Pages:
Author

Topic: Risk of ASIC proliferation (Read 3723 times)

legendary
Activity: 952
Merit: 1000
January 20, 2013, 10:56:38 AM
#22
When FPGA came out, did all the GPU miners pack up and call it quits, leaving only FPGa units? Nope.
Yeah the network hashrate didn't jump up a giant percentage when FPGA units joined
Well that's because FPGAs were only marginally faster then GPUs. Their main advantage was in power savings. A closer analogy would be when GPUs started mining, and completely dominated the CPU mining network.



See that giant rise around the middle of 2010? That's when Poclbm was released.

Wrong chart, crazyates. That linear chart shows the effects of the price bubble mid 2011.

The chart below shows the effects you mention (look for the huge vertical leap in hashrate mid 2010):


I choose that one cuz I've never been a fan of log scale charts, but you're right that it shows the difference much better. Thank you!
legendary
Activity: 916
Merit: 1003
January 20, 2013, 10:50:03 AM
#21
One of the things I always find ironic is how I "discovered" BTC and started mining with a 200 Mh/s 6770 GPU right at the bottom of that first retracement in hash rate and difficulty in Dec 2011.  It turns out I caught the almost exactly perfect time to enter as a new miner when difficulty took a big dip.

Difficulty will soon take off hockey stick fashion after the current dip.
donator
Activity: 2058
Merit: 1007
Poor impulse control.
January 20, 2013, 10:34:45 AM
#20
When FPGA came out, did all the GPU miners pack up and call it quits, leaving only FPGa units? Nope.
Yeah the network hashrate didn't jump up a giant percentage when FPGA units joined
Well that's because FPGAs were only marginally faster then GPUs. Their main advantage was in power savings. A closer analogy would be when GPUs started mining, and completely dominated the CPU mining network.



See that giant rise around the middle of 2010? That's when Poclbm was released.

Wrong chart, crazyates. That linear chart shows the effects of the price bubble mid 2011.

The chart below shows the effects you mention (look for the huge vertical leap in hashrate mid 2010):

full member
Activity: 143
Merit: 100
January 20, 2013, 10:30:07 AM
#19
As the ASIC miners start mining, and "outdated" miners go offline, does this not pose a risk to the network in the sense that there are fewer overall miners?

The network as it stands now is far more distributed amongst a greater number of computers & hardware mining. If the majority of these go offline due to obsolescence, in essence there would be a centralization of power to a far fewer number of miners.

Although the network hash rate will be significantly higher, the number of people involved will greatly reduce. Does anyone see a risk in this? Perhaps if the ASICs develop hardware issues in the future (I hope not) and cease mining operations, I would think the network would be very vulnerable to an attack. Granted, difficulty would drop and "regular" miners would return in a balancing act.

Is the efficiency and hash rate increase worth the trade off of a less distributed mining network?

The way I see, if the do exist then latest-gen ASIC's are going to be kept by the people that made them, for obvious reasons.

Mining will be done by a few corporate bodies and I suppose the entrepreneurs will shift focus to the service side of things.

Yup completely on target... well except for the fact that Avalon has started to ship (at least demo units are shipping).
legendary
Activity: 3318
Merit: 4606
diamond-handed zealot
January 19, 2013, 01:04:31 AM
#17
@niko

Insofar as investment is an expression of hash power I do not disagree.

I think it is naive to think that raw hash/second is not important, but your point that investment is an abstraction of hashrate is a good one.

I do not think we really disagree, the point is to stay ahead of centralized powers, let us not get caught up too much in language.
legendary
Activity: 916
Merit: 1003
January 18, 2013, 08:55:43 PM
#16
bcpokey and bonkey, while I acknowledge there is some risk in my little $150 purchase, I believe I will have my Jalapeno some time in March or perhaps April since I ordered late.  It is not out of any religious fervor for BFL but simply that their story rings true.  When reps from the forums visit BFL we'll further see that they really do have a facility and aren't just some pimply faced script kiddies in their moms' basements having a few yucks.
hero member
Activity: 784
Merit: 502
January 18, 2013, 06:56:08 PM
#15
As the ASIC miners start mining, and "outdated" miners go offline, does this not pose a risk to the network in the sense that there are fewer overall miners?

The network as it stands now is far more distributed amongst a greater number of computers & hardware mining. If the majority of these go offline due to obsolescence, in essence there would be a centralization of power to a far fewer number of miners.

Although the network hash rate will be significantly higher, the number of people involved will greatly reduce. Does anyone see a risk in this? Perhaps if the ASICs develop hardware issues in the future (I hope not) and cease mining operations, I would think the network would be very vulnerable to an attack. Granted, difficulty would drop and "regular" miners would return in a balancing act.

Is the efficiency and hash rate increase worth the trade off of a less distributed mining network?

The way I see, if the do exist then latest-gen ASIC's are going to be kept by the people that made them, for obvious reasons.

Mining will be done by a few corporate bodies and I suppose the entrepreneurs will shift focus to the service side of things.
hero member
Activity: 602
Merit: 500
January 18, 2013, 06:00:43 PM
#14
I bought a Jalapeno for slightly more than I paid for my el-Cheapo 6770 board.

If a piker like me can afford an ASIC anyone can.  I think there will be more ASIC users than OP thinks.

(Slightly OT, but only time will tell if running the Jalapeno is even worth it when the difficulty jumps 10X - 20X.  If not, it was a minor loss for me.)

A few points, you didn't buy a jalapeno, you bought a promise of a jalapeno. Mild but important difference.

As to whether it will be worth it to run, it of course depends on your definition of "worth it", btc price and elec price, but it will profitable into difficulty 300x give or take.
legendary
Activity: 916
Merit: 1003
January 18, 2013, 04:12:42 PM
#13
I bought a Jalapeno for slightly more than I paid for my el-Cheapo 6770 board.

If a piker like me can afford an ASIC anyone can.  I think there will be more ASIC users than OP thinks.

(Slightly OT, but only time will tell if running the Jalapeno is even worth it when the difficulty jumps 10X - 20X.  If not, it was a minor loss for me.)
full member
Activity: 143
Merit: 100
January 18, 2013, 04:01:02 PM
#12
My opinion (doesn't matter much, a btc noob)...

Yes ASICS will pretty much become the defacto requirement for mining.  Yes it will bump out the "slap extra parts" mining rigs... but I wonder how much of the current hash rate fits in that category (probably not much)?

Miners with mining "rigs" will upgrade, just like they did buying 3x+ the same video card to setup their current rig9s).  The price ranges for ASICS seem to be pretty close to what a GPU mining rig would be, probably a little more but not obscenely more.  There are $150 and up options, easily accessible to the GPU rig people.

The big question is once the hashrate settles after ASICS release, will the same investment into ASICS produce roughly the same income as GPU did (more of a investment recovery timeframe)?  I assume difficulty and profitability will balance all that out and asics will just be another mining upgrade to keep up with the jones'.
legendary
Activity: 952
Merit: 1000
January 18, 2013, 12:13:16 PM
#11
When FPGA came out, did all the GPU miners pack up and call it quits, leaving only FPGa units? Nope.
Yeah the network hashrate didn't jump up a giant percentage when FPGA units joined
Well that's because FPGAs were only marginally faster then GPUs. Their main advantage was in power savings. A closer analogy would be when GPUs started mining, and completely dominated the CPU mining network.



See that giant rise around the middle of 2010? That's when Poclbm was released.
sr. member
Activity: 388
Merit: 250
Save A Life, Adopt a Pet Today!
January 17, 2013, 10:46:42 PM
#10
So where can I buy FPGAs?

You might not want to since they are never going to ROI anymore.

But since (most) can only be using for mining, then their value should drop proportionally - look and the bfl singles vs. mmq's  the BFL holds its value only because it can be traded for its original cost - so people are still buying them (2nd hand of course) for roughly what they are worth - whereas mmq's dropped much further in value to roughly $300 because that was (until the bitcoinasic debacle) what they would be worth on a trade.  I suspect that their cost will drop in propotion to their value or close enough that running 4 or 5 mmq's at 3-4 GH will have the ability (albeit over a longer period) to recoup their investment.  Not everyone mines solely for profit.

Just like when newer cpu's come out the older ones drop in price etc..  And to be honest, yes, gpu's killed cpu mining - for the most part but i still mine a farm of decent cpu's on machines tasked for other purposes that effectively have no direct cost for the mining electricity to the tune of about 1 7970 gpu worth of hash power - and they have in the last 2 months solved 2 blocks - so it won't completely kill them.

I do believe that the sale of FPGA's will become more of a secondhand sales item much like older computers and parts are on ebay..

Just my 2 cents fwiw
legendary
Activity: 1526
Merit: 1001
January 17, 2013, 08:39:37 PM
#9
So where can I buy FPGAs?

You might not want to since they are never going to ROI anymore.
full member
Activity: 238
Merit: 100
January 17, 2013, 12:44:39 PM
#8
So where can I buy FPGAs?
newbie
Activity: 24
Merit: 0
January 17, 2013, 09:34:17 AM
#7

When FPGA came out, did all the GPU miners pack up and call it quits, leaving only FPGa units? Nope.

Yeah the network hashrate didn't jump up a giant percentage when FPGA units joined

No, but they also did not increase hashing power on the scale that ASIC will either. Where are all the CPU miners today? GPU mining pushed them into obsolescence. When FPGA miners were released, people realised the end is neigh for GPU, and it seems as if ASIC will drive the nails into the coffin for GPU.

Perhaps we'll have botnet operators to thank for the backup distributed hashing power of CPU and GPU, if all else fails..

This is likely a non-issue, the thought occurred to me the other day and I'm looking for more powerful minds to hash out the details  Smiley
hero member
Activity: 896
Merit: 532
Former curator of The Bitcoin Museum
January 17, 2013, 09:15:00 AM
#6

When FPGA came out, did all the GPU miners pack up and call it quits, leaving only FPGa units? Nope.

Yeah the network hashrate didn't jump up a giant percentage when FPGA units joined
hero member
Activity: 756
Merit: 501
There is more to Bitcoin than bitcoins.
January 17, 2013, 01:24:24 AM
#5
we need to continually increase network hash to stay ahead of the banks and their puppet regimes.

This is a misconception. Network hash rate in itself is irrelevant. What matters is the total investment in the mining business (and hobby). Based on the price of best available technology (CPUs, GPUs, FPGAs, some day ASICs) expressed in, for example, USD per GHash, this investment can be used to calculate network hashrate - but that's simply a consequence of the investment, not a relevant indicator of anything.

To rephrase your statement: total investment into mining hardware needs to grow to stay ahead of any potential adversaries. Note that the size of mining industry depends on expected ROI and the block reward plus transaction fees (expressed in bitcoins). For example, assuming a one-year ROI, it is around 1.4 million coins. This may be worth several tens of millions of US dollars. That is what keeps Bitcoin network secure. One of many smart underpinnings of Bitcoin is that this quantifiable incentive to secure the network correlates with the value of bitcoins (size of economy), which in turn may correlate with the incentive for an established and short-sighted business to attack the network. The less valuable bitcoins are, the easier it is to attack the network, but there is less interest in doing so. The more valuable they are, the more likely they are to be perceived as a threat, but then the mining stakes are higher and it is more expensive to stage an attack via this channel.

Besides, attacking the network would really be an idiotic attempt to disrupt Bitcoin; there are much cheaper and much more effective ways to attempt this. My opinion is that any sort of attempt by anyone would be idiotic, as Bitcoin technology can be useful for anyone: for governments, banks, terrorists, freedom fighters, retirees, kids, travellers, non-profits - anyone. Think Internet.

Now, back to the OP - I agree that shrinking or centralization of the mining power can be a problem (simply due to random events trigerring catastrophic disruptions in small networks), but I disagree that shrinking or centralization will be happening as a consequence of ASICs.
full member
Activity: 155
Merit: 100
January 17, 2013, 12:17:10 AM
#4
As the ASIC miners start mining, and "outdated" miners go offline, does this not pose a risk to the network in the sense that there are fewer overall miners?

No.

When FPGA came out, did all the GPU miners pack up and call it quits, leaving only FPGa units? Nope. Those interested joined in with FPGA. The same will happen with ASIC. Everyone will switch to FPGA/ASIC miners. Even a casual miner could afford a Jala to tinker around with. Yes you're going to lose some people (those that only do it because they happen to have video cards). They'll either continue anyway (because their parents pay electricity, and they don't care about mining at a loss), or they'll drop out. Either way, they constitute a very small and rather insignificant portion of the pool. In terms of security, having 100k users (95% of which are super casual and only contribute 5% of the total hash rate) is not that mscurried secure than 5k users with even distribution of power. The security comes from total network hash distribution, not total distribution of participants.
legendary
Activity: 3318
Merit: 4606
diamond-handed zealot
January 16, 2013, 08:56:28 PM
#3
I agree, dammed if you do dammed if you don't

we need to continually increase network hash to stay ahead of the banks and their puppet regimes.
Pages:
Jump to: