1GH/s, 20w, $700 (was $500) — Butterflylabs, is it for real? (Part 2)

[RandyFolds]

Yep, but I'm sick to death of the immature bullshit in this forum and so I am going to roll around in it like a dog with a dead fish.

You are getting pretty butt-hurt about this. Perhaps you should just shut the fuck up and wait for your holy grail to arrive in 4-6 weeks. No one is forcing you to respond, and as you've added nothing, no one will care if you stop.

[RandyFolds]

I do guess it is some form of "OMG this is the current holy grail" (I make that statement seriously not as a joke) but in reality how much better is it than anything else currently available?
I've seen a few posts comparing them, but interestingly enough those posts are not consistent.
The Anti-BFL's post specs that say they are not much better.
The Pro-BFL's post comments saying the new reduced specs are way better.

Can we also get some agreement of how it compares to current specs for other FPGA's?

Here you go: http://bitcoinfpga.com/

Good link. I hadn't seen this yet.

[kano]

I do guess it is some form of "OMG this is the current holy grail" (I make that statement seriously not as a joke) but in reality how much better is it than anything else currently available?
I've seen a few posts comparing them, but interestingly enough those posts are not consistent.
The Anti-BFL's post specs that say they are not much better.
The Pro-BFL's post comments saying the new reduced specs are way better.

Can we also get some agreement of how it compares to current specs for other FPGA's?

Here you go: http://bitcoinfpga.com/

Thanks, that page sums it up well.
I presume there isn't anyone on either 'side' that disagrees with those details?

But wow - those spec comparisons are surprising.
I really didn't think it was that different.

BFL's is the best MH/$ of them all and worst MH/W of them all.
The specs they originally quoted would have put them in the lead in all respects ... ... ...

The up front costs are quite a lot more for the others compared to BFL.

Long term, for anyone who has to pay for electricity, they are the worst but if you get it free/fixed cost or steal electricity they are the best.
However, the running cost difference may be small enough to not be a concern if it effectively adds up to less than the up front cost difference over something between 6 to 12 months.
Interesting.

And in case it wasn't obvious for the understanding-challenged person who replied further up, I really don't care if you lose all your money.
My questions are based on my reading this whole annoying thread (annoying due to the extremely high level of crap in it) and interest in getting some FPGAs myself in the not too distant future and wondering why people are so positive about BFL given the circumstances that are clearly obvious.
I guess if they do get the power usage down it may still be a new holy grail.

Also, as I have already stated, but I guess I need to point out to him again coz it may be difficult for him to understand what I wrote, the earlier comparisons posted in this thread where QUITE inconsistent depending on who posted them.

[yochdog]

This sentance cracked me up considering the operating costs I have:

It looks likely that miners using GPUs will need to migrate to FPGAs to remain profitable during 2012.


I don't NEED to do shi*t to remain profitable.  Doing the mental math, this guy thinks I should sink over $10,000 into new FPGA hardware to replicate my current farm.

I don't think so. 



PS- For my farm to become un-profitable, the difficulty MUST stay at 1,250,000 while the price per BTC declines to 60 cents.  I seriously doubt that will happen. 

[RandyFolds]

This sentance cracked me up considering the operating costs I have:

It looks likely that miners using GPUs will need to migrate to FPGAs to remain profitable during 2012.


I don't NEED to do shi*t to remain profitable.  Doing the mental math, this guy thinks I should sink over $10,000 into new FPGA hardware to replicate my current farm.

I don't think so. 



PS- For my farm to become un-profitable, the difficulty MUST stay at 1,250,000 while the price per BTC declines to 60 cents.  I seriously doubt that will happen. 

I know I asked a bunch of months back, but you got that hydroelectric shit, or what? That is some dirt cheap power.

[rjk]

My questions are based on my reading this whole annoying thread (annoying due to the extremely high level of crap in it)

Yes unfortunately the bullshit-to-useful content ratio is fairly bad.

According to the table on the bitcoinfpga website, if we ignore the BFL product, Icarus comes out number 1 in the mhash/watt department, followed by Ztex. However, Ztex is number one with efficiency (not including rph because so far he isn't selling anything), followed by Icarus. So, ignoring BFL, Icarus is the best if you have free power, and Ztex is the best for those that need to pay for power.

Now, when the issues with thermal runaway are sorted out, I think BFL will be able to deliver closer to 900 mhash at 60 watts. This would still be less efficient than any other solutions, but if you have free power it would return its cost the fastest. Out of the other FPGAs (not including rph), only the X6500 is dedicated to bitcoin, like BFL. The others are designed as development boards, and therefore have good utility outside of Bitcoin.

[yochdog]

This sentance cracked me up considering the operating costs I have:

It looks likely that miners using GPUs will need to migrate to FPGAs to remain profitable during 2012.


I don't NEED to do shi*t to remain profitable.  Doing the mental math, this guy thinks I should sink over $10,000 into new FPGA hardware to replicate my current farm.

I don't think so. 



PS- For my farm to become un-profitable, the difficulty MUST stay at 1,250,000 while the price per BTC declines to 60 cents.  I seriously doubt that will happen. 

I know I asked a bunch of months back, but you got that hydroelectric shit, or what? That is some dirt cheap power.

Yes, we are a bunch of fish-haters out here, so we have big-ass dams that produce our power.  The cool thing is snow is free, so we don't need to buy coal or natural gas.  Our cost is .06 per KW/h. 

Also, about 2/3's of my farm is running in office space where the Utilities are included with rent.....That comes out even cheaper than .06. 

[yochdog]

My questions are based on my reading this whole annoying thread (annoying due to the extremely high level of crap in it)

Yes unfortunately the bullshit-to-useful content ratio is fairly bad.

According to the table on the bitcoinfpga website, if we ignore the BFL product, Icarus comes out number 1 in the mhash/watt department, followed by Ztex. However, Ztex is number one with efficiency (not including rph because so far he isn't selling anything), followed by Icarus. So, ignoring BFL, Icarus is the best if you have free power, and Ztex is the best for those that need to pay for power.

Now, when the issues with thermal runaway are sorted out, I think BFL will be able to deliver closer to 900 mhash at 60 watts. This would still be less efficient than any other solutions, but if you have free power it would return its cost the fastest. Out of the other FPGAs (not including rph), only the X6500 is dedicated to bitcoin, like BFL. The others are designed as development boards, and therefore have good utility outside of Bitcoin.

I really like how they use a 6870 for comparison....judging by early assessments of he 7970, GPU efficiency just took a huge leap forward. 

[DeathAndTaxes]

Where do people get the idea that if you have free power you would pick a less efficient FPGA?  Um GPU?

What is still hard to understand is why are the chips in BFL board so inefficient.

I mean 7970 gets 4.2MH/W and that is with an unoptimized miner.  7990 and optimized miner could be 5MH/W.  An FPGA design getting only 10MH/W is simply "weird".  Eventually someone will get one and be brave enough to crack that cover and we will finally learn what this incredibly cheap and yet at the same time horribly inefficient chip is.

[kano]

Actually I'm just looking at if from a FPGA comparison at the moment.

Until both this exists (still not clearly known when) and some useful figures for 7970 and 7990 have been out for a while, it's not really possible to compare them - but of course I'd also not make any decisions before that either.

[xzion]

Where do people get the idea that if you have free power you would pick a less efficient FPGA?  Um GPU?

perhaps FPGA actually works out cheaper when you take into account not needing an extra PC for every 4 FPGA's, unlike GPU setups. for large scale anyway.

[xDGDZEx]

Where do people get the idea that if you have free power you would pick a less efficient FPGA?  Um GPU?

perhaps FPGA actually works out cheaper when you take into account not needing an extra PC for every 4 FPGA's, unlike GPU setups. for large scale anyway.

The appeal to me would mostly be less hassle for maintenance (presumably), as well as less noise, space, heat etc. To me, gpu's are still, and for a while still, will be competitive options, but not everyone has the space and ability to have multiple noisy mining rigs heating their room 24/7.

[simonk83]

For me its really simple.  Electricity is expensive here, and I can get the same hashrate with two of these boxes at a third, almost quarter, of the power usage.  Plus no more noise, heat, or failing fans from the 4 gpus I have now.

I did look at the other FPGAs, but I'd need 4, at least, to equal what two of these can do.

No brainer.

[suppp]

http://bitcoinfpga.com says ship date is 2012-01-23, it's correct info?

[grue]

My questions are based on my reading this whole annoying thread (annoying due to the extremely high level of crap in it)

Yes unfortunately the bullshit-to-useful content ratio is fairly bad.

According to the table on the bitcoinfpga website, if we ignore the BFL product, Icarus comes out number 1 in the mhash/watt department, followed by Ztex. However, Ztex is number one with efficiency (not including rph because so far he isn't selling anything), followed by Icarus. So, ignoring BFL, Icarus is the best if you have free power, and Ztex is the best for those that need to pay for power.

Now, when the issues with thermal runaway are sorted out, I think BFL will be able to deliver closer to 900 mhash at 60 watts. This would still be less efficient than any other solutions, but if you have free power it would return its cost the fastest. Out of the other FPGAs (not including rph), only the X6500 is dedicated to bitcoin, like BFL. The others are designed as development boards, and therefore have good utility outside of Bitcoin.

what happened to the homemade pan reflowed, $1/MH FPGA?

[cuz0882]

This sentance cracked me up considering the operating costs I have:

It looks likely that miners using GPUs will need to migrate to FPGAs to remain profitable during 2012.


I don't NEED to do shi*t to remain profitable.  Doing the mental math, this guy thinks I should sink over $10,000 into new FPGA hardware to replicate my current farm.

I don't think so. 



PS- For my farm to become un-profitable, the difficulty MUST stay at 1,250,000 while the price per BTC declines to 60 cents.  I seriously doubt that will happen. 

I know I asked a bunch of months back, but you got that hydroelectric shit, or what? That is some dirt cheap power.

Yes, we are a bunch of fish-haters out here, so we have big-ass dams that produce our power.  The cool thing is snow is free, so we don't need to buy coal or natural gas.  Our cost is .06 per KW/h. 

Also, about 2/3's of my farm is running in office space where the Utilities are included with rent.....That comes out even cheaper than .06. 

Sounds expensive to me, its only .0202 per watt here.

[Enigma81]

Sounds expensive to me, its only .0202 per watt here.

You pay $20.20 per kilowatt-hour?  Seems a bit unprofitable to mine bitcoins...

[rjk]

what happened to the homemade pan reflowed, $1/MH FPGA?

I believe that's the one marked "rph" on the website. He isn't selling them as far as I know.

[rjk]

Sounds expensive to me, its only .0202 per watt here.

You pay $20.20 per kilowatt-hour?  Seems a bit unprofitable to mine bitcoins...

Did you read backward? He said $0.0202 kw/h.
EDIT: That's a damn good rate btw.

[notme]

Sounds expensive to me, its only .0202 per watt here.

You pay $20.20 per kilowatt-hour?  Seems a bit unprofitable to mine bitcoins...

Did you read backward? He said $0.0202 kw/h.

No, he said ".0202 per watt"... like you quoted... learn 2 read