ASIC = The end of decentralized mining

[notme]

FPGA hardware will advance at about the same rate as GPU hardware.  Even if they only put more cores per chip, they should see the same improvements as GPUs which are doing the same: more cores per chip.  The work you're referring to is all software, has already been done, and the situation is nearly identical to GPU mining software.  Both GPU programming and FPGA programming is somewhat specialized, but I know where I got my undergraduate degree we had to code for FPGAs as part of the degree program, but not GPUs.

I just can't help but feel that the GPU market (as a whole) will move faster than the FPGA market due to demand and market size.

This will not be true as long as high frequency trading is allowed.  Investment banks provide a lot of demand in the FPGA market.

[dave3]

I think to a certain degree, ASIC mining will balance out.  The thing is, the market for ASIC mining devices is limited and the amount of money miners are willing to invest is limited.  While I'd expect ASIC prices to drop over time, I'm not convinced they'll drop so quickly as to make ASIC mining unprofitable.

But... with ASICs things do become more unsettled and uncertain.  While I may invest in some ASICs, it probably won't be as much as I would have put into FPGAs before BFL announced they were going to make an ASIC.

[aqrulesms]

Personally I think people are hyping way too much over ASICs and the huge "leap" in profits they will gain, while in reality they wouldn't.

There will always be only 7200 BTC created per day (excluding block reward drops).  At the most you are simply benefiting short term.  In the long term you would make just about the same as a GPU miner due to difficulty increases from widespread adoption of ASICs and the high prices for ASICs.

Just look at how unrealistic it is for many FPGAs out there right now already due to their high cost.

The biggest increase you would see from ASICs would be power consumption, but expect hash rates to be around the same levels as the FPGAs and GPUs.

[dave3]

Yeah, I'm not expecting higher profits from ASICs.  It's just to keep up and stay in the mining game.

[lemonz]

There will always be only 7200 BTC created per day (excluding block reward drops).  At the most you are simply benefiting short term.  In the long term you would make just about the same as a GPU miner due to difficulty increases from widespread adoption of ASICs and the high prices for ASICs.

I don't think anyone is calculating long term returns with ASICs based on today's difficulty - that would just be silly.  But in the long term, sustainable GPU mining (if possible) will never be as profitable as mining with an ASIC.  The good news however, is that come the end of the year, half of the coins will have already been mined

[bulanula]

I have no intentions on buying FPGA or ASICs

Same here. Don't have faith in this boat ... I think ASIC is the final push to sinking

[dave3]

Right now bandwidth usage isn't a big deal, but with ASICs, it could become an issue.  Especially if you're on a DSL line or something with a limited uplink.  If 1GH is ~3.5kbps, I guess that would make 100GH ~350kbps.

[SgtSpike]

Right now bandwidth usage isn't a big deal, but with ASICs, it could become an issue.  Especially if you're on a DSL line or something with a limited uplink.  If 1GH is ~3.5kbps, I guess that would make 100GH ~350kbps.

Pretty sure anyone with 100 GH/s isn't going to have a fit over having to pay for a better internet service than DSL.

[zvs]

When you make all-encompassing statements like "your seriously going to sit here and try and say that GPU mining is more profitable than FPGA," then I'll take it literally.  If you preface your statement with "Unless your electricity is free, blah blah blah," then I probably wouldn't bug you.

Also, have you done a recent analysis on how long it takes for an FPGA to out-profit a GPU setup at different electric rates?  I doubt it is just at $0.00/kwh where it would make more sense to buy GPU's instead of FPGA's...

A quick analysis using the following figure on bitcoinx.com using default values and the following:

GPU: 7970
Watts: 208 624/3 [wall draw / 3 for a 3 card system]
$570 ((470 * 3)+300)/3 [((card cost * 3) + base cost of a rig) / 3]

FPGA: USB-FPGA Module 1.15y
Watts: 38
$950

$/kw/h - cost per kw/h
GPU - days to pay off
FPGA - days to pay off

$/kw/h   GPU   FPGA
 $-         398   447
 $0.01    414   449
 $0.02    430   451
 $0.03    448   453
 $0.04    467   456
 $0.05    489   458
 $0.06    512   460
...
 $0.12    732   473


The number appears to be between 3 and 4 cents.  These are all meaningless though, since difficulty / price will not stay where it is, but does give an indication as to currently where that sweet spot lies.

but also totally ignores equipment depreciation

(or appreciation, in the case of some ppl buying 5830's)

[notme]

All this BS about ASIC early adopters being screwed is assuming asics will only be available in massive sizes that cost thousands of dollars.  What's stopping someone from putting out a smaller device that only pulled a gigahash or so for $100?  That's what I would do if I were selling them, but then again I wouldn't be in it for the money.  I just like building shit.

[SgtSpike]

All this BS about ASIC early adopters being screwed is assuming asics will only be available in massive sizes that cost thousands of dollars.  What's stopping someone from putting out a smaller device that only pulled a gigahash or so for $100?  That's what I would do if I were selling them, but then again I wouldn't be in it for the money.  I just like building shit.

I think that's exactly what BFL is planning to do - offer a smaller, affordable unit to those who don't have much to spend.

[BR0KK]

Then why do they take singles back for refund ?
I'll expect a basic ASIC somewhere between 900 to 1k € (1500$)

[SgtSpike]

Then why do they take singles back for refund ?
I'll expect a basic ASIC somewhere between 900 to 1k € (1500$)

Whose to say that a single single couldn't buy 6 mini-ASICs?

[P4man]

All this BS about ASIC early adopters being screwed is assuming asics will only be available in massive sizes that cost thousands of dollars.  What's stopping someone from putting out a smaller device that only pulled a gigahash or so for $100? 

Makes no difference, at those per GH prices they will sell boatloads and difficulty will just go up even faster and you will still have close to no chance to recover your investment. 

Not that I see why would BFL charge just $100 for 1GH if people are willing to pay > $600 today.

As for the internet connection dave mentioned; thats of course a complete non issue. Having a 100GH mining rig will consume the exact same amount of bandwidth as 1 GH today,  once difficulty goes x100. Difficulty has gone x100000 in the past 24 months, but that hasnt affected bandwidth required for mining.

[BR0KK]

If i were bfl I wouldn't sell them to cheap .....

I'm expecting 1gh at double or triple € as per single.

[SgtSpike]

If i were bfl I wouldn't sell them to cheap .....

I'm expecting 1gh at double or triple € as per single.

I bet you'll be wrong about that one... want to bet on it?

[BR0KK]

Can't afford it because I invested to much into FPGS:( (I'm seriously out of money or BTC ATM ;().

[notme]

Having a 100GH mining rig will consume the exact same amount of bandwidth as 1 GH today,  once difficulty goes x100.

This is not true if you mine at a pool that expects difficulty 1 shares, as most do.

[ice_chill]

Lets wait for the announcement, due today

[DeathAndTaxes]

This is reasonable, and only time will tell.  I just can't help but feel that the GPU market (as a whole) will move faster than the FPGA market due to demand and market size.  Of course technology across the board will benefit from advancements, but I believe the larger and more profitable markets will see it first.

Very very very unlikely.  AMD just made a shift to 7000 series and performance (relative to 5000 series) was lackluster.   AMD won't be making a new generation for at least 18 months.   

Current 45nm FPGA compete well with GPU in all but the lowest cost areas.  28nm FPGA will eliminate any remaining advantage and ASICs will make GPU mining look as stupid as CPU mining is now.

GPU mining is dead.  Time will only worsen the relative comparison.

Note this comes from someone running a 15GH/s farm all GPU farm.