Quote from: anonymous

Infinium-8 just released is the first cryptonote to have inflation but no other features

They have chosen an interesting inflation peramater
Block_reward = log2(difficulty) * 2^40

Essentially the block reward increases linearally with difficulty.
The inflation on this coin will still effectivally go to zero but not nearly as fast as those with only a fixed block reward like dogecoin.

log2 (2^x) = x

log10(10^x) = x

log2(10^x) = x / log10(2)

Thus we see that exponential growth (K^x) is converted to linear growth using a log of any base, not just logK. The base K is only a constant factor (as shown by the / log10(2)).

So they convert an assumed exponential growth in hashrate to a linear growth in nominal debasement rate.

Edit: they are increasing the nominal debasement rate, not decreasing it as most coins do. If the hashrate growth eventually slows, then the percentage rate of debasement must decline asymptotically towards 0, but early on the curve is much less favorable to early adopters and thus more fair and theoretically better for the network effects.

Edit#2: it rewards adding more hashrate to the system which is good for security. But a doubling of the hashrate at much higher nominal levels has a much less increase on nominal debasement, e.g. log(4) / log(2) = 2 and log(16) / log(8 ) = 1⅓.

Quote from: smooth on July 27, 2014, 04:44:21 AM

Quote from: AnonyMint on July 27, 2014, 04:04:57 AM

I admire this idea [blockchain as scratchpad] as a short-time way to help insure CPU-only. I need to study it more to see if I can find any flaw.

It certainly is not CPU-only even short term as there have been private GPU miners with a significant advantage since shortly after launch (and public ones recently).

Implicit in my thinking was the GPU resistance comes from the AES instructions, as I had indicated to you in our prior public discussions. I as speaking to the ASIC resistance, while addressing the slow speed of Monero's Pow hash.

Note I don't expect tromp's Cuckoo hash to be GPU resistant, because it can be (even if sublinear) parallelized. I don't know if anyone has tested parallelizing it on more than the two dozen cores he tested?