What this guy said.

I just did a quick comparison against poclbm for me.

On my dedicated 5870:
poclbm     SDK 2.1: ~424
phatk       SDK 2.1: <420
phatk imp SDK 2.1: ~432

poclbm     SDK 2.4: <424
phatk       SDK 2.4: ~424
phatk imp SDK 2.4: ~437

So on SDK 2.1 your improvements made it so phatk was better than poclbm in 2.1, and way better in 2.4.

The init optimizations gave me a minor boost of ~0.5 MHs over 2011-07-11.

As someone said, this is just 3001 of 61017 accounts. About 1700 of all accounts were unsalted.

It is very obvious, if you do the maths, that this cannot have been brute-forced.
A 15 character mix has a pretty good length advantage, purely on the number of combinations. And since there seems to be quite a few of that kind and above, I'll base calculations on that.
The most basic mix of characters would be numbers and lowercase letters, for a total of 36 different possibilities. That gives a 15 char long password a total of 36^15 or 2.21*10^23 combinations.
Since the bitcoin network is a very good indicator of hashing power, if we translate the current capacity (12.571 Thash/s) to only 5870 cards, that gives 31292 @ 400 Mhash/s.
The best numbers my cards would give was 3.9 billion combinations/s each, which gives the bitcoin network a total power of 122567 billion hashes each second, or ~1.23*10^14.

Simple mathematics then gives: 2.21*10^23 comb. / 1.23*10^14 comb./s = 1.80*10^9 seconds, or 57.2 years.
To get anything realistic for brute-forcing, they would need 21.4 million 5870 cars, which brings it down to one month for a single 15 char length password with only numbers and lowercase letters.
One of the passwords was 24 letters like that, which would take longer than the calculated age of the universe using the bitcoin network!

Conclusion, this is either malware/virus, some form phishing attack or, though unlikely, hash collision.

I find it hard to believe they brute-forced my password, along with all the rest, as it is long and secure.
A good password should be at least 15 alphanumeric characters, which at 1 billion comparisons a seconds takes 7 million years to test all combinations. It would take a humongous amount of computing power to crack that in a few days, even if you split it up amongst tens of millions of machines.
And that's just for one 15 character length password, and each character adds 36 times the number of combinations.
If you're using non-alphanumeric characters, like @,$ etc it takes exponentially longer to crack.

A hashing calculation: at one billion hash comparisons a second, a 15 character alphanumeric password will take roughly 7000000 years to compare all combinations.

No, running the result through the hash function again is a very common way to make it safer.
What you basically do is, if they can figure it out, is force them to do multiple hashes for every comparison. This is a direct counter to brute-force attacks.
If they have to spend one second for EVERY combination comparison, then it will be basically impossible for them to brute-force anything.
There are even specific algorithm which are designed just for this.

If you take the above calculation, if they can only do one comparison per second instead of a billion, then it will take 7000000000000000 years instead to test all combinations for one password.

Dual salts would also be good, where one of the salts set in source code. If only the database is compromised, the passwords are safe.

I'm not that worried, my password is quite long and secure.

Make sure the config looks like this for MSI Afterburner:

[ATIADLHAL]
EnableUnofficialOverclocking   = 1
UnofficialOverclockingEULA   = I confirm that I am aware of unofficial overclocking limitations and fully understand that MSI will not provide me any support on it
UnofficialOverclockingMode   = 1
AccessibilityCheckingPeriod   = 0

And then when you hit the underclock limit, restart the application. It will then position itself at the previous clock as the middle, and you can downclock further.

260W @410MH/s on a single 5870.
Effective energy usage though is 150W as I only mine on my regular computer, which I use all the time anyway. And it's idle usage is 110w.

Getting a second 5870 later today, so will update with new numbers then. But my guess is that It also will be adding another 150W @~400 after I modify clocks.

89 degrees feels like it's getting near the edge of what's safe, so you should downclock the memory on both cards to save power and lower temperature.

Switch the master cards again, and the application should update the slider for the new clock setting.

What temperature does the cards reach? If there is a case of overheating, the card will donwclock itself automatically to prevent damage, lowering hashrate in the process.

Removing the crossfire bridge and using dummy plugs would be better, as then you can clock them individually using for example MSI Afterburner. That's how I plan to run my dual 5870 setup.

If the AC is of portable kind, where you have a hose that expels hot air through the window, it will be a win-win since this also creates a negative pressure, pulling cool air from the rest of the house.

Otherwise if the house is cool, then it's actually any way that's good. The question would be if you want to be able to control the temperature in the room yourself, or want to rely on the temperature of the rest of the house.

The fans would create a lower pressure in the room blowing the air out, resulting in air being sucked in from somewhere. Most likely it'll be from other parts of the house that's also hot.

An AC would be the best bet, since it actually cools the air and blowing it in, not merely shuffles air around.

You should underclock the memory. My 4870 went down 10 degrees by lowering from 900 to 190.
And if you hit the sweet spot, you'll gain a few MH/s. For me it's at 190, for 5xxx I've read 1/3 of core.

Try to use an extra fan to help move air away from it.


That only helps if the air can be moved away, otherwise the temperature will still build up.

What damages a card is too much heat, which wears the components out. Cards have automatic temperature limits at where the cards underclock to prevent damage.
The other thing that can damage cards is too much voltage. While more voltage leads to more heat, voltage itself can be too much for the components to handle, literally frying them.

So underclocking puts less stress on components, and actually increases their lifespan. And as have been said, the thing that can happen is that the drivers malfunction, and have to be reset.

And it can also happen because the actual chip is one of the lesser ones in a batch.

I can suggest you try to overvolt the cards a small amount, if you have enough leeway temperature wise, to see if it helps in stability when overclocking.


If they indeed are that bad, then you should definitely recommended you use the 1200w PSU in your gaming rig.

I have seen both 1 and 2, and apparently the difference is that 1 allows the clocks to change if the card runs idle, for automatic downclocking, while 2 forces it to stay at the designated clock.