Take it easy, I was mostly kidding about the house getting burned down

Very few people appreciate how much magic there is to that metal box with wires.
Most go something like "Huh? Is there actually any magic in that sad little gray thing?"

80 PLUS TITANUIM certification has been released for 220V.
It mandates that the PSU does not waste more than 4% power at 50% load. That's 96% efficiency.
Whoever is able to get their design certified, will have to have cast insane amounts of magic putting that PSU together
Link: http://www.plugloadsolutions.com/80PlusPowerSupplies.aspx

Good catch P4 but that should be inconsequential!
Provided the protective circuitry worked correctly, the unit should have seen it was in trouble and should have powered down cleanly.

See my edited previous post for some PSU suggestions.
I'm suggesting the XFX model (which I own) but if you do your research, you'll find other PSUs based on the same Seasonic-designed platform.
Armed with the basics of how PSUs work, you should be able to find even better units, built on some even more recent designs.

Elementary school science here:
Volts * Amps = Watts,
therefore:
22A * 12V =  264W
30A * 5v = 150W

This PSU of yours has a terrible layout for mining:
First 12V rail provides power for all modular cables and mobo connector whereas the other for the hardwired CPU connectors.
Therefore, if your GPUs are going to draw more power than that measly 264 Watts, you're going out of spec.

Warning, warning: when Hardware Secrets editor tried overloading this particular PSU, it BURNED DOWN on them.
Full review here: http://www.hardwaresecrets.com/article/Enermax-Liberty-DXX-500W-Power-Supply-Review/303/8

When looking for a small (550W) mining-grade power supply, I'd recommend you try this baby: http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=225
All this unit's +12V juice goes through just one rail, which could give you up to 530 Watts of power provided the minor rails (5V and 3.3V) aren't loaded.
Here's a review of the 650W and 750W models: http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=216
I've been running the XFX 750W PSU for 8 months and have absolutely nothing bad to say about it.

Here is a 10 page long article explaining some of the PSU-fu: http://www.hardwaresecrets.com/article/327
I suggest everyone interested in learning some PSU basics follow this link I posted and familiarize themselves with the basic concepts.
Take the time and you won't burn your PSU, your rig, or your house.
I'll be checking up on this thread and will try to answer any questions.

Mother of God, why? Are you aware what you're asking for?
Not only wasn't your Xoom designed for this kind of abuse and will likely overheat, cpuminer is heavily optimized for x64 and x86 architectures and you want to run it on an ARM device?
That's rubbish.

My UPSs fall flat on their face when asked to support my mining rigs. Only my network infrastructure and work stations are protected.
Anything capable of sustaining a 800 Watt miner is going to be ridiculously expensive.
Fortunately, none of my machines has that Biostar bug.

That power button extender you linked to simply allows one to use any network cable for trivially extending the power button.
I'm thinking of using the router's GPIO to achieve the same goal. That would require user ssh login to the router and running a script.
I'm gonna have to disable my SD-card mod as well (the SD card uses the GPIO lines for data transfers).

No promises and don't hold your breath. My rigs are built using Gigabyte mobos so this will be a low-priority task.

I have all the necessary parts, I can build one just like it. Combined with a hacked, dd-wrt enabled router...
I might work something out over the Weekend  

Jesus H Christ, that's what you're getting when buying Biostar?

I can think of Mc-Gyvering a solution based on a few basic electronic components.
A 220V relay will trip on power loss triggering the clock signal generator. An IC waits 15 seconds during which time the said relay is cut off from external power. After 15 seconds, the relay is released.
For reliability's sake I'd actually use two relays in parallel so that if one of them acts up your mining rig isn't rebooted.
Does anyone have the motivation to actually build such a doohickey?

If you never heard of netcat, you'll be glad to learn about cryptcat Kano.

Cryptcat uses Bruce Schneier's Twofish encryption algorithm to secure the connection. Sweet.
If you forget to provide a passphrase, the hard-coded default is "metallica". How can one NOT love open source software?

And how should an automated script know that a user had just made a mistake, pray tell?
The user's mind is a PRIVATE class without any interface, the exchange can't see whether what you did was what you meant to do.
The exchange should therefore NOT assume the role of order police, checking all orders to see whether they might or might not be a mistake on the user's part.

EDIT:: Do you think you could fix those nasty typos in the thread title?

Not so fast, Cowboy
http://msdn.microsoft.com/en-us/library/windows/desktop/ms682586%28v=vs.85%29.aspx

Pay close attention to the folowing point:
If a DLL with the same module name is already loaded in memory, the system uses the loaded DLL, no matter which directory it is in. The system does not search for the DLL.


Copying the file into application directory is not enough if the library from SDK 2.6 has already been loaded into memory. System reboot is mandatory.
Only then can you guarantee that the correct library will be used.
There are of course ways to forcefully unload a library without rebooting the machine but they are geeky, thus hard to recommend to the general public.

Robert, this "another program" as you put it, 7-zip, is what currently happens to be the best free, open source, multi-platform archive manager available.
You'll be glad to have it installed so stop being so grumpy

Shame on Microsoft for not having touched their built-in compressor/decompressor for ages.

Bullshit spread by individuals who forget to shut up when they have no idea what they're talking about.

A hd6770 (juniper-xt) is basically nothing more than a re-branded 5770 with some minor bios tweaks completely orthogonal to mining. So are 68xx cards (barts).
VLIW4 starts with cayman (hd6950).
There is a difference in shared count but nothing as drastic as people make it sound.

EDIT:: I was asked to post an example
At stock clocks:
hd 5850: 725 MHz * 1440 SP = 1,044,000
hd 6870: 900 MHz * 1120 SP= 1,008,000

The stock 6870 is about 3,5% weaker. That's not actually a bad choice, especially considering that it can be purchased new with full manufacturer's warranty.
Anyone buying a used card risks spending their money on a device having a only a month of life left in the tank.

People associate usernames with certain individuals, not with user_id or date_registered values.
With your approach however, the implication that a username translates into any particular person is false.

How can I tell that eg. ArtForz is actually the same individual who I think he is?
I'd need to check the Date Registered value. Good luck expecting anyone to go through so much trouble.

I already provided an example, let's elaborate on it:
(1) Assume that a well-respected member (again, let the example be ckolivas) decides to remove his account.
(2) A malicious user reuses the freed username.
(3) Other forum users are unaware that their PMs and donations are targeted at the correct account name but the wrong person.

If you still don't see the point, fine. Do account management the 1990s way if you so please.

Some outstanding cgminer-fu from P4! The force is strong in you, my man.

I used this approach to find the maximum acceptable clocks when upgrading my home server into a half-decent mining rig.
As I like to have full control over the GPUs, once I found those clocks, I set them in cgminer's config file, then set the overheat temp at target_temp +3°C.
This way, if any of the fans goes belly up, the card will aggressively attempt to cool itself down and the noise will bring the issue to your attention.

Those would be POWERED pci to pcie extenders of course, Chefnet?

The spec allows up to 75W for pci-e but no more than 25W for good-old pci. Supplying external power is therefore mandatory.
Just wanted to bring this potential point of failure to attention.

5xxx series availability issues? How about going with 6770 cards then?
They're very cheap VLIW5 cards, priced at slightly over 2/5 of what a 6950 costs here in my neck of the woods.

I've been running a couple of Sapphire 6770/512 MB cards for 7 months with no failures, each of those little beasties overclocked to 995/300 without any overvolting resulting in slightly over 230 MHash/card. Originally, I clocked them at 1000/300 but that would result in a couple HW errors/day (no hangs). 1010/300 was unstable.


I couldn't help but laugh and pick the "BFL PROFESSIONAL product" in your survey meaning should BFL ever release a professional product, go for it

As to receiving free power as part of one's benefits, hello you Americans, welcome to Europe
I'm just wondering, what do the benefits of coffin manufactury workers are? I'm betting oblong.

Thanks for clarifying that rjk. I admit I was being too vague.
There still are implications though... you're giving up some of the control to a third party.

Never-ever trust a client admitting to storing private keys server-side, not after the MyBitcoin fiasco!



For anyone interested in the matter, here is the RFC document establishing a standard for the transmission of IP datagrams on avian carriers http://www.faqs.org/rfcs/rfc1149.html
The following extension is even better introducing QOS! http://www.faqs.org/rfcs/rfc2549.html

On a more serious note: if you're in the middle of fraking Amazonian Forest, a thin client might be the solution for you. The blockchain remains in the cloud and does not have to be downloaded and synced.
Naturally, there are security implications to this approach as nothing in life is free.

While it is true that the RFC for IP over Avian Carriers never tackled environmental issues like your -20°C weather, heavy snowstorms, hunting season, and the like, our ground-breaking proprietary extensions to the protocol should make the delivery possible: we send your bird carriers with FedEx!

Implementing out protocol extension, the transmission link will be established...
Unless it won't, see this guy's video: http://www.youtube.com/watch?v=6wpkbYL7wFA&feature=related
That may or may not have been one of our ostrich messengers sent out to deliver an urgent marriage proposal made by one of our clients via Internet Messaging.
126 bytes sent, 0 received, and we never saw that bird again. As a matter of fact, I don't recollect hearing from that particular client again either...

Should your organization prove unable to implement our extension for any reason, we also offer reliable, if somewhat slower, solutions based on the vanilla version of the standard.
Please use the mii-tool-avian to reduce the desired link speed and we'll send penguin carriers right away.
Persistent buggers, those penguins. Just make sure to have enough fish when your blockchain is received