WTH Net DIff ??? bipolar much



  • what is with the difficulty i have seen it as high at 64k and as low as 32k today just randomly looking at my miner. its always fluxed but have never seen it swing so far in just a matter of a few blocks


  • admin

    @Gibsalot It's been doing this for a few days, the difficulty is all over the map.



  • @haitch that diff is calculated by the pool right? The one that shows in the miner...
    How is it calculated?
    I think it varies so much because the pool calculate it with base in the best DL and number of DL submitted...

    So my line of thought is that pools submitting less DLs due to small targetDLs can be the cause of the issue because less DLs are being submitted, and only the best ones... I'm not sure if this is the way it is calculated.



  • @gpedro i was under the impression it was a direct result of the last block time vs target block time of 4 min , and the dif is raised or lowered depending if the previous winning block was under or over the 4 min target ... effectivly the dif is the varrible that keeps our average 4 min block time dispite the total hash power growing or shrinking.


  • admin

    @gpedro The difficulty is calculated by the whole network, not individual pools. It's my understanding is that it adjusts to target 4 minute blocks, if there is a sequence of fast blocks, it'll go up, a sequence of slow blocks and it'll drop.



  • @Gibsalot said in WTH Net DIff ??? bipolar much:

    what is with the difficulty i have seen it as high at 64k and as low as 32k today just randomly looking at my miner. its always fluxed but have never seen it swing so far in just a matter of a few blocks

    I know! Ugh! Somewhere there was a post about the global size of total hard drive space dedicated to Burst mining growing rapidly in the last few weeks. So basically lots of new miners and/or lots of Petabyte or near Petabyte sized rigs being added to the network. So the network hashratehas gone way up.

    I found the following online regarding Bitcoin:

    *The Bitcoin difficulty started at 1 (and can never go below that). Then for every 2016 blocks that are found, the timestamps of the blocks are compared to find out how much time it took to find 2016 blocks, call it T. We want 2016 blocks to take 2 weeks, so if T is different, we multiply the difficulty by (2 weeks / T) - this way, if the hashrate continues the way it was, it will now take 2 weeks to find 2016 blocks.

    For example, if it took only 10 days it means difficulty is too low and thus will be increased by 40%.

    The difficulty can increase or decrease depending on whether it took less or more than 2 weeks to find 2016 blocks. Generally, the difficulty will decrease after the network hashrate drops.* -from a post on StackExchange authored by Meni Rosenfeld.

    Since it is often quoted that Burst blocks are mined on average every 4 minutes, I'm guessing the Burst difficulty must follow a similar procedure to modulate or throttle the mining with regards to more or less hashrate.



  • @haitch @Gibsalot Maybe the blockchain should look for more blocks to calculate an average, It may be just using last block and that is pretty uncertain, if it uses an average of lets say a couple of weeks it should be a lot more stable 😉

    EDIT: just saw @Evo post hehehe Looks like what I said is what is used in BTC, maybe Burst uses only just one block and should use an average 😉


  • admin

    @gpedro said in WTH Net DIff ??? bipolar much:

    @haitch @Gibsalot Maybe the blockchain should look for more blocks to calculate an average, It may be just using last block and that is pretty uncertain, if it uses an average of lets say a couple of weeks it should be a lot more stable 😉

    EDIT: just saw @Evo post hehehe Looks like what I said is what is used in BTC, maybe Burst uses only just one block and should use an average 😉

    I don't think it looks at just the last block, but I think it needs to look at more blocks in order to smooth out the variations.



  • @haitch , process control is a discipline that I spent many years doing. If you try to control a process variable (blocks/day) to a setpoint (360) by a controlled variable (net difficulty) too tightly you may end up with swings in the controlled variable to maintain the process close to the setpoint. If you relax the requirements on the process you end up with less volatility on the controlled variable at the expense of the process not maintaining tight to setpoint. This is exacerbated by the fact that our process is not analog/continuous, but it is random in nature. If you close a valve by 10% you immediately get a corresponding reduction in fluid flow that is predictable immediately. Raising net difficulty to drive DL higher to spread block generation may still result in random blocks at lower times than before the change took place. Additionally, the load on the system (actual capacity) moves in step changes (as opposed to continuous) as people see mining as a lucrative pursuit or fall off the wagon as prices/block reward decrease.


  • admin

    @rds Great explanation of control processes, but I'm not one of the devs in the wallet. They are the the ones that need to look at changes to the balancing algorithm.



  • The classic control algorithm for an analog system would be to look at the actual block/day and measure the deviation between that and 360. The deviation would feed into a proportional/integral/derivative control system to provide an intermediate output. Each component (P, I, or D) can be scaled to provide the best response for optimum control to load changes. That intermediate output would be biased by a feedforward signal that would be developed by monitoring the latest block times. For this process though, it would have to be some average over at least a few blocks because of the total randomness of any one DL (block generation time).



  • well at over 62k diff even my 21TB is going multi block with no found deadlines...... kind of sucks



  • Depends on the pool, a low deadline pool will act to multiply the difficulty as it goes up.

    Have to consider either adding TB's or dropping to a higher minimum deadline pool to stay competitive if you're not getting DLs anymore.



  • @captinkid im on one of the highest deadline pools , only 2 pools that take higher deadlines that i know of. the mobile and the new pool Ping started
    im on @tross pool , ran buy the guy that keeps fighting to have deadlines raised on all pools so smaller miners can join big pools and be able to submit deadline's.



  • Just have to see how it works out, the way people are dumping drives into burst right now, it might force pools to raise their deadlines because there won't be anyone left who can use them otherwise.



  • looking at my analytics it would seem japan has been the most recent addition to our miner pool
    yesterday that shows the us.

    0_1497866186079_ccca04e7-1c42-4384-82ab-f31dd7a8a51f-image.png

    top that off i seem to be gaining about 100 miners daily 😕



  • This post is deleted!