defxor
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 07, 2011, 06:56:46 PM |
|
You fucking idiot.
It's nice to see you too ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) ArtForz himself, the man who supposedly spend a shitload of his own cash, to pursue as close to an ASIC as he could get (the sASIC he mentioned) doesn't believe ASICs are viable... ... and crush your FPGA argument at the same time? I don't agree with ArtForz though, with a _low_ BTC price ASICs make sense. FPGAs already. I'm not posting calculations!
Understandable. After all, those calculations would likely disprove your own point ![Wink](https://bitcointalk.org/Smileys/default/wink.gif)
|
|
|
|
Synaptic
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 84
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 07, 2011, 10:20:05 PM |
|
You fucking idiot.
It's nice to see you too ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif) ArtForz himself, the man who supposedly spend a shitload of his own cash, to pursue as close to an ASIC as he could get (the sASIC he mentioned) doesn't believe ASICs are viable... ... and crush your FPGA argument at the same time? I don't agree with ArtForz though, with a _low_ BTC price ASICs make sense. FPGAs already. I'm not posting calculations!
Understandable. After all, those calculations would likely disprove your own point ![Wink](https://bitcointalk.org/Smileys/default/wink.gif) ArtForz is hobbyist. Vladamir is a businessman. Which one is using FPGAs?
|
|
|
|
newMeat1
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 07, 2011, 10:37:15 PM |
|
Hobbyist or not, which one has made more money? OK, I don't know. They've both probably made a lot.
With these low prices, which one stands to make more in the future?
|
|
|
|
logansryche
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 08, 2011, 09:27:54 PM |
|
This reminds me of the debate a few months back when they said mining off a sound card was possible. Interesting read.
|
|
|
|
nmat
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 09, 2011, 12:02:49 AM |
|
Well, these guys say they are developing an ASIC miner. According to this it will be ready by middle 2012
|
|
|
|
netrin
Sr. Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif) ![*](https://bitcointalk.org/Themes/custom1/images/star.gif) ![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 322
Merit: 251
FirstBits: 168Bc
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 09, 2011, 12:33:24 AM Last edit: September 09, 2011, 12:52:19 AM by netrin |
|
Let's recap shall we? Winter saw lows of $0.07 and highs of $1.00 Spring saw lows of $0.50 and highs of $9.00 Summer saw lows of $8.00 and highs of $32.00 Autumn sees lows of $6.15 and expected highs between $25 and $100 ![](https://ip.bitcointalk.org/?u=http%3A%2F%2Fbitcoincharts.com%2Fcharts%2Fchart.png%3Fs%3D2010-10-11%26e%3D2011-10-01%26i%3DWeekly%26c%3D1%26l%3D1%26t%3DS%26noheader%3D1%26height%3D500%26width%3D720%26m%3DmtgoxUSD&t=665&c=-PPwhdzAm6wvrg)
|
|
|
|
nmat
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 09, 2011, 12:57:34 AM |
|
|
|
|
|
netrin
Sr. Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif) ![*](https://bitcointalk.org/Themes/custom1/images/star.gif) ![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 322
Merit: 251
FirstBits: 168Bc
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 09, 2011, 01:35:54 AM |
|
I take it you don't like logarithms. ![](https://ip.bitcointalk.org/?u=https%3A%2F%2Fi.imgur.com%2FqqtEf.png&t=665&c=6xt-6a-CnKNeSQ)
|
|
|
|
Synaptic
Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 84
Merit: 10
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 09, 2011, 05:55:00 AM |
|
I take it you don't like logarithms. ![](https://ip.bitcointalk.org/?u=https%3A%2F%2Fi.imgur.com%2FqqtEf.png&t=665&c=6xt-6a-CnKNeSQ) LOL. nice.
|
|
|
|
netrin
Sr. Member
![*](https://bitcointalk.org/Themes/custom1/images/star.gif) ![*](https://bitcointalk.org/Themes/custom1/images/star.gif) ![*](https://bitcointalk.org/Themes/custom1/images/star.gif)
Offline
Activity: 322
Merit: 251
FirstBits: 168Bc
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 09, 2011, 01:34:27 PM |
|
I can't take credit for the image. While I think the point is invalid (logs are more descriptive), the joke is funny.
|
|
|
|
Fjordbit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 16, 2011, 04:39:37 PM |
|
I really disagree with the way Synaptic has presented himself in this debate, but I do agree with his basic premise. There are some real problems with FPGA as it stands right now. If things improve with respect to $/Mhps, then the equation changes, but I don't see how it's really viable. In fact, without a rise in btc price (or a drop in difficulty), I don't see how buying any hardware now is viable. Here is my math and reasoning (since Synaptic doesn't seem capable of producing any).
Right now the X6500 FPGA costs 610 and will deliver 180 Mhps. I have a spreadsheet I use to calculate out ROI and at least for me the ROI puts it at around 40 months. The problem is in about 15 months the block reward will be cut from 50 to 25. This actually makes the ROI 25 months longer. At the time of recast, there will actually be 50 months left. But wait, in that time, there is another block reward reduction. So assuming a constant difficulty and price, it doesn't seem that the current FPGAs can pay themselves off. People say that at lower btc prices FPGA make more sense but when I lower my btc price to $4, the FPGA ROI goes to 55 months. At $3, it's 76 months. This doesn't include the reward decreases.
Another factor here is that the hardware is all continuously improving. What do you think GPUs will look like in 40 months? What about other FPGAs? There's really no easy way to trade up FPGA hardware, but you can pretty easily trade up GPUs (sell on ebay, buy new GPU). When you buy a current FPGA, you lock into a specific amount of MHps. You could sell your FPGA to other miners, I suppose, but my feeling here is that this will more strictly follow performance relative to the latest chips because there isn't any ancillary use (gaming) for these boards. Because the hardware is used as well, you will need to discount it and it's possible that this discount will just take up any profit margin you had on that board. This is all theoretical, but it seems likely to me.
At these prices, GPU mining isn't much better. My mining rig is running at 60 months ROI. But then again I've been mining for the last two months and have been selling at some good prices, so when I take out what I've profited, it goes into the 30s. If I take out what I can sell my cards for on ebay today, it goes to under 10 months. So it still makes some sense for me to mine. But I can't recommend it to someone else unless they want to add a card to their machine for gaming and will mine while idle.
I don't think that people are being realistic about the price drop and the effect it has on long term profitability. Maybe people will begin to understand and this will result in a pull back in difficulty, but so far this hasn't been evident.
|
|
|
|
molecular
Donator
Legendary
Offline
Activity: 2772
Merit: 1019
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 16, 2011, 04:52:55 PM |
|
Right now the X6500 FPGA costs 610 and will deliver 180 Mhps.
I know some people here don't give a damn what ArtForz believes. Well, take this as his opinion then: (this was on 2011/9/16 on #namecoin) <ArtForz> I wouldnt be surprised if someone got 300Mh/s+ out of a S6-LX150
So the parameters of the calculations might change in the future.
|
PGP key molecular F9B70769 fingerprint 9CDD C0D3 20F8 279F 6BE0 3F39 FC49 2362 F9B7 0769
|
|
|
Fjordbit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 16, 2011, 07:08:55 PM |
|
Right now the X6500 FPGA costs 610 and will deliver 180 Mhps.
I know some people here don't give a damn what ArtForz believes. Well, take this as his opinion then: (this was on 2011/9/16 on #namecoin) <ArtForz> I wouldnt be surprised if someone got 300Mh/s+ out of a S6-LX150
So the parameters of the calculations might change in the future. 300Mh/s puts the ROI at 26 months, which is really 37 months given the reward decrease. In many cases a 37 month ROI would be great, but there is a severe amount of risk here. That is the ROI today at today's prices and today's difficulty. It's doubtful this will hold up for even 15 months let alone 37. And my main point is that as the btc price decreases, this ROI decreases too. People are saying that low prices are where FPGA shines, but it really just gets worse.
|
|
|
|
molecular
Donator
Legendary
Offline
Activity: 2772
Merit: 1019
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 16, 2011, 07:35:06 PM |
|
Right now the X6500 FPGA costs 610 and will deliver 180 Mhps.
I know some people here don't give a damn what ArtForz believes. Well, take this as his opinion then: (this was on 2011/9/16 on #namecoin) <ArtForz> I wouldnt be surprised if someone got 300Mh/s+ out of a S6-LX150
So the parameters of the calculations might change in the future. 300Mh/s puts the ROI at 26 months, which is really 37 months given the reward decrease. In many cases a 37 month ROI would be great, but there is a severe amount of risk here. That is the ROI today at today's prices and today's difficulty. It's doubtful this will hold up for even 15 months let alone 37. And my main point is that as the btc price decreases, this ROI decreases too. People are saying that low prices are where FPGA shines, but it really just gets worse. The argument is, that at really low prices (< $1), the difficulty (also part of the roi calculation) should drop significantly. It all depends on wether or not it really would drop at such low prices. I still hope we wont find out.
|
PGP key molecular F9B70769 fingerprint 9CDD C0D3 20F8 279F 6BE0 3F39 FC49 2362 F9B7 0769
|
|
|
Fjordbit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 16, 2011, 09:59:57 PM Last edit: September 16, 2011, 10:11:34 PM by Fjordbit |
|
The argument is, that at really low prices (< $1), the difficulty (also part of the roi calculation) should drop significantly. It all depends on wether or not it really would drop at such low prices. I still hope we wont find out.
I know the argument, and the argument is speculation. I don't think it's something to build a business model after. If difficulty does come down or price does go up or cost/Mhps does go down, then it can change the equation for these things, but making assumptions about what might happen is just gambling. And for the record, if the difficulty drops to 800,000 and the price to $3, then the X6500 has an 18 month (really 21 month) ROI. There is nothing stable enough about BitCoin that I would enter into it with an 21 month ROI on dedicated hardware. When I bought my rig, the ROI was just under 3 months. If the price goes to $2, the ROI is 28 month (41 months). This is using the 300 Mhps number which isn't guaranteed. There is nothing about low prices that is good for FPGA.
|
|
|
|
grod
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 16, 2011, 10:26:31 PM |
|
One other thing to put into your calculations.
The current board seats two FPGAs which cost $158 retail from Digikey. The power circuitry, and, really, the rest of the board should be manly enough to seat six or eight of those. Basically we're a small rework away from a $1400 board that could churn out as much as 1.8 gigahash (but today would yield around 1.4).
That's down there with 3x5830 rigs on a price/mhash, but would use 1/10th the power. Also, with 6 chips per board it'll be easier to hit quantity discounts (lop another 200 off). $1k for a 1.8 ghash board, anyone?
This could be done TODAY, with TODAY's pricing.
|
|
|
|
Nagle
Legendary
Offline
Activity: 1204
Merit: 1002
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 16, 2011, 11:39:27 PM |
|
$1k for a 1.8 ghash board, anyone? As a rule of thumb, figure a profitable retail price on electronics as 4x the parts cost. 2x in real volume, like a cell phone. What's your time worth?
|
|
|
|
grod
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 17, 2011, 11:25:03 PM |
|
$1k for a 1.8 ghash board, anyone? As a rule of thumb, figure a profitable retail price on electronics as 4x the parts cost. 2x in real volume, like a cell phone. What's your time worth? And yet, you can buy a 1 chip board for around $410, a two chip board for $610. Retail. Assuming digikey is a viable retail operation you can guess volume production costs using wholesale pricing could make the final product viable at retail cost of the parts. I won't deny this is not reality today. Reality today is 180 mhash for $610. Possibilities are there for much, much more however. Within a time frame much smaller than "bitcoin at $100."
|
|
|
|
Fjordbit
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 18, 2011, 06:36:15 PM |
|
The current board seats two FPGAs which cost $158 retail from Digikey. The power circuitry, and, really, the rest of the board should be manly enough to seat six or eight of those. Basically we're a small rework away from a $1400 board that could churn out as much as 1.8 gigahash (but today would yield around 1.4). This could be done TODAY, with TODAY's pricing.
The ROI on this is certainly impressive, but I don't really get how this can be used as part of a business model unless the product were offered on the market or the person running the business knows how to make this mod. I'm still not seeing your numbers though. If we use the current going rates of $210 for the main board and then $200 for each chip and each chip is at 90Mhps, then a 6 chip board is $1410 and does 540 MHps. The 1.8 Ghps board would be 20 chips, which would be $4210. If you got the chips for your retail price, it would be $3160. I think you skipped over something that shows it's possible today at today's pricing. Maybe you were using 150Mhps per chip (the 300Mhps rumor), but that's still $1896 for the chipset from Digikey, so I still don't get it.
|
|
|
|
molecular
Donator
Legendary
Offline
Activity: 2772
Merit: 1019
|
![](https://bitcointalk.org/Themes/custom1/images/post/xx.gif) |
September 19, 2011, 07:30:30 AM |
|
The argument is, that at really low prices (< $1), the difficulty (also part of the roi calculation) should drop significantly. It all depends on wether or not it really would drop at such low prices. I still hope we wont find out.
I know the argument, and the argument is speculation. I don't think it's something to build a business model after. I agree. I also agree on your further arguing that diff 800,000 / price $3 is not low enough.
|
PGP key molecular F9B70769 fingerprint 9CDD C0D3 20F8 279F 6BE0 3F39 FC49 2362 F9B7 0769
|
|
|
|