Mining with GPUs for GPUs?

Have you heard about GPUcoin (GPUC)? It's an interesting idea that could have been a killer and made GPUC mining something worth doing... except the whole alt-coin market is going soft, and with less profitability comes lower demand for GPUs, which leads to lower prices. In short, after peaking at prices of $420 for R9 280X, $550+ for R9 290, and $700+ for R9 290X, we're starting to see a return to sanity on GPU pricing. So let me start off with some quick GPU pricing links that might be of interest:

Radeon R7 240: Newegg starts at $65
Radeon R9 270: Amazon starts at $209; Newegg starts at $179
Radeon R9 270X: Amazon starts at $229; Newegg starts at $199
Radeon R9 280: (Can't find on Amazon); Newegg starts at $289
Radeon R9 280X: Amazon starts at $329; Newegg starts at $349
Radeon R9 290: Amazon starts at $439; Newegg starts at $459
Radeon R9 290X: Amazon starts at $569; Newegg starts at $579

The reason those prices are important is that we can now buy some of those GPUs using GPUC -- but how much does it end up saving...or costing!? Here are the current GPUC prices for the R7 240 (don't bother...), R9 270, and R9 280X. Yes, the R9 280X they have on offer is a nice Sapphire TOXIC, but performance won't be all that different from other R9 280X GPUs if you know what you're doing. Right now, you'll have to pay about 32 satoshis per GPUC in order to buy enough coins for a GPU (though that could obviously change). The good news is that you could now build a capable miner for about $1500 instead of $1900, though at the current rates you'd still be looking at four months or more to break even on the investment. But what about buying graphics cards with GPUC?

[Note: Prices and difficulty for GPUC are jumping around like jackrabbits, so take the following as a snapshot in time rather than a "this is the way it will always be" statement.]

• R9 280X: 2,520,301 GPUC = $470. OOPS! The exact same GPU is in stock at Newegg for $399.

• R9 270: 1,227,839 GPUC = $229. Oops again. Yeah, we're still talking pricing higher than Amazon/Newegg right now, by $20 to $50.

• R7 240: 174,482 GPUC = $32.50, so that's a pretty good price.

My guess is the people at GPUcoin ordered right before prices on most AMD GPUs dropped. The faster GPUs end up costing more than you'll now pay at retail, while the R7 240 wasn't as desirable for mining so you can pay less than retail. Unfortunately, R7 240 will only hash at around 100KH, so while it would be $75 less for the GPUs than buying retail, you're maxing out your motherboards with only 600KH or so. Meanwhile, the minimum cost of a mobo, CPU, PSU, RAM, storage, and case is going to be around $350, so you're spending $550 total for 600KH where a complete $1500 rig might do 2100KH (i.e. better price/performance overall).

The only way this gets worked out and GPUC thrives is if they can go and purchase a bunch more GPUs at well below the MSRP (which would mean below $299 for the 280X). Then they can take a bit of a loss on the ones they already purchased and end up delivering GPUs at a reasonable price. It would actually be super cool to go buy new hardware with GPUC (or BTC), especially if the price was below normal online pricing, but that would take a lot of work and a big warehouse to store inventory I suspect.

I should note that the GPUcoin store also has an open market where users can sell items. This is getting closer to the ideas above, but it's more like eBay with additional risk -- you send your GPUC, then have to wait/hope you didn't get scammed. And if you don't receive the product you orders, TOO BAD! Mwahahaha.... Yeah, I'd be a little leery at the least. If you're curious about pricing, though, someone is offering a 2MH scrypt rig for 10M GPUC -- that works out to about $1850. Buying the same hardware online (5x R9 270, ASRock H81 mobo, 4GB RAM, Intel Celeron CPU, 1000W PSU, 5 x16 risers, and a power button) will set you back about $1500. So sadly, no, there are not good deals to be had there right now it seems. And if you actually want to buy 10 million or more GPUC fast, you'd probably pay at least 25% more than the price I've just quoted.

As an alternative, you can already go through Gyft.com and buy a $10-$2000 Amazon.com gift card with Bitcoins, and then you could buy GPUs that way -- or anything else you might want to buy at Amazon. (And you get 3 Gyft points back for paying with BTC, compared to 2 points with Paypal or 1 point with credit cards -- basically a 3% "cash" back reward.) See the problem with all these new coins claiming to improve on the existing options? Some will likely succeed, but there's a lot of "reinventing the wheel" going on.

If you're struggling to find good coins to mine, my newsletter is always available -- and a quick side note to subscribers is that if you haven't received today's full newsletter, send me a note and I'll make sure you get it (or potentially let you know your subscription is expired). Another alternative is putting your rigs up for rental on LeaseRig where often there are people willing to pay a bit more than what you might mine normally to borrow your rig and chase some new coin (or boost hashing power for their new pool, etc.)

There was plenty more information that only went out to subscribers, along with the above commentary; want to see what you're missing? Start with a 2-day $5 trial, and if you upgrade to a week ($10) or month ($30) I'll count your initial payment towards the total! Let me sweeten the pot for all subscribers as well with this offer: when I hit 100 active subscriptions (I have fewer than 50 right now), I'll start sending out a daily newsletter with current charts to all active subscribers. I'm also looking into other ways to provide data in a real-time fashion (i.e. sharing a Google Spreadsheet). In short, the more people subscribe, the more it becomes worthwhile for me to spend time on the newsletters.

Standard Miner Profitability, One Crazy Weekend, Subscriptions

I've just sent out the latest copy of my newsletter, which covers the aftermath of this rather crazy weekend for cryptocurrencies. What happened and who were the winners and losers? Subscribe and find out! $5 for one issue, $10 for a week (typically three issues, but really however many I write in seven days), and $30 for the month of posts. Email me if you have questions. I'll even make you a deal: start at either the one day or one week subscription, and if you decide you want the full month I'll count your initial donation towards the $30 -- so $25 to go from one letter to the full month, or $20 to go from one week to the full month.

I've been working on a more formalized posting of how to subscribe, and at some point I need to get a link to it put in a prominent position on the blog, but for now donations of one of these coins should suffice. You can do the math to determine roughly how much $5/$10/$30 is, or go with the rates I list below (which might end up being too cheap/expensive if the prices change much):

BTC: 153qS9Ze32hnV3fwirZLWNka4wBAowc21E (0.01/0.02/0.05 BTC)
DOGE: DD9iTWf8diPkvKdB8roPJepTyp6BGVQtct (6500/12900/38500 DOGE)
DRK: Xd3EaCJg6G8ZnGuKkpvwyRMwyHzbaRDnob (7/13/39 DRK)
LTC: LfCLyykrNFftzpdWejR73hf478ZtBzQ9jE (0.32/0.64/1.9 LTC)
MAX: mf5DXTLiZFCnJC2x13MXSyigyUjmBnrwjG (27/53/158 MAX)
MRC: 1Ctnz6cHcMYiF9fz2pyd6orFuo1mDhKdWj (810000/1620000/4840000 MRC)
VTC: VaNuRCj73JVAwR1YMnt8CXaqoiPgykiMTk (4/8/22 VTC)

Moving on, I've listed roughly the same PC quite a few times on my blog, but I'm going to go ahead and call this my "baseline mining PC" for now. It consists of the following parts:

Baseline Mining System

GPUs	3 x Radeon R9 280X 3GB	$1,200
Motherboard	Gigabyte GA-990FXA-UD3/UD5 AM3+	$137-$166
Processor	AMD FX-8320 Vishera	$154
Memory	2x4GB DDR3-1600 (desktop)	$77
Power Supply	2 x Rosewill Capstone 750W 80 Plus Gold	(Two) $200
Storage	2.5" 60GB Kingston V3 SSD	$59
Risers	3 x x16 to x16 and/or x1 to x16 powered	(Three) $13/$9 each
Case	Build it out of wood or PVC pipes	~$40
Total (price range)		$1906-$1935

My recommendations for mining are based off of that particular build, and the goal has mostly been to try and choose coins that can pay that off in four months or less. Note that it's possible to add a fourth GPU quite easily, and the two 750W PSUs have enough power to do that, so technically you could spend about $2320 for a system that's 33% faster than the above and it would end up being ab it more efficient. Anyway, right now, the market is in a bit of a slump, so the best I can do as around $13-$15 per day. That means it will take more like 4-5 months to break even, provided prices don't drop -- but prices could also jump up again.

Consider LTC; Litecoin launched and went up to around $4.50 per LTC last April, then it quickly fell to the $2-$3 range and stayed there for much of the year, with a drop down to as low as $1.11 at one point. Many were proclaiming the end of LTC...and then November happened and LTC shot up to $48. Those who had strong hands and held until the bubble came out with some amazing profits; even if you sold at $20-$25 (which is what I did with the LTC I still had), it's hard to complain. Well, we're now repeating that with most cryptocurrencies. Sure, prices are higher on LTC and other coins, but so is the difficulty, meaning for every dollar spent on a GPU you're looking at three months to earn it back. Save the coins, though, and I suspect in another year many people will be looking back thinking, "I wish I had kept all my VTC until it hit $50!"

Anyway, if you want to join in the fun, at least some of the prices for hardware have come down. The same system as above last month cost around $2020, so $100 less now. GPU prices are still way over MSRP (by 30% or more), which oddly enough appears to be more of a US phenomenon. I have friends in Europe that are able to get GPUs for less than what I have to pay, where normally it's the exact opposite. Their power costs are still a lot higher, so I suppose that's a small consolation.

Email Subscriptions - Sample and Introduction

I mentioned in the previous post that I was considering trying out an email subscription model for some of my content. I got some takers on the idea -- thanks to all of you! -- and so I composed and sent out my first email today. To give you an idea of what would go into the email subscriptions, I am going to post the majority of that email here -- minus the specific mention of coins as well as my "What to Mine" table showing the calculations and the analysis. If you'd like the full email, just drop me a note and a donation ($5 or equivalent in any of the major cryptocurrencies for now will do -- if needed, I can get you an address for any other coins you might want to donate).

---

Welcome to my first official subscriber email on what to mine. Disclaimer: this first email might be (okay, *is*) a bit longer, and I'm trying to decide on a good subscription rate. I said "donate $5 or more and I'll tell you what I'm mining", but I'd like to hear your thoughts on this. Should I have individual, weekly, and monthly rates? (I don't want to do yearly, as I can't guarantee I'll be doing this consistently for that long!) If so, how often should I update, and what should the rate be? I'm sort of thinking three updates per week might be a good start -- Monday, Wednesday, and Friday. I could do $5 for a single update, $10 for the week, or $30 for the month. Does that sound like a reasonable structure, or should I charge more/less? If you would, please email me back with your thoughts. Also, there's a lot of other "stuff" to discuss as I kick things off; I'll add some of this to the blog, minus the "what to mine" table and lists of specific coins. And with that out of the way....

I already emailed a few people earlier this past weekend, but of course prices, difficulty, etc. change regularly. The short summary is that right now, the most profitable coins tend to be those that are not using standard scrypt. This is hardly surprising, considering there are well over 100 "major" (i.e. on an exchange) scrypt coins now. The trick is figuring out which of the alt-coins are going to be worth something in the long run. My thoughts are that for a coin to truly succeed, it needs to be something different/new/useful. LTC is the first scrypt coin, and it's still doing okay (though clearly having trouble of late). VTC, MAX, DRK, XPM, QRK, etc. were all really great to mine early and hold. Sadly, I'm not even sure which coin was the very first scrypt-jane offering (YAC?), but hopefully you get the point. Don't underestimate the power of marketing either -- a "dumb" coin with the right marketing hook can take off in a big way; it might die out just as quickly, but the momentum of belief can be slow to change (DOGE, AUR, and other coins fall into this category).

Now, as I see it there are a couple ways of going about mining. One is to look for the truly special coins and hop on those and hold for a while -- it's what I've done with VTC, DRK, and to a lesser extent MRC. Truth be told, I don't particularly care for the rewards structure of MRC, but so far it's doing okay; the coming month will or so will be telling as the rewards will start to decline in a few days, and price will either stay where it is and/or stagnate, or the "decreasing value" of mining will start to increase the price. It really can't go much lower, as it's currently trading at 3-4 satoshi, but we'll see. If it goes down, I "lose" potentially half to three quarters of my MRC value, but if it goes up to 10-20 satoshi (which is my hope!) I could realize some nice profits. The other option is to simply try to mine any new coin, hold a chunk of them for a while to see if price goes up, and you'll get some winners and a lot of losers -- or you could try to cash out as fast as possible if the coin is listed on an exchange, and potentially make some decent returns. This requires a lot more time and dedication, so most of us don't go that route I think.

I suppose I should quickly put in here that a "satoshi" is termed as 0.00000001 coins. The code for Bitcoin originally defined that as the minimum value you could trade, and if BTC prices go up into the tens of thousands or more perhaps they will one day update the client to allow for smaller transactions, but for now that's where we are. All of the other coins have stayed with the same core code structure in this area, so the smallest amount you can buy/sell of any coin is one satoshi.

Going with the above, that means any currency traded against BTC has a minimum value of 1 satoshi BTC (around $0.0007). That's not much, but with coins like DOGE, MRC, FLAP, etc. issuing billions of coins, and mining bringing in millions of coins per day, that quickly adds up. I have to admit that I'm still shocked that DOGE ever reached 50 satoshi, let alone 280 (or the current 170 or so satoshi). If any of the "billions minted" coins I'm holding hit even 50 satoshi, I'm looking at potential gains of 20X or more. But I think DOGE is a bit of an anomaly, so I wouldn't count on a repeat performance from most clones. If you're interested in trying "day trading" of cryptocurrencies, you could also try to do the old buy low/sell high approach on some of these low-value coins -- if you can buy at one satoshi and sell at two, you double your coins every time the buy/sell cycle repeats. However, the exchanges fill bids/asks in a first come, first filled basis, so as an example if there are already bids for a billion CAGE at 1 satoshi and another billion asks for CAGE at 2 satoshi, you could be waiting a long time for your orders to get filled. (And of course some coins trade against LTC or DOGE as a way to overcome the 1 satoshi minimum issue.) And if a coin ever reaches the point where no one is interested in paying even 1 satoshi per coin, you're left holding the bag.

With all of the above out of the way, let me note that I'm only one guy, and I already have a full-time job, so I can't hope to track everything. I've wondered if perhaps I could build a large enough readership/subscriber base that I could actually make this a full-time job and perhaps invest more time/effort into some programming and such to automate pulling data for more coins, but that's a big pile of work to tackle right now. In short, I'm going to tell you what coins I *have* tracked in a second, but there's a very real (and even very likely) chance that I've missed something good. If you know of something better and want to give me a heads up, that would be great, and if if I agree I'll even give you a free week of email updates, but if you want to keep your "trade secrets" I won't blame you. :-)

So what am I tracking (and in some cases, mining)? Basically I have a bunch of coins where I've downloaded the wallets, synced up, and can reasonably easily check difficulty and pricing. That in turn let's me determine profitability. It's like Coinwarz, except I track most things manually -- and I can also track the newest coins. It is a bit of a pain, and actually mining on any of the coins requires registering on a new pool (unless solo mining is viable, which is typically only in the first hours of a coin's life), creating workers, editing your miner configuration files, etc. Again, it's a pain, so I usually only switch what I mine daily, and there are quite a few coins where I downloaded the client and perhaps even registered with a pool and did some mining before I decided, "Nope -- this coin is going nowhere." Here's the current list, with exchange prices, difficulty, and reward values current as of (around) 11:00AM PST, March 3...

[SNIP!]

---

The missing content includes a table of the coins I'm tracking (and felt were worth including) along with several more paragraphs of analysis. And as a final comment, if you're buying any hardware and Amazon has prices that you like (and you're in the US), shop there via my Amazon Affiliate Links and let me know what you ordered (and when) via email. For every $100 ordered via Amazon, I'll credit you with the equivalent of $5 towards these email updates -- so $200 ordered gets you a week at my current pricing, $600 would get you a month, etc. You can also shop via my Newegg Affiliate Link, but the commission there is so low (1%) that I can only give $5 credit per $400 ordered. Then again, that's just one R9 280X so it can't hurt. :-)

Rest assured, the email updates won't replace me writing on this blog. There are still many other topics I want to tackle, when I get the time. Hopefully that will be later this week....