Ethereum GPU mining?

Page 28 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

MrTeal

Diamond Member
Dec 7, 2003
3,554
1,658
136
2. On taxation/sanctions - we are all from different places here, but - if I mine currency, I believe I will have to pay tax as soon as I sell the currency - it'll be considered under capital gains tax where I live. But regulation of virtual currency seems rather minimal - is this something you guys are concerned about? There have been criminal prosecutions in the USA e.g., of that teenager who posted the Bitcoin/ISIS guide. If I do not know who I am selling to, or buying from, am I at risk of circumventing UN Resolutions? Any news on virtual currency regulation?
Not necessarily.

https://www.cryptocompare.com/wallets/#/cards

You should be able to find a card that allows you to transfer the funds directly to it to use it as a plastic or virtual debit/payment card.

For example, I chose the first one for EUR:
AdvCash EUR
http://advcash.com/en/solutions/card/

How are they going to prove that 600 EUR on your "credit card" is taxable income? That's a lot of tax payer dollars that are better diverted towards tax evasion on Panama Papers. :D

I'm pretty sure he was asking what the regulations and tax implications are, not so much how he can get away with tax evasion. :p

@Aristotelian, with BTC it depends on your jurisdiction. From a couple guys I've talked to about their tax situation in the US (they were in Texas), the guidance they received was that the value of the Bitcoin is assessed at the time it is produced, and the more granular the data is the better. Easy way to do that is to just log it on a daily basis using a daily average price. That would be considered income, with (if you're running a business that can claim deductions) your costs in that period being an expense. When you go to sell it, any difference between what you sell it for and what its value was at the time it's produced is a capital gain/loss.

Now, I'm not an accountant or tax lawyer, and this does not constitute professional advice especially given it from a couple years ago in a field that moves at lightening speed.
 

Madpacket

Platinum Member
Nov 15, 2005
2,068
326
126
So tired...

Just finished milk crate miner #2

http://imgur.com/a/kpCwl.

Okay so mining overnight was a success on Milk Crate Miner #2.

Some comparisons between the two:

Milk Crate Miner #1 consists of 4xMSI 390's and consumes roughly 975-1050W on average from the wall. This using a very efficient EVGA G2 1300W power supply (Leadex Superflower desig). This using a Sempron 140 CPU and same model motherboard (different revision) and 4GB of RAM. Everything's undervolted / underclocked.

Milk Crate Miner #2 consists of 4xPowercolor 390's running on a EVGA 1050GS (Gold rated SeaSonic) and pulls roughly 900-1000W on average from the wall. If we assume 87% efficiency its sucking around 870W in total. I was a little worried going with the SeaSonic 1050GS over another Superflower design but I can put that to rest as I still comfortably have a 150-175W before fully maxing out the supply. Also the air leaving the power supply is just slightly warmer than air temp. This power supply is very well built and it cost a good $100+ CAD less than then 1300G2.

The MSI's crate miner is more heavily undervolted. -100mv on all cores but running but running higher core speeds of 1111 Mhz.

The Powercolor crate is running -87mv across all cores but only running at 1060Mhz across all cores. I haven't indivudally tweaked each card so there's more performance to be had yet.

All that being said I think I can safely assume at this point the Powercolor cards are a little more power efficient than the MSI's. The Powercolor rig is also running a more power hungry CPU and for whatever reason this revision of the motherboard (rev 5 instead of 4) won't let me underclock the CPU properly, I can only turn off all but one core and it still runs at 3.4Ghz instead of the set 2.0Ghz. Small difference though with all the undervolting going on.

As for performance they're pretty close but I'll have a better idea over the next few days looking at averages.

I also have 4 Asus 390's but I don't think I'll be building a third milk crate miner. The Asus cards seem to have the most consistent hashing but likely with better tweaking I can get the MSI and Powercolor's working better. All of this takes a a lot of time and I'm ready to work on a new project.

This has been a lot of fun but I'm kinda maxing out my power and available space in my house, also don't want the cops thinking I'm running a grow op :)
 

Staple

Member
Jun 9, 2013
30
0
0
Really nice builds here. I'm still amazed at the amount of money people are investing considering the value of eth/increasing diff not to mention power costs.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
Really nice builds here. I'm still amazed at the amount of money people are investing considering the value of eth/increasing diff not to mention power costs.

Power costs are dirt cheap in Ontario, Canada. On weekends, it's 8.3 cents per kWh. Averaging out the entire week works out to about $0.1075 CAD per kWh. Let's say 1kWh power usage with 4x390s is $2.58/day = $79.98 CAD a month. 4x 390s will make roughly 1.3 Ether a day @ $7.8 USD = $10.14 USD x 1.27 = $12.88 Canadian. Monthly revenue is $12.88 x 31 = $399 CAD.

Profit = $399.28 - $79.98 = $319 Canadian.

With 4x Polaris 10 cards, it may be possible to get a similar output with 1/2 the power usage. The profit would then be $399 - $40 = $360 CAD and chances are Polaris 10 cards will cost considerably less than an R9 390 costs in Canada. This is one of the major reasons I don't want to expand the operation until I see prices and hashing rate on Polaris 10 cards. If Polaris 10 costs $100-150 less per card than a 390 and has a similar hash rate in a 110W power usage, it's game time!*

If 4x 390s make roughly $320 CAD profit a month, if each Polaris 10 card costs $100 less, than waiting until 1st or 2nd week of June ends up similar to buying 4x 390s now and mining for a month or so. Of course if Polaris 10 prices suck and/or its output is crappy, then that's $319 or so lost in profits in May.
 
Last edited:

Staple

Member
Jun 9, 2013
30
0
0
Power costs are dirt cheap in Ontario, Canada. On weekends, it's 8.3 cents per kWh. Averaging out the entire week works out to about $0.1075 CAD per kWh. Let's say 1kWh power usage with 4x390s is $2.58/day = $79.98 CAD a month. 4x 390s will make roughly 1.3 Ether a day @ $7.8 USD = $10.14 USD x 1.27 = $12.88 Canadian. Monthly revenue is $12.88 x 31 = $399 CAD.

Profit = $399.28 - $79.98 = $319 Canadian.

With 4x Polaris 10 cards, it may be possible to get a similar output with 1/2 the power usage. The profit would then be $399 - $40 = $360 CAD and chances are Polaris 10 cards will cost considerably less than an R9 390 costs in Canada. This is one of the major reasons I don't want to expand the operation until I see prices and hashing rate on Polaris 10 cards. If Polaris 10 costs $100-150 less per card than a 390 and has a similar hash rate in a 110W power usage, it's game time!*

If 4x 390s make roughly $320 CAD profit a month, if each Polaris 10 card costs $100 less, than waiting until 1st or 2nd week of June ends up similar to buying 4x 390s now and mining for a month or so. Of course if Polaris 10 prices suck and/or its output is crappy, then that's $319 or so lost in profits in May.

But you have to look at difficulty. Even with a 20% decrease each month... you'd have a hard time maintaining profits in just a couple of months. If you just bought cards you'd never see profit.

https://etherscan.io/charts/difficulty
https://etherscan.io/charts/hashrate
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
The party is over when Proof of Work ends and Proof of Stake begins. Then it's a whole new ball game.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126

That doesn't make any sense. Even today cards make money and they will tomorrow. I clearly outlined the calculations too. At current rate it's easily possible to make over $300 in profits in May alone. Even if rewards halve, that's $150 in profits a month. Ok so $300 in May, $150 in June, $150 in July, $150 in August = $750 at least. There is no indication that rewards will be halved June 1st. I just checked my wallet => In the last 24 hours, 3 Hawaii cards made 1.09 Ether. So my estimate that 4 Hawaii cards make 1.3 Ether is actually lower than it would be if someone started tomorrow. Besides, deals on more affordable R9 280X cards pop up from time to time (i.e., $150 USD R9 280X).

A lot of us here will use the GPUs for gaming after crypto-currency fails. If someone is trying to do this to become a millionaire, ya it's not that type of gig. However, if I have to choose between a $300-800 AMD and an NV card and the former makes $, I'd be an idiot not to buy that one. As a result, as long as some high-end AMD card makes $ until the last day, NV cards make no sense. All I know is the naysayers have been around since 2008 and the train is still going.

The party is over when Proof of Work ends and Proof of Stake begins. Then it's a whole new ball game.

That's estimated to happen in more than 550 days from now.

"Ethereum to switch to Proof-of-Stake before November 1, 2017 This market resolves to Yes, if the Ethereum network switches from Proof-of-Work to Proof-of-Stake consensus as part of the Ethereum Serenity live release before November 1, 2017. The resolution date of this market may be accelerated. Bets matched after the antedated resolution date will be voided."
Source: http://www.coindesk.com/

That's enough time to pay off a Tri-Fire Vega 10 + i7 6700K/6800K/7700K PC. Worst case if everything fails, it's very easy to sell off GPU hardware. The reason I took a small pause is to re-assess the market around June but after I am probably buying more. There is plenty of time left to make $ imo.
 
Last edited:

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Well, technically it says before Nov 1st 2017 and that the resolution date may be accelerated. So nobody really knows. That's the longest possible time we'll have PoW. Some folks think it's happening in Dec 2016 or thereabouts. We won't know 'til we get there.
 

Staple

Member
Jun 9, 2013
30
0
0
That doesn't make any sense. Even today cards make money and they will tomorrow. I clearly outlined the calculations too. At current rate it's easily possible to make over $300 in profits in May alone. Even if rewards halve, that's $150 in profits a month. Ok so $300 in May, $150 in June, $150 in July, $150 in August = $750 at least. There is no indication that rewards will be halved June 1st. I just checked my wallet => In the last 24 hours, 3 Hawaii cards made 1.09 Ether. So my estimate that 4 Hawaii cards make 1.3 Ether is actually lower than it would be if someone started tomorrow. Besides, deals on more affordable R9 280X cards pop up from time to time (i.e., $150 USD R9 280X).

A lot of us here will use the GPUs for gaming after crypto-currency fails. If someone is trying to do this to become a millionaire, ya it's not that type of gig. However, if I have to choose between a $300-800 AMD and an NV card and the former makes $, I'd be an idiot not to buy that one. As a result, as long as some high-end AMD card makes $ until the last day, NV cards make no sense. All I know is the naysayers have been around since 2008 and the train is still going.



That's estimated to happen in more than 550 days from now.

"Ethereum to switch to Proof-of-Stake before November 1, 2017 This market resolves to Yes, if the Ethereum network switches from Proof-of-Work to Proof-of-Stake consensus as part of the Ethereum Serenity live release before November 1, 2017. The resolution date of this market may be accelerated. Bets matched after the antedated resolution date will be voided."
Source: http://www.coindesk.com/

That's enough time to pay off a Tri-Fire Vega 10 + i7 6700K/6800K/7700K PC. Worst case if everything fails, it's very easy to sell off GPU hardware. The reason I took a small pause is to re-assess the market around June but after I am probably buying more. There is plenty of time left to make $ imo.

Well, you're talking about four 390s...which is more than just for gaming. I was talking about people making rigs for mining with multi cards...I don't see people making back their investment let alone turning a profit at this point.

Four 390s ($300 each) plus another $300 for mobo/cpu and PSU (and a PSU would actually probably run more for four 390s) is $1500.

Four 390s would probably generate about $375 USD per month. If we assume a 20% reduction each month due to network hashrate/diff increase thats:

month1: $375
month2: $300
month3: $240
month4: $192
month5: $154

Total: $1,260

Four 390s drawing 1kw from the wall will cost about $72 USD per month if we assume $.1kw/h which is pretty common US pricing. So in 5 months that's $360 in electric costs.

At 5 months that puts it at $1,260 (and this ASSUMES that ETH doesn't drop in value at all) minus $360 in electric=

$900 total which is still $600 short of recovering the initial investment.

On top of that, in the US you are taxed on the value of the currency at the time of mining. So $900x.35 in taxes=$315.

So your real income after 5 full months assuming nothing changes at all with ETH value is actually...

$585

And the investment was nearly triple that at $1,500. And this is 5 months of watching the rigs and dealing with the heat during the summer etc.

So sure, if someone wants to buy a couple of gaming cards and mine when they're not using them I guess that's fine...but it makes no sense to build a mining rig at this point.
 

Madpacket

Platinum Member
Nov 15, 2005
2,068
326
126
Well, you're talking about four 390s...which is more than just for gaming. I was talking about people making rigs for mining with multi cards...I don't see people making back their investment let alone turning a profit at this point.

Four 390s ($300 each) plus another $300 for mobo/cpu and PSU (and a PSU would actually probably run more for four 390s) is $1500.

Four 390s would probably generate about $375 USD per month. If we assume a 20% reduction each month due to network hashrate/diff increase thats:

month1: $375
month2: $300
month3: $240
month4: $192
month5: $154

Total: $1,260

Four 390s drawing 1kw from the wall will cost about $72 USD per month if we assume $.1kw/h which is pretty common US pricing. So in 5 months that's $360 in electric costs.

At 5 months that puts it at $1,260 (and this ASSUMES that ETH doesn't drop in value at all) minus $360 in electric=

$900 total which is still $600 short of recovering the initial investment.

On top of that, in the US you are taxed on the value of the currency at the time of mining. So $900x.35 in taxes=$315.

So your real income after 5 full months assuming nothing changes at all with ETH value is actually...

$585

And the investment was nearly triple that at $1,500. And this is 5 months of watching the rigs and dealing with the heat during the summer etc.

So sure, if someone wants to buy a couple of gaming cards and mine when they're not using them I guess that's fine...but it makes no sense to build a mining rig at this point.

+1 listen to Staple ;D

We don't need any more miners :)
 

beginner99

Diamond Member
Jun 2, 2009
5,208
1,580
136
it makes no sense to build a mining rig at this point.

Note that I have not build a mining rig but what you forget is that a) for some this is a hobby and hobbies can cost and consume time and might seem idiotic to others (eg the tinkering). The later is why I don't do it plus heat in summer (no AC here...). Plus b) that many of the hard core miners are betting on ETH to rise over time. Be it now or in 5 years. If it takes of like bitcoin and is worth $500 a coin, you suddenly have a huge a mount of money.
 

poofyhairguy

Lifer
Nov 20, 2005
14,612
318
126
I am kinda scared to buy a card just for mining right now, but my plan is to buy every card I need this year (I need a Polaris 11 bad and maybe a 10 if it is better than a 390x at 1080p) as soon as they launch and not sell my old card until mining isn't useful.
 

Shmee

Memory & Storage, Graphics Cards Mod Elite Member
Super Moderator
Sep 13, 2008
7,381
2,415
146
I am DLing 16.4.2 now to give them a try.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
I am kinda scared to buy a card just for mining right now, but my plan is to buy every card I need this year (I need a Polaris 11 bad and maybe a 10 if it is better than a 390x at 1080p) as soon as they launch and not sell my old card until mining isn't useful.

I make mining very simple.

If I need to upgrade, AMD upgrade costs X - Y earnings a month, while NV card costs X - either 0 earnings or less than Y earnings a month. In that case, since AMD cards end up costing basically $0 over time and do so faster than Nv cards do, NV cards aren't an option until this changes. It's as objective as it gets. If you can get a free Mercedes, would you pay full price for a BMW?

The poster above me tried to position mining as a 1 way street - investment and tried to calculate that it's impossible to even break even. There is another way to look at it. For example, I had older PC parts and I wanted a full system upgrade. I sold my old parts to get a Skylake 6700K system. That means the $ spent on upgrading doesn't count towards system cost since I was going to do that anyway before my old Sandy Bridge DDR3 didn't dive into the ground with resale. Same with GPUs. Eventually they become outdated and need upgrading. The way I look at it is I will be upgrading no matter what -- so might as well get an AMD card that pays for a fraction of itself or for most of it. I've used this strategy since 2008 with HD4890 and paid $0 for GPU upgrades ever since. Obviously since the payouts are still very good, if I were building a new rig or upgrading, I would go for 2-3 Polaris 10 anyway. This is because they are making $$$ in the same system. For you think one Polaris 10 $299-349 will pay for itself in 6-7 months, then 2-3 Polaris 10 cards will pay for themselves in the same timeframe. The risk is extremely limited since even after 2 months or mining, you are already winning. For Canadians we also get an automatic 27% uplift in currency value. The only questions right now are Polaris 10's price, exact launch date and MH?

Otherwise, come June and it'll be X-Polaris cards XF vs. X-GP104 cards SLI. One makes more money than the other. That means if you buy NV, financially it's even worse. For that reason there is little logic that I've seen thrown around in the last 8 years. It's usually excuses from team green faithful. Their argument is they'd rather pay $1000 GTX580 SLI, $1000 for 680 SLI, then $1400 for 780 SLI then $1300 for 980Ti SLI vs. getting 'equivalent' AMD setups for free ... Because reasons. Start adding up all the GapU upgrade costs every gen, and it turns into thousands, more for multi-GPU setups. Believe me the same arguments we read for the last 8 years, even back when a single 4870/4890 made 0.5-1 bitcoin a day!

Once mining fails, I'll reassess and move on. But until then, every time I upgrade, the NV option is always more $, always.

TL;DF: As long as you are looking at mining as a bonus for planned upgrades in the first place, there is no risk at all. It's essentially AMD vs. NV where one pays for itself far faster. That means if you have 2-3 PCs with older cards that need GPU upgrades, AMD wins automatically. If you are thinking of building 2-3 or more mining systems in hopes to make lots of $, and those rigs were not planned upgrades or you didn't calculate the risks, that's completely different.

Note that I have not build a mining rig but what you forget is that a) for some this is a hobby and hobbies can cost and consume time and might seem idiotic to others (eg the tinkering). The later is why I don't do it plus heat in summer (no AC here...). Plus b) that many of the hard core miners are betting on ETH to rise over time. Be it now or in 5 years. If it takes of like bitcoin and is worth $500 a coin, you suddenly have a huge a mount of money.

Great points. It's like on the CPU sub-forum people who have 2008-2010 rigs are trying to argue that new Intel stuff isn't worth getting since it's too expensive for them. Yet, they don't want to do the work reselling old parts to help them offset the cost. If you don't want to put in the hours into your hobby, it's either not your hobby, or you make so much $, that your time is way more valuable. With PCs, many people upgrade regardless of mining. You can view it as "I want to mine --> need new hardware OR I need new hardware for my PC --> might as well mine on the side for fun when at work/sleep).
 
Last edited:

VeryCharBroiled

Senior member
Oct 6, 2008
387
25
101
[...] OR I need new hardware for my PC --> might as well mine on the side for fun when at work/sleep).

this is the reason for my r9 390.

havent gamed for ages. I recently decided to get back into it. my old 6870 in the daily driver (used for btc mining back in the day) couldnt cut it with modern games.

whatever I make in eth is off the price of the card.

no brainer.
 
Last edited:

Shmee

Memory & Storage, Graphics Cards Mod Elite Member
Super Moderator
Sep 13, 2008
7,381
2,415
146
Good points, and I expect the ethereum price to go up when PoS starts. Even rubycoin did this. Also, just a heads up, 16.4.2 still has the lesser hashrate, but it does have some good driver fixes for games, specifically with CF involved, when compared to 16.2
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
AMD needs to fix that problem with their OpenCL performance in newer driver sets. What gives?
 

poofyhairguy

Lifer
Nov 20, 2005
14,612
318
126
I make mining very simple.

If I need to upgrade, AMD upgrade costs X - Y earnings a month, while NV card costs X - either 0 earnings or less than Y earnings a month. In that case, since AMD cards end up costing basically $0 over time and do so faster than Nv cards do, NV cards aren't an option until this changes. It's as objective as it gets. If you can get a free Mercedes, would you pay full price for a BMW?

The poster above me tried to position mining as a 1 way street - investment and tried to calculate that it's impossible to even break even.

It doesn't make sense to not take into account the salvage value of the hardware. Even if the new GPUs are amazing all of this hardware will still have SOME value that can be recovered via selling it to someone.

Seems to me the math is:

(Y earnings + X salvage earnings when GPU is sold) - (Z purchase cost + W power costs) = Profit

The X factor of salvage profit is key, as one can expect the value of the mining hardware to drop over time until it hits some rock bottom point all GPUs hit. So for example, someone buying a 390 today might could get at least $240 of that value if they turned around and sold it, but if Polaris comes out and kicks Fury ass for $299 then that 390 might only be worth $160. Suddenly $80 worth of value is wiped out, which is about a month's worth of Eth mining profit for a 390. Holding onto a 390 for the month of May until Polaris comes out might not pay off once that decline in value is taken into account. Someone selling it a day before that drop gets to double dip though, if such things can be timed.

On the other hand there is no guarantee that Polaris will be great mining cards out the gate. On the contrary new AMD drivers (which the new cards will use) are worse for mining, and when looking at average hashrates it seems that the mining software is best optimized for GCN 1.0 and 1.1 so who knows how well GCN two removed from that will do. So we might see something like the 390 beating the Fury killing (in games) $299 Polaris 10 just in mining, which could extend the value of older hardware. Hell Polaris might not even work at all at first for mining or be severely gimped, which makes it hard to bet the farm on some near-term Polaris future.

The real problem is there won't be the wiggle room to act with caution. If Polaris 10 really is a 980 ti for $299 then it will be hard to buy at MSRP for months, and those who do get them will have to act like the Tesla pre-order people and just blindly buy the first card they can from the first batch and not wait for reviews or even someone to confirm mining works. By the time it's weeks later and we confirm mining works and we get full reviews you won't be able to buy one at MSRP unless Polaris is a dud or supply is unlike any card launch in modern history, which then has to go into the personal calculation.

All together what this made me realize is that I need to sell my 390x sooner rather than later to avoid the depreciation decline. If I can time it right I possibly don't lose much on that card in total, but I need to get it on Ebay before Computex to maximize resale value. That would leave me without any gaming card and open me up to the risk that Polaris sucks at mining or is very hard to find at first, which is why I bought a 280x for cheap last night. Not only will the 280x increase my total mining power for this month (before the difficulty gets even worse) but it will be a good enough gaming card for me to sit on until I actually get a Polaris 10 card. If Polaris 10 is a great miner I will sell the 280x and eat 50% of its value (which is better than 50% of a 390x and can be accounted for with at most three months of 280x mining), if it's a poor miner I will buy a second cheaper 28nm card and put together a dedicated mining rig out of parts I have laying around to keep the ball rolling.
 
Last edited:

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
^ That's a good strategy. It also means you keep mining with the 280X with minimal loss of mining efficiency over the 390X. I still think it's possible to mine with a 390/390X for another 2 or maybe 3 weeks before selling it. AMD could launch new cards at E3, not at Computex. Last year they decided to shift their focus to E3 with their R9 300 and Fury launches:

http://wccftech.com/amd-officially-launches-radeon-fury-x-nano-650/

This year the show is on June 14-16:
https://www.e3expo.com

Launching Polaris 10 at E3 has 5 benefits for AMD.

1) Focus new GPU launch around games. People buy GPUs for games so launching PC hardware at a game show could be viewed as even better than at Computex from a marketing point of view.

2) Allows AMD an extra 2 weeks to clear out stock on Tonga and Hawaii cards.

3) Allows AMD to see GP104 prices and performance around May 31-June 4th. This is huge because many times AMD got burned completely with NV launching after and dominating. This time AMD knows NV wants to maximize profits with higher end GP104. If AMD sees the lowest GP104 prices at $329, they will have almost 2 weeks to perfect their prices based on GP104 benchmarks. Imo, it's actually smarter for AMD to wait for GP104's reviews and prices because they lose 2 weeks of sales but gain a strategic understanding on where the competition lands.

4) By seeing in which games GP104 excels, AMD can work for 1.5-2 weeks straight to improve drivers for GCN 4.0 so that they can close any major performance gaps. First impressions count big for new GPU launches.

5) Since Polaris 10/11 are aimed at so many markets and mainstream segments, volume requirements to meet OEM demands would allow AMD extra time to build Polaris inventory.

I have no idea when Polaris 10 will launch though; just speculating Computex for GP104 vs. E3 launch for AMD.

As you said, it's pretty hard to time it. If Polaris 10 ends up launching only June 14-16th, that's an extra 1.5 months that a Hawaii card can mine.

Whoever on here gets the first Polaris 10 and posts MH numbers and power usage stock and overclocked will do us a big favour indeed :)

I do agree though that if you are worried about resale value of a 390X, probably safe to sell it in May. What's interesting to note is that prices of 390/390X cards have hardly budged and we are just days away from May. Then again, NV waited until the week of 970/980 launch to start major price drops on 770-780Ti SKUs.
 
Last edited:

SimianR

Senior member
Mar 10, 2011
609
16
81
As much as I want to be confident in Polaris 10/11 and its ability to be competitive with NVIDIA's offerings, I think whoever launches 14nm chips first will see some huge gains in marketshare, however temporary those may be. It's been pretty obvious over the last couple months that AMD has cleared some inventory because of the Ethereum mining craze, unless sales due to mining continue to surge, AMD should probably cut it's losses now and start clearing what inventory it has left instead of waiting till E3 to launch (assuming they even have the ability/inventory of Polaris chips to launch before E3 - something tells me that it won't be clearing inventory that delays them from launching it).
 

poofyhairguy

Lifer
Nov 20, 2005
14,612
318
126
^ That's a good strategy. It also means you keep mining with the 280X with minimal loss of mining efficiency over the 390X. I still think it's possible to mine with a 390/390X for another 2 or maybe 3 weeks before selling it. AMD could launch new cards at E3, not at Computex.

I figure we will get leaked benchies long before any concrete announcement, and via those benchies it will be easy to approximate what kind of decline I can expect for my 390x. One good thing is I checked Ebay and almost every auction for a 390x the last month has gone for $35 more than I paid, so I have a lot of wiggle room until a Polaris 10 card hits Newegg.

The 280x I bought for $30 more than what I sold my 7970 Ghz last year for, but hey what are you going to do? Looking back the move from that 7970 Ghz to a 970 was the worst GPU upgrade of my entire life if you count the lost mining time.

As you said, it's pretty hard to time it. If Polaris 10 ends up launching only June 14-16th, that's an extra 1.5 months that a Hawaii card can mine.

Isn't Computex like May 31st? That makes for a two week difference if I get close to that date, which isn't such a massive loss of time. What scares me more than anything is a sharp difficulty rise for mining, that is why I am getting the 280x now to maximize what I can do in May.

Whoever on here gets the first Polaris 10 and posts MH numbers and power usage stock and overclocked will do us a big favour indeed :)

I am planning on buying one of the first ones available no matter what cards I have sold or not sold. I need a 750 ti replacement BADLY for my Mini ITX rig, so I plan to get both a Polaris 10 and a Polaris 11. The only thing that will stop me is a huge letdown for Polaris 10, like it matches a 390 for $200 kind of letdown.

What's interesting to note is that prices of 390/390X cards have hardly budged and we are just days away from May.

It is VERY obvious mining is affecting price. I get the Newegg deals email daily and they were so proud of a $299 390 in the recent one, aka a GPU that would have gone for $15+ less for the last six months. This gives me hope that second hand prices will hold even after mid-May (for example) Polaris 10 performance leaks because people caught up in mining might not be paying attention to gaming.

Really it's all a bunch of speculation, my game is to keep my costs low and get as much as I can while I can. I have read enough about Ethereum that I feel it maybe has a bright future, so really all I care about is mining coins right now for less than what I can buy them for some future payout. At the current rate on the 390x with my power costs I am at around $2.25 per Eth before counting hardware, and I want to bank away about 50 at least before mining isn't doable on GPUs. That means I have about $200-250 bucks of wiggle room to be lost via hardware depreciation, extra power use, etc. before I would have been better off just buying the Eth at current market rates. The only way I see myself going far past that amount is if I screw up and put money into places that cannot be recouped easy, like getting some 1000w power supply I will never use in a post-mining era.

The way I see it the 280x is a much better deal for mining than the 390x ever was (about one half the cost for 2/3rds the mining power) so I will just buy another if Polaris 10 can't mine and still be ahead of a single 390x.
 
Last edited:

Madpacket

Platinum Member
Nov 15, 2005
2,068
326
126
I went a little crazy on mining gear for Ether however most of it was bought with BTC earnings so for me a reinvestment of earnings was made. This allowed me to take a little more risk so regardless of the performance of Polaris 10 I will continue to mine until it becomes unprofitable for an extended period of time or PoS is established. To me there's more risk in selling my cards prematurely and hoping Polaris will work properly out of the gate and availability will be adequate. I suspect it'll take a bit of time for the developers to get things working right on Polaris considering the OpenCL app was designed on a 7850 and hasn't been updated in ages.

Sure the price of ether today is low but it has a good chance of shooting up given the current maturity level of the platform. I would kick myself if I sold all my mining gear prematurely and then ether shot up to 20 or 30 dollars later this year.

The first DAO platform is launching by a capable company "slock.it" likely in the next few days so we should see a lot of practical DAPPS being made which I suspect will payoff big time if they get the DAO working as intended. We should see a bunch of new use cases for the platform emerge as incentive to build good DAPPS will be driven by DAO's.

Philips is now adopting Ethereum Blockchain technology for the medical industry which is a good sign the platform has legs outside of commerce (where BTC rules) and that's just one use case. They are using it for storing medical records to reduce a huge number of silos and redundancy. A global medical database with much better security and integrity built in. Cool stuff.

We should all be pretty excited on what this Blockchain tech will allow, even outside of an investment vehicle.
 
Last edited:

wege12

Senior member
May 11, 2015
291
33
91
Recently when I start my Mist wallet, it will attempt to download all the newest blocks but it just sits there not downloading a single one. I can't seem to figure out what the issue is. Any suggestions?

Thank you