00000011ee1088ab58ade91da86761478799b685699b30bc649c83f13725bc6d 000000037a35ebf60c59619042ece8ae71fecf4d3146098c966f820482d33955 0000001253d4dbf6b7b34079a70ef0357a7b12e83a38f29946effbc863f194f3 0000000b4f9d1ba885832b1b9dd4f6183483a27a4762a48114c00a812f1b42dd 00000011240c362604a624fc4469413f36ff5cb823c0733d570d54c7916a913f 000000065d332b249b5fc8068177776b3dddb073d308ae5b090147c41477e351
3 Likes
Someone DM’d me their validation times on a RPi 5 running Umbrel and Core 30.2.
First run
2026-04-08T14:05:16Z [bench] - Connect block: 22887.59ms [6818.44s (22.79ms/blk)]
2026-04-08T14:06:47Z [bench] - Connect block: 21731.38ms [6840.17s (22.86ms/blk)]
2026-04-08T14:07:33Z [bench] - Connect block: 21029.07ms [6861.20s (22.93ms/blk)]
2026-04-08T14:08:40Z [bench] - Connect block: 21478.89ms [6882.68s (23.01ms/blk)]
2026-04-08T14:09:41Z [bench] - Connect block: 23007.94ms [6905.69s (23.08ms/blk)]
2026-04-08T14:11:08Z [bench] - Connect block: 23214.31ms [6928.90s (23.16ms/blk)]
Second run
2026-04-08T22:04:35Z [bench] - Connect block: 20975.20ms [6952.16s (23.23ms/blk)]
2026-04-08T22:05:26Z [bench] - Connect block: 23250.04ms [6975.41s (23.31ms/blk)]
2026-04-08T22:06:49Z [bench] - Connect block: 20488.56ms [6995.89s (23.38ms/blk)]
2026-04-08T22:07:41Z [bench] - Connect block: 23458.84ms [7019.35s (23.46ms/blk)]
2026-04-08T22:08:35Z [bench] - Connect block: 22032.76ms [7041.39s (23.53ms/blk)]
2026-04-08T22:10:07Z [bench] - Connect block: 20499.21ms [7061.88s (23.60ms/blk)]
Third run
2026-04-09T09:03:14Z [bench] - Connect block: 22365.10ms [7086.41s (23.68ms/blk)]
2026-04-09T09:04:43Z [bench] - Connect block: 22936.59ms [7109.35s (23.75ms/blk)]
2026-04-09T09:05:37Z [bench] - Connect block: 22544.76ms [7131.89s (23.83ms/blk)]
2026-04-09T09:06:47Z [bench] - Connect block: 22968.04ms [7154.86s (23.90ms/blk)]
2026-04-09T09:08:20Z [bench] - Connect block: 21931.17ms [7176.79s (23.98ms/blk)]
2026-04-09T09:10:11Z [bench] - Connect block: 22856.47ms [7199.65s (24.05ms/blk)]
1 Like
The second round with an i7 870 @ 2.93GHz and Bitcoin Core v30.2 00000011ee1088ab58ade91da86761478799b685699b30bc649c83f13725bc6d 23 000000037a35ebf60c59619042ece8ae71fecf4d3146098c966f820482d33955 17 0000001253d4dbf6b7b34079a70ef0357a7b12e83a38f29946effbc863f194f3 26 0000000b4f9d1ba885832b1b9dd4f6183483a27a4762a48114c00a812f1b42dd 15 00000011240c362604a624fc4469413f36ff5cb823c0733d570d54c7916a913f 14 000000065d332b249b5fc8068177776b3dddb073d308ae5b090147c41477e351 25
2 Likes
second run
raspberry pi 4 B - Bitcoin Core v30.2.0
from when block was reconstructed to when tip was updated
| block hash | delay (seconds) |
|---|---|
00000011ee1088ab58ade91da86761478799b685699b30bc649c83f13725bc6d |
73 |
000000037a35ebf60c59619042ece8ae71fecf4d3146098c966f820482d33955 |
74 |
0000001253d4dbf6b7b34079a70ef0357a7b12e83a38f29946effbc863f194f3 |
77 |
0000000b4f9d1ba885832b1b9dd4f6183483a27a4762a48114c00a812f1b42dd |
76 |
00000011240c362604a624fc4469413f36ff5cb823c0733d570d54c7916a913f |
74 |
000000065d332b249b5fc8068177776b3dddb073d308ae5b090147c41477e351 |
73 |
2 Likes
I ran a custom P2P client during last weeks slow blocks on Signet event that measured block propgation duration and block validation duration. I’ve written about my methodology and results in detail in Block propagation and validation duration during slow-to-validate blocks on Signet - Observations - Bitcoin Network Operations Collective - but wanted to share some of the findings here:
-
normal block propgation on signet is usally done in under 1s. 50th percentile has the block at about 200ms, and the 90th percentile is 600ms
-
the slow-to-validate blocks propagated a lot slower: 40s for 50th percentile and nearly 2 minutes until the 90th percentile has annouced the block.
-
obviously, the slow-to-validate blocks are slower than the normal blocks. This plot visualizes this:
-
It takes the median node I was conencted to 176ms to validate a normal block, but nearly 20s to validate a slow block:
-
Looking at the nodes indiviually, the median node experiences a 160x slowdown on these blocks. An according to @AntoineP, these are far from the worst blocks one can construct:
The numbers put my host into perspective: It’s probably close to the 90th percentile and a comparativly slow host. I forgot how old CPUs from 2019 are..
With around 53s, your thin-client is around the 80th percentile range. Maybe a bit above.
With about 2s, you are below the 5th percentile!
Your NanoPC-T6 is probably somewhere around the 40th percentile with 16s.
With around 25s, you are a bit slower than the median. Probably 60th percentile.
Probably around 5th percentile at around 3s. Little bit slower than @svanstaa.
With 55s, you are a tiny bit slower than the 80th percentile on that M3 Pro. @Emzy’s thin-client is faster! Better stick with that NanoPC-T6 for verifying slow-to-validate signet blocks.
That’s around 15th percentile with close to 5s valdation time.
I think that’s about 4th percentile. Tiny bit faster than @ViniciusCestarii but slower than @svanstaa.
Very close to the median with 22s. A tiny bit slower.
You cut of the headings of your table, but it seems like your close to the 25th percentile with 9s.
Just above median with 22s.
Probably higher than 90th percentile with about 100s.
Remember, I measured this against listening nodes and I think many of the ones here aren’t listening. The duration distribution might look different on the full network.
5 Likes