[ANN] [XMG] MAGI | CPU mining | mPoW | mPoS | [MagiPay]

[ex33s]

Sounds great! Can't wait to test it out. Already compiled the test wallet

For whatever reason, I just can't compile it on Ubuntu 16... downgraded to Openssl1.0.2g but always fails on bignum errors...
Will await build versions and just extract.

Thanks Joe for all work.

I Get that too but on Manjaro dist. If you really want to compile but get bugnum errors like these.

Code:

error: cannot convert ‘const CBigNum*’ to ‘const BIGNUM*

This is how I fixed it, I compiled a standalone OpenSSL that works besides the systems Openssl libs

This is only for when you compile m-wallet and on a linux system
If you do not feel comfortable doing these steps, don't do it.
I make no guarantee that this will work and / or not break your system, use caution.

This is for Linux only

Goto https://www.openssl.org/source/ and download openssl-1.0.2l.tar.gz
Extract the files from the tar.gz

Configure OpenSSL with the following
I'm using the path /opt/openssl-1.0.2l as the example here (Also the same on how I installed it on my system)

Code:

./config --prefix=/opt/openssl-1.0.2l --openssldir=/opt/openssl-1.0.2l -static

--prefix and --openssldir needs to be the same location, you can change location as you wish but I recommend that you place it where you can find  easy it and where it will not overwrite systems OpenSSL (Caution! Overwrite system OpenSSL might break the system)

After the config is complete you need to compile and test it with the folloing commands

Code:

make
make test

When that's completed we need to install it to the directory that you configured with --prefix

Code:

make install

After that is completed you can verify it with this command

Code:

/opt/openssl-1.0.2l/bin/openssl version

And you should see something like this

Code:

OpenSSL 1.0.2l  25 May 2017

Now that you have a nice standalone version of Openssl besides you now you need to tell qmake to user this version of OpenSSL instead of system default

Then it's time to compile the wallet, I'm using Qt5 and qmake-qt5 in this example

Code:

qmake-qt5 m-wallet.pro OPENSSL_LIB_PATH=/opt/openssl-1.0.2l/lib OPENSSL_INCLUDE_PATH=/opt/openssl-1.0.2l/include
make

And after that you should have a nice compiled version of m-wallet with openssl-1.0.2l instead of the system default

[edward0181]

Great info. Many thanks

[malafaya]

The source code for test is updated here:

https://github.com/magi-project/magi/tree/v1.4.3-test

I have to rebuild the chain which has been taking one day long; as soon as I get the chain, I'll post a link for download and then we can start the test.

At this time, this test takes care of 1) and 2). Depending on the time available, we'll see when 3) is to be merged.

This is just for testing, right? The main chain still operates with the "old" rules (plain v1.4.3)...

[ex33s]

The source code for test is updated here:

https://github.com/magi-project/magi/tree/v1.4.3-test

I have to rebuild the chain which has been taking one day long; as soon as I get the chain, I'll post a link for download and then we can start the test.

At this time, this test takes care of 1) and 2). Depending on the time available, we'll see when 3) is to be merged.

This is just for testing, right? The main chain still operates with the "old" rules (plain v1.4.3)...

Correct, that is only for testing.

[Pool42]

Thanks for the test update !

[Windozxpert]

Bitcoin is growing because hashrate and difficulty are growing. This way magi hashrate and diff could grow too.

I don't agree.
This results in only people with big, bigger, biggest hashrate could mine XMG coin.
That's absolutely not the way.

Indeed, the limit would punish everyone just because a few people would be mining with their botnets. A few botnets or farms start mining, the limit is reached, PoW is disabled, the smaller miners are prevented from mining. This idea gives control to the botnets.

[HR]

AFAIK, it is not possible for the network, to distinguish a pool from a solo miner.

Especially if we're talking about a botnet! 

[HR]

How exactly is pow exploited ? I'm new to this coin.

XMG POW is unique in the idea that high network hashrate generate less POW blockreward

the origin of that idea was to remove a arms race in mining gear and allow each normal dude with a single CPU in his PC take part in mining

but clever people found out that there is a timewindow in the diff ajusting mining reward adjusting waves
where diff is high and rewards are high too

thats because the adjusting algorhytm of mining rewards reduction have a delay

so what happens is big miners with botnets or access to company unused CPU capacity or even private GPU miners
mine fast a few blocks with high rewards until rewards adjusting kicks in start to decrease the rewards
then stop mining and the normal 24/7 miners have to mine the high diff low reward blocks

this leads to a big share of POW coin rollout goes into hands of few who mine in waves
and the many 24/7 miners have to mine the low reward high diff blocks alone

i trust in joes abilities to fix POW and come closer to the original vision

Is this still where we are?

Why not a diff adjustment like DigiShield or DGW (Dark Gravity Wave), everything slows down as hashrate rises, a vice versa, and let it be?

[edward0181]

Sounds great! Can't wait to test it out. Already compiled the test wallet

For whatever reason, I just can't compile it on Ubuntu 16... downgraded to Openssl1.0.2g but always fails on bignum errors...
Will await build versions and just extract.

Thanks Joe for all work.

I Get that too but on Manjaro dist. If you really want to compile but get bugnum errors like these.

Code:

error: cannot convert ‘const CBigNum*’ to ‘const BIGNUM*

This is how I fixed it, I compiled a standalone OpenSSL that works besides the systems Openssl libs

This is only for when you compile m-wallet and on a linux system
If you do not feel comfortable doing these steps, don't do it.
I make no guarantee that this will work and / or not break your system, use caution.

This is for Linux only

Goto https://www.openssl.org/source/ and download openssl-1.0.2l.tar.gz
Extract the files from the tar.gz

Configure OpenSSL with the following
I'm using the path /opt/openssl-1.0.2l as the example here (Also the same on how I installed it on my system)

Code:

./config --prefix=/opt/openssl-1.0.2l --openssldir=/opt/openssl-1.0.2l -static

--prefix and --openssldir needs to be the same location, you can change location as you wish but I recommend that you place it where you can find  easy it and where it will not overwrite systems OpenSSL (Caution! Overwrite system OpenSSL might break the system)

After the config is complete you need to compile and test it with the folloing commands

Code:

make
make test

When that's completed we need to install it to the directory that you configured with --prefix

Code:

make install

After that is completed you can verify it with this command

Code:

/opt/openssl-1.0.2l/bin/openssl version

And you should see something like this

Code:

OpenSSL 1.0.2l  25 May 2017

Now that you have a nice standalone version of Openssl besides you now you need to tell qmake to user this version of OpenSSL instead of system default

Then it's time to compile the wallet, I'm using Qt5 and qmake-qt5 in this example

Code:

qmake-qt5 m-wallet.pro OPENSSL_LIB_PATH=/opt/openssl-1.0.2l/lib OPENSSL_INCLUDE_PATH=/opt/openssl-1.0.2l/include
make

And after that you should have a nice compiled version of m-wallet with openssl-1.0.2l instead of the system default

Worked like a charm, many thanks.

[ex33s]

Sounds great! Can't wait to test it out. Already compiled the test wallet

For whatever reason, I just can't compile it on Ubuntu 16... downgraded to Openssl1.0.2g but always fails on bignum errors...
Will await build versions and just extract.

Thanks Joe for all work.

- snip -

Worked like a charm, many thanks.

No worries, glad to help out

[starmman]

Good news about the test build - I'll update cryptopia, they'll wan't know an ETA of live - is there a rough estimate?

[edward0181]

Already 18 minutes no pow blocks found...

[ex33s]

Already 18 minutes no pow blocks found...

PoW blocks are coming in, but that wallet I helped you to compile is only for testing and not stable?

[edward0181]

Correct,
isn't working yet

[rodrigodims]

xmgpool i lost 24 coins from your pool...what i need do?

PM your username, wallet address, let me check.

i lost more tan 700, and i cant see my historic transactions, you clean the Database ?
how can i get my coins back ?

[joelao95]

The source code for test is updated here:

https://github.com/magi-project/magi/tree/v1.4.3-test

I have to rebuild the chain which has been taking one day long; as soon as I get the chain, I'll post a link for download and then we can start the test.

At this time, this test takes care of 1) and 2). Depending on the time available, we'll see when 3) is to be merged.

This is just for testing, right? The main chain still operates with the "old" rules (plain v1.4.3)...

Yes, it's on a test, no touch of the main chain. Once we confirm no forks, we'll migrate to main chain.

[joelao95]

Is this still where we are?

Why not a diff adjustment like DigiShield or DGW (Dark Gravity Wave), everything slows down as hashrate rises, a vice versa, and let it be?

Yes, trying on the diff adjustment that would be ideal as solution in part to the issue.

[Myth888]

Dont get me wrong i am supporting any solution to make this coin as decentralized as possible specially for mining POS and POW. But let me share a few things on my mind.

1. 0 POW blocks and low rewards are almost the same effect.
2. Instant adjustment- immediate next block penalties weather low reward or POW cancel will give big hash miners another variable to contend with. They can flash mine then leave the next block but it will be more troublesome for them and it would be painful if by some chance someone else got the block before them. This would actually add a deterrent over the current implementation of slow adjustment spread over several blocks.

Then i would like to present an alternate solution that can be used.

1. let magi reward fluctuate and be random to some degree but following an average depending on network hash for the last 12 hrs or so. The randomness of the reward but with a target average would benefit miners who mine long term and not botnets who only jump when reward is high and leave as soon as it is low. Because rewards go up and down they are left with one choice. Continuously farm or not to farm at all.

[The Frisian]

1. let magi reward fluctuate and be random to some degree but following an average depending on network hash for the last 12 hrs or so. The randomness of the reward but with a target average would benefit miners who mine long term and not botnets who only jump when reward is high and leave as soon as it is low. Because rewards go up and down they are left with one choice. Continuously farm or not to farm at all.

This is really a jolly good idea. Maybe just thinking a bit (or byte) more, but this can be the start of a definite solution, good thinking.

[Myth888]

1. let magi reward fluctuate and be random to some degree but following an average depending on network hash for the last 12 hrs or so. The randomness of the reward but with a target average would benefit miners who mine long term and not botnets who only jump when reward is high and leave as soon as it is low. Because rewards go up and down they are left with one choice. Continuously farm or not to farm at all.

This is really a jolly good idea. Maybe just thinking a bit (or byte) more, but this can be the start of a definite solution, great.

ive been busy lately so havent been on the forum. But been thinking of a way to contribute an idea that could possibly fix this POW hash rate. But this idea of mine is very complicated to actually code. You will need to code a randomizer which will also have an total average. Im not saying Joe cant do it. But I imagine it to be a daunting task. Computers really dont like random numbers.

and like i said in the past rather than complain its best that we try to be constructive and think of ideas to solve something instead of just restating a problem that everyone already knows about.