Artificial Value of Older AMD/ATI GPU's

[dlasher]

Disclaimer: I say this as someone who has many many Ghash in AMD GPU's...... I'm in the same boat if I ever try to sell my cards..

In my humble opinion, when people put mining hardware up for sale, they need to understand both their audience, and some reality about old hardware.

1. Sellers/Buyers : People here are going to buy for the investment value. Buyers looking at the potential earning of the hardware between now and December. No one is speculating further than past both the reward drop, and the potential next-generation hardware (asic/minirigs/etc)... there's no magic 8-ball. If a buyer can't make money on the hardware between now and December 1st, it's not worth buying. (Whether or not you believe BFL can deliver a 40Ghash box for $1300, you have to factor in the possibility they might be telling the truth... 40Ghash boxes for $1300 mean ebay flooded with worn out GPU's... and if 1Thash boxes come online in October, difficulty increases will stop GPU mining WELL before the reward drop..)

2. Cards : When you're listing 58xx cards, you're talking about 3-4 year old GPU hardware.. and not just 3-4 year old cards, but cards that have endured 10-years worth of GPU load in the 3-4 years they've been alive.... rode hard, put away wet... and likely zero warranty. Speaking also as a gamer, no gamer is going to buy a 5830/5850/5870 or even a 5970 when the 78xx/79xx series cards are, here, warrantied, lower power, lower heat, and, faster at games.

If you're selling 3Ghash worth of a mining hardware, and you're asking $3000, you're dreaming...

• 3 G/hash nets you approximate 1.5btc per day.. there are around 165 days left till the drop.. That's 247 btc, before power..
• Most people are going to spent at least $2 a day (or .3 btc) to power that 3ghash, which eats out 50-ish BTC from that 247 btc.. (That's cheap power btw)
• .. leaving you with around 200 btc... at $6.5 that's $1300 best case
• the person buying the hardware doesn't want to break even, they want to make money.. so that box isn't worth $1300, it's worth what ever amount less than the risk the person wants to take. $1000 maybe?

The MiniRig, which is now a shipping product, has a $0.61/M/hash upfront cost, with a fraction of the power cost (Approx 16kw in GPU's for 25Ghash vs 1400w for a Mini-rig.).. Are you really saying your hardware is worth more $ per M/hash than a Minirig?


If you have hardware up for sale right now, please don't take this personally.. but take a moment to think...This is the reality of the now..

[tosku]

I'm not saying you are wrong, but something like dual HD5850:s still packs quite a punch for gaming.To a gamer, they are still worth something. Not the original value, of course, but something:

The chart is taken from here: http://www.guru3d.com/article/radeon-hd-7970-crossfire-review/13

[dlasher]

I'm not saying you are wrong, but something like dual HD5850:s still packs quite a punch for gaming.To a gamer, they are still worth something. Not the original value, off course, but something:

You are correct.. I'm not saying they are completely worthless.. I am saying they're not worth more than $0.61/mhash, best case, and really more like $0.30/mhash...

[crazyates]

You can still pay off any investment in a 5xxx card between now and the reward half. After about the end of July or mid August, I don't see anyone adding any more GPUs into the network (because at that point, the investment wouldn't pay itself off unless you had free electric).

And ya, I still game on a 5870. You'd be surprised what they can do, even on new games.

[eroxors]

Your "humble opinion" sounds more like a rant... if you are so unhappy about the asked prices, offer something lower or don't buy... You're essentially thread crapping on half of the computer hardware board.

[Jezzz]

Some corrections to your math ...

3ghs will earn closer to 1.748 btc per day at the current difficulty.  Not 1.5 btc.  For a total of 286.67 BTC.

Also, your power assumption is quite high.  I pay $0.11/kwh.  Which translates into $0.96 per day or 24.37 BTC over 165 days @ $6.5 per btc.  Not 50 btc.

So using these figures you get ...
Total income after power:  262.3 BTC ($1,704.95)

This is $0.57/mhash

From there you can deduct for the profit the buyer needs to make for the sale to make sense.  But it isn't $0.30/mhash.

[Definit]

maybe this guy is wanting to sell all his stuff for a low price?

sounds like a steal to me.

[dlasher]

3ghs will earn closer to 1.748 btc per day at the current difficulty.  Not 1.5 btc.  For a total of 286.67 BTC.

According to BitcoinX calc, 3000 mhash at present difficulty is indeed 1.7477 BTC a day. However, that assumes non-PPS mining. PPS mining takes anywhere from 3% - 10% off the top depending on where you mine  PPS mining, downtimes, ddos, etc, and I think 1.5 btc a day is a more realistic expectation to average over that time period.

Also, your power assumption is quite high.  I pay $0.11/kwh.  Which translates into $0.96 per day or 24.37 BTC over 165 days @ $6.5 per btc.  Not 50 btc.

Sorry mate, your power calc is way off.

power cost = ((wattage at socket * 24) / 1000 ) * kw/h rates

3000 Mhash of GPU will run you somewhere around 1500W at the socket, depending on whether you split it across 1 machine or more. 1500W at the socket, at $0.11/kw/hr is about $4 USD per day (not counting service delivery fees) .

((1500W * 24) / 1000) * 0.11) = $3.96/day, or about 0.6BTC a day, is 100BTC over the 165 remaining days.

So using these figures you get ...
Total income after power:  262.3 BTC ($1,704.95)

This is $0.57/mhash

Using your #'s for production (which I think are high) and your numbers for power (which are also a bit high) it translates into 286 btc minus 100btc for power, or 186 BTC.... 186 BTC x $6.50 = $1209. (which is worse than the initial #'s I posted)

[dlasher]

Your "humble opinion" sounds more like a rant... if you are so unhappy about the asked prices, offer something lower or don't buy... You're essentially thread crapping on half of the computer hardware board.

I apologize if it seems that way... math is math is math.

FPGA's are here, ASIC's are coming.. it's a new day.

The GPU had a good run, but it's almost over. Either mine the heck out of your GPU's, which likely already paid for themselves, or sell them for a price that reflects their new value... but their new value is not equal to FPGA's....

[crazyates]

3ghs will earn closer to 1.748 btc per day at the current difficulty.  Not 1.5 btc.  For a total of 286.67 BTC.

Taking into account future difficulty rises, I don't think 1.5BTC/day is unreasonable.

Also, your power assumption is quite high.  I pay $0.11/kwh.  Which translates into $0.96 per day or 24.37 BTC over 165 days @ $6.5 per btc.  Not 50 btc.

And I think your power assumptions are low. Extremely low. 3GH/s would be what, 4x 5970 at stock volts? 7x 5870 at stock volts? You're talking at least 1kw for your 3GHs rig (and that's being extremely conservative).

So if you're getting 11c/kwh, how are you gonna mine on a 1kw rig, and only pay 96c electric? My math for 24 hours @ 11c/kwh on a 1kw rig is $2.64. A 1.1kw rig is almost $3 a day.

My math = 67 BTC for the full 165 days. However, like the OP said, you can expect to mine ~250BTC in that time period. As I stated before, you can still come out profitable, but it depends on a variety of factors.

For example, I've purchased some of BFL's ASICs, but Ive also recently got some 5970s for dirt cheap, so I'm gonna make another rig where I mine at a colleague's datacenter where they get near free electric (1.5c/kwh). By the time the reward half comes, I still plan on recouping my costs put into the rig, and making a profit.

Buying a 4x 5970 rig for $1300 just to make $1200 (after electric)? Not worth it. Buying a 3x 5970 rig for $700 to make $1100 (after electric)? Worth it.

[sadpandatech]

3ghs will earn closer to 1.748 btc per day at the current difficulty.  Not 1.5 btc.  For a total of 286.67 BTC.

According to BitcoinX calc, 3000 mhash at present difficulty is indeed 1.7477 BTC a day. However, that assumes non-PPS mining. PPS mining takes anywhere from 3% - 10% off the top depending on where you mine  PPS mining, downtimes, ddos, etc, and I think 1.5 btc a day is a more realistic expectation to average over that time period.

I am at 4.2GH and at a 5% PPS pool am pulling 2.34BTC per day.

edit; point is that is 1.755 at 3GH at a 5% poo. So jezz is right, atleast on that aspect.

[Jezzz]

Sorry mate, your power calc is way off.

Agreed.  Not sure what I did there.  $3.96/day is accurate.  My apologies. 

Though I'm not sure why you prefer to use PPS as your btc/day rate when there are better alternatives.

[dlasher]

Though I'm not sure why you prefer to use PPS as your btc/day rate when there are better alternatives.

Right you are..  but that's yet another unpredictable in this equation. If you're trying to decide "should I spend any more $$/btc on GPU mining" one should try to work with known factors. >100% PPS pools are nifty, mostly profitable, and ... volatile?

[dlasher]

I am at 4.2GH and at a 5% PPS pool am pulling 2.34BTC per day.
edit; point is that is 1.755 at 3GH at a 5% poo. So jezz is right, atleast on that aspect.

You're right.. I tend to round down, since, so does life, at least to me.

To be exact, 95% of 1.7477 is 1.660 BTC per day, not 1.5.

1.660 x 165 days = 273.95 btc

0.60 x 165 days = 99 btc for power

Still leaving you will 174.95 BTC over the period, or $1137.18 @ $6.50 per BTC.

Not hardware worth paying even $1200 for if you're trying to make money off the purchase from coin earned alone.

[sadpandatech]

I am at 4.2GH and at a 5% PPS pool am pulling 2.34BTC per day.
edit; point is that is 1.755 at 3GH at a 5% poo. So jezz is right, atleast on that aspect.

You're right.. I tend to round down, since, so does life, at least to me.

To be exact, 95% of 1.7477 is 1.660 BTC per day, not 1.5.

1.660 x 165 days = 273.95 btc

0.60 x 165 days = 99 btc for power

Still leaving you will 174.95 BTC over the period, or $1137.18 @ $6.50 per BTC.

Not hardware worth paying even $1200 for if you're trying to make money off the purchase from coin earned alone.

hehe, I'm making the numbers I quoted 'after' the 5%. so it would be 1.755 after 5%.
Sorry, we are kinda getting off track of your original idea with that but just wanna make sure all our numbers are on the same page.

With the current market value of my hardware if I bought it today, I am at 173 days before electricity. Not bothered to recalc the elec because I paid off myself a long time ago. Last check though I believe I was paying less than $2 per BTC..(new citation needed on my watt usage as that was when BTC was just over $5)

edit; it's ~255 with electricity. Just under .11 per KWH here.  edit; Should also note my situation may not be typical as I am undervolted all to heck.

[superfastkyle]

ugh 1000w is not 1000kwh at 24/7 its 720kwh for 30 days. Why is it so hard for people to understand this? So 1500 watts cost $118.80 to power a month at 0.11 cents a kwh

[dlasher]

ugh 1000w is not 1000kwh at 24/7 its 720kwh for 30 days. Why is it so hard for people to understand this? So 1500 watts cost $118.80 to power a month at 0.11 cents a kwh

Right you are.

daily power cost = ((wattage at socket * 24) / 1000 ) * kw/h rates

((1500W * 24) / 1000) * 0.11) = $3.96/day = $118.80 for 30 days.

[crazyates]

ugh 1000w is not 1000kwh at 24/7 its 720kwh for 30 days. Why is it so hard for people to understand this? So 1500 watts cost $118.80 to power a month at 0.11 cents a kwh

I'm a little confused, because no one ever mentioned 1000kwh in their math...