Return-Path: Received: from smtp1.linuxfoundation.org (smtp1.linux-foundation.org [172.17.192.35]) by mail.linuxfoundation.org (Postfix) with ESMTPS id C3356B5F for ; Tue, 28 Feb 2017 16:43:33 +0000 (UTC) X-Greylist: whitelisted by SQLgrey-1.7.6 Received: from mail-wr0-f171.google.com (mail-wr0-f171.google.com [209.85.128.171]) by smtp1.linuxfoundation.org (Postfix) with ESMTPS id B6040157 for ; Tue, 28 Feb 2017 16:43:31 +0000 (UTC) Received: by mail-wr0-f171.google.com with SMTP id u48so12768674wrc.0 for ; Tue, 28 Feb 2017 08:43:31 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc; bh=NjhbQkqqH2kYTq7jx3+FHNeN/PFvb5e3P0A99bFULPA=; b=PRvFsImxEL6Z+qj9mx7S5mRWfIrkTqnP7dTz8tFoeRHNJOMS9OYO+grKjF2J36MEpw YRGxtSlTs3lHlaS2D4vlUExYTZ5zVL2tTZaCMKiTY129Fi+QOPAVuQhq1jSAAjz8w2nf AT4Ojh6heA12YGgJLW3CdITjTXbxw33NET8YIvrhX46nNq6kU0Rrexew3yoPm/oTPjhI qaHSyvO2J2za/RLUtRmWrJXHtjpbGVC316sy0M4CMsUslwbTIeDWWBjXtH4jWMxsL3eu v2tJZY1oQeAFflMgjezfDaDMEvCq9asbMVs+Nx56t3geVd3n3gn9FJQPpRpzsKxtCuhK ZaTQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:cc; bh=NjhbQkqqH2kYTq7jx3+FHNeN/PFvb5e3P0A99bFULPA=; b=MN2qa/3iN+k3/QcRPQPhn7H52a+nZo2MYxi5z1cXC5hjKZsDkbiaTXR7ZYg343XMmM d+4EwskWJVpPhhKv9Q7qx2f2K9JzjvgSS1gQgaQ0apIJr62BkKWH9xuazTg56ONO/UlK zBo0Q7NNYR45tc4XKzh+/tjJUbY6c6F4mntMgQzSAY1iew7u0fjAOzLVnQ0GBGnaedJq zAEQ/ngHp4jS2zpj7CBp+mRrYyhf0/pmzVSIUC2RKtET9NK/WfD85DkF5hKlPSoLowP2 VQWClYc3qtlc6tKvE3xpDdmzfGOFHbzJvaNHjrIqJWQfWZbVYkcwtsFbwTkx+YcddwvR 0qQw== X-Gm-Message-State: AMke39k2LJM3ajkSSQtNEgqZOQNUr6Pw45Zu53DrTjZ7g3p2zfzqfcEueytNtsFXIwoL+Msv43BZbsgUX/sO+A== X-Received: by 10.223.136.253 with SMTP id g58mr3558481wrg.10.1488300210170; Tue, 28 Feb 2017 08:43:30 -0800 (PST) MIME-Version: 1.0 Received: by 10.80.142.9 with HTTP; Tue, 28 Feb 2017 08:43:29 -0800 (PST) In-Reply-To: References: <20170223235105.GA28497@savin.petertodd.org> <20170224010943.GA29218@savin.petertodd.org> <20170224025811.GA31911@savin.petertodd.org> <20170224031531.GA32118@savin.petertodd.org> <20170224043613.GA32502@savin.petertodd.org> <20170225041202.GA11152@savin.petertodd.org> From: "G. Andrew Stone" Date: Tue, 28 Feb 2017 11:43:29 -0500 Message-ID: To: Bram Cohen , Bitcoin Protocol Discussion Content-Type: multipart/alternative; boundary=001a11461c7258d6f3054999e53d X-Spam-Status: No, score=-1.5 required=5.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, FREEMAIL_FROM, HTML_MESSAGE, RCVD_IN_DNSWL_NONE, RCVD_IN_SORBS_SPAM autolearn=no version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on smtp1.linux-foundation.org Subject: Re: [bitcoin-dev] A Better MMR Definition X-BeenThere: bitcoin-dev@lists.linuxfoundation.org X-Mailman-Version: 2.1.12 Precedence: list List-Id: Bitcoin Protocol Discussion List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 28 Feb 2017 16:43:33 -0000 --001a11461c7258d6f3054999e53d Content-Type: text/plain; charset=UTF-8 I can understand how Bram's transaction double sha256 hashed UTXO set patricia trie allows a client to quickly validate inputs because the inputs of a transaction are specified in the same manner. So to verify that an input is unspent the client simply traverses the patricia trie. It also makes sense that if transaction inputs were specified by a [block height, tx index, output index] triple we'd have a much more size-efficient transaction format. This format would make look up pretty simple in Peter's pruned time-ordered TXO merkle mountain range, although you'd have translate the triple to an index, which means we'd have to at a minimum keep track of the number of TXOs in each block, and then probably do a linear search starting from the location where the block's TXOs begin in the MMR. (The ultimate option I guess is to specify transaction inputs by a single number which is essentially the index of the TXO in a (never actually created) insertion-ordered TXO array...) But since transactions' prevouts are not specified by [block height, tx index, output index] or by TXO index, I don't understand how an insertion ordered TXO tree can result in efficient lookups. Can you help me understand this? On Sat, Feb 25, 2017 at 1:23 AM, Bram Cohen via bitcoin-dev < bitcoin-dev@lists.linuxfoundation.org> wrote: > On Fri, Feb 24, 2017 at 8:12 PM, Peter Todd wrote: > >> >> So to be clear, what you're proposing there is to use the insertion order >> as >> the index - once you go that far you've almost entirely re-invented my >> proposal! >> > > I'm not 'proposing' this, I'm saying it could be done simply but I'm > skeptical of the utility. Probably the most compelling argument for it is > that the insertion indexed values are much smaller so they can be compacted > down a lot resulting in using less memory and more locality and fewer > hashes, but your implementation doesn't take advantage of that. > > >> Your merkle-set implementation is 1500 lines of densely written Python > > > The reference implementation which is included in those 1500 lines is less > than 300 lines and fairly straightforward. The non-reference implementation > always behaves semantically identically to the reference implementation, it > just does so faster and using less memory. > > >> with >> almost no comments, > > > The comments at the top explain both the proof format and the in-memory > data structures very precisely. The whole codebase was reviewed by a > coworker of mine and comments were added explaining the subtleties which > tripped him up. > > >> and less than a 100 lines of (also uncommented) tests. > > > Those tests get 98% code coverage and extensively hit not only the lines > of code but the semantic edge cases as well. The lines which aren't hit are > convenience functions and error conditions of the parsing code for when > it's passed bad data. > > >> By >> comparison, my Python MMR implementation is 300 lines of very readable >> Python >> with lots of comments, a 200 line explanation at the top, and 200 lines of >> (commented) tests. Yet no-one is taking the (still considerable) effort to >> understand and comment on my implementation. :) >> > > Given that maaku's Merkle prefix trees were shelved because of performance > problems despite being written in C and operating in basically the same way > as your code and my reference code, it's clear that non-optimized Python > won't be touching the bitcoin codebase any time soon. > > >> >> Fact is, what you've written is really daunting to review, and given it's >> not >> in the final language anyway, it's unclear what basis to review it on >> anyway. > > > It should reviewed based on semantic correctness and performance. > Performance can only be accurately and convincingly determined by porting > to C and optimizing it, which mostly involves experimenting with different > values for the two passed in magic numbers. > > >> I >> suspect you'd get more feedback if the codebase was better commented, in a >> production language, and you have actual real-world benchmarks and >> performance >> figures. >> > > Porting to C should be straightforward. Several people have already > expressed interest in doing so, and it's written in intentionally C-ish > Python, resulting in some rather odd idioms which is a bit part of why you > think it looks 'dense'. A lot of that weird offset math should be much more > readable in C because it's all structs and x.y notation can be used instead > of adding offsets. > > >> In particular, while at the top of merkle_set.py you have a list of >> advantages, >> and a bunch of TODO's, you don't explain *why* the code has any of these >> advantages. To figure that out, I'd have to read and understand 1500 >> lines of >> densely written Python. Without a human-readable pitch, not many people >> are >> going to do that, myself included. >> > > It's all about cache coherence. When doing operations it pulls in a bunch > of things which are near each other in memory instead of jumping all over > the place. The improvements it gets should be much greater than the ones > gained from insertion ordering, although the two could be accretive. > > > _______________________________________________ > bitcoin-dev mailing list > bitcoin-dev@lists.linuxfoundation.org > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev > > --001a11461c7258d6f3054999e53d Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
I can understand how Bram's transaction= double sha256 hashed UTXO set patricia trie allows a client to quickly val= idate inputs because the inputs of a transaction are specified in the same = manner.=C2=A0 So to verify that an input is unspent the client simply trave= rses the patricia trie.

It also makes sense that if trans= action inputs were specified by a [block height, tx index, output index] tr= iple we'd have a much more size-efficient transaction format.=C2=A0 Thi= s format would make look up pretty simple in Peter's pruned time-ordere= d TXO merkle mountain range, although you'd have translate the triple t= o an index, which means we'd have to at a minimum keep track of the num= ber of TXOs in each block, and then probably do a linear search starting fr= om the location where the block's TXOs begin in the MMR.=C2=A0 (The ult= imate option I guess is to specify transaction inputs by a single number wh= ich is essentially the index of the TXO in a (never actually created) inser= tion-ordered TXO array...)

But since transactions&#= 39; prevouts are not specified by [block height, tx index, output index] or= by TXO index, I don't understand how an insertion ordered TXO tree can= result in efficient lookups.=C2=A0 Can you help me understand this?

On Sat, Feb 25, 2017 at 1:23 AM, Bram Cohen via= bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org> wrote:
On Fri, = Feb 24, 2017 at 8:12 PM, Peter Todd <pete@petertodd.org> wr= ote:

So to be clear, what you're proposing there is to use the insert= ion order as
the index - once you go that far you've almost entirely re-invented my<= br> proposal!

I'm not 'propo= sing' this, I'm saying it could be done simply but I'm skeptica= l of the utility. Probably the most compelling argument for it is that the = insertion indexed values are much smaller so they can be compacted down a l= ot resulting in using less memory and more locality and fewer hashes, but y= our implementation doesn't take advantage of that.
=C2=A0
Your merkle-set implement= ation is 1500 lines of densely written Python

The reference implementation which is included in those 1500 line= s is less than 300 lines and fairly straightforward. The non-reference impl= ementation always behaves semantically identically to the reference impleme= ntation, it just does so faster and using less memory.
=C2=A0
with
almost no comments,

The comments at the top= explain both the proof format and the in-memory data structures very preci= sely. The whole codebase was reviewed by a coworker of mine and comments we= re added explaining the subtleties which tripped him up.
=C2=A0
and less than a 100 = lines of (also uncommented) tests.

T= hose tests get 98% code coverage and extensively hit not only the lines of = code but the semantic edge cases as well. The lines which aren't hit ar= e convenience functions and error conditions of the parsing code for when i= t's passed bad data.
=C2=A0
By
comparison, my Python MMR implementation is 300 lines of very readable Pyth= on
with lots of comments, a 200 line explanation at the top, and 200 lines of<= br> (commented) tests. Yet no-one is taking the (still considerable) effort to<= br> understand and comment on my implementation. :)

Given that maaku's Merkle prefix trees were shelved bec= ause of performance problems despite being written in C and operating in ba= sically the same way as your code and my reference code, it's clear tha= t non-optimized Python won't be touching the bitcoin codebase any time = soon.=C2=A0
=C2=A0

Fact is, what you've written is really daunting to review, and given it= 's not
in the final language anyway, it's unclear what basis to review it on a= nyway.

It should reviewed based on s= emantic correctness and performance. Performance can only be accurately and= convincingly determined by porting to C and optimizing it, which mostly in= volves experimenting with different values for the two passed in magic numb= ers.
=C2=A0
= I
suspect you'd get more feedback if the codebase was better commented, i= n a
production language, and you have actual real-world benchmarks and performa= nce
figures.

Porting to C should be = straightforward. Several people have already expressed interest in doing so= , and it's written in intentionally C-ish Python, resulting in some rat= her odd idioms which is a bit part of why you think it looks 'dense'= ;. A lot of that weird offset math should be much more readable in C becaus= e it's all structs and x.y notation can be used instead of adding offse= ts.
=C2=A0
I= n particular, while at the top of merkle_set.py you have a list of advantag= es,
and a bunch of TODO's, you don't explain *why* the code has any of = these
advantages. To figure that out, I'd have to read and understand 1500 li= nes of
densely written Python. Without a human-readable pitch, not many people are=
going to do that, myself included.

It's all about cache coherence. When doing operations it pulls in a = bunch of things which are near each other in memory instead of jumping all = over the place. The improvements it gets should be much greater than the on= es gained from insertion ordering, although the two could be accretive.


_______________________________________________
bitcoin-dev mailing list
bitcoin-dev@lists.= linuxfoundation.org
https://lists.linuxfoundation.org= /mailman/listinfo/bitcoin-dev


--001a11461c7258d6f3054999e53d--