Stale / rejected share rate?

[tictok]

Hi all -

This isn't the best sub-forum for this, but its the only place I can post right now.

I've been running my miner in Slush's pool since the early hours of the morning at approx 420Mhash/sec

I now have 4005 accepted shares and 89 rejected.

Those 89 rejected / stale shares are the equivalent of 2.25% of all shares.

Been fairly new I'm wondering if that is an acceptable rate?

I know the difference between stale and plain bad shares, but using Phoenix I don't think there's anyway to differentiate between the two.

How many stale / bad / rejected shares do you get?

Do you think I need to worry about 2.25% and if so could I improve the situation any?

It seems a shame to burn electricity on rejected shares if I can do something about it, even if its only a few...

Cheer

[organofcorti]

Stale/invalid shares can be caused by issues to do with your GPU (eg too much heat or OC) and issues to do with the pool. I've noticed I get many more stale on eligius then deepbit, and I get lots of "miner idle" messages with deepbit. I keep going back to Eligius because the overall hashrate is better. Just try a few different pools and see if your invalid% changes.

FYI my % invalid on eligius has been about 2% recently, on deepbit about 0.5%.

[SomeoneWeird]

Hi all -

This isn't the best sub-forum for this, but its the only place I can post right now.

I've been running my miner in Slush's pool since the early hours of the morning at approx 420Mhash/sec

I now have 4005 accepted shares and 89 rejected.

Those 89 rejected / stale shares are the equivalent of 2.25% of all shares.

Been fairly new I'm wondering if that is an acceptable rate?

I know the difference between stale and plain bad shares, but using Phoenix I don't think there's anyway to differentiate between the two.

How many stale / bad / rejected shares do you get?

Do you think I need to worry about 2.25% and if so could I improve the situation any?

It seems a shame to burn electricity on rejected shares if I can do something about it, even if its only a few...

Cheer

That's not bad, but it's not the best either. Try turning your fans up or trying a different pool.

[tictok]

Thanks for your answers!

I have overclocked a fair bit, so maybe that's causing a few dodgy shares, although underlocking the memory speed has helped keep power consumption down and the temp around a stable 75ºC @ 70% fan speed, so I'm confident my GPU is cool enough. Could slowing the memory speed down have had a negative effect on quality of share result?

Maybe re-flashing my bios with a less aggressive OC would help?

For a newby with my kind of hash rate (420Mh/S), who's pool would you recommend for a steady payout and less stale shares?

I guess its a balance as the more people in a pool, the quicker blocks get completed, but I guess that also means more stale shares?

So would a smaller pool mean less stale shares? And then also longer block times?  But more reward per person for completing a block mean overall payout / reward would be comparative to a larger pool?

Does that make sense?

From what I can gather Slush and Deepbit are the two big ones?
Who else do you recommend I look at? Obviously trust plays a bit part in the decision and being new thought that following the crowd would be the way to go.


Cheers
tictok

[organofcorti]

shorter ping times seem to correlate with lower stales for me. try pinging them.

Underclocking mem too far can increase stale/invalid and temp - on my 6990 if I set core and mem at 920 and 920, the 2 gpus are stable at 77-78 degrees. Downclock mem to 440 and temp spikes and so do invalids. Watch for that!

[czz]

I've noticed it depends on the pool.

I'm looking at my stats for last 1500 shares for each pool I'm using, and here are the results:
With BTCGuild, it's 16 of 1500,
with DeepBit, it's 6 of 1500,
and with bitcoins.lc it's 40 of 1500.

So it may really depend on the pool.

Two cards run @ 1GHz core/300 MHz RAM, and one runs at @ 950/300. When overclocked too much, they don't increase the rejection rate, but the miner just hangs instead

[JinTu]

I've noticed it depends on the pool.

I'm looking at my stats for last 1500 shares for each pool I'm using, and here are the results:
With BTCGuild, it's 16 of 1500,
with DeepBit, it's 6 of 1500,
and with bitcoins.lc it's 40 of 1500.

So it may really depend on the pool.

Two cards run @ 1GHz core/300 MHz RAM, and one runs at @ 950/300. When overclocked too much, they don't increase the rejection rate, but the miner just hangs instead

I noticed that while BTC Guild was undergoing maintenance two weeks ago my stale rate shot up for my Phoenix miner, but not for the Ufasoft miners. As soon as the maintenance was finished the Phoenix stale rate dropped back down to next to nothing with no setting changes on my part.