In my experience, upgrading from gen4 to gen5 did not affect verification time on local data as much as I expected it would. CPU being under utilized was evidence that something else was the bottleneck (RAM clock?). I didn't measure times but I remembered they were about the same. Granted the gen 4 index happened 2 years previous. Would two years of blocks wipe out any gains? CPU utilization seemed consistent between the two.
blockstream-electrs reindex between pcie4 and 5 did make a big difference (full mode).
