If someone mentioned it already, point me to it. What you just did is PoW with variable time/space tradeoff as used in cryptanalysis and password crackers. In current PoC, the split ratio it permanently fixed, disfavouring PoW heavily. It's been actually discussed few years ago in #bitcoin as PoW alternative. For each bit of nonce storage shaved off, the readout can brute force it in a time-space tradeoff. Assuming we get 2^22 of 32bit nonces in a scoop number/1TB readout where we hash 2^22 times to get 2^22 candidate deadlines, we could store 2^22 31bit nonces, and hash 2^23 times during a readout... with 1Gh/s GPU you can get rid of 10 bits or so. Yes, it's a logarithmic relation, but profitable to PoW up to a point (1/3 split PoW/Poc). The problem is that one needs more and more faster PoW as PoC gets larger, potentially up to 50/50 split of the tradeoff (when using ASICs for the POW part).

Ironically, this solves your point 1, as POC2 is capped by PoW requirement introduced.

Bottomline, plotting will be very slow using groestl as it is. Forget about slowing it down even more with that PoW during plotting, the throughput is not up for it. Salsa kernel or even siphash (1GByte/s) could bring it to speeds of current plotting, with 6-8 zero bits plot-PoW (1GByte/s & 8 bit pow -> 4MByte/s). Ideally, this step can be avoided altogether, and just use large totalscoops instead, though it brings engineering difficulties to make memory usage during plotting sane.

why?

I seem to be having an issue with the lotteries, when I go to http://localhost:8125/atlotteries.html it gives me file missing error, but the file is clearly there. What is it that I'm not doing correctly

Can I use my atom cpu + 2GB to mine some plotted TB SATA hdds?

Thanks

Thanks, I was worried if cpu speed if somehow a big factor in hash speed and I guess the same with ram?

Just planning to use my windows home server for burst mining.

We have noticed that a lot more people are able to answer mining questions now than before - that's nice But ideally we'd not have any questions if the guides were good enough.

We would like to update crowetic's guide but have little time doing so. Here it is: https://docs.google.com/document/d/1ytq47AmQZ1rzjoFuMW33c82JkuJ1L61Pdp95ZLbi3-k/

So here's an offer: if you feel like you could improve the text, email burst@countermail.com, include your BTT nick and some links to answers you've given in the thread or at BurstForum.com, and I'll give editing rights on the document. If you contribute constructively to the guide, I will send at least 5000 BURST to you as a token of appreciation. If you do a lot of quality work, the payment will be considerably higher.

We want the best possible guide


If not a minimum of three people email their interest and are considered qualified, I'll make it open for all to edit but will not pay anyone.

---

The same conditions applies to TetraHect0rCannabinol's linux guide found here: https://bt.irlbtc.com/view/731923.msg8299637#msg8299637 It will also be on google docs.

Thanks!



***EDIT: this is cross-posted at BurstForum, and the post here will be deleted eventually trying to move of the activetly to BurstForum.com, as there are much better possibilities at that forum to separate topics and questions.
***EDIT2: If you have try to email already I'm sorry, but I've corrected the email address now: burst@countermail.com and nothing else.

Impressive! A couple of things:
- a video guide is being produced already, so if I were you I would not start that project at now.
- in the post I said that We want the best possible guide! We could use your document and crow's - and you get paid and might get ppl helping you. How about that?