Status as of Monday, June 12, 2017

Here's a few notes for today:

• I completed all development and initial testing for the random share reduction improvement and 2-factor authentication. Chris will perform final testing of these things tonight.
• In the next release, I'm going to look into figuring out a way to prevent connections from IP addresses who have never connected to the pool before, or to allow only a certain number of new connections. Hopefully this limitation will improve the experience for customers. Meanwhile, mining continues normally, as the system is designed to prioritize mining above all other functions when resources are limited.
• The database is currently backed up 27GB of share data. It will take a while to catch up, and if the database-operator crashes, mining will continue but Chris will have to make manual corrections himself.
• After this release, the next task will be to continue improving CPU performance, as I suspect that CPU will again become the limiting factor.
• We reviewed past records and determined that in November 2014, during the Elacoin surge, the system encountered performance difficulties at about 6 GH/s scrypt, while now the system handles about 700 total GH/s. This shows how the future of computing performance improvements lies with software, as we have been able to achieve a 120x improvement over three years.

the 6gh/s vs 700gh/s statistic you mentioned, is that based on the same hardware, minus the new installed server obv. oris that including hardware upgrade over the last 3 years plus software improvements to upgrade efficiency?

How far back does that 27GB go, in terms of real time? A measure of minutes? hours?

GregoryGHarding wrote:the 6gh/s vs 700gh/s statistic you mentioned, is that based on the same hardware, minus the new installed server obv. oris that including hardware upgrade over the last 3 years plus software improvements to upgrade efficiency?

Only about one doubling was caused by hardware, if that. The remaining 7 doublings were caused by software improvements.

Hardware is not a long-term solution to most computer science problems; the only way that we can resolve these issues is through software.

Steve Sokolowski wrote:

GregoryGHarding wrote:the 6gh/s vs 700gh/s statistic you mentioned, is that based on the same hardware, minus the new installed server obv. oris that including hardware upgrade over the last 3 years plus software improvements to upgrade efficiency?

Only about one doubling was caused by hardware, if that. The remaining 7 doublings were caused by software improvements.

Hardware is not a long-term solution to most computer science problems; the only way that we can resolve these issues is through software.

i agree with you steve, software efficiency is where its at, and as new ways of doing things are uncovered or made available, the software changes makes big differences.i was just curious.

Steve,

I have been thinking about this problem of limiting the new connections... We cant stop growth so why not capitalize on it...

Make miners that sign up after July 1st subject to 75% of current member's earnings. This way they will have to spend a lot of time here in order to earn the most. You could of course on the backstroke give all existing members a 10% bump as a loyalty bonus

Just throwing things out there.
Tmopar

tmopar wrote:Steve,

I have been thinking about this problem of limiting the new connections... We cant stop growth so why not capitalize on it...

Make miners that sign up after July 1st subject to 75% of current member's earnings. This way they will have to spend a lot of time here in order to earn the most. You could of course on the backstroke give all existing members a 10% bump as a loyalty bonus

Just throwing things out there.
Tmopar

or just pocket the 25% lol

GregoryGHarding wrote:

tmopar wrote:Steve,

I have been thinking about this problem of limiting the new connections... We cant stop growth so why not capitalize on it...

Make miners that sign up after July 1st subject to 75% of current member's earnings. This way they will have to spend a lot of time here in order to earn the most. You could of course on the backstroke give all existing members a 10% bump as a loyalty bonus

Just throwing things out there.
Tmopar

or just pocket the 25% lol

This is a great idea for existing miners, but it isn't quite fair to everyone else. Given that you have an account already, I can see why you would appreciate it, of course.

I don't think that we will need to limit things for right now, given that we have 1.05TH/s capacity at least, and we are only at 777GH/s at present. We'll reevaluate the situation if 250 more GH/s arrives on the pool.

Steve Sokolowski wrote:

GregoryGHarding wrote:

tmopar wrote:Steve,

I have been thinking about this problem of limiting the new connections... We cant stop growth so why not capitalize on it...

Make miners that sign up after July 1st subject to 75% of current member's earnings. This way they will have to spend a lot of time here in order to earn the most. You could of course on the backstroke give all existing members a 10% bump as a loyalty bonus

Just throwing things out there.
Tmopar

or just pocket the 25% lol

This is a great idea for existing miners, but it isn't quite fair to everyone else. Given that you have an account already, I can see why you would appreciate it, of course.

I don't think that we will need to limit things for right now, given that we have 1.05TH/s capacity at least, and we are only at 777GH/s at present. We'll reevaluate the situation if 250 more GH/s arrives on the pool.

I agree that it is tilted to earlier adopters but many things in this are done this way, think of the Eth presale, block sizes are larger in the beginnings of all coins etc.

I can see them returning to 100% earnings within 3 months... this means that we would have at least some mitigation against a flood in demand amounting to a denial of service. That basically is what is going on, when too many new people show up the system goes down for the existing loyal miners. This percentage makes other pools slightly more profitable for them but if they choose to mine here we all benefit a little extra and they are paying for the new infrastructure.

Think of it more as a toll on a road... People want to cross a bridge but it simply wont build itself. The users must pay a fee to cover the construction and interest on the loans etc. This is very much akin to what I am describing, basically charging a toll for the first few months and reducing the tolls on the early contributors. In the end everyone will benefit because the windfall from these users could be directly allocated to new infrastructure.

I cant speak for everyone but I point 100% of my scrypt hashing power here and so maybe I am a little biased. This is a good thing for you guys though and something that should be fostered. It is possible for competition to emerge any day. If you make newcomers work for their extra percentage as well as giving existing customers a bonus they will both feel vested and continue here also.

Tmopar

Steve Sokolowski wrote:

GregoryGHarding wrote:

tmopar wrote:Steve,

I have been thinking about this problem of limiting the new connections... We cant stop growth so why not capitalize on it...

Make miners that sign up after July 1st subject to 75% of current member's earnings. This way they will have to spend a lot of time here in order to earn the most. You could of course on the backstroke give all existing members a 10% bump as a loyalty bonus

Just throwing things out there.
Tmopar

or just pocket the 25% lol

This is a great idea for existing miners, but it isn't quite fair to everyone else. Given that you have an account already, I can see why you would appreciate it, of course.

I don't think that we will need to limit things for right now, given that we have 1.05TH/s capacity at least, and we are only at 777GH/s at present. We'll reevaluate the situation if 250 more GH/s arrives on the pool.

just a thought but you want to keep in mind, if youre going to re-eval when the usage goes up, we risk another site/mining issue with the organic DoS, in my opinion its best to think ahead, build a system, turn it on when it gets to that point. no down time, happy customers, loyalty matters