Post
Topic
Board Announcements (Altcoins)
Re: [ANN] [PASC] PascalCoin, deletable blockchain & bank account system [PASA]
by
D8V1D
on 27/01/2017, 21:08:00 UTC
Hey everyone, updated the CUDA GPU miner with some of KlausT's improvements: https://github.com/Vorksholk/PascalCoin-CUDA/releases/tag/v1.04

No promise that the new miner can actually mine blocks, but if anyone gets a block with this, please let the rest of us know. About 360% performance improvements, I'm getting around 900 MH/s on an overclocked 1080.

I've also seen some people worried about the PascalCoin wallet saying "Error: Proof of work is higher than target payload." This isn't a problem--the way my miner works (and presumably some others that are floating around now, although I haven't looked at the code for them) is by submitting shares that meet a target of 24000000. Note that we are currently on track for a target of 31000000 soon, which is 8192 times higher, so you should expect to see, on average, 8192 of those errors per actual block mined. The miner is submitting shares to the wallet, effectively. These shares don't count for anything unless they meet the target.


You make a miner but say "No promise that the new miner can actually mine blocks".

WTF is this, how do you not know, if the miner that you made actually works?

It's possible that one of the optimizations (namely the funnelshift, can't imagine an unroll causing a problem) could introduce a bug by not behaving the way I expect. The way to truly know a miner works is to mine a block with it (although you can be pretty certain that it works by manually checking hashes).


So If it doesn't mine a block, we'll never know if it actually working.

People will be mining for days, saying they can't find a block, not knowing if the miner actually works.

What a waste of time and electricity.

That's the game of mining for you. Smiley


That's like saying. I made a really fast car, 360% performance improvement, goes 900 mph.
But no promises that the new car will actually work.