Socket AM3 45W Single-Core Desktop Processor, $76

2x   Crucial Ballistix 2GB 240-Pin DDR3 SDRAM DDR3 1333 Desktop Memory Module, $19

2x   CORSAIR Enthusiast Series TX850 V2 850W ATX12V v2.31/ EPS12V v2.92 80 PLUS BRONZE Certified Active PFC Power Supply, $210

2x   Kingston Storage Locker 8GB USB 2.0 Flash Thumb Drive, $16

2x   APC BX1500G 1500 VA 900 Watts Power-saving Backup-UPS, $400(optional, but recommended)

2x   (pre-order) bASIC 72GH/s ASIC devices, $2200

---------------------------------------------------------------------------------------------------

                                     $5,210



Slightly over your budget, but you can leave out the UPS battery backups and shave off the $400.
This would get you setup with nearly 4GH/s worth of gpu mining equipment to play with while waiting for 144GH/s worth of ASIC chips. Once ASICs arrive, sell the GPUs and order more ASICs.
Hope this helps.

[crazyates]

Ah if you wanted to keep your ASIC orders included in the 5 grand, you could pick up 4 BFL Singles, which would give you about 3.3GH/s. This is about comprable to kwoody's suggestion of 6 x 7970s (6 x 630MH/s = 3.7GH/s), but they're WAY easier to set up.

4x used Singles = 2400

4x SC upgrades = 2800

This gives you a slightly smaller, but MUCH simpler and easier to set up, farm that uses way less power, and then sets you up for 240GH/s of pre-orders.

[Miner99er]

Ah if you wanted to keep your ASIC orders included in the 5 grand, you could pick up 4 BFL Singles, which would give you about 3.3GH/s. This is about comprable to kwoody's suggestion of 6 x 7970s (6 x 630MH/s = 3.7GH/s), but they're WAY easier to set up.

4x used Singles = 2400

4x SC upgrades = 2800

This gives you a slightly smaller, but MUCH simpler and easier to set up, farm that uses way less power, and then sets you up for 240GH/s of pre-orders.

This!

[meowmeowbrowncow]

Buy shares in a ASIC mfg that self mines.  Nothing beats mining with at cost hardware.

[creativex]

Seems like funding the competition.

[ATC777]

Buy shares in a ASIC mfg that self mines.  Nothing beats mining with at cost hardware.

Brilliant suggestion. I've been thinking of that as a way of adding some "deferred hashing power" to our company's portfolio (we're a conglomerate -- Bitcoin is a new frontier for us).

[BR0KK]

I personally recommend not to buy at the moment .... Especially not fpga or ASIC (neither shares of them)

For playing and learning a simple rig (1-2 decent mining GPUs -- 5870; 5970; 6990, 7970)

Just wait and see how that ASIC circus will unfold next year (good or bad) and then if you are still interested in BTC, make a good and well thought decision.....

Do not buy vaporware and therefore into dept (if ASIC fails or you have to wait to long to get one) Remember: Especialy for asic, you will be last in line! Diff will be high when you receive yours.... Profit?! ... No one knows

[ATC777]

Thanks for all this great advice!  

I think this is what I want to do... I simply want to build the most cost-effective GPU or FPGA rig possible with a maximum of $1,000.00 invested, and I want to build it soon. The only purpose of it is to do a live demonstration for our company shareholders and show them that this whole "Bitcoin thing" (which is totally new to them) is real, legit and it works. I want them to see that with the proper setup mining can potentially be profitable -- perhaps very profitable (we'll just have to wait and see). So now I have some questions...

First, about FPGA rigs ::

• Who are the best/most realiable FPGA vendors for Bitcoin mining?
• How/where do you buy this equipment?
• Where can I learn more about assembling FPGA setups and getting them running?
• Where can I learn more about designing and building custom FPGA chipsets?

Secondly, about GPU setups ::

• Where do you order your parts/equipment?
• What's the best way to design a modular "farm" I can add GPUs to at a whim?
• What mobos, RAM, CPUs and other hardware is best for the rest of the system?
• Where can I find learning resources about building effective GPU farms?

EDIT ::

I see that BFL offers some FPGA rigs, but I can't see any way to actually order them. It always re-directs me to the pre-order page for their upcoming ASIC products... wtf?

[BR0KK]

FPGA:
Cost to high and literally no resale value at aftermarket.
Energy efficient

Ztex Quad 1.15x @ ~860MH/s (LX150)
Enterpoint Quad Spatan 6 LX150 (not shure if producing new units)
X6500 (not shure if is producing new units)
Icarus Dual Spartan 6 LX150 (discontinued)
Lancelot Dual "" (never actually released)
Bfl Single 700 - 800 MH/s

Recommendation (personal)
Ztex since I have a lot of them

[ATC777]

FPGA:
Cost to high and literally no resale value at aftermarket.
Energy efficient

Ztex Quad 1.15x @ ~860MH/s (LX150)
Enterpoint Quad Spatan 6 LX150 (not shure if producing new units)
X6500 (not shure if is producing new units)
Icarus Dual Spartan 6 LX150 (discontinued)
Lancelot Dual "" (never actually released)
Bfl Single 700 - 800 MH/s

Recommendation (personal)
Ztex since I have a lot of them

But where on earth do you buy this stuff, and which vendors are safe/trustworthy?

[BR0KK]

Interwebz is wide just google for the names of the boards...


Ztex: Is based in EU and trustworthy. I (Ztex is located in Germany so am i) can get a board; if in stock; in under 3 days. But they are a bit expensive!
Enterpoint. Eu based (i think England). Cheaper then Ztex. Not so trivial to configure...
Icarus /  Lancelot: Since Discontinued or never released you can only get them here (forum, bitmit). Developer is based in China!
X6500: Might be based in America.? unsure if they are still producing boards?
BFL: Somhwere in the US .... It depends witch sources you want to believe. They deliver a Single that is quite good (search for threads here) , but It comes with a catch!
Read further about their "business ethics" and "shady" Business activities (Regarding first single and nowadays ASIC).

Recomendation is: (Not to by them because when (IFIFIF) ASIC hit you'll loose money)
Try to spot them used on BITMIT or forum


Obtaining a decent mining capable GPU should be self-explanatory

If i would want to try out BTC, i wouldn't buy specialized (potentially obsolete and not resell-able, or nonexistent unicorn blood powered) hardware at the moment. Buy (if you don't have one already) a decent mining GPU and start at a big pool. Then read further into BTC and try different things. Do not under any circumstances get into a "gold rush fever" that leads you to buy hardware you can't afford. Do not invest in anything when you are not  1k % certain to loose it all!

^may contain personal opinion  about stuff

[mokahless]

I have no experience with business or large scale operations.

However, A GPU setup is not viable because of what the company needs to present. A small efficient setup would display the possibility of expansion. Waiting for ASICs is stupid because then you delay the business moving forward. Make ASICs a fundamental part of your presentation for shareholders.
Currently, the only viable FPGA seems to be the ZTEX 1.15y. I would prefer a setup using x6500s but they are just not as available.
Order everything directly from ZTEX: 2x 1.15y boards, 1x power supply, 1x 5 way splitter, 2x usb cables. total: 1918 Euros. For the sake of quick calculations, let's call this $3000 of our budget after shipping, taxes and let's add in a power meter and raspberry pi or similar low-powered, low cost... let's call them "cluster controllers" with a usb hub supporting up to 5 at one time.
I'm assuming this company is already established and has some kind of working office.
Measure power, present results, analyze profitability, present analysis of bitcoin network, ASIC adoption and profitability over time. Convince them that things will not become less profitable over time and will either remain constant or increase with expansion.
The other $2000 goes towards this however the company analyzes the cost of work and presentation.
That's as specific as I can be.
Just posted because I thought it was something cool to think about.

[BR0KK]

Some corrections:

The pi is far from being a perfect mining controller :/ believe me I've tried!
Better use a crappy ATOM Netbook.

You need two power supply's if you want to run them with wall plugs. Each board draws about 40w so only one NT (65w) is not sufficient. Even a 90w ps won't work for long.
These ps will die over time! I had three out of 5 failing on me after 5 month (not the best quailty; they get hot, ideal load is about three single 1.15x boards not a quad; need cooling ---> if they get to hot or to old your board will be malfunction).

Why the hell is a GPU setup not viable for displaying?

[mokahless]

Some corrections:

The pi is far from being a perfect mining controller :/ believe me I've tried!
Better use a crappy ATOM Netbook.

You need two power supply's if you want to run them with wall plugs. Each board draws about 40w so only one NT (65w) is not sufficient. Even a 90w ps won't work for long.
These ps will die over time! I had three out of 5 failing on me after 5 month (not the best quailty; they get hot, ideal load is about three single 1.15x boards not a quad; need cooling ---> if they get to hot or to old your board will be malfunction).

Why the hell is a GPU setup not viable for displaying?

Good catch on the power supply. My post was rushed and I don't really care about doing too much research since I just posted for fun.
It's just my opinion; A GPU setup presents a lower hashrate, using more floor space and a higher power consumption with a lower initial investment (supposedly. I haven't checked costs in a while). If I wanted to invest in something long-term as a business, I would be more interested in the most efficient. I assume this supposed company is interested in large-scale, meaning a working small-scale model of something that can be scaled up would be much more convincing.

[crazyates]

Funny thing about Bitcoin: it's very predictable in some ways, and very un-predictable in others.

You can quite easily choose to reduce any variance to zero with PPS, and you can tell almost exactly what a certain hashrate will earn over a 2 week diff recalculate. You could tell your investors every 2 weeks "I earned 7.45 BTC the past 2 weeks, and I expect to earn 7.68 BTC over the next to weeks".

The problem becomes when trying to decide difficulty more than 2 weeks, 1 month, or even 3 months out. And even if you can come up with a decent estimate (which you can get pretty close), you have no idea about (or control over) what the price will be by then. It could go up 25%, or down 75%. It's very temperamental.