Pages:

Author

Topic: please delete (Read 2135 times)

ABCbits

legendary

Activity: 2870

Merit: 7490

Crypto Swap Exchange

Quote from: larry_vw_1955 on September 25, 2021, 07:21:23 AM

Quote from: ?? on ??

FYI, i never said Bitcoin would collapse if you do that.

No but you hinted that if someone did this then it might cause changes to have to be made to bitcoin. Which presumably would be those proposed by the OP which are indeed somewhat radical in nature. Some guy uploads a 2.4 GB video to bitcoin and everyone else has to start signaling to the network that they still want their bitcoins or else they get returned to the miners. Sounds about right...

Basically what i wanted to said is Bitcoin community might reconsider about UTXO indexing (on full node implementation, not Bitcoin protocol), such as,
1. Should client use list which contain of transaction with main/sole purpose of storing arbitrary data?
2. Should all UTXO indexed?
3. If non-indexed UTXO appear on mempool, should the client perform manual lookup (really slow) or partially skip process of transaction verification?

Besides, there's no correlation someone uploading arbitrary data and ownership of someone's Bitcoin.

Quote from: larry_vw_1955 on September 25, 2021, 07:21:23 AM

But he does have more degrees than Vitalik. And he did invent bitcoin so who you gonna trust?

Quote

https://bitcoinmagazine.com/business/op-ed-how-many-wrongs-make-wright

n0nce

hero member

Activity: 882

Merit: 5834

not your keys, not your coins!

Quote from: larry_vw_1955 on September 26, 2021, 11:56:52 PM

Quote from: n0nce on September 26, 2021, 10:28:53 AM

Not sure where you are getting your figures but something like the Core i9-11900K has a base frequency of only 3.50 GHz.

Hey Sherlock, you know what? It goes even lower when idle. Usually we look at the boost clocks, because that's what a processor runs at when you give it work to perform.

Also, as we can see from 3950X to 5950X the all-core load (blender render) improved a lot and only pulled 120W in the latest version; as I said, modern CPUs are targeted at anywhere between 95 and 120W.

source: https://www.youtube.com/watch?v=72AHENDeTEI

I'm aware Intel (which is not the leader of CPU performance right now) can be overclocked to pull much more than that, but we should compare apples to apples, in my opinion.

The 5950X boosts up to over 5GHz without overclock (so within the given TDP), for an example. You're wrong in thinking performance doesn't keep increasing.

For more info:
https://www.technipages.com/base-clock-and-boost-clock-explained
https://en.wikipedia.org/wiki/Dynamic_frequency_scaling

Quote from: larry_vw_1955 on September 26, 2021, 11:56:52 PM

Well yeah I mean if you go all the way back to 1975 but I'd say in the last 10 years, the cost of computing has not gone down at all. Not a single penny.

https://eu.usatoday.com/story/money/2019/06/18/cost-of-a-computer-the-year-you-were-born/39557187/

A PowerBook G4 cost $5,000 in 2001, while today a MacBook M1 costs $1,000, so it seems to me prices did go down. Also consider that 5K is adjusted for inflation - in 2019. It is probably even more in 2001 dollar valuation.

Another source that suggests performance does continue to increase:
https://ourworldindata.org/technological-progress

This is now a HUGE tangent from the thread and I'd like to end it for the sake of the original topic.

Dabs

legendary

Activity: 3416

Merit: 1912

The Concierge of Crypto

CSW is a little off topic here. He is a fraud. Can't sign proof. And someone else signed proof that he is a fraud using addresses he claims he owned. But that's off topic and should be discussed in the scam threads. There's a few around there.

larry_vw_1955

sr. member

Activity: 1190

Merit: 469

Quote from: n0nce on September 26, 2021, 10:28:53 AM

Not sure where you are getting your figures but something like the Core i9-11900K has a base frequency of only 3.50 GHz.

Quote

Power consumption / efficiency is mostly limited by cooling; PC cooling designs haven't changed much in the last, idk, 10-20 years, so since then most of the consumer CPUs have been and still are targeted to run at 95-125W max.

If you're talking about something like a Core i9-11900K then that thing goes up to nearly 300 watts peak.

Quote

It's also obvious that cost of computing has been ever decreasing; while it was absolutely not normal for everyone to afford a computer in 1975, I now know barely anyone who has just one computer (I count modern mobile phones and tablets as computers as well) anymore.

Well yeah I mean if you go all the way back to 1975 but I'd say in the last 10 years, the cost of computing has not gone down at all. Not a single penny.

Quote

You really like to beat a dead horse, don't you? Just fyi, this gets tiring really quick. If you keep reiterating this nonsense in every second post, I (and most others) can't continue reading every reply, it just gets too repetitive and obnoxious.

Sorry about that. But to me, the doctor belongs getting credit for his invention. I used to think he was a fraud too but that's because I just listened to what other people said about him. I never took the time to actually listen to what he had to say and he's got alot of videos out there on youtube. So people can decide for themself. I already decided for myself who satoshi is.

Quote

Saying that chips are not improving over time is just untrue.

Well I agree they are still improving over time but not at the rate they once did. Like if you compare 2005 to 2020 progression to 2020 to 2035 it's not going to be as dramatic. And that will likely continue with an exponentially decreasing improvement factor. Unless there is some huge breakthrough like photonic computing.

Quote

I believe that we've made our point with the many pages of discussions. I would encourage you to at least consider the posts and DYOR. It is not a constructive discussion if you were to be so insistent on your views and refusing to acknowledge the facts.

I've said pretty much everything I needed to say. And I appreciate all the input from you and others. What you said about large block sizes is probably correct.

ranochigo

legendary

Activity: 3038

Merit: 4418

Crypto Swap Exchange

Quote from: larry_vw_1955 on September 26, 2021, 06:09:15 AM

yeah these days improvements come from improving cpu architecture and adding on more cores but not by pushing the laws of physics about how many transistors you can fit into a given space or how fast you can turn them on and off. that's over and done. the thing is though we have to be careful about using Instructions per cycle as the only metric. power consumption is important too. and cost of the cpu. all those things matter. to really say that things are improiving we need higher performance (more IPC), lower power consumption and also lower cost (on an inflation adjusted basis of course). you got all 3?

TDPs of CPUs are mostly within quite a fixed range. 7nm Zen 3 are more efficient than 7nm Zen 2 and 12nm Zen +. As we're getting to smaller node size, architectures are more efficient (performance to power ratio). In fact, all of the criteria that you've listed all hold true. CPUs are cheaper relative to their performance, better performance and basically the same power consumption. Saying that chips are not improving over time is just untrue.

Quote from: larry_vw_1955 on September 26, 2021, 06:09:15 AM

craig wright is satoshi so he's in a special category Grin

I certainly hope that you're not trolling but it feels like I'm talking to a wall at this point in time.

Quote from: larry_vw_1955 on September 26, 2021, 06:09:15 AM

Big blocks encourage very low transaction fees. which is a good thing for users. I mean I think Craig himself said that bitcoin is not for hodling it's for using it like money, spending it. That dude knows his crypto.

Big blocks encourage centralization, by making is far less practical for users to run nodes and relying on SPV or centralized servers instead. Excessively big blocks makes the network less secure, where users are seeing their own versions of the chain. With on-chain scaling, you are still limited to many factors that governs it, be it transaction finality due to the stale rates or just the block intervals. By introducing excessively big blocks, you are scaling Bitcoin by solving something and sacrificing the practicality of it.

I believe that we've made our point with the many pages of discussions. I would encourage you to at least consider the posts and DYOR. It is not a constructive discussion if you were to be so insistent on your views and refusing to acknowledge the facts.

n0nce

hero member

Activity: 882

Merit: 5834

not your keys, not your coins!

Quote from: larry_vw_1955 on September 26, 2021, 06:09:15 AM

we have to be careful about using Instructions per cycle as the only metric. power consumption is important too. and cost of the cpu. all those things matter. to really say that things are improiving we need higher performance (more IPC), lower power consumption and also lower cost (on an inflation adjusted basis of course). you got all 3?

I have to chime in on this: Nobody just cares about IPC. It's just a good way to keep constantly improving performance right now. Also pushing frequency isn't dead yet either; with now mainstream CPUs running at 5GHz easily (Intel 11th gen + Ryzen 5000), while this was considered a super heavy overclock mere years ago.

Power consumption / efficiency is mostly limited by cooling; PC cooling designs haven't changed much in the last, idk, 10-20 years, so since then most of the consumer CPUs have been and still are targeted to run at 95-125W max.

It's also obvious that cost of computing has been ever decreasing; while it was absolutely not normal for everyone to afford a computer in 1975, I now know barely anyone who has just one computer (I count modern mobile phones and tablets as computers as well) anymore.

Quote from: larry_vw_1955 on September 26, 2021, 06:09:15 AM

Quote

Inventing is not something that I would agree with. Bitcoin SV and Bitcoin are practically the same, save for a few parameter changes.

I doubt that. They seem very different from one another. Now I do have to agree that anyone that forks bitcoin didn't really invent bitcoin but of course craig wright is satoshi so he's in a special category Grin

Syke

legendary

Activity: 3878

Merit: 1193

Quote from: larry_vw_1955 on September 26, 2021, 06:09:15 AM

/sigh. No, it hasn't stopped.

https://www.anandtech.com/show/16823/intel-accelerated-offensive-process-roadmap-updates-to-10nm-7nm-4nm-3nm-20a-18a-packaging-foundry-emib-foveros

larry_vw_1955

sr. member

Activity: 1190

Merit: 469

Quote from: ranochigo on September 25, 2021, 11:57:42 PM

No, you're comparing apples to oranges.
...
We are not pursuing higher clock speeds, only better IPCs. Your argument is that the we expected CPUs to be running at higher clock speeds over time, which is untrue. In this thread, we are talking about the increment in the performance of the chips, which is proven true for the past few years.

Quote

Inventing is not something that I would agree with. Bitcoin SV and Bitcoin are practically the same, save for a few parameter changes.

I doubt that. They seem very different from one another. Now I do have to agree that anyone that forks bitcoin didn't really invent bitcoin but of course craig wright is satoshi so he's in a special category Grin

Quote

Again, Apples to Orange comparison. Bitcoin SV has 157 nodes[1], while Bitcoin has 8200 nodes[2]. It doesn't seem to be a problem because those who are using it are obviously supporters of it. You don't have to consider the actual feasibility, which lies in quite a few aspects (block propagation time, block processing time, storage limitation, etc) if you think that is the way a crypto is supposed to function. Bigger blocks are possible, excessively big blocks are possible as well, but there are quite a few drawbacks in the latter.

Shymaa-Arafat

full member

Activity: 228

Merit: 156

Quote

Basically what i wanted to said is Bitcoin community might reconsider about UTXO indexing (on full node implementation, not Bitcoin protocol), such as,
1. Should client use list which contain of transaction with main/sole purpose of storing arbitrary data?
2. Should all UTXO indexed?
3. If non-indexed UTXO appear on mempool, should the client perform manual lookup (really slow) or partially skip process of transaction verification?

I'm suggesting to partition the UTXO set, so that u insert/delete/verify/search for proofs dealing with each pile separately; clearly all piles will be smaller than the total set

ranochigo

legendary

Activity: 3038

Merit: 4418

Crypto Swap Exchange

Quote from: larry_vw_1955 on September 25, 2021, 11:42:43 PM

Wrong. I care about it. It really does have an effect on computer performance. Maybe you meant to say "no one in industry cares about the clock speed" well i might even disagree with you about that too. They care but there's nothing they can do about it.

No, you're comparing apples to oranges. In a similar architecture, yes a higher clock speed results in a better performance, simply by the logic of having similar IPCs. The point here is, if you are comparing a 6GHz Core 2 Quad from those days and a 2GHz i5 from today, the i5 wins, hands down simply by the architectural improvement. We are not pursuing higher clock speeds, only better IPCs. Your argument is that the we expected CPUs to be running at higher clock speeds over time, which is untrue. In this thread, we are talking about the increment in the performance of the chips, which is proven true for the past few years.

It isn't difficult to achieve higher clock speeds, if you design a CPU specifically just for it but it won't be efficient.

Quote from: larry_vw_1955 on September 25, 2021, 11:42:43 PM

Well because bitcoin sv is handling huge volumes of data with gigabyte sized blocks and we were talking about blockchain growth and size issues here with bitcoin. Doesn't seem to be a problem for sv.

I also find it fascinating that Craig Wright invented bitcoin sv. Cheesy

Inventing is not something that I would agree with. Bitcoin SV and Bitcoin are practically the same, save for a few parameter changes.

Again, Apples to Orange comparison. Bitcoin SV has 157 nodes[1], while Bitcoin has 8200 nodes[2]. It doesn't seem to be a problem because those who are using it are obviously supporters of it. You don't have to consider the actual feasibility, which lies in quite a few aspects (block propagation time, block processing time, storage limitation, etc) if you think that is the way a crypto is supposed to function. Bigger blocks are possible, excessively big blocks are possible as well, but there are quite a few drawbacks in the latter.

[1] https://blockchair.com/bitcoin-sv/nodes
[2] https://blockchair.com/bitcoin/nodes

larry_vw_1955

sr. member

Activity: 1190

Merit: 469

Quote from: ranochigo on September 25, 2021, 08:25:48 AM

If I earn $1000 from my fees and the storage costs $10 per month, I probably wouldn't complain either.

Exactly! I mean who wouldn't take a deal like that?

Quote from: larry_vw_1955 on September 23, 2021, 10:55:51 PM

Back in 2009, they probably thought cpus would be running at 20Ghz in 2021. Enough said.

For what it's worth, that was never the case. No one cares about the clock speed, because that is irrelevant to CPU processing speed.

Quote from: Shymaa-Arafat on September 25, 2021, 08:16:39 AM

Full nodes are the ones who maintain the UTXOS set and to delete a subset there must be a consensus among them all to do it and consider a TX containing one of those invalid, a soft fork I think.

I was just talking about deleting spam out of the utxo set on an individual basis. They might not be able to verify some blocks if that happened but they could certainly mine new ones.

Shymaa-Arafat

full member

Activity: 228

Merit: 156

Ah, u mean non-standard UTXOS, but that was not the main original topic here maybe it got into it & I skippeded some replies in the middle.
-Still, they could have done it in the OP_Return part without keeping them in the UTXOS set, but probably such techniques were studied after 2013.

n0nce

hero member

Activity: 882

Merit: 5834

not your keys, not your coins!

Quote from: Shymaa-Arafat on September 25, 2021, 01:31:30 PM

Quote

you mean you wouldn't mine a transaction like https://mempool.space/tx/54e48e5f5c656b26c3bca14a8c95aa583d07ebe84dde3b7dd4a78f4e4186e713, then I don't understand why. You get your fees, so why would you not mine it? A miner doesn't care much about anything else usually.

Why would someone make such a TX with this cost (and 938 UTXOS most of them are 1Satoshi) unless it was a dust hack?
It's from 2013 so if it was a hack, it must have been documented somewhere

It's not a hack, it's the Bitcoin whitepaper PDF encoded as hex and split up into 938 addresses.
In 2013 it wasn't that expensive. I think in beginning of 2013 the BTC price was around 100 bucks and they paid 0.59BTC in transaction fees, so it cost around $59 to permanently save the whitepaper in the blockchain forever, sounds like a good deal to me Cheesy

Encoding data into addresses like this is one way to save information into the Bitcoin blockchain, which is the topic we collectively derailed into in the last few days.. Cheesy

You can also check this page which extracts characters which were encoded in some Bitcoin blocks, which is another way to encode data into the blockchain (but just text, not binary like pdf or movie): https://bitcoinstrings.com/

By the way, there's an ASCII image in block 3:

Code: (https://bitcoinstrings.com/blk00003.txt)

File: blk00003.txt

---BEGIN TRIBUTE---
#./BitLen
:::::::::::::::::::
:::::::.::.::.:.:::
:.: :.' ' ' ' ' : :
:.:'' ,,xiW,"4x, ''
: ,dWWWXXXXi,4WX,
' dWWWXXX7" `X,
lWWWXX7 __ _ X
:WWWXX7 ,xXX7' "^^X
lWWWX7, _.+,, _.+.,
:WWW7,. `^"-" ,^-'
WW",X: X,
"7^^Xl. _(_x7'
l ( :X: __ _
`. " XX ,xxWWWWX7
)X- "" 4X" .___.
,W X :Xi _,,_
WW X 4XiyXWWXd
"" ,, 4XWWWWXX
, R7X, "^447^
R, "4RXk, _, ,
TWk "4RXXi, X',x
lTWk, "4RRR7' 4 XH
:lWWWk, ^" `4
::TTXWWi,_ Xll :..
=-=-=-=-=-=-=-=-=-=
LEN "rabbi" SASSAMA
1980-2011
Len was our friend.
A brilliant mind,
a kind soul, and
a devious schemer;
husband to Meredith
brother to Calvin,
son to Jim and
Dana Hartshorn,
coauthor and
cofounder and
Shmoo and so much
more. We dedicate
this silly hack to
Len, who would have
found it absolutely
hilarious.
--Dan Kaminsky,
Travis Goodspeed
P.S. My apologies,
BitCoin people. He
also would have
LOL'd at BitCoin's
new dependency upon
ASCII BERNANKE
:'::.:::::.:::.::.:
: :.: ' ' ' ' : :':
:.: _.__ '.:
: _,^" "^x, :
' x7' `4,
XX7 4XX
XX XX
Xl ,xxx, ,xxx,XX
( ' _,+o, | ,o+,"
4 "-^' X "^-'" 7
l, ( )) ,X
:Xx,_ ,xXXXxx,_,XX
4XXiX'-___-`XXXX'
4XXi,_ _iXX7'
, `4XXXXXXXXX^ _,
Xx, ""^^^XX7,xX
W,"4WWx,_ _,XxWWX7'
Xwi, "4WW7""4WW7',W
TXXWw, ^7 Xk 47 ,WH
:TXXXWw,_ "), ,wWT:
::TTXXWWW lXl WWT:
----END TRIBUTE----

Shymaa-Arafat

full member

Activity: 228

Merit: 156

Quote

Why would someone make such a TX with this cost (and 938 UTXOS most of them are 1Satoshi) unless it was a dust hack?
It's from 2013 so if it was a hack, it must have been documented somewhere

n0nce

hero member

Activity: 882

Merit: 5834

not your keys, not your coins!

Quote from: larry_vw_1955 on September 24, 2021, 11:29:34 PM

Quote from: n0nce on September 24, 2021, 09:12:19 PM

Try it! It will be very expensive if it will even work at all! Cheesy

Even short videos are orders of magnitude larger than a few pages of PDF.

I don't know what your definition of "orders of magnitude" is but I doubt it. Obviously someone wanting to store some video on the blockchain needs to use an optimal compression and frame rate for their video to save money. I mean that goes without saying. Craig Wright figured it out though in that you don't have to worry about any of that on BSV, just upload the full HD video and you're good to go. But I think I could store a small video on the bitcoin blockchain using that method linked to I'm sure of it. By the way not to pimp SV too hard but I'm sure, let me capitalize that. SURE that they make it easy to download the video too which for bitcoin it would be a huge pain in the ass to do. Craight Wright being a doctor you understand he thought of everything...

It sounds like you really want me to try uploading a video to bitcoin though.

I don't have my own definition, I use the one that, well, everyone uses, which is:

Quote from: larry_vw_1955 on September 25, 2021, 07:21:23 AM

But if I was a miner and I saw a huge number of utxos that were just dust amounts and I did further analysis and saw that the addresses for those utxos didn't have any other utxos i would delete them. Grin

If you mean you wouldn't mine a transaction like https://mempool.space/tx/54e48e5f5c656b26c3bca14a8c95aa583d07ebe84dde3b7dd4a78f4e4186e713, then I don't understand why. You get your fees, so why would you not mine it? A miner doesn't care much about anything else usually.

Shymaa-Arafat

full member

Activity: 228

Merit: 156

Quote from: ranochigo on September 25, 2021, 10:09:31 AM

I don't agree with the proposal. UTXO size is not a concern for us, not to the extent that would make us resort to draconian measures to intentionally censor and invalidate certain UTXOs. It isn't difficult to store the entire UTXO within your ram, and it certainly isn't impossible to have dynamic caching of it through your ram as well.

Me neither on deleting people money, yet I did suggested to separate those unlikely to be invoked from any search or Merkle structure to shorten the path length. If they ever invoked, no problem their Merkle proofs will be even much shorter.

ranochigo

legendary

Activity: 3038

Merit: 4418

Crypto Swap Exchange

Quote from: Shymaa-Arafat on September 25, 2021, 09:45:49 AM

A transaction containing one of those UTXOS will be invalid for those running the new version, valid for those running the old version... isn't that soft fork?

Hmm, okay I misunderstood what you meant. By not considering those UTXOs, it can be a soft fork because the majority always prevails. Unless there is a divergence between them, for which there would be a messy situation.

I don't agree with the proposal. UTXO size is not a concern for us, not to the extent that would make us resort to draconian measures to intentionally censor and invalidate certain UTXOs. It isn't difficult to store the entire UTXO within your ram, and it certainly isn't impossible to have dynamic caching of it through your ram as well.

garlonicon

copper member

Activity: 821

Merit: 1992

Pawns are the soul of chess

Quote

isn't that soft fork?

Yes, it is. Even all transactions except coinbase could be invalid in a soft-fork: https://petertodd.org/2016/forced-soft-forks.

Shymaa-Arafat

full member

Activity: 228

Merit: 156

A transaction containing one of those UTXOS will be invalid for those running the new version, valid for those running the old version... isn't that soft fork?

ranochigo

legendary

Activity: 3038

Merit: 4418

Crypto Swap Exchange

Quote from: larry_vw_1955 on September 22, 2021, 05:32:59 AM

They can be forced to though. if necessary. there are ways to do that. but i guess that's up to bitcoin sv whether they need to. as long as my homevideos are on there forever, i'll be good to go.

If I earn $1000 from my fees and the storage costs $10 per month, I probably wouldn't complain either.

Quote from: larry_vw_1955 on September 22, 2021, 05:32:59 AM

Well that brings up a good question. Who exactly is paying for it? not that I'm worried about it you understand, just curiious.I mean I had a choice either put my homevideos on youtube but then you never know what might happen to youtube. they might not be around in 100 years.

Do you know what pools are?

Quote from: larry_vw_1955 on September 23, 2021, 10:55:51 PM

Back in 2009, they probably thought cpus would be running at 20Ghz in 2021. Enough said.

For what it's worth, that was never the case. No one cares about the clock speed, because that is irrelevant to CPU processing speed. Moore's law is about the transistor density on chips which has been true for quite a while, and even in the recent generations of CPU, the instructions per clock has advanced far greater than that of the CPU of those days.

Just to address some of the points raised in the thread. As I've said, the point about not invalidating/deleting UTXO has to do with the ethical and practical concerns, where limits have to be arbitrarily defined and that will never happen. Truncating Bitcoin in that sense is NOT impossible, in case you've misunderstood me. It is totally possible, but you're going to be assuming that the pre-defined checkpoint is valid. As of now, it is not an issue because we don't have that sort of problem yet, nor do people want to hassle with it to solve an issue that might never really be one.

I'm not sure why Craig Wright is really in this conversation at all. He was NEVER proven to be Satoshi and the list of addresses he provided turned out to be someone else's[1]. A PhD in Computer Science isn't really that uncommon, and it really does nothing to show their honesty or moral compass. Signature validation are fairly expensive, if you were to actually look it up and compare it with an average computer running a Bitcoin node.

[1] https://web.archive.org/web/20200525113441/https://paste.debian.net/plain/1148565

Quote from: Shymaa-Arafat on September 25, 2021, 08:16:39 AM

Hard fork.