Mining at 7566.57 MH/s... should I go solo, etc?

[ensign_lee]

Even big pools like Slush's pool still experience variance. BTCGuild has grown rapidly to > 125 Ghash/sec and it still experiences rounds ranging from 11 seconds to 12 hours.

Why would a 7.5 Ghash/sec user not benefit from reducing his variance like the rest of us? He'll still experience a fair amount of variance?

[1713481120]

1713481120

[1713481120]

1713481120

[1713481120]

1713481120

[1713481120]

1713481120

[1713481120]

1713481120

[1713481120]

1713481120

[bahnfire]

@Clavulanic: I am running these stock with fan cooling only. I am expanding watercooling to all soon (the watercooled machine is 400MH/s per GPU and running around 60'C).  Stock (which mine are, except for 2x cards) are expected to be at 300Mh/s, so I do not understand why I should be way above 7.5GH/s.

If you have any suggestions on how to "fine tune" my operation - let me know. It is always easier to tell someone else to destroy their $800 card instead of your own :-)

[Bitsinmyhead]

Push cards higher...
Make more BTC...
Buy new cards if cards start getting destroyed...
Still would make more than running at those girly stock speeds...
Those cards built for 80C+...
I get almost 1/3 of your Mhash on my 3x 5970 setup...

[JayC]

@Clavulanic: I am running these stock with fan cooling only. I am expanding watercooling to all soon (the watercooled machine is 400MH/s per GPU and running around 60'C).  Stock (which mine are, except for 2x cards) are expected to be at 300Mh/s, so I do not understand why I should be way above 7.5GH/s.

If you have any suggestions on how to "fine tune" my operation - let me know. It is always easier to tell someone else to destroy their $800 card instead of your own :-)

Where to begin...  first Linux, I'd install it, phoenix, and use 2.4 SDK and phatz kernel.  I'd experiment with underclocking the memory on the card, you should be able to find a sweet spot, and at the same time reduce heat production (you don't need the memory speed for mining, by reducing it, you don't hurt anything on the card).

I would expect you should be able to get 8.4+ Gh/s without having to overclock the core.

[Xer]

@Clavulanic: I am running these stock with fan cooling only. I am expanding watercooling to all soon (the watercooled machine is 400MH/s per GPU and running around 60'C).  Stock (which mine are, except for 2x cards) are expected to be at 300Mh/s, so I do not understand why I should be way above 7.5GH/s.

If you have any suggestions on how to "fine tune" my operation - let me know. It is always easier to tell someone else to destroy their $800 card instead of your own :-)

Where to begin...  first Linux, I'd install it, phoenix, and use 2.4 SDK and phatz kernel.  I'd experiment with underclocking the memory on the card, you should be able to find a sweet spot, and at the same time reduce heat production (you don't need the memory speed for mining, by reducing it, you don't hurt anything on the card).

I would expect you should be able to get 8.4+ Gh/s without having to overclock the core.

Phatk and memory underclocking does 0 or negative differance on the 69xx gpus. do read up some before spitting out that.

[JayC]

@Clavulanic: I am running these stock with fan cooling only. I am expanding watercooling to all soon (the watercooled machine is 400MH/s per GPU and running around 60'C).  Stock (which mine are, except for 2x cards) are expected to be at 300Mh/s, so I do not understand why I should be way above 7.5GH/s.

If you have any suggestions on how to "fine tune" my operation - let me know. It is always easier to tell someone else to destroy their $800 card instead of your own :-)

Where to begin...  first Linux, I'd install it, phoenix, and use 2.4 SDK and phatz kernel.  I'd experiment with underclocking the memory on the card, you should be able to find a sweet spot, and at the same time reduce heat production (you don't need the memory speed for mining, by reducing it, you don't hurt anything on the card).

I would expect you should be able to get 8.4+ Gh/s without having to overclock the core.

Phatk and memory underclocking does 0 or negative differance on the 69xx gpus. do read up some before spitting out that.

I have 6950's running and did see a slight bump in using phatz (4 or 5 Mh/s), and underclocking the memory doesn't hurt performance in my experience and does seem to help reduce heat.  You notice I didn't state that he would see a Mh/s bump from reducing the memory clock, only that he should be able to find a sweet spot where it doesn't hurt Mh/s and does reduce heat.

[Jaime Frontero]

linux, yes.

and take the MemClocks down, to reduce heat.  with AMD CC 11.5 and SDK 2.4 (required with 6xxx series cards for best performance), lowered MemClocks may not increase the M/h rate (might even lower it by a percent... OH NOES!), but it will dramatically reduce heat output.

try DiabloMiner - it's just as "quick to test and play with", if not quicker; and usually faster.

with linux, use AMDOverdriveCtrl for clocking - it's not as crazy powerful as some of the windows utilities, but it'll do; and you're much less likely to take the card past a point where you can damage it.

if you're really set on spending major dough to put water-cooling on all those machines, then do  it smart - get longer tubes, and use external (not case-mounted) radiators.  hang the radiator and fan bundle outside a window on brackets.  then you won't need to spend anywhere near as much on AC.

[rezin777]

Question - why does common concensus seem to be that he should go and mine solo wiht 7.5 Ghash/sec?

It's not like his volatility doesn't lessen a ton like most of the rest of us, and a even a 2% fee (there are several with lesser fees) isn't a huge cut. Why wouldn't he want to mine in a pool?

1: The Bitcoin network is much more robust if it is distributed as it was intended to be, instead of being centralized into several large pools.
 
2: Any fee is huge if you have that kind of mining capability, and it most definitely cuts into your bottom line. Paying a fee to avoid variance with that kind of hashing power would be like throwing money away. The day will come when only the most efficient miners can make a profit, and the ones that start maximizing their potential now will certainly stick around longer than those who don't.

[Bitsinmyhead]

Anyone saying he should go solo...
Probably never have studied statistics...
Or are just some degen gambler...

Go read up on Poisson distribution...
You be suprised how huge variance is...
He could mine for week++ with no blocks...
At this difficulty level...

Actually going solo...
Could break his operation...
Unless he is degen gambler...
3% for reduced variance is nothing here...

[JayC]

Anyone saying he should go solo...
Probably never have studied statistics...
Or are just some degen gambler...

Go read up on Poisson distribution...
You be suprised how huge variance is...
He could mine for week++ with no blocks...
At this difficulty level...

Actually going solo...
Could break his operation...
Unless he is degen gambler...
3% for reduced variance is nothing here...

More importantly, there are pools out there with 0% or .5% fees...  It's worth the effort to use a pool even 8Gh/s.

[grndzero]

Anyone saying he should go solo...
Probably never have studied statistics...
Or are just some degen gambler...

Go read up on Poisson distribution...
You be suprised how huge variance is...
He could mine for week++ with no blocks...
At this difficulty level...

Actually going solo...
Could break his operation...
Unless he is degen gambler...
3% for reduced variance is nothing here...

If he mines 8.5=10 Gh/s his average for hitting a block is around 1.5 days. Going a full week or more without hitting a block would be statistically insane. If I had half speed I would go solo instantly.

As Vladamir said, the hardest part is not sitting there watching it every minute and worrying that it's not getting anything.

[Clavulanic]

@Clavulanic: I am running these stock with fan cooling only. I am expanding watercooling to all soon (the watercooled machine is 400MH/s per GPU and running around 60'C).  Stock (which mine are, except for 2x cards) are expected to be at 300Mh/s, so I do not understand why I should be way above 7.5GH/s.

If you have any suggestions on how to "fine tune" my operation - let me know. It is always easier to tell someone else to destroy their $800 card instead of your own :-)

That is ridiculous to watercool all of your cards, with that money you could build several more rigs. Hell, it'd probably be cheaper to buy a window AC unit and blow that sucker at the systems. Though I will say watercooling would be a great way to have all of your heat exhaust outside of your house if you have any rigs in your house.

Anyways, the notion around this forum that overclocking will destroy your cards is also ridiculous, do some research, visit some forums and find out how it's done. There are correct ways and incorrect ways to approach overclocking your cards. There are also things that should be done to your cards. Reseat the heatsinks with better TIM, space them properly and have good airflow.

[Jaime Frontero]

Anyone saying he should go solo...
Probably never have studied statistics...
Or are just some degen gambler...

Go read up on Poisson distribution...
You be suprised how huge variance is...
He could mine for week++ with no blocks...
At this difficulty level...

Actually going solo...
Could break his operation...
Unless he is degen gambler...
3% for reduced variance is nothing here...

If he mines 8.5=10 Gh/s his average for hitting a block is around 1.5 days. Going a full week or more without hitting a block would be statistically insane. If I had half speed I would go solo instantly.

As Vladamir said, the hardest part is not sitting there watching it every minute and worrying that it's not getting anything.

agreed.

i was solo when difficulty was 90 whatever.  when it hit 127,000, i put half my miners in a pool.  at 157,000, i put 'em all in a pool.  i'm running at 2.1 Gh/s right now.  if i get up over 5 Gh/s, i'll go solo again, as long as difficulty stays under 350k.

pool vs. solo is all about the temperament of the miner.  got patience?

vladimir is correct.

[Bitsinmyhead]

Why don't you invest all your life savings in mining...
It's because of the risk involved...
Risk comes from future BTC price...
Future difficulty levels...
And how you run at solving blocks...

Reduce one of these risks...
And you can invest more...
Pay 3% to reduce varience...
Will make you be able to invest more...

Your risk adjusted profit will be better...
Thats where it all is...
You guys should study maths/economics...
Unless he is very willing to take risk...
For all his life savings...
It would be crazy going solo...

Even with 10Ghs at this difficulty level...
If he finds block every 1.5 days...
He still has 1% of going week without block...
Hardly statistically insane...
Poisson process, something every miner should know...

[kebumaha]

I think we have the mother of all "..." syndromes in here 

[IlbiStarz]

I was thinking about this, can someone answer..

Even if he has 7.5Gh total, each card is still only giving him 600Mh. Since not all the cards are working on the same block, wouldn't it actually take something like 20 days (according to the calculator) to solve like 24 blocks? Instead of one a day. In reality, each core of each card is still working by itself on a block right..?

[Freakin]

... o_0

[grndzero]

I was thinking about this, can someone answer..

Even if he has 7.5Gh total, each card is still only giving him 600Mh. Since not all the cards are working on the same block, wouldn't it actually take something like 20 days (according to the calculator) to solve like 24 blocks? Instead of one a day. In reality, each core of each card is still working by itself on a block right..?

um, no. Each GPU gets it's own work and they all work toward solving the current block.

[keybaud]

If a 100 GH/s pool has variance people moan about, a solo miner is going to get even worse variance. Solo if you want the chance of getting more bitcoins, but accept the fact that there is an equal chance you'll get less.

The average number of bitcoins you get in a pool is similar to the number you get solo. The only real difference is the variance and any fees.