Scrypt-N

Adaptive N-Factor

What is Adaptive N-Factor, and why should you care?

When Litecoin was first developed, the idea was to resist ASIC development by changing the proof-of-work algorithm used by Bitcoin (SHA256) for a different one (Scrypt) which had higher memory requirements. It’s expensive to put memory on an ASIC, so the Litecoin developers believed that increasing the memory requirements would hold off development of custom hardware for mining.

It worked for a while. Even though Bitcoin ASICs are very well established now (and mining Bitcoin with anything other than an ASIC is basically impossible as a result) we are only just starting to see the first Scrypt ASICs now (Gridseed GC3355) which are capable of mining Litecoin and it’s clones.

There’s a parameter in the scrypt algorithm called N, N must be a power of two greater than one. N in Litecoin is 1024 (2^10 = 1024)

N-Factor is log_2(N), N-Factor in Litecoin is = 10.

To execute the scrypt algorithm, memory of approximately 128*r*N bytes is needed. r in Litecoin = 1. Using that information, we can calculate how much memory Litecoin requires – around 1280 bytes per thread.

In Litecoin and it’s clones, N is a fixed number – it will always be 10. Now that ASICs are coming out to mine this algorithm, they will always be able to mine it.

Vertcoin changes the N parameter over time. Currently, N=11 in Vertcoin. Because memory requirements increase according to a schedule, and if you were to successfully build an ASIC to mine when N=11, it would be useless as soon as the change to N=12 happened (February 2016). As such, the huge development costs involved in ASIC production would never be paid back. Changing N over time puts financials/practicality barriers in the way, strongly discouraging ASIC design – who is going to invest money developing an ASIC for something where that ASIC will be useless shortly after it’s release? It’s still technically possible to eventually make an ASIC that can mine a given N-factor, but the cost to develop them won’t be worth the effort since they’ll need to be completely remade for the next N.

The schedule of N-factor changes is:

| N | Memory | Timestamp | Date/Time | |

|———-|——–|————-|——————————-|—|

| 2048 | 256 kB | 1389306217 | Thu, 09 Jan 2014 22:23:37 GMT | |

| 4096 | 512 kB | 1456415081 | Thu, 25 Feb 2016 15:44:41 GMT | |

| 8192 | 1 MB | 1506746729 | Sat, 30 Sep 2017 04:45:29 GMT | |

| 16384 | 2 MB | 1557078377 | Sun, 05 May 2019 17:46:17 GMT | |

| 32768 | 4 MB | 1657741673 | Wed, 13 Jul 2022 19:47:53 GMT | |

| 65536 | 8 MB | 1859068265 | Tue, 28 Nov 2028 23:51:05 GMT | |

| 131072 | 16 MB | 2060394857 | Tue, 17 Apr 2035 03:54:17 GMT | |

| 262144 | 32 MB | 2463048041 | Sun, 19 Jan 2048 12:00:41 GMT | |

| 524288 | 64 MB | 2999918953 | Fri, 23 Jan 2065 06:49:13 GMT | |

| 1048576 | 128 MB | 3536789865 | Wed, 28 Jan 2082 01:37:45 GMT | |

| 2097152 | 256 MB | 5684273513 | Mon, 16 Feb 2150 04:51:53 GMT | |

| 4194304 | 512 MB | 7831757161 | Sat, 07 Mar 2218 08:06:01 GMT | |

| 8388608 | 1 GB | 9979240809 | Thu, 25 Mar 2286 11:20:09 GMT | |

| 16777216 | 2 GB | 16421691753 | Fri, 19 May 2490 21:02:33 GMT | |

| 33554432 | 4 GB | 22864142697 | Sun, 15 Jul 2694 06:44:57 GMT | |

source: http://vertcoinforum.com/index.php?topic=5.0

Scrypt-N