Hi Eric,

> Libbitcoin significantly outperforms bitcoind on a 16GB Raspberry Pi5.

Independent benchmarks so far don’t confirm better performance of Libbitcoin4
even when compared to Bitcoin Core v30 even with 32 GB RAM: 
https://blog.lopp.net/2025-bitcoin-node-performance-tests/

I won’t run those benchmarks because in the likely case that they don’t match
up with your marketing slide, you will probably make new fake claims of me not
being honest again, which I am getting very tired of. Instead, I will ask the
community to impartially run benchmarks and share them in a constructive manner.

One important thing to add to that: these benchmarks need to be run without the
network being a significant factor obviously. This is where the config of the
nodes clearly diverges and where (to my understanding at least) Libbitcoin
behaves very differently from Bitcoin Core. First of all, from the config it’s
very clear that you are using 100 outbound connections, which is of course, an
order of magnitude more than what Bitcoin Core uses. But in order to get these
100, Libbitcoin will connect to up to 500 nodes in parallel and let them race
against each other and only keep the ones that are fastest to respond. I could
not identify any other mechanisms to ensure peer diversity of any kind
(correct me if I am wrong here).

This is terrible for many reasons but I will pick the IMO most important ones:
First, making this many connections is not sustainable for the network as a
whole. If every node would be configured like this a new node would have an
extremely hard time finding any connections because the inbound slots would
likely all be exhausted all the time. Second, just taking the nodes that are
the fastest to respond and ignoring everything else is just begging to be
eclipse attacked. Or if there is no attacker, at least to be connected to
only to nodes in the nearest AWS data center.

So these two examples make clear where Bitcoin Core and Libbitcoin really
diverge: Bitcoin Core cares deeply both about the broader network health as
well as the security of the individual node. Performance is great of course
but it usually takes a backseat compared to these other two. Libbitcoin only
cares about performance and that is certainly also a fine approach in
isolation, though personally I wouldn’t run a node that prioritizes that
above network health and security. But what is not fine IMO is to try to dunk
on Bitcoin Core for marketing purposes using only performance numbers and
completely glossing over all these other considerations that Bitcoin Core is
taking into account and you are ignoring. Really the only reason Libbitcoin
can even realistically make 100 outbound connections quickly enough for your
benchmarks to even run at all is because of the very conservative
configurations Bitcoin Core has chosen and which you are repeatedly mocking
on Twitter. And I think it’s also pretty funny that you are accusing me of
trying to off-load hosting cost to the P2P network when you yourself are
exploiting it in a manner that is completely unsustainable.

If you are interested in fixing your eclipse attack vulnerability and related
issues, I would recommend you take a look at ASmap: 
https://github.com/bitcoin-core/asmap-data

FWIW, I hate that this conversation has veered so far off-topic, but most of
this ML thread never actually discussed the BIP proposal and I need to respond
to what you are putting out here, including your latest ad-hominem attacks
(“snake oil”, “delusional”). I have also responded in the PR now since you
were trying to put words in my mouth again and I guess I need to repeat
myself over and over now again because, otherwise, my opinion might have
changed in the meantime: 
https://github.com/bitcoin/bitcoin/pull/35054#issuecomment-4530515268

Best,
Fabian

On Sunday, May 24th, 2026 at 8:29 PM, Eric Voskuil <[email protected]> wrote:

> Responding to a few earlier comments.
>
>> The same applies to the 32 GB RAM minimum noted for libbitcoin... which
>> already excludes a large share of consumer hardware outside high-income
>> regions...
>> Following the trajectory of the last few years to its conclusion, the
>> minimum is likely to be 64 GB next year, then 128 GB the year after, and
>> validating with libbitcoin will soon require the latest generation of
>> Nvidia chips. Validation time will keep going down on that path, but the
>> user base it serves will keep shrinking.
>
> Libbitcoin significantly outperforms bitcoind on a 16GB Raspberry Pi5.
>
> https://x.com/evoskuil/status/2058574669708443756
>
> All it takes is changing the config settings that target larger machines. 
> Despite 16 years of optimization and money poured into bitcoind, and its 
> various clones and ports, it underperforms on both modest and high end 
> hardware. We are a small team of volunteers, just getting started on 
> performance tuning and optimization.
>
>> I am obviously exaggerating somewhat to make a point:
>
> You aren't just exaggerating, you sound delusional.
>
>> it is easy to construct a plausible-sounding slippery slope argument,
>
> "plausible-sounding" is one way to put it, and yet the truth of it was 
> clearly demonstrated 13 days after you dismissed it.
>
>> including one about libbitcoin's hardware trajectory. And I don't think
>> you are able to prove today that libbitcoin will not follow this path in the
>> future. That is exactly why I don't find your similar argument persuasive.
>
> Well it only took me 13 days for the stunning arrival of the very scenario 
> you denied. And the hardware fantasy that you fabricated to make this point 
> actually applies to bitcoind architectures, not Libbitcoin. I would call that 
> unpersuasive.
>
>> And in all seriousness, I think at least some Bitcoin implementations should
>> aim to be accessible with low bandwidth and minimum hardware requirements
>> compatible with widely available consumer hardware outside high-income 
>> regions.
>> That is what many of the developers I work with are aiming for.
>
> No, that is what we are doing. What you are doing here is called snake oil. 
> My suggestion is you roll the clock back a decade and do the work that needed 
> to be done in order to avoid this predicted outcome.
>
> e
>
> --
> You received this message because you are subscribed to the Google Groups 
> "Bitcoin Development Mailing List" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected].
> To view this discussion visit 
> https://groups.google.com/d/msgid/bitcoindev/929d44f7-42d8-4670-8cf6-d01d44c36c2en%40googlegroups.com.

-- 
You received this message because you are subscribed to the Google Groups 
"Bitcoin Development Mailing List" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/bitcoindev/qASxp5RkHoZo6U9wmZuUlemOWYZkMfXSIDOSz-Km8opsbZDlWDZ8UP7DFLTNrEZE0XNaHjGHArFm6rT4KpV_viBWb1ybXf7SLdL-AMZmmOg%3D%40protonmail.com.

Reply via email to