Topic: [1050 TH] BitMinter.com [1% PPLNS,Pays TxFees +MergedMining,Stratum,GBT,vardiff] - page 55. (Read 837101 times)
May 29, 2014, 03:54:57 PM
No,no,no...fox has real shitty wireless internet in his area, I was on teamspeak with him when he disappeared...a big storm ripped through here last night too, headed out his way too. Wife just told me we have another big storm rolling in here too, Im going to have to shut down too.
May 29, 2014, 03:23:56 PM
Huh......Koi leaves the pool yesterday and today it appears that Fefox is gone. Wonder what's up?
??
I just got another S2 to add to the pool, but that's a farcry from making up the difference in hashing power of these two miners.
??I just got another S2 to add to the pool, but that's a farcry from making up the difference in hashing power of these two miners.
it is permanent ? No ones know so far, let's see and wait.
May 29, 2014, 03:04:36 PM
Huh......Koi leaves the pool yesterday and today it appears that Fefox is gone. Wonder what's up???
I just got another S2 to add to the pool, but that's a farcry from making up the difference in hashing power of these two miners.
??I just got another S2 to add to the pool, but that's a farcry from making up the difference in hashing power of these two miners.
May 29, 2014, 01:50:46 PM
Hey what gives. The last post and bottom screen capture shows me mining at 7,870 and I have 8,300. Only joking 
one of my Terraminers switched to back up.
Doc's approximate stats are pretty much spot on for me too.
Jacko
one of my Terraminers switched to back up.Doc's approximate stats are pretty much spot on for me too.
Jacko
May 29, 2014, 12:37:28 PM
The server can handle a few hundred thousand more connections with no problem. There are fewer connections and fewer people/workers mining than in a very long time.
All Bitminter servers have hardware suitable for the job they do. Configurations, including Linux kernel parameters, have been tweaked to run well with the hardware in use and to handle high loads.
I never said higher bitcoin difficulty should mean that your hashrate is reduced. Your hashrate is unaffected.
Other miners are not seeing the problem you have. See for example what philipma1957 just posted.
When we reached 1 PH/s the load on the server was lower, not higher, than before.
Now imagine on your gaming server you had 10 000 people playing and everything is fine. Later, with no changes to hardware or software, you have 5 000 people playing and one of them is having a problem. If your conclusion is that this one person has a problem because your game cannot handle 5 000 people playing, then your conclusion is probably wrong.
The web server estimates your hashrate based only on accepted work. The most common reasons that you would see a lower hashrate on the web server than your actual hashrate is rejected work or hardware errors. Since you are not getting much rejected work, have you checked how much is lost due to hardware errors?
All Bitminter servers have hardware suitable for the job they do. Configurations, including Linux kernel parameters, have been tweaked to run well with the hardware in use and to handle high loads.
I never said higher bitcoin difficulty should mean that your hashrate is reduced. Your hashrate is unaffected.
Other miners are not seeing the problem you have. See for example what philipma1957 just posted.
When we reached 1 PH/s the load on the server was lower, not higher, than before.
Now imagine on your gaming server you had 10 000 people playing and everything is fine. Later, with no changes to hardware or software, you have 5 000 people playing and one of them is having a problem. If your conclusion is that this one person has a problem because your game cannot handle 5 000 people playing, then your conclusion is probably wrong.
The web server estimates your hashrate based only on accepted work. The most common reasons that you would see a lower hashrate on the web server than your actual hashrate is rejected work or hardware errors. Since you are not getting much rejected work, have you checked how much is lost due to hardware errors?
or we are now under 1ph and numbers are 420-430 on the underside of that number.
the only major under reporting issue your pool has ever had for me was red fury usb sticks. they consistently went too low. other then that all your numbers have been fine.
I mined with you only from aug 2012 until nov 2013.
In nov 2013 I finally was able to teach myself cgminer bfgminer cpuminer along with ltc mining and I now mix mining with you cex.io btcguild and wemineltc.
You pool seems as good as any. It also is very newbie friendly.
I tell you what. I have another antminer s-1 on btcguild. I will move it to bitminter. this one is in my house and is clocked to 200gh see photo
this is a more difficult test for the Doc's server.
the 3 at my friends place of business are at 420gh and are on a verizon modem dsl.
my 200gh is on a cablevision modem one is Lakewood NJ USA one is Howell NJ USA..
All gear is pointed at the Doc's pool all have the same worker name in my account. I was at 420-430gh
I should go to 620-630-gh see second photo I do. I will post again in a day to show long numbers not just 10 minutes.
and here you go 630 gh using 4 ant miner s-1's .
3 set on underclocking total 430 gh (Lakewood, NJ USA) verizon dsl hi speed
1 set on overclocking 200gh (Howell , NJ USA) cablevision mid speed
May 29, 2014, 12:06:11 PM
The server can handle a few hundred thousand more connections with no problem. There are fewer connections and fewer people/workers mining than in a very long time.
All Bitminter servers have hardware suitable for the job they do. Configurations, including Linux kernel parameters, have been tweaked to run well with the hardware in use and to handle high loads.
I never said higher bitcoin difficulty should mean that your hashrate is reduced. Your hashrate is unaffected.
Other miners are not seeing the problem you have. See for example what philipma1957 just posted.
When we reached 1 PH/s the load on the server was lower, not higher, than before.
Now imagine on your gaming server you had 10 000 people playing and everything is fine. Later, with no changes to hardware or software, you have 5 000 people playing and one of them is having a problem. If your conclusion is that this one person has a problem because your game cannot handle 5 000 people playing, then your conclusion is probably wrong.
The web server estimates your hashrate based only on accepted work. The most common reasons that you would see a lower hashrate on the web server than your actual hashrate is rejected work or hardware errors. Since you are not getting much rejected work, have you checked how much is lost due to hardware errors?
All Bitminter servers have hardware suitable for the job they do. Configurations, including Linux kernel parameters, have been tweaked to run well with the hardware in use and to handle high loads.
I never said higher bitcoin difficulty should mean that your hashrate is reduced. Your hashrate is unaffected.
Other miners are not seeing the problem you have. See for example what philipma1957 just posted.
When we reached 1 PH/s the load on the server was lower, not higher, than before.
Now imagine on your gaming server you had 10 000 people playing and everything is fine. Later, with no changes to hardware or software, you have 5 000 people playing and one of them is having a problem. If your conclusion is that this one person has a problem because your game cannot handle 5 000 people playing, then your conclusion is probably wrong.
The web server estimates your hashrate based only on accepted work. The most common reasons that you would see a lower hashrate on the web server than your actual hashrate is rejected work or hardware errors. Since you are not getting much rejected work, have you checked how much is lost due to hardware errors?
May 29, 2014, 09:59:43 AM
I see, this helps me understand a bit better. However, I used to build games for a living, now I just hire. I understand the difficulty and variance curve, I've built whole architectures of difficulty-to-level curves where variance is usually dynamically coded. I don't know Bitcoin THAT much but I understand game and web development quite a bit which both aren't too far off from each other.
Since you insist, though, here's the numbers I cringe at: Let you see what my eyes see (and potentially others who may never speak) so may you understand me better and the like:
Before 1PH/s - 388.025 AVG
http://screencast.com/t/Chsjv3L2U0
After 1PH/s - 313.23 AVG
http://screencast.com/t/xHqxfwx9
No matter what anyone says, it makes no sense that increased difficulty results in slow hash rates, just means blocks take longer if you don't compensate with more power. It also doesn't make any sense that one would lose. Also, how many concurrent connections are you allowed to have? This is one thing some webmasters know nothing of. You may have 100GB/mo of transfer but if you only have 100 simultaneous connections on your package, it doesn't help if you're running a gaming type website where there are thousands of connections.
This is one thing most hosts and ISP's don't tell people about is the simultaneous connections your server comes with because most web hosts don't cater to game development companies. I learned that the hard way when players complained too much about system lag, I only had a server that could handle 500 connections but had nearly limitless everything else. Upon research, tech support decided the server I had couldn't handle the amount of simultaneous connections. You may have hit that cap. Yeah, I expect this to be wide open also. Streaming and Fox servers are the most robust for anything net extensive.
On that test, I had joined a few shifts before the 1PH/s gap breach. I watched my my miners' interface often and never seen them go below 194GH/s, that's the current low record for me. If that's not bottle-necking or at the very least throttling of some kind, then I don't know what it is.
Also, difficulty SHOULDN'T have an effect on the amount of hashing power your server takes from me. If it does, I want to see that explanation and logic behind the explanation.
When I said difficulty had no effect, it was on the test, in that each pool I tested, it was submitted to the same battery of tests and monitoring. Each pool all had as greater chance than the other to make it worth my while. I could write a whole report on pools... or maybe I should, share my research with the whole community and let everyone rips each other's hair out. Therefore, as you see, the pre 1PH/s is what is closest to what it should be and is what is closest to other pools avg rate too.
Of course, someone who has 50TH/s in their back pocket probably don't give a crap but maybe smaller mining operations and new miners do? Maybe you were here for so long, you have no idea what type of market it is today to enter now if you don't want to spend all your savings on it, compared to when you started, right? More small businesses, like mine, who enter, the more decentralized the network will remain because I have resources and knowledge no one else probably has in one package around here to help build a better place for all, don't you? If you could give your miners on average 30GH/s more at any given time would you? of course, who why not try to figure out why there's bottlenecking going on past 1PH/s. It is what it is, there is definitely bottlenecking going on past 1PH/s and if it doesn't get fixes, other miners who don't speak publicly will just leave for another one.
In the end, for me, that's quite a significant "variance", in my book and I think it's something the pool owner should look into, I mean, you don't have to, you're your own boss. I bet I'll go to BMTR right now and notice the difference since pool speed is much below 1PH/s now. I at least want to finish this current 24 test of Eligius though, about 4 hours.
I'll switch back over and let you know if there's a difference. I'd be willing to bet, I'll be back to the prior averages. If that's the case, the lack of available simultaneous connections argument becomes more and more the prime suspect, to me anyways.
Anyways, I must go, the spring smells wonderful and the sun is shinning for once. I'll come back tomorrow and post results.
PS: You might not think so, but I am trying to help you. if I was a pool operator and my system wasn't running as I expect and led me customers to expect also, I'd want to know because I wouldn't want to lose miners or waste their power... but that's me, I have extremely high standards for myself and my customers. So.... I was putting my feet in your shoes.
PSS: My reject rate is always below 0.10%...
Peace Out
Since you insist, though, here's the numbers I cringe at: Let you see what my eyes see (and potentially others who may never speak) so may you understand me better and the like:
Before 1PH/s - 388.025 AVG
http://screencast.com/t/Chsjv3L2U0
After 1PH/s - 313.23 AVG
http://screencast.com/t/xHqxfwx9
No matter what anyone says, it makes no sense that increased difficulty results in slow hash rates, just means blocks take longer if you don't compensate with more power. It also doesn't make any sense that one would lose. Also, how many concurrent connections are you allowed to have? This is one thing some webmasters know nothing of. You may have 100GB/mo of transfer but if you only have 100 simultaneous connections on your package, it doesn't help if you're running a gaming type website where there are thousands of connections.
This is one thing most hosts and ISP's don't tell people about is the simultaneous connections your server comes with because most web hosts don't cater to game development companies. I learned that the hard way when players complained too much about system lag, I only had a server that could handle 500 connections but had nearly limitless everything else. Upon research, tech support decided the server I had couldn't handle the amount of simultaneous connections. You may have hit that cap. Yeah, I expect this to be wide open also. Streaming and Fox servers are the most robust for anything net extensive.
On that test, I had joined a few shifts before the 1PH/s gap breach. I watched my my miners' interface often and never seen them go below 194GH/s, that's the current low record for me. If that's not bottle-necking or at the very least throttling of some kind, then I don't know what it is.
Also, difficulty SHOULDN'T have an effect on the amount of hashing power your server takes from me. If it does, I want to see that explanation and logic behind the explanation.
When I said difficulty had no effect, it was on the test, in that each pool I tested, it was submitted to the same battery of tests and monitoring. Each pool all had as greater chance than the other to make it worth my while. I could write a whole report on pools... or maybe I should, share my research with the whole community and let everyone rips each other's hair out. Therefore, as you see, the pre 1PH/s is what is closest to what it should be and is what is closest to other pools avg rate too.
Of course, someone who has 50TH/s in their back pocket probably don't give a crap but maybe smaller mining operations and new miners do? Maybe you were here for so long, you have no idea what type of market it is today to enter now if you don't want to spend all your savings on it, compared to when you started, right? More small businesses, like mine, who enter, the more decentralized the network will remain because I have resources and knowledge no one else probably has in one package around here to help build a better place for all, don't you? If you could give your miners on average 30GH/s more at any given time would you? of course, who why not try to figure out why there's bottlenecking going on past 1PH/s. It is what it is, there is definitely bottlenecking going on past 1PH/s and if it doesn't get fixes, other miners who don't speak publicly will just leave for another one.
In the end, for me, that's quite a significant "variance", in my book and I think it's something the pool owner should look into, I mean, you don't have to, you're your own boss. I bet I'll go to BMTR right now and notice the difference since pool speed is much below 1PH/s now. I at least want to finish this current 24 test of Eligius though, about 4 hours.
I'll switch back over and let you know if there's a difference. I'd be willing to bet, I'll be back to the prior averages. If that's the case, the lack of available simultaneous connections argument becomes more and more the prime suspect, to me anyways.
Anyways, I must go, the spring smells wonderful and the sun is shinning for once. I'll come back tomorrow and post results.
PS: You might not think so, but I am trying to help you. if I was a pool operator and my system wasn't running as I expect and led me customers to expect also, I'd want to know because I wouldn't want to lose miners or waste their power... but that's me, I have extremely high standards for myself and my customers. So.... I was putting my feet in your shoes.
PSS: My reject rate is always below 0.10%...
Peace Out
May 29, 2014, 08:08:27 AM
It looks like you don't understand two basic mining concepts, called "variance" and "difficulty". These are the two most important concepts for a new miner to learn and understand. I recommend researching these two concepts before investing in any mining hardware. And please learn from intelligent people like Organofcorti and Meni Rosenfeld, not the people who post misinfirmation on this forum on purpose.
But I will try to respond to your post below:
I don't understand what you mean. There is plenty of free filehandles on the server, we could easily handle a couple hundred thousand more connections. There is also plenty of free bandwidth. I don't see any kind of bottle necks. The load is actually much lower than it has been in the previous 12 months. In the past we have handled loads MUCH MUCH higher than what we see today.
The load on the server is much lower now than when we were below 1 PH/s. There are much fewer connections, much lower CPU usage and much lower memory usage. This is because there are fewer people mining. We can easily handle an unlimited hashpower. It's the number of people/workers mining that creates the load, and the load keeps dropping lower and lower all the time, as the smallest miners give up.
The server is almost falling asleep at this point.
During the GPUMAX days there were loads challenging me to improve the Bitminter server software to properly handle it when GPUMAX turned on the firehose. Eventually, after many improvements, Bitminter became one of their favorite pools because it could handle extreme loads. Today with the stratum protocol and with fewer people mining, the load that mining pools are seeing is very small and any newbie programmer could write software to handle it on cheap and slow server hardware.
This is a modern and non-beta pool. And we are using 100% of the hashing power.
Luck goes up and down. But we are very close to the average expected number of blocks found.
As the difficulty goes up you need more hashpower to find the same number of blocks per day.
No surprise there. It is as expected.
If you think a difficulty increase of 10-15% makes no difference then you don't understand what the difficulty is and you need to research it further. I've heard it from some miners before: "I don't think difficulty actually has any impact on mining". If you have such thoughts then you need to consult your local psychiatrist and/or mathematics professor.
The server does not know your hashrate. The website tries to make an estimate of your hashrate based on how many proofs of work you sent in during the last couple minutes. This is very inaccurate.
You could take a look at the estimated hashrate for a shift rather than the live hashrate esimates. The shift estimates are more accurate as they are estimated over more time. Have a look at https://bitminter.com/shifts
The current server has very low bandwidth usage, very low CPU usage, and very low memory usage. I'm not sure what kind of server upgrade would help. Actually, I know that no server upgrade would help. An increasing bitcoin difficulty isn't balanced out by upgrading a mining server. You need more hashpower, that's the only thing that can do it, and it has nothing to do with the mining server.
What kind of bottle-necking are you talking about? The load on the server is extremely low. The server is basically falling asleep from boredom because of the low load. There is no bottle neck.
If you are getting over 10% rejected work then something is seriously wrong. The average reject ratio for all users in the pool right now is about 0.13-0.14%. If you are not actually getting rejected work, but just saw your 400 GH/s display as 350 GH/s at the website for a moment, this is normal - the website live hashrates are (inaccurate) estimates of your real hashrate.
Please understand this: the server load and network load is almost zero. We find fewer blocks now than some months ago because the difficulty has increased very quickly.
As a miner one of the most important topics for you to research and understand is "difficulty".
To newbie miners: please stop thinking server load is proportional with hashrate. The server load comes from the number of workers mining in the pool. Even the biggest pools today can run on a single server if their software is efficient.
To experienced miners: yes, it is true, fewer and fewer people are mining. Mining is becoming more and more centralized. Small miners give up when their hashpower becomes insignificant and buying significant hashpower is out of their budgets. It is an unfortunate development. This may be important for the future of bitcoin.
But I will try to respond to your post below:
You need to UPGRADE your server, one PH/s I lose too much hashing power because of traffic bottle-necking, too many connections. Tried your pool for a week today and I am disconnecting today too due to having too great of a hash power loss compared to when I joined when you were below 1PH/s. I'll assume you're hosting online, most hosts can do an upgrade in an hour or so.
I don't understand what you mean. There is plenty of free filehandles on the server, we could easily handle a couple hundred thousand more connections. There is also plenty of free bandwidth. I don't see any kind of bottle necks. The load is actually much lower than it has been in the previous 12 months. In the past we have handled loads MUCH MUCH higher than what we see today.
The load on the server is much lower now than when we were below 1 PH/s. There are much fewer connections, much lower CPU usage and much lower memory usage. This is because there are fewer people mining. We can easily handle an unlimited hashpower. It's the number of people/workers mining that creates the load, and the load keeps dropping lower and lower all the time, as the smallest miners give up.
The server is almost falling asleep at this point.
During the GPUMAX days there were loads challenging me to improve the Bitminter server software to properly handle it when GPUMAX turned on the firehose. Eventually, after many improvements, Bitminter became one of their favorite pools because it could handle extreme loads. Today with the stratum protocol and with fewer people mining, the load that mining pools are seeing is very small and any newbie programmer could write software to handle it on cheap and slow server hardware.
That's why you're having such a hard time to find blocks as of late, you're not using the pool's complete hashing potential I am afraid. I observed about 10%-20% average hashing power loss using your pool compared to other modern, non-beta pools I have tested. Has nothing to do with the recent difficulty increase, made really little to no difference.
This is a modern and non-beta pool. And we are using 100% of the hashing power.
Luck goes up and down. But we are very close to the average expected number of blocks found.
As the difficulty goes up you need more hashpower to find the same number of blocks per day.
No surprise there. It is as expected.
If you think a difficulty increase of 10-15% makes no difference then you don't understand what the difficulty is and you need to research it further. I've heard it from some miners before: "I don't think difficulty actually has any impact on mining". If you have such thoughts then you need to consult your local psychiatrist and/or mathematics professor.
I noticed the difference the minute BMT went back over 1PH/s and kept growing on it. I monitor my workers like a hawk, so I know in real time what's going on, at all times of the day and when your stats show almost 50GH/s difference from my miner's interface to yours for most of the time, I start asking questions because that's rather significant, in my book anyways. Connection is steady though.
The server does not know your hashrate. The website tries to make an estimate of your hashrate based on how many proofs of work you sent in during the last couple minutes. This is very inaccurate.
You could take a look at the estimated hashrate for a shift rather than the live hashrate esimates. The shift estimates are more accurate as they are estimated over more time. Have a look at https://bitminter.com/shifts
Your server needs a hardware upgrade and maybe very well, bandwidth.
The current server has very low bandwidth usage, very low CPU usage, and very low memory usage. I'm not sure what kind of server upgrade would help. Actually, I know that no server upgrade would help. An increasing bitcoin difficulty isn't balanced out by upgrading a mining server. You need more hashpower, that's the only thing that can do it, and it has nothing to do with the mining server.
Sorry but I can't mine here with only having 400GH/s with a 50GH/s loss due to bottle-necking. People buy machines that clock at 50GH/s, right now it's like I have one of those and your pool's rejecting the whole thing.
What kind of bottle-necking are you talking about? The load on the server is extremely low. The server is basically falling asleep from boredom because of the low load. There is no bottle neck.
If you are getting over 10% rejected work then something is seriously wrong. The average reject ratio for all users in the pool right now is about 0.13-0.14%. If you are not actually getting rejected work, but just saw your 400 GH/s display as 350 GH/s at the website for a moment, this is normal - the website live hashrates are (inaccurate) estimates of your real hashrate.
Please understand this: the server load and network load is almost zero. We find fewer blocks now than some months ago because the difficulty has increased very quickly.
As a miner one of the most important topics for you to research and understand is "difficulty".
To newbie miners: please stop thinking server load is proportional with hashrate. The server load comes from the number of workers mining in the pool. Even the biggest pools today can run on a single server if their software is efficient.
To experienced miners: yes, it is true, fewer and fewer people are mining. Mining is becoming more and more centralized. Small miners give up when their hashpower becomes insignificant and buying significant hashpower is out of their budgets. It is an unfortunate development. This may be important for the future of bitcoin.
Just read this here is my take.. I have 3 s-1's on your pool.
They are down clocked and under volted . on my end they read a total of 420 gh to 430 gh pull 565 watts. here is how your pool shows their hash every shift is good but one.
the one 'bad' shift of 396gh was caused when the miners were moved from my garage to my friends office. (power is 4 cents a k-watt at his place)
May 29, 2014, 07:37:04 AM
Does anyone know what would be an optimal worker's setting, of course, based on worker's "capability"?
General advice, set your minimum difficulty to your hashrate in GH/s divided by 1.4. Info on difficulty, work submits per minute and the easy mode perk:
https://bitcointalksearch.org/topic/m.3142713
Than, what's Stratum extranonce2?
The stratum extranonce2 setting is useful for subpools. Don't worry about it. Most people have no use for it.
May 29, 2014, 04:41:30 AM
Does anyone know what would be an optimal worker's setting, of course, based on worker's "capability"?
Than, what's Stratum extranonce2?
Thanks
Than, what's Stratum extranonce2?
Thanks
May 29, 2014, 03:36:01 AM
Quote
To experienced miners: yes, it is true, fewer and fewer people are mining. Mining is becoming more and more centralized. Small miners give up when their hashpower becomes insignificant and buying significant hashpower is out of their budgets. It is an unfortunate development. This may be important for the future of bitcoin.
Dr Haribo
Unqote
Dear Doc
Im mining since November 13 in your pool and read almost every post in the forum since. Even I dont understand everything I read I understood that this is a great and very helpful forum and you are taking every effort to keep it as friendly and helpfull as you can.
Thanks for that, I will stay mining with bitminter for that.
Here are some of my thoughts on the future of mining:
At the current grow rate bitcoin mining kills it itself with a rat race for hashrate and watts.
In a very foreseeable future it will be unprofitable, even for somebody paying 6c/KWh.
In my opinion the future of mining lays in the spread of specialised computers hosting a blockchain and a wallet and a miner ( say a blockerupter 335 MHs ).
If all bitcoin users would install this into their equipment the system would function self sufficient and the costs and rewards would be distributed evenly among the community.
This would create a rat race between major computer companies to give the most reward to their customers.
And the rat race starts again.
So enjoy to make money, but dont forget that everything earned on one is a loss for someone else.
I think that the real charme of bitcoin is a anarchistic one. I like it.
Joerg
May 28, 2014, 08:40:24 PM
Thanks for the primer, Doc.
Now for the most pressing question: what happened to Koi and their 220 TH/s?
Now for the most pressing question: what happened to Koi and their 220 TH/s?
I hope he\they return. The current questions will, by direct experience, be answered.
We will then know for a certainty that the
Too bad we have to learn the hard way. Sure, I like a bigger piece, but pie will be smaller now. Smaller rewards, more frequently,
is more encouraging for me, though.
May 28, 2014, 07:29:39 PM
To newbie miners: please stop thinking server load is proportional with hashrate. The server load comes from the number of workers mining in the pool. Even the biggest pools today can run on a single server if their software is efficient.
To experienced miners: yes, it is true, fewer and fewer people are mining. Mining is becoming more and more centralized. Small miners give up when their hashpower becomes insignificant and buying significant hashpower is out of their budgets. It is an unfortunate development. This may be important for the future of bitcoin.
To experienced miners: yes, it is true, fewer and fewer people are mining. Mining is becoming more and more centralized. Small miners give up when their hashpower becomes insignificant and buying significant hashpower is out of their budgets. It is an unfortunate development. This may be important for the future of bitcoin.
Just backing these statements up. BTC Guild today is using less bandwidth per month today than ever, even going back to June 2011. For reference, there are single unit ASICs today with more power than the entire Bitcoin network had back in June 2011. The load is also down across the board. Fewer miners in total with higher average speeds than what we've seen in the past.
Stratum scales with the number of connections. The speed of miners is irrelevant on any pool with vardiff, which ALL major pools have. A miner running at 4 GH/s at diff4 and a miner running at 4 TH/s running at diff 4096 will put the same bandwidth and CPU load on the server as each other.
May 28, 2014, 07:02:29 PM
...which will no doubt push out the shift duration
But our shares will go back up, so it should balance itself out. I guess only time will tell.
Who knows, maybe Koi is just having technical difficulties. Anyone know how longs they've been gone?
May 28, 2014, 06:51:18 PM
...which will no doubt push out the shift duration
May 28, 2014, 06:20:29 PM
Thanks for the primer, Doc.
Now for the most pressing question: what happened to Koi and their 220 TH/s?
Now for the most pressing question: what happened to Koi and their 220 TH/s?
May 28, 2014, 06:12:12 PM
Nice job, DrH. Everybody should bookmark this post for when these questions come up again in the future as they most certainly will.
May 28, 2014, 05:39:25 PM
It looks like you don't understand two basic mining concepts, called "variance" and "difficulty". These are the two most important concepts for a new miner to learn and understand. I recommend researching these two concepts before investing in any mining hardware. And please learn from intelligent people like Organofcorti and Meni Rosenfeld, not the people who post misinfirmation on this forum on purpose.
But I will try to respond to your post below:
I don't understand what you mean. There is plenty of free filehandles on the server, we could easily handle a couple hundred thousand more connections. There is also plenty of free bandwidth. I don't see any kind of bottle necks. The load is actually much lower than it has been in the previous 12 months. In the past we have handled loads MUCH MUCH higher than what we see today.
The load on the server is much lower now than when we were below 1 PH/s. There are much fewer connections, much lower CPU usage and much lower memory usage. This is because there are fewer people mining. We can easily handle an unlimited hashpower. It's the number of people/workers mining that creates the load, and the load keeps dropping lower and lower all the time, as the smallest miners give up.
The server is almost falling asleep at this point.
During the GPUMAX days there were loads challenging me to improve the Bitminter server software to properly handle it when GPUMAX turned on the firehose. Eventually, after many improvements, Bitminter became one of their favorite pools because it could handle extreme loads. Today with the stratum protocol and with fewer people mining, the load that mining pools are seeing is very small and any newbie programmer could write software to handle it on cheap and slow server hardware.
This is a modern and non-beta pool. And we are using 100% of the hashing power.
Luck goes up and down. But we are very close to the average expected number of blocks found.
As the difficulty goes up you need more hashpower to find the same number of blocks per day.
No surprise there. It is as expected.
If you think a difficulty increase of 10-15% makes no difference then you don't understand what the difficulty is and you need to research it further. I've heard it from some miners before: "I don't think difficulty actually has any impact on mining". If you have such thoughts then you need to consult your local psychiatrist and/or mathematics professor.
The server does not know your hashrate. The website tries to make an estimate of your hashrate based on how many proofs of work you sent in during the last couple minutes. This is very inaccurate.
You could take a look at the estimated hashrate for a shift rather than the live hashrate esimates. The shift estimates are more accurate as they are estimated over more time. Have a look at https://bitminter.com/shifts
The current server has very low bandwidth usage, very low CPU usage, and very low memory usage. I'm not sure what kind of server upgrade would help. Actually, I know that no server upgrade would help. An increasing bitcoin difficulty isn't balanced out by upgrading a mining server. You need more hashpower, that's the only thing that can do it, and it has nothing to do with the mining server.
What kind of bottle-necking are you talking about? The load on the server is extremely low. The server is basically falling asleep from boredom because of the low load. There is no bottle neck.
If you are getting over 10% rejected work then something is seriously wrong. The average reject ratio for all users in the pool right now is about 0.13-0.14%. If you are not actually getting rejected work, but just saw your 400 GH/s display as 350 GH/s at the website for a moment, this is normal - the website live hashrates are (inaccurate) estimates of your real hashrate.
Please understand this: the server load and network load is almost zero. We find fewer blocks now than some months ago because the difficulty has increased very quickly.
As a miner one of the most important topics for you to research and understand is "difficulty".
To newbie miners: please stop thinking server load is proportional with hashrate. The server load comes from the number of workers mining in the pool. Even the biggest pools today can run on a single server if their software is efficient.
To experienced miners: yes, it is true, fewer and fewer people are mining. Mining is becoming more and more centralized. Small miners give up when their hashpower becomes insignificant and buying significant hashpower is out of their budgets. It is an unfortunate development. This may be important for the future of bitcoin.
But I will try to respond to your post below:
You need to UPGRADE your server, one PH/s I lose too much hashing power because of traffic bottle-necking, too many connections. Tried your pool for a week today and I am disconnecting today too due to having too great of a hash power loss compared to when I joined when you were below 1PH/s. I'll assume you're hosting online, most hosts can do an upgrade in an hour or so.
I don't understand what you mean. There is plenty of free filehandles on the server, we could easily handle a couple hundred thousand more connections. There is also plenty of free bandwidth. I don't see any kind of bottle necks. The load is actually much lower than it has been in the previous 12 months. In the past we have handled loads MUCH MUCH higher than what we see today.
The load on the server is much lower now than when we were below 1 PH/s. There are much fewer connections, much lower CPU usage and much lower memory usage. This is because there are fewer people mining. We can easily handle an unlimited hashpower. It's the number of people/workers mining that creates the load, and the load keeps dropping lower and lower all the time, as the smallest miners give up.
The server is almost falling asleep at this point.
During the GPUMAX days there were loads challenging me to improve the Bitminter server software to properly handle it when GPUMAX turned on the firehose. Eventually, after many improvements, Bitminter became one of their favorite pools because it could handle extreme loads. Today with the stratum protocol and with fewer people mining, the load that mining pools are seeing is very small and any newbie programmer could write software to handle it on cheap and slow server hardware.
That's why you're having such a hard time to find blocks as of late, you're not using the pool's complete hashing potential I am afraid. I observed about 10%-20% average hashing power loss using your pool compared to other modern, non-beta pools I have tested. Has nothing to do with the recent difficulty increase, made really little to no difference.
This is a modern and non-beta pool. And we are using 100% of the hashing power.
Luck goes up and down. But we are very close to the average expected number of blocks found.
As the difficulty goes up you need more hashpower to find the same number of blocks per day.
No surprise there. It is as expected.
If you think a difficulty increase of 10-15% makes no difference then you don't understand what the difficulty is and you need to research it further. I've heard it from some miners before: "I don't think difficulty actually has any impact on mining". If you have such thoughts then you need to consult your local psychiatrist and/or mathematics professor.
I noticed the difference the minute BMT went back over 1PH/s and kept growing on it. I monitor my workers like a hawk, so I know in real time what's going on, at all times of the day and when your stats show almost 50GH/s difference from my miner's interface to yours for most of the time, I start asking questions because that's rather significant, in my book anyways. Connection is steady though.
The server does not know your hashrate. The website tries to make an estimate of your hashrate based on how many proofs of work you sent in during the last couple minutes. This is very inaccurate.
You could take a look at the estimated hashrate for a shift rather than the live hashrate esimates. The shift estimates are more accurate as they are estimated over more time. Have a look at https://bitminter.com/shifts
Your server needs a hardware upgrade and maybe very well, bandwidth.
The current server has very low bandwidth usage, very low CPU usage, and very low memory usage. I'm not sure what kind of server upgrade would help. Actually, I know that no server upgrade would help. An increasing bitcoin difficulty isn't balanced out by upgrading a mining server. You need more hashpower, that's the only thing that can do it, and it has nothing to do with the mining server.
Sorry but I can't mine here with only having 400GH/s with a 50GH/s loss due to bottle-necking. People buy machines that clock at 50GH/s, right now it's like I have one of those and your pool's rejecting the whole thing.
What kind of bottle-necking are you talking about? The load on the server is extremely low. The server is basically falling asleep from boredom because of the low load. There is no bottle neck.
If you are getting over 10% rejected work then something is seriously wrong. The average reject ratio for all users in the pool right now is about 0.13-0.14%. If you are not actually getting rejected work, but just saw your 400 GH/s display as 350 GH/s at the website for a moment, this is normal - the website live hashrates are (inaccurate) estimates of your real hashrate.
Please understand this: the server load and network load is almost zero. We find fewer blocks now than some months ago because the difficulty has increased very quickly.
As a miner one of the most important topics for you to research and understand is "difficulty".
To newbie miners: please stop thinking server load is proportional with hashrate. The server load comes from the number of workers mining in the pool. Even the biggest pools today can run on a single server if their software is efficient.
To experienced miners: yes, it is true, fewer and fewer people are mining. Mining is becoming more and more centralized. Small miners give up when their hashpower becomes insignificant and buying significant hashpower is out of their budgets. It is an unfortunate development. This may be important for the future of bitcoin.
May 28, 2014, 05:21:27 PM
Seems like a nice pool.
Definitly checking it.
Definitly checking it.
May 28, 2014, 11:56:32 AM
I am starting to think the cause of this is because FW or something
Not sure what it is. Let's see if it helps when the new server version is up and running.
---
Hey Doc, some advice/tips from a webmaster/game developer...
You need to UPGRADE your server, one PH/s I lose too much hashing power because of traffic bottle-necking, too many connections. Tried your pool for a week today and I am disconnecting today too due to having too great of a hash power loss compared to when I joined when you were below 1PH/s. I'll assume you're hosting online, most hosts can do an upgrade in an hour or so.
That's why you're having such a hard time to find blocks as of late, you're not using the pool's complete hashing potential I am afraid. I observed about 10%-20% average hashing power loss using your pool compared to other modern, non-beta pools I have tested. Has nothing to do with the recent difficulty increase, made really little to no difference.
I noticed the difference the minute BMT went back over 1PH/s and kept growing on it. I monitor my workers like a hawk, so I know in real time what's going on, at all times of the day and when your stats show almost 50GH/s difference from my miner's interface to yours for most of the time, I start asking questions because that's rather significant, in my book anyways. Connection is steady though.
Your server needs a hardware upgrade and maybe very well, bandwidth.
Sorry but I can't mine here with only having 400GH/s with a 50GH/s loss due to bottle-necking. People buy machines that clock at 50GH/s, right now it's like I have one of those and your pool's rejecting the whole thing.
I hope you take from what I just said, lots of people pay good money for my advice. But again, my knowledge on mining is still rather rudimentary but my knowledge on making money, computers and web design is spot on =D
I mean, it's not just your pool with bottle-necking but I actually like BitMinter... well used to thus so far anyways. Maybe it's just me? I doubt it, your pool does stand out in the bottle-neck column. That's why i don't hash with GHash anymore, bottlenecks and gaps were too great during the day.
Peace out
Ive ALWAYS made more here....
Jump to: