[ANN] profit switching auto-exchanging pool - www.middlecoin.com

[phzi]

no but if you link me to the source but sounds like your using linux you should have tools to compile that already installed.

https://github.com/veox/cgminer
^ best branch so far I would say.

[bal3wolf]

no but if you link me to the source but sounds like your using linux you should have tools to compile that already installed.

https://github.com/veox/cgminer
^ best branch so far I would say.

I looked at it after compiled and the source seems like all he has done is removed asic and btc mining stuff and just made it for scypt mining and a few displaying changes, no optimizing yet older vers of cgminer still work better on middlecoin with the changed to the tcp_keepidle tcp_keepintvl.


compiled it with and without the changes to tcp_keepidle tcp_keepintvl lol my www is getting alot of hits now good thing i get 70gigs a month should be plenty even with alot of hits.

no changes just compiled the source
http://bal3wolf.centelia.net/MINING/cgminer-veox-3.7.2.zip

changed the tcp_keepidle tcp_keepintvl to values for middlecoin
http://bal3wolf.centelia.net/MINING/cgminer-veox-3.7.2-middle.zip

[Kalroth]

I have one question: why is the EU server physically located in America?

It's located in Ireland, which is pretty far from America last I checked.

My traceroute goes from Telia London to Telia Dublin to Amazon Data Services Ireland anyways.

EDIT: If we're throwing custom cgminers out here then take a look at http://k-dev.net/cgminer/
I've implemented a few R9 290(X) fixes, lowered pool switch back to one minute as well as the keepidle changes.
See the change log for more. There are both windows binaries and single source code files for *nix.

[daynomate]

are you looking to launch any server instances in the Sydney region as well as there are a number of Australian miners who would appreciate a regional hub to use.

First I'm going to fix up the EU server, then I'm setting up either Toyko, Singapore, or Hong Kong. Then I'll look at other places like Australia.

Singapore would still be great for us Aussies... Should mean roughly 100ms ping

[njaiswal]

Now running cgminer-veox-3.7.2-middle.zip from bal3wolf with load balancing between US and EU servers (running from the UK), no drop in hash rate now at my end at least.

[daynomate]

Guys I wrote a quick and dirty calculator that estimates profit of this pool and thought I'd share.

It extracts h2o's profit data and extrapolates that trend. Obviously we can't predict the future trend, but this should get you a short term ballpark.

https://docs.google.com/spreadsheet/ccc?key=0Ahtw8TH0C2qIdEJSYTRmZEpZRXBVRzlGQTg0QWdmRkE&usp=drive_web#gid=0

Thanks for that - looks interesting. I took that same data and put into excel and added a moving 30-day trend line. It had dropped over time but was recently climbing again so I took a guesstimate at about the same as your figure of 17mBTC per day @ 1Mhash/s

[SimonG]

no but if you link me to the source but sounds like your using linux you should have tools to compile that already installed.

https://github.com/veox/cgminer
^ best branch so far I would say.

I looked at it after compiled and the source seems like all he has done is removed asic and btc mining stuff and just made it for scypt mining and a few displaying changes, no optimizing yet older vers of cgminer still work better on middlecoin with the changed to the tcp_keepidle tcp_keepintvl.


compiled it with and without the changes to tcp_keepidle tcp_keepintvl lol my www is getting alot of hits now good thing i get 70gigs a month should be plenty even with alot of hits.

no changes just compiled the source
http://bal3wolf.centelia.net/MINING/cgminer-veox-3.7.2.zip

changed the tcp_keepidle tcp_keepintvl to values for middlecoin
http://bal3wolf.centelia.net/MINING/cgminer-veox-3.7.2-middle.zip

Do you know why it sometimes switches coin to dogecoin even when its not the most profitable? After I manually restart its back mining something with difficulty 70.

I get 1/2 of shares submitted on US compared to what I get on eu server.

[1/3/2014 11:37:00 AM]  CGMiner (1052): Network difficulty is now 64.
[1/3/2014 11:37:12 AM]  CGMiner (1052): Network difficulty is now 67.
[1/3/2014 11:37:43 AM]  CGMiner (1052): Network difficulty is now 71.
[1/3/2014 11:37:53 AM]  CGMiner (1052): Network difficulty is now 73.
[1/3/2014 11:39:24 AM]  CGMiner (1052): Network difficulty is now 73.
[1/3/2014 11:40:24 AM]  CGMiner (1052): Network difficulty is now 77.
[1/3/2014 11:40:24 AM]  CGMiner (1052): 1 getwork failure(s) reported.
[1/3/2014 11:40:55 AM]  CGMiner (1052): Network difficulty is now 79.
[1/3/2014 11:41:45 AM]  CGMiner (1052): Network difficulty is now 82.
[1/3/2014 11:42:26 AM]  CGMiner (1052): Network difficulty is now 84.
[1/3/2014 11:43:57 AM]  CGMiner (1052): Network difficulty is now 80.
[1/3/2014 11:44:37 AM]  CGMiner (1052): Network difficulty is now 82.
[1/3/2014 11:45:08 AM]  CGMiner (1052): Network difficulty is now 85.
[1/3/2014 11:45:38 AM]  CGMiner (1052): Network difficulty is now 87.

[eric89]

are you looking to launch any server instances in the Sydney region as well as there are a number of Australian miners who would appreciate a regional hub to use.

First I'm going to fix up the EU server, then I'm setting up either Toyko, Singapore, or Hong Kong. Then I'll look at other places like Australia.

Singapore would still be great for us Aussies... Should mean roughly 100ms ping

Yes!

Singapore please!

[Kluge]

I have one question: why is the EU server physically located in America?

It's located in Ireland, which is pretty far from America last I checked.

My traceroute goes from Telia London to Telia Dublin to Amazon Data Services Ireland anyways.

EDIT: If we're throwing custom cgminers out here then take a look at http://k-dev.net/cgminer/
I've implemented a few R9 290(X) fixes, lowered pool switch back to one minute as well as the keepidle changes.
See the change log for more. There are both windows binaries and single source code files for *nix.

The geoIP database I use says it's in Seattle. Weird.

What's up with all those cgminer clones? What's the point? Last time I heard, the original was pretty well optimized.

There's a bug in the pool where the pool hangs when it switches coins. It's not particularly significant, but it can be an underlying cause of more significant complications, particularly with CGWatcher since it'll probably be configured to start issuing restarts of cgminer or the PC which can make the relatively quick hang time more significant (cgminer or CGwatcher may also be configured to switch to the less-profitable eu pool when the US pool is detected as non-working which may also make a mole hill into a mountain of a problem). The custom builds were originally being posted because they alter settings to minimize or eliminate the effects of this bug while h2o is working on the server-side fix.

[MiMiMiner]

Question: I want to point all my servers with a  total of 160.000 khs to the VS server. But, we use linux and i cant find any compiled versions that are good to use with middlecoin.

Any help?

[ripsnort]

So a few pages back someone was complaining about how h2o holds the coins before exchanging them... Although I agree that h2o does a decent job when it comes to exchanging, I too see a problem with this, with an example.

I check my balance/unexchanged/immature every hour or so. I checked it and my balance was .0030 (after payout of course), my unexchanged was .03195, and my immature was .01233. Now I check it, and my balance is still .0030, my unexchanged has gone down to .0254 and my immature has gone down to .0054. I understand with the immature coins you can't exchange them, but since my mature unexchanged coins didn't get auto-traded, I lost .006 btc. I understand this isn't much, but if it effects me, then I'm sure it's effected others that had a higher unexchanged balance.

I saw the same behavior.  I think it may have something to do with timing of exchanging coins and updating of middlecoin JSON.  See spikes in my graph below, which suggest that nothing is "lost", since graph resumes approximately where it left off after downward spike.  My graphs are different from coinGeek's because I plot summation of Immature (purple shading), Unexchanged (yellow shading) and Exchanged (green shading).  So on my graph, there should always be an upwards slope.  The spike and then recovery could impy that Unexchanged coins were sent to be exchanged, but JSON was written before Exchanged balance was updated.  It may also have something to do with the Immature coin balance temporarily dropping to zero, though I'm not sure why.  Regardless, the balances appear correct in is following JSON updates.

Bottom line:  I don't believe you "lost" any unexchanged coins.

https://i.imgur.com/7ZA93x0.png

[Subo1977]

H20 great job!
Please fix this though http://middlecoin2.s3-website-us-west-2.amazonaws.com/profitgraph.html
Its been down for more than 2 weeks

bump

[Biggen]

Instead of all this load sharing and recompiling nonsense.  Why don't you guys just set the U.S. server as the only server in CGminer and don't set a failover?  Cgminer won't give any of these problems if only one server is listed and the --failover-only flag isn't set.

Then when both servers are fixed, you can go back and set the EU as a failover server??

I know this means that if the US server were to go down than your miners would be offline but it appears H20 runs a pretty tight ship around here so the odds of that happening look to be pretty low.

[MiMiMiner]

Instead of all this load sharing and recompiling nonsense.  Why don't you guys just set the U.S. server as the only server in CGminer and don't set a failover?  Cgminer won't give any of these problems if only one pool is listed.

Then when both servers are fixed, you can go back and set the EU as a failover server??

Is that true? What ping do you have?

[envy2010]

So a few pages back someone was complaining about how h2o holds the coins before exchanging them... Although I agree that h2o does a decent job when it comes to exchanging, I too see a problem with this, with an example.

I check my balance/unexchanged/immature every hour or so. I checked it and my balance was .0030 (after payout of course), my unexchanged was .03195, and my immature was .01233. Now I check it, and my balance is still .0030, my unexchanged has gone down to .0254 and my immature has gone down to .0054. I understand with the immature coins you can't exchange them, but since my mature unexchanged coins didn't get auto-traded, I lost .006 btc. I understand this isn't much, but if it effects me, then I'm sure it's effected others that had a higher unexchanged balance.

I saw the same behavior.  I think it may have something to do with timing of exchanging coins and updating of middlecoin JSON.  See spikes in my graph below, which suggest that nothing is "lost", since graph resumes approximately where it left off after downward spike.  My graphs are different from coinGeek's because I plot summation of Immature (purple shading), Unexchanged (yellow shading) and Exchanged (green shading).  So on my graph, there should always be an upwards slope.  The spike and then recovery could impy that Unexchanged coins were sent to be exchanged, but JSON was written before Exchanged balance was updated.  It may also have something to do with the Immature coin balance temporarily dropping to zero, though I'm not sure why.  Regardless, the balances appear correct in is following JSON updates.

Bottom line:  I don't believe you "lost" any unexchanged coins.

Sometimes the exchange APIs go down causing the unexchanged/immature coin values to not display correctly. You don't lose any value when that happens, and it should come back to a higher value when the exchanges come back online.

Also, coins tend to increase in value as often as they drop in value during the time before exchange. In fact, from my experience, the exchange frequently results in more BTC than estimated (which a good autotrading program should). The payout is always timely, so I have no issues with any of this.

[Biggen]

Is that true? What ping do you have?

Its very true.  Try it yourself.  Cgminer 3.7.2 happily switches to new coins with no issues with no failover set.

My ping to the US server 100ms.

[envy2010]

Is that true? What ping do you have?

Its very true.  Try it yourself.  Cgminer 3.7.2 happily switches to new coins with no issues with no failover set.

My ping to the US server 100ms.

How long do your rigs idle during coin switches? Mine were seeing 5-10% downtime because no work was available during switches on the US server.

[Waramp22]

Is anyone getting lockups with the eu server? My computer started locking up after ~15 minutes of mining. Im not sure if its cgminer related or what because i havent installed anything or browsed the internet on my miner. All it does it mine and something has happened to it.

[Biggen]

How long do your rigs idle during coin switches? Mine were seeing 5-10% downtime because no work was available during switches on the US server.

It depends on how many switches are happening at once.  I was watching CGminer a second ago and the US Middlecoin server switched to three different coins in the span of about 5 seconds.  That kinda threw CGminer for a loop (said lost connection) but it recovered after about 5 seconds or so.  If the coin switches are slower, then CGminer keeps chugging along with no issues.

I'm not sure why Middlecoin switches coins that quickly though sometimes.  I'm assuming that is part of the problem that H20 is going to correct.  I can't believe there would be a reason to switch from one coin to another in the span of less than a minute to make much of a difference as far as profits are concerned.

[MiMiMiner]

Is that true? What ping do you have?

Its very true.  Try it yourself.  Cgminer 3.7.2 happily switches to new coins with no issues with no failover set.

My ping to the US server 100ms.

I am trying it right now on Fedora, CGminer 3.7.2. But my hashrate per card drops from 870 to about 800...