All pools require rounding at some point in the process.  It depends on which point in the process and how deep they go into decimals for how that affects users.

BTC Guild PPS rounds down at the 16th decimal place for display, and the satoshi level when sending a payout (required obviously).  However, all shares are stored in a way that the payout amounts are calculated live without rounding until the final number.  This means there is no rounding done on individual share payout rates, so instead of losing out on rounding per-share, you will never be behind a full satoshi in total.

BTC Guild PPLNS method does require rounding, which is rounded DOWN at the 14th decimal place.

No, Namecoins will not be distributed on the test server (which is why it's being offered at 0% PPS for BTC).  A full post will be made once the test server successfully mines a BTC block [indicating no bugs in the new MM code].  That post will include roll-out plans/timeline and how NMC will be awarded to miners/what stats to expect on the frontend.

No charts available, but I have some internal commands I can query to see total shares submitted to the test server so far.  As of this morning the test server has received 13,492,615 shares and solved 3 namecoin blocks, but 0 bitcoin blocks.  That's not unexpected (BTC difficulty is 5x higher).  It means the test server has been mildly unlucky (you'd expect 4 NMC solves by that many shares).

More power means the test will end faster (more likely to solve a block), so it's always welcome if you're willing to update your miners [BUT DONT FORGET A FALLBACK SERVER!].



0.25% isn't abnormal, a lot of it has to do with miner settings (higher intensity = more stales).  Fun thing to know, every PPS pool I've seen that "pays for stales", doesn't treat all stales equally.  Your "invalid" shares on 50BTC are almost certainly stale shares that were deemed too late to be counted as valid.

Quoting this again (new thread page) so anybody who missed it is aware that the test server is available and connection details are included above.

Current plan is the test server will remain online until it solves a Bitcoin block (since it will solve at least 1 NMC block in that time).  If this happens while I'm asleep, then the server will shut down once I wake up and see the block.  At the current difficulty, this may last 24+ hours depending on how much hash power is directed to the test server.

Stratum+Merged Mining Test Server
All mining done on the test server is paid at 0% Fee PPS

As promised, here it is!  The test server is now running on the live Bitcoin+Namecoin networks, and users are encouraged to redirect some hash power so we can see if it solves any blocks properly.  It will show up in the 'Payments' section as the current difficulty+1, with a 0% fee to avoid any interference with current difficulty payments.

To connect to the test server, point your STRATUM mining software at:  198.154.98.210  port 7777


The test server may experience outages, unannounced reboots, or even be turned off with very little notice.  Make sure you have a backup pool server configured so you don't end up losing mining time when the test server is not available.

Sorry for the extra delays.  Since there's always a chance this test code ends up having a bug which prevents it from solving normal BTC blocks, I've been having to double check everything to make sure these shares won't count towards the 25 million used for a PPLNS Shift [that way the rest of the pool isn't penalized due to a potential software glitch].  I've also been doing a minor code audit just to see if there's anything I can fine tune (some parts of the code were written almost a year ago, and for a better protocol than Stratum which was dropped to prevent competing protocols from killing adoption).

Merged Mining Progress Update

Next series of tests are done on testnet.  I constantly "crashed" the namecoin daemon in order to test how the stratum code functioned.  As expected, 0 impact was seen on the bitcoin side.  The server continued along happily as if nothing was wrong (other than my logs going nuts due to constant RPC connect errors on namecoind).

Live testing server will be up in the next 2-3 hours hopefully.  Once the livenet testing is under way, I will post a formal explanation of how namecoins will be distributed via Stratum, and what to expect on the website frontend.  The biggest focus I've had during this update to the code is that namecoin should not have ANY impact on performance.  This includes inside the pool server as well as on the database server.

Merged Mining Progress Update

Now successfully merged mining on testnet!  Now fixing up some of the code (I was using a dummy namecoin coinbase rather than one I can actually claim coins from).  I won't say for certain the MM-Stratum server will be available for public testing tonight, but it's highly likely.  Keep checking the forum, I will be continuing to provide progress updates as the server nears live network testing.

Decided to sleep on the problem rather than stay up til dawn last night.  Woke up this morning and found my error!  Testnet mining in progress right now, will update later.


Small warning to EU users:  An antivirus company has reported my servers as hosting/running a trojan to the dedi provider.  Since both the dedicated server provider and antivirus companies are complete morons, there is a chance the servers get shut down and I have to move DNS to the US if that happens.  I've responded to the abuse email and sent an email to Symantec.  I honestly thought by now AV companies had enough of a clue to understand the difference between a C&C server and a Bitcoin server, but apparently they're still clueless on the distinction.

In the event that DNS is required to change, you shouldn't experience much downtime.

Getting very close now, running into some issues with namecoind claiming "nBits below minimum" work.  Once that's gone it *should* be working.  Hoping to have it done tonight still (even if it requires a half gallon of coffee and "tonight" being redefined to "before dawn").

As stated previously, the information for the public test server on the live network will be made available in the forums as soon as its online.

Another update on Stratum+merged mining server progress:  I'm expecting to have a working implementation (tested with bitcoin+namecoin testnet-in-a-box servers) sometime tomorrow.  Once everything is working, I will make an updated version designed to work on the live networks, which should only take a few minutes.  I will be setting up a special test server on livenet tomorrow/Monday, which will be set to record shares at current difficulty+1 and a 0% fee, and will be asking for volunteers to mine on this test server.


Once the test server is up, I will post the details for connecting and anybody is free to give it a try.  The test server will get shut down without much notice, so make sure you have another BTC Guild server set as your backup before switching over!

For the test server, you will NOT earn NMC, but you will get 0% PPS on the Bitcoin network for the duration of the testing.  Since it is a test server, obvious warnings are applicable:  Have backup servers set up, monitor your workers, and if you see shares stop counting, STOP mining and move back to the main server because I cannot *guarantee* restoration of missing shares if something breaks on the test server [thread locking/mysql failure conflicts COULD happen but shouldn't].

Just so you know, "part time miners" are not penalized under PPLNS.  I believe they're not punished under DGM either, but I've never wrapped my head around DGM.  The only system that actually penalizes miners for running part time is slush's score method.

Under PPLNS (BitMinter and BTC Guild), it might look like you are penalized at first.  This is because PPLNS applies payments to shares for a few hours after they have been submitted.  However, turning your miner off doesn't change how much those shares will be paid.  I believe DGM has a similar "charge up" (I've seen it referenced on ozcoin in the past).


That said, PPS has an appeal.  Even at higher fees, it is nice to see steady income, although you miss out on the roller coaster ride when luck is really good.

7.5% PPS, yes.  Only 3% on PPLNS, which pays txfees (~1.5% bonus).  Also adding NMC to PPLNS soon which is a ~4% bonus at the moment as well.  The only time BTC Guild has been unstable in over a year iis during attacks or server migrations [which are results of attacks].

Merged mining still in progress.  Lots of distractions during the week that pushed back a working implementation.  Still hopeful to have it working on testnet this weekend and likely deployed on the live Stratum servers early next week.

Hmm, I bet the cookies don't refresh themselves.  I'll take a look and make it set the cookies to something like 1-year in the future if that's the issue.

The # of items in data tables should be saving as a cookie so that it remembers your settings when alternating pages.  I'll take another look at the dashboard to see if something is forcing those back to default size.  The way it works it the whole table loads with the page, then the javascript condenses it and splits it into pages, which is why on a slow load you might see the whole table before it shrinks.

The most common report of this is from GUI Miner users that are having Hardware Errors/Verification Failed errors.  These are local problems, meaning the miner is having issues and isn't even sending the shares to the pool at all.  You can confirm if this is the problem by going to View -> Show Console, and watching that for a few minutes.  The most common reason for this is due to bad driver versions.  I am not able to give much guidance on what the best versions to use are, I haven't mined on GPUs in a very long time.

Merged mining doesn't cause instability unless the implementation is poor.  It can cause minor (<0.1%) drops in BTC due to added latency when preparing new work after a longpoll.  However, I'm specifically implementing this in a way where namecoin efficiency is sacrificed to let the pool run without skipping a beat even if namecoind is frozen/slow/crashed.  The only efficiency change will be work messages from the pool require ~50 bytes of extra data inside the coinbase.


UPDATE:  Phase 1 went well.  I've got the pool creating the bitcoin side of merged mining (coinbase message that includes the mm signal + namecoin block hash) while still successfully creating blocks.  Tomorrow I'll be working on building the NMC block with the MM headers included.  It's very likely this can be done by the weekend unless I run into unexpected problems.  Everything is in place so the entire namecoin side is running in a separate thread that will not block the primary pool/bitcoin threads.

It's a plan at the moment.  I posted in the IRC room that I may have a test server available by this weekend.  I'm actually hoping to get this running on a local bitcoin+namecoin testnet tonight, then it's just a matter of making sure namecoind crashing/hanging won't cause the server to hang/wait.