These have been sitting on a shelf since June-July. I don't know if they came in the same shipment that we're using in production right now, but so far we haven't hit one that we ordered that wasn't unlockable.

I don't want to oversell this aspect of it, please don't buy it if you're going to come back yelling at me that they won't unlock to 4 cores. I'm selling them as 2 cores, and I can't determine what they really are capable of without ruining their "new in box" state.

Be very careful, none of their security/ratings logos are clickable. In the upper right corner, there's a McAffe SECURE logo. Real ones are clickable and go to the McAffe site to show the results. That one isn't clickable. Searching for them on McAffe's site doesn't return anything. Their resellerratings.com and "top rated by google" images don't go anywhere either.

They also claim to have 488 5970s in stock (you can put 488 in your cart, but 489 says you exceeded their stock limit). It's kinda unlikely a tiny store like this could afford to stock a quarter million dollars of a single video card.

We'd love to buy more of these (we're using cards for non-bitcoin related opencl work, bitcoin is what i'm doing in the off-hours), but every time I've looked at that company they've seemed too shady. I'd love to be proved wrong though, I'd buy a ton.

Opened 5770s and all 5830s are sold out now.

Current price is $8.84 from the last trade i saw. That's $159.12. Priority mail shipping will be ~$10, so that leaves me with $150 or so. the wiki lists price as $140. The cheapest new one I see on Amazon right now is $180 with 1 left.

So I don't think the pricing is all that bad, especially considering it's almost impossible to find now, brand new, and I'm accepting bitcoins.

Updated prices again, and marked that the fans are sold out now.

Okay, but last time I'm doing this. And yes, not everything on the list above is in the picture, it's still on the dock right now.

See this thread where I sold several 5770 cards and this thread where I sold USB drives in bulk as proof that I'm not ripping people off.

I have the following goods available for sale.

Sapphire Radeon 6990 4GB 100310SR



New in box, never opened, full manufacturer 2 year warranty applies.

Radeon 6990s are sold out everywhere, and brand new cards are becoming impossible to find.

90BTC - I have 10 available 6 LEFT Free shipping to anywhere in the US. Outside the US, please contact me.
That's only $702 at current exchange pricing, significantly below retail, and shipping is included!


Sapphire Radeon 5830 1GB 100297-2L

This is one of the most best bang for the buck cards you can get, 285Mhash is easy to get. This card is almost impossible to find now as well, being guilty of probably higher performance per dollar than AMD intended.

Newegg page for this product

New in box, unopened.

18BTC - I have 2 available SOLD OUT Free shipping to anywhere in the US. Outside the US, please contact me.



Samsung Series 5 3G Chromebook in Arctic White with carrying sleeve

New in box, never opened, also with some kind of designer carrying sleeve accessory.

Amazon is selling these for $449

Includes built in 3G modem, where you can sign up for a no charge 100MB/month plan for 2 years. The laptop itself is nicely hackable, you can drop to a shell and play to your heart's content.

40BTC - I have 1 available Free shipping to anywhere in the US. No point in selling outside the US, since verizon won't activate it and it's locked to them.
That's roughly $312 at current exchange pricing.


XFX Radeon 5770 cards



You should get somewhere between 185 and 220Mhash/sec with these depending on your cooling and luck.

I sold a bunch of these in a previous thread and we'd like to clear out our remaining stock now. Amazon is currently selling them for $134.

Unopened brand new cards:

9BTC - I have 10 available SOLD OUT These boxes are just barely too thick to fit into a large flat rate priority mail box. If you don't mind me taking the card out of the box and repacking them, I can do free shipping within the US. If you're outside the US or if you want them totally pristine in their retail boxes, PM me for a shipping quote.

Opened, tested and used for a few weeks. You get the card and accessories, but no box:

7.5BTC - I have 5 available SOLD OUT AGAIN Free shipping to anywhere in the US. Outside the US, please contact me.



MSI 890FXA-GD70 Motherboard

MSI's page about this board

Easily run 4 double-wide cards per motherboard. Easy overclocking, and allows unlocking hidden cores on X2 and X3 CPUs to make them X4.

Unopened, brand new.

22BTC - I have 3 available 2 LEFT NOW Free shipping to anywhere in the US. Outside the US, please contact me.



AMD X2 250 3.0GHz CPU

With the above motherboard, we've had 100% success in unlocking all 4 cores using the above motherboard, but I can't guarantee your results. With all but one system, we've been able to convert these to X4 processors and overclock to 3.4GHz or higher.

Unopened, brand new, comes with AMD heatsink/fan.

6.5BTC - I have 2 available 1 LEFT NOW Free shipping to anywhere in the US. Outside the US, please contact me.



If you're interested in any of these, please PM me for a unique address to send them to, and include your shipping address. If the exchange rate changes, I may alter the prices above, but once you've paid the price is locked in. If I receive your payment while I'm still at the office, your order goes out in the next business day, with USPS priority mail tracking number. Bitcoin payments only, no paypal or similar. Lets get the economy going by actually using BTC.

i0coind also seems to have a bug regarding it's NTP implementation.

One of the changes between i0coin and bitcoin was that it was trying to keep the clock more accurate by polling an NTP time server every 5 minutes. It looks like the code to do this isn't very well thought out, and will hang if the ntp server it randomly picks doesn't respond. It doesn't retry, and it doesn't give up after a while, it just freezes.

Since it's choosing a different NTP server each time from a random pool of volunteer NTP servers on the internet, it's not improbable that one of them will stop responding occasionally. Eventually all iocoind processes will hang due to this.

ipv6 allocations aren't made according to geographical region, and even then once a block of IPs has been assigned to an ISP/hosting company/corporation/whatever, they're free to use that block anywhere they want.

There are regional "registries" that are responsible for handing out blocks within their region (there's a north american registry called ARIN, one for europe called RIPE, etc) but as long as you have some tangentially related business presence within their region you can request space from any of them. They get assigned large blocks that they break up and give to people (like us) who request them. But once they're assigned, we can use them globally and unless we choose to make it obvious where a specific address is being used, it's very hard to tell.

For example, my company has a large single ipv6 allocation, which we've subnetted into giving addresses to our customers in the US, Europe and Asia. People will make databases of where they think addresses are by bulk WHOIS requests and trying to decipher hostnames in traceroutes, but that's no different than how ipv4 works now.

Further example: ARIN was given 2001:4800:* through 2001:49FF:*, which they broke up handed out to a bunch of different companies/ISPs. You can tell that one of those blocks came originally from ARIN to some company with a connection to North America, but just because the blocks came from that space doesn't mean the end user who got assigned that address is in the US or anything.

For us (slightly altered IPs so nobody gets any funny ideas):

IPv4: We have 99.5.16.0 through 99.5.32.255 assigned to us. 99.5.17.14 might be in the US, but we might have routed 99.5.22.(anything) to our datacenter in Amsterdam.
IPv6: We have 2001:4840:0000:0000:0000:0000:0000:0000 - 2001:4840:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF. Some of those addresses might be pointed at our datacenter in the US, some in Tokyo.

You can't tell just by looking at the numbers, and if we want to make it hard to tell we could.


Also, with the P2P nature of bitcoin, you don't need to connect to every person you want to send money to. As long as you can connect to someone, who can connect to someone, who can connect to someone who can eventually get the transaction to the recipient, it still works. If you're in the US and the government compiles a mostly accurate database of all Chinese IPs and blocks them, unless EVERY OTHER COUNTRY blocks china as well, your transaction will still get to China.

What you're seeing is the result of one "share" being roughly as difficult to find as the block itself. (1 share = 1 block exactly, when the difficulty was 1, a tiny handful of shares = 1 block when it was 4, etc). It's basically a race to be the first to get new work, and submit an answer than everyone else is trying to submit at once.

The difficulty is at 16 right now, and is about to flip over to 64 in a few minutes. Once that happens things will start becoming a lot more stable. When it hits 256 in an hour or two, things should be pretty close to normal. Just be patient!

(I haven't earned a share from i0guild either, but this was pretty much exactly what was expected to happen)

If you are using the CLI, can you show me what "getinfo" says? what block count are you up to on your client? Or if you're using the windows GUI, it should say at the bottom. Are you seeing any confirmations yet?

(trying to figure out if there's actually a fork or not)

True, but if you go far enough back they're showing full confirmations on blocks that don't match what my client sees. I think they've forked off into their own world entirely.

Either btcguild is lagging really far behind, or the chain has already been forked going considerably far back. They're still showing themselves as winning every block, which I can see isn't the case. Their block count from their pool server is also lower than what my client is showing.

I think what I mentioned a few posts above yours. Before your block was fully confirmed (>100-120 confirmations), an alternate block chain (probably BTCguild's own view of the winners) managed to get more confirms than the chain your block was in.

And for those of you wondering why a single pool getting >50% of the network's hash rate, consider this an example of what they could do.

I'm just watching because I was curious to see how the network/protocol would behave in this situation.

It looks like from look at btcguild's block page, they believe they're finding ALL the blocks from their perspective. In periods where they're finding >50% of the blocks, they're effectively coming up with their own block chain where they're the winner to everything. When they don't manage to muscle out everyone else, they end up losing their whole forked chain.

Is this right? Anyone with more experience with how the block collision scheme works could probably answer this better, but it's kinda interesting to watch the debug log showing the battles.

I do still have 5 unopened cards left, and a pile of opened cards left. I've been talking to a couple of people about buying the whole lot of them for a bulk deal, but nothing finalized yet (busy with my day job).

If you're interested in buying one, send me a PM and I promise I'll respond within the next 24 hours.

Unlikely, but if I do I'll make a post here and send you a message.

7 unopened cards left now, and a handful of opened cards.

Last chance before these go up on ebay!

I dropped the prices on these further to try to tempt one of you into buying the whole lot!