I did a benchmark IBD a long time ago. It's about 5% for assumevalid 0. It's massive for certain forms of massive. 😆

The fact that everything else is twice as fast will probably have a decent impact as well.

And it seems they upgraded the power systems quite a bit, which was causing a lot of trouble on Pi 4 with dodgy power supplies.

So, even though it not perfect due to closed binaries, it looks like a really nice upgrade on paper.

https://github.com/bitcoin/bitcoin/pull/24115#issuecomment-1026969684

Reply to this note

Please Login to reply.

Discussion

On a RockPro64:

```

processor : 5

BogoMIPS : 48.00

Features : fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid

CPU implementer : 0x41

CPU architecture: 8

CPU variant : 0x0

CPU part : 0xd08

CPU revision : 2

```

Bitcoin core I wrote long ago says:

```

2022-09-28T11:51:15Z Using the 'arm_shani(1way,2way)' SHA256 implementation

```

Using SATA SSD disks on a real SATA3 connection (not USB) on LUKS (RockPro64 has the encryption extensions) still takes around a week to complete.

s/Bitcoin core I wrote long ago/Bitcoin core log I got long ago/