7 March 2018
Z
10:43
Zack
It is possible to write code for syncing when you don't have external ip. but that kind of code is messy, because you keep asking the peer every few seconds "do you have any new blocks"??
All those little messages can add up to a lot of bandwidth if it is not done carefully.
Especially since some of them are mining, and miners want to know about the new block within a fraction of a second of it being available.
K
10:45
Kenny
yeah, i think in the future that probably needs to be implemented in some form, but for now opening ports works. Much bigger fish for ya to fry right now. Keep up the good work!!!!
MH
11:02
Mandel Hoff
In reply to this message
Thank you - updating now
Z
11:03
Zack
In reply to this message
great
MH
11:04
Mandel Hoff
Updated. Restart looks good. I hear my gpu's spinning back up already.
MF
11:11
Mr Flintstone
In reply to this message
Nice!!!
Z
11:12
Zack
historical hashrate. every 50 blocks.
[17,22,20,24,22,17,21,26,16,14,21,13,22,14,22,14,18,15,14,18,8,10,10,18,17,18,24,17,14,17,18,16,17,13,16,22,14,17,18,15,34,30,43,208,609,703,870,914,831,1142,1142,1075,1075,962,1142,1075,1075,1142,1142,962,762,870,1075,914,1075,962,870,962,870,1075,172,1142,795,1306,831,1306,914,962,870,962,2185,3351,4347,5663,2957,4145,3848,4324,4062,4759,5624,8042,9036,12185,7733,4103,7180,7733,5361,10582,6815,12765,1824,7963,5828,3070,3848]
11:15
in 100s of megahashes per second.
11:19
Deleted Account
15 second one block now?
Z
11:20
Zack
212 seconds per block is the average of the last 10 blocks.
P|
11:23
Paul N. | BAND (πŸ…‘) - I'll never DM you
@Mandelhoff Is it possible to get my Veo out of your pool before reaching 1 Veo?
11:23
I have 12 pub keys in your pool
NM$L joined group by link from Group
MH
11:28
Mandel Hoff
In reply to this message
It currently is not possible. It is very likely this will be a feature added eventually (likely estimate - a few weeks).
P|
11:33
Paul N. | BAND (πŸ…‘) - I'll never DM you
Lol. Considering that using 12 pubkeys was a work around so that I could use 20 cards mining simultanously on your pool, I'd say that's a little disappointing reply to hear.
11:33
But, ok. Please make sure that feature delivers ;)
Z
11:34
Zack
too bad his server isn't open source. I think I could write code for sending veo between accounts on his server pretty quickly.
MH
11:36
Mandel Hoff
Yes, that may be the best way to fix the multiple accounts issue. Load keys to wallet - I was thinking it might be hard to verify a user, but key loading is already available via the wallet.
11:37
I was thinking of using ip address, but that's not very secure. Easy to forge too.
11:38
I like the sync fix so far - pool is finding lots of blocks now.
Z
11:38
Zack
the javascript light wallet has code for signing and verifying transactions.
You could make something that is similar to a spend transaction in amoveo. but with a tag saying "mandel's server".
11:38
In reply to this message
great
11:45
Deleted Account
In reply to this message
Anyone?
MH
11:59
Mandel Hoff
The gpu miner works with multiple - but you have to kick off one process per device. Set the device Id argument on the command line: 0, 1, 2...
tool.exe {address} {deviceId}
11:59
No clue on overclocking settings.
12:36
Deleted Account
In reply to this message
Does anyone know why the miner is shutting down my windows miners?
MH
12:40
Mandel Hoff
If you are using this, I'm not sure why. Many people are running multiple miner tools on a single box.
https://github.com/Mandelhoff/AmoveoMinerGpuCuda/releases/tag/1.0.0.3
N
12:51
NM$L
It's too difficult to mine 1 amoveo
12:51
Deleted Account
In reply to this message
Yeah that's what I'm using. Strange
12:51
Is there a way to get the miner to run multiple gpu devices in a single command?
12:52
Or do we need to run a seperate command for each GPU/I'd?
MF
12:53
Mr Flintstone
u could script it
12:53
In reply to this message
Block reward is a governable variable
12:54
but I would keep in mind that 1 represents an arbitrarily placed decimal point
12:54
like in bitcoin
12:55
@zack I don’t think the # of satoshis per veo is something in protocol?
N
12:55
NM$L
12:55
Is it normal?
MF
12:56
Mr Flintstone
what is your benchmarked hash rate
MH
12:56
Mandel Hoff
In reply to this message
Separate command per device.
N
12:56
NM$L
In reply to this message
gtx 1060
MF
12:57
Mr Flintstone
in gh/s
N
12:57
NM$L
822.08 MH/s now
12:57
It's unstable
12:58
Is there a roadmap?
MF
12:59
Mr Flintstone
you can test your hash rate from the command line using Zack ’s miner
12:59
I believe you also only get an unpaid balance on that pool when they find s block
MH
13:00
Mandel Hoff
In reply to this message
13:00
Yes, you've only been credited on one block so far. You have some more blocks that are unconfirmed where you will get credit coming soon.
N
13:00
NM$L
why the hashrate changes so high to so low?
13:01
MH
13:01
Mandel Hoff
It's an estimate based on your share difficulty solutions.
N
13:01
NM$L
240 Mh/s is normal?
MH
13:01
Mandel Hoff
Yes, that's what I get on my 1060gtx
N
13:02
NM$L
I start mining about 1 hour, just get 0.001 unpaid balance
13:03
around 0.024 per day?
MH
13:03
Mandel Hoff
You'll get more - like I showed - you have several blocks reward shares coming , just not yet confirmed.
N
13:04
NM$L
Ok. Thanks
MH
13:04
Mandel Hoff
There's a lead-miner who is throwing insane money at the network. We are all competing against that entity. He's running about $250k in hardware.
N
13:04
NM$L
You are the pool admin.
N
13:20
NM$L
what's the problem?
13:20
I
13:21
Iridescence
I think that's mine
13:21
XD
13:22
debug info is written to debug.txt
13:22
oh
13:22
you forgot an end point
13:22
miner_gpu:start().
13:22
note the period at the end
N
13:30
NM$L
In reply to this message
Thanks, I'm mining by 1080
13:32
I wanna upgrade my speed
I
13:32
Iridescence
PM me for upgrades
BB
13:35
Brian Brian
Iri you gonna open source your upgrade ever?
I
13:36
Iridescence
If I have an incentive to
13:36
Deleted Account
is your upgrade the same thing that was posted on reddit and added to mandel's pool?
I
13:36
Iridescence
Nope.
13:38
Deleted Account
haha, the post was deleted
I
13:38
Iridescence
...
13:38
regardless of upgrade, I will claim that the numbers on my github are not made up and are from real recorded results
13:38
Deleted Account
(Was just that you dont have to sha the entire 64 bytes each iteration)
I
13:39
Iridescence
so it is easy to compare if Mandel's miner has reached parity with my upgraded version
MH
13:39
Mandel Hoff
Yes, that's the change. "Save the partial context".
Z
13:41
Zack
In reply to this message
it would make the coins you already have more valuable if you open sourced it.
N
13:43
NM$L
pls open source it
I
13:44
Iridescence
In reply to this message
how so?
Z
13:44
Zack
How is it in your incentive to give Kassel even more control?
13:44
If the difficulty goes up, so does the value of the existing coins.
If you share your software, the difficulty will go up.
I
13:45
Iridescence
The difficulty will go up due to Kassel anyway
Z
13:45
Zack
The difference between 90% control and 95% control is double the effort for Kassel.
13:46
If only one person has good software, the difficulty wont go as high. it is like monopoly pricing
N
13:47
NM$L
what's the problem?
13:47
13:51
what's the best software now for mining on ubuntu?
14:01
Z
14:01
Zack
what?
N
14:01
NM$L
I don't wanna pay for upgrade
Z
14:01
Zack
what are you going to upgrade?
14:01
Deleted Account
Jamhan, cant complain much when you're getting it for free anyway
14:02
the options are currently, Mandels CPU & GPU miners for windows
14:02
and Iridescence GPU linux miner
14:02
you could always try to compile Mandel's on linux, and then contribute that to the community
N
14:04
NM$L
I am not professonal of coding.
14:06
Deleted Account
Never a better time to learn than now
N
14:09
NM$L
14:10
how to fix it?
14:10
Z
14:10
Zack
I have never installed any of these gpu miners. Don't ask me.
I
14:12
Iridescence
yeah i think it is your C miner
N
14:12
NM$L
yes, your C miner
Z
14:13
Zack
It looks like your pubkey isn't formatted right
14:13
or maybe it is because the server is down.
14:14
Deleted Account
is mandels pool down? seems 404 error
K
14:15
Kenny
im mining to it
14:15
and website seems up
Z
14:16
Zack
now my pool is up again too
MH
14:18
Mandel Hoff
Node crashed a few times. Should be up now. Thank you for the alert.
K
14:18
Kenny
In reply to this message
imo it is pointless to use the c miner. if you have a gpu you can use one of the gpu miners. i dont have a link, but iridescence has one that works for linux. madel has one for windows that you can only use at http://amoveopool.com/miner that can be found here https://github.com/Mandelhoff/AmoveoMinerGpuCuda/releases
N
14:19
NM$L
I give up, waiting for a easy gpu miner on linux.
K
14:19
Kenny
@Iridescence do you have a link for your miner?
14:20
Deleted Account
mined for 3days only get 0.7veo
14:20
with my 1050ti
N
14:20
NM$L
you should upgrade the software
N
14:21
NM$L
In reply to this message
Thanks. my win10 with 1060 is using your software.
14:22
In reply to this message
my 1060, mining just 2hours with 0.12402281 unpaid balance.
14:23
looks much better than you
14:26
Deleted Account
In reply to this message
what is your hashrate
N
14:26
NM$L
240Mh
14:27
Deleted Account
i only get 180mh
CB
14:27
CeloStarter Bandits
if i use aws.. it stops working.. any hint
14:27
?
14:27
all dlls are in place
14:27
Deleted Account
but how could that being so large difference
CB
14:27
CeloStarter Bandits
its just stops working
N
14:28
NM$L
what price now?
15:25
Deleted Account
too hard for mining
N
15:35
NM$L
@Mandelhoff When can I get paid balance
I
15:50
Iridescence
ATTENTION everybody who is using my cuda miner please git pull to the latest version immediately
15:50
there was a severe perf bug that was fixed
G
17:38
Gonzalo
In reply to this message
In comparison with Bitcoin on its first days (with 50BTC reward per block) 0,7 VEO = 35 BTC, so not too bad. Just HODL!
Z
18:10
Zack
Looks like $36 per Veo. there are about 8000 tokens. so the market cap is around $288k
18:13
It seems to me that if I leave Amoveo open in a terminal via SSH, it doesn't crash. but if I disconnect SSH, it crashes within 20 minutes.

I am guessing that the print statements are acting like short delays, and these delays are preventing some race condition.

Maybe I am noticing a pattern where there is none.
CryptoMach joined group by link from Group
C
18:31
CryptoMach
Hi guys - anyone know what port number to put for the pool?
N
18:39
NM$L
when my balance will get paid?
18:40
Z
18:40
Zack
In reply to this message
Probably 8085
That's what my pool uses
C
18:41
CryptoMach
Thanks Zack :)
Z
19:07
Zack
I think we should do a hard fork on the mining algorithm.

We will flip the header around, so the first 32 bytes are nonce, and the last 32 are the hash of the header.
19:08
I will start programming this, but I will wait for community feedback before making a final decision.
G
19:15
Gonzalo
Zack why this flip?
Z
19:16
Zack
This is how Iri and Kassel have such a big advantage, and they aren't sharing their advantage.
19:19
I just realized that hash2integer is about 8 times more comparisons than necessary. This could probably give us a few percentage improvement.
G
19:20
Gonzalo
πŸ‘
19:24
Deleted Account
In reply to this message
just
s
19:43
scott_l | DeFi Pulse
In reply to this message
can we stop attacking community members without evidence?

what kassel claimed to u in chat is not accurate. iri's improved code also has not produced blocks.
19:44
despite performing to spec
Z
19:44
Zack
Both of them admit to having better code, and they admit to not sharing. This isn't an attack, I am stating facts.
s
19:46
scott_l | DeFi Pulse
but you said they were both monopolizing mining
19:46
and we dont know if kassel actually has better code
19:47
very possible is using fpga's or stolen hardware where cost is not relevant... but definitely seems to ignoring blocks found by others
Z
19:50
Zack
In reply to this message
I am the one running the mining pool. So if someone is ignoring blocks, it is me.
I pushed an update about 7 hours ago that significantly improved this problem, and you can see that Mandel's pool is finding many more blocks since this upgrade. but block propagation is hard.
k
19:51
kdmlll
Although I would be happy to see their advantage evaporate, their advantage will not last. They do not have a monopoly on the knowledge to optimise the algorithm. Your effort would be beneficial only in the immediate term.
Z
19:51
Zack
yes. it would help for the next 2 weeks.
k
19:52
kdmlll
You work yourself very hard
19:53
I already made this solution in Erlang
19:54
Deleted Account
I welcome the change as we should choose a memory intensive mining algo as to avoid dual mining that could wreck Amoveo...
Z
19:55
Zack
In reply to this message
I worry that a memory intensive mining algorithm would be hard to make ASICS for.
19:57
Deleted Account
I dont know much about Asics but we can do some research, anyway we could hard fork it in the future once Amoveo is settle for ASICS.
Z
19:57
Zack
In reply to this message
if there is no economic incentive to build an ASIC, then no one will build it.
s
20:01
scott_l | DeFi Pulse
In reply to this message
could u create a restarted amoveo fork with a memory hard algo, and let the mkt decide which is more valuable?
P|
20:02
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
Why would you want ASIC again?
Z
20:06
Zack
In reply to this message
The transition from one technology is dangerous.
If we resist the transition, it is like pushing a problem further into the future.

I think it is best to tackle these problems while the network is small, and we have less to lose.
P|
20:12
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
If that's your goal, then why don't make Amoveo's hashing exactly the same as Bitcoin? Nobody will have an edge because everyone's got access to Antminers if they have money.
Z
20:12
Zack
In reply to this message
Because then the bitcoin miners could shut us down on a whim.
20:14
In reply to this message
Maybe eventually.
For now my hands are full just trying to keep Amoveo alive.
P|
20:14
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
But then the first person who makes ASIC for Amoveo could arguably do the same, isn't it?
20:15
and like you said, we don't really want that transition
s
20:15
scott_l | DeFi Pulse
In reply to this message
yea that makes sense
N
20:24
NM$L
20:24
am i limited?
20:26
Deleted Account
Does the Windows desktop miner work? How much cpu and memory does it typically consume and how long typically before it mines first block?
G
20:44
Greg
In reply to this message
knowing this, we can't just update the miners?
Z
20:46
Zack
In reply to this message
We have a weird situation where 256**2 bits of nonce are in a block of their own. This makes the ideal solution a little more memory hard than we want. It would make it far more expensive to build an ASIC.
20:48
The new hash algorithm is around 6 times shorter than bitcoin's, so we will need 6 times less logic gates on the ASIC. Which means we can make more ASICS from a single dye.
20:49
https://electronics.stackexchange.com/questions/7042/how-much-does-it-cost-to-have-a-custom-asic-made
Interesting stuff on ASIC manufacture costs.

Seems like there are multiple levels of technology all called ASIC.
G
20:53
Greg
so you really want custom ASICS for Amoveo? why not just using a good GPU algo?
Z
20:54
Zack
$47 per Veo.
G
20:54
Greg
everyone can buy GPUs, there is multiple manufacturer + there is already an enormous individual base. all of this combined seems good decentralization to me
OK
20:55
O K
In reply to this message
Are you using tmux on server? Mine runs smooth after I detach
G
20:55
Greg
In reply to this message
based on one auction! I don't think it works like that πŸ€”
Z
20:55
Zack
In reply to this message
it crashed while I was attached. I guess it was an imaginary pattern.
20:56
In reply to this message
it is more than 1% of the supply.
G
20:56
Greg
we need to wait the trade to settle between the 2 parties
Z
20:56
Zack
yes, true.
MF
20:56
Mr Flintstone
In reply to this message
also there have been OTC deals in the 30-70 dollar range
G
20:57
Greg
In reply to this message
good start πŸ‘πŸ»
20:58
In reply to this message
the best bid is yours!
MF
20:58
Mr Flintstone
not for long
20:58
i fear this will get ugly
JS
21:00
Jon Snow
@Jbreezy0 is the true whale
MF
21:00
Mr Flintstone
Kassel is the one true whale
21:00
we are but disciples
P|
21:07
Paul N. | BAND (πŸ…‘) - I'll never DM you
oops, I missed the 5 minues mark
21:07
😳
21:08
Deleted Account
It's betting already possible?
Z
21:09
Zack
In reply to this message
yes, but I don't have any markets on my server. you can lauch and amoveo node and make markets.
JS
21:09
Jon Snow
Congrats @Jbreezy0
I?
21:09
IMPLENIA - Crypto Fund πŸ‡ΈπŸ‡¨ πŸ‡¨πŸ‡­
In reply to this message
who is kassel?
P|
21:09
Paul N. | BAND (πŸ…‘) - I'll never DM you
Congrat! @Jbreezy0
21:10
I thought we could have fought till 1 BTC tho lol
21:10
Deleted Account
In reply to this message
there is no ASICS for many coins
21:11
Deleted Account
Kassel is large miner somewhere in Germany
21:12
Deleted Account
Kassel is acity in germany.
JS
21:12
Jon Snow
What’s the current % of hashing power coming from Kassel?
N
21:13
NM$L
21:13
Does my hashrate limited by 300000000
I?
21:18
IMPLENIA - Crypto Fund πŸ‡ΈπŸ‡¨ πŸ‡¨πŸ‡­
In reply to this message
is he here in chat?
Z
21:37
Zack
The trade was a success. It looks like Amoveo is worth about $57.40
21:38
This makes our market cap about $452 000
21:38
144 block rewards, about 1 days worth, is about $8200
N
21:39
NM$L
nice
MF
21:48
Mr Flintstone
In reply to this message
thanks, though if you are willing to pay up to 0.01 btc per veo I am sure there are some ppl will sell
21:49
and now perhaps you don’t have to pay that much
P|
21:50
Paul N. | BAND (πŸ…‘) - I'll never DM you
It's not all about owning veo myself, you know
Z
21:50
Zack
Should I auction off bitcoin next, and see how many Veo people are willing to pay?
P|
21:50
Paul N. | BAND (πŸ…‘) - I'll never DM you
it's about kicking start amoveo as well
MF
21:50
Mr Flintstone
fair
21:51
In reply to this message
interesting idea
P|
21:51
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
I won't be participating lol
21:52
you won't be taking my veo any time soon
21:52
πŸ˜‰
MF
21:54
Mr Flintstone
yeah 0% chance I sell any
Z
21:55
Zack
everyone has a price
P|
21:55
Paul N. | BAND (πŸ…‘) - I'll never DM you
I do, if you plan to reset everything at any moment
21:56
Deleted Account
lets reset everything!
P|
21:56
Paul N. | BAND (πŸ…‘) - I'll never DM you
I'd sell in a heartbeat
21:56
lol
21:56
Deleted Account
too much went to secret miners
21:56
i want more veo but dont feel too committed to poject with the little i have :P
MF
22:06
Mr Flintstone
In reply to this message
yeah this is the same for me obvs
22:07
In reply to this message
0% chance was probably a little exaggerated lol
Z
22:08
Zack
Can someone confirm that the batch size for SHA256 is 55 bytes, so if our header is less than 55 bytes we can do it in a single batch?
AS
22:10
Aizen Sou
Ur pool is down again, Zack?
Z
22:11
Zack
yes. thank you
AS
22:14
Aizen Sou
You're welcome. I just want to check my mined balance πŸ˜…. Thanks for fixing it
G
22:25
Greg
Zach, could you please make a new estimate on the probability that the chain survives the end of the month?
Z
22:29
Zack
The situation with Kassel, and the results of the auction are both positive.
But my node has crashed more than a dozen times in the last day, which is very negative.
Hopefully I will find a solution to this soon.

I think for now, we have a 60% chance to still know who's tokens are who's at the end of the month, but we have only a 30% chance that the network will be usable for more than 1/2 the time.
The network might get frozen for days or weeks if we can't figure out what is wrong soon.
22:31
Today I will search the merkle trees, maybe we aren't garbage collecting something.
I probably should have checked them first, they are the most likely to have this kind of problem.
G
22:33
Greg
interesting, thanks for this update
P|
22:34
Paul N. | BAND (πŸ…‘) - I'll never DM you
Zack Again, thanks for the hard work man. You're doing great, and we're here to root for you. (and yes I am greedy, and that's why I'm still here ;) )
G
22:35
Gonzalo
πŸ‘πŸ‘
22:36
Deleted Account
yeah good job we need to make prediction market dream real
Z
22:37
Zack
Maybe if I launch on a server with more ram, it will delay the problem a couple weeks
P|
22:37
Paul N. | BAND (πŸ…‘) - I'll never DM you
Zack how much ram you're running it on now?
Z
22:38
Zack
500 mb
P|
22:38
Paul N. | BAND (πŸ…‘) - I'll never DM you
GG Zack
Z
22:39
Zack
gg? its not over yet
P|
22:39
Paul N. | BAND (πŸ…‘) - I'll never DM you
my CPU miners' provider could only allow minimum ram of 86 GB on that machine
Z
22:39
Zack
haha
22:39
does Amoveo crash on your node?
P|
22:40
Paul N. | BAND (πŸ…‘) - I'll never DM you
it was on day 1 or 2, I'm not running the node anymore
22:40
but I don't recall it crashing πŸ˜…
G
22:42
Gonzalo
My node is running smooth with only 2GB, no crashes at all, but also not serving as a pool
Z
22:42
Zack
In reply to this message
Thanks Gonzalo, good to know
G
22:43
Gonzalo
πŸ‘
MF
22:46
Mr Flintstone
do you think the node is crashing because of communication with other nodes/ miners?
Z
22:48
Zack
Amoveo is supposed to consume 250 mb of ram while syncing, which it is doing correctly.
It is supposed to consume <50mb while maintaining consensus, but it is actually taking 250+ mb
Ant then suddenly something tries to allocate 145+ more mb in one moment, and it crashes.
22:48
I guess there is a memory leak
Z
23:19
Zack
Maybe it is in governance. Because the governance tree pointer keeps increasing even though we haven't modified anything.
SS
23:25
S S
Zack I noticed that each you put a value in trie, all stems are recreated. Could that be a reason for memory leak?
Z
23:26
Zack
I should delete the unused code from that directory.
We only insert using batch now.
SS
23:26
S S
Oh okay
Z
23:27
Zack
Thank you for looking. It is helpful.
SS
23:27
S S
Np
Z
23:31
Zack
test 18 and 19 in test_txs.erl show that we can have thousands of blocks completely filled with create account or spend txs, and it doesn't cause a problem.

So I guess it can't be the Merkel trees.

I guess it has something to do with communicating with other nodes.
23:32
maybe the mining pool is making me do some complicated work and overflowing a mailbox on one of the gen_servers.
23:35
potential_block:read(). might be causing the problem then.
23:39
maybe it called "tree_data:garbage()" an operation that takes a long time. and while waiting for that, some other gen_server that needs to read from potential_block also gets frozen, so it can't process it's mailbox, and the mailbox overflows.
23:40
garbage is a call type, not a cast type, so this could be it.
23:43
So I guess a good short term solution would be to increase the frequency that the garbage collector runs, that way it doesn't take as long.
I
23:46
Iridescence
My miner is having trouble with the mining pool, so my pool isn't finding blocks anyway. =P
Z
23:47
Zack
maybe I should put a delay in the mining pool.
I
23:48
Iridescence
I'm still in the middle of debugging
P|
23:48
Paul N. | BAND (πŸ…‘) - I'll never DM you
23:48
seems down
Z
23:49
Zack
yes, it is. I am about to push a fix.
I
23:49
Iridescence
Part of my problems seem to stem from using files to pass info between the Erlang program and the C miner
Z
23:49
Zack
thanks for saying
23:52
In reply to this message
yeah, that probably is not a good design choice on my part. I don't know much C.
P|
23:52
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
Yeah. I was curious as well. Won't communicating via TCP be sufficient/less error-prone?
Z
23:54
Zack
I think I should put a limiter in the mining pool, so nodes can only contact me every 1 second or so.
That way it wouldn't overwhelm erlang when we are garbage collecting
OK
23:57
O K
Zack I can make you a user account with my VPS that you could use for troubleshooting if it would help, my node has been relatively stable but it's not used as a mining pool (even though the mining pool is activated)
Z
23:59
Zack
In reply to this message
thanks for offering
8 March 2018
Z
00:03
Zack
increasing garbage collection frequency didn't seem to help
00:10
im installing on a bigger server.
Tv
00:10
Tarrence van As
@Iridescence after updating the gpu miner my hash rate went from 150MH/s to 70MH/s
I
00:12
Iridescence
Haha
00:13
I don't think you got the update from me =)
MF
00:13
Mr Flintstone
In reply to this message
I hope this can work as stop gap solution while we try to fix issue
00:13
I hope nodes can be run with low ram requirements
I
00:14
Iridescence
But if you are indeed seeing worse performance I can help you troubleshoot
00:15
Oh I get what you mean, I was confused over what you meant by update
Z
00:17
Zack
159.65.120.84
Here is my second mining pool.
OK
00:26
O K
Are you familiar with claymore miner @Iridescence
00:29
He makes money by close-sourcing the miner, releasing to public, having it mine to user configured address for something like 59/60 minutes, and mine to his address for 1 min
00:29
I don't know the exact specifications
00:30
But then at least we could try your optimized miner and see the results for ourselves on our own hardware
Z
00:37
Zack
how do you hide your address in a compiled file?
Store each byte in a different variable?
OK
00:38
O K
I don't know that his addresses were secret
00:38
I don't know the answer, I'm sorry
00:38
That's his business model though
Z
00:38
Zack
if you find the address in the compiled file, you can replace it with your own, and then you wont be paying him.
OK
00:39
O K
If we miners were very tech savvy, we would just optimize our own miners
Z
00:39
Zack
someone is already using the new mining pool
00:41
http://159.65.120.84:8080/wallet.html
Here is it's location.
If my first node goes down, you can use this one.
OK
00:44
O K
Whoever is coding the miners to work with default style pool might consider a backup define(Peer line for if one peer goes down
Z
00:44
Zack
good idea.
AS
00:54
Aizen Sou
just adding another define(Peer) line in miner.erl ?
Z
00:55
Zack
I don't understand the question.
OK
00:59
O K
The miner would have to recognize that the first peer was unreachable, and then switch to the backup peer
01:00
Just adding another line would probably break something
AS
01:00
Aizen Sou
my question was how to define backup pool, Zack
P|
01:00
Paul N. | BAND (πŸ…‘) - I'll never DM you
now only BPWN... pumping out new blocks
AS
01:00
Aizen Sou
just simple like this:
01:00
-define(Peer, "http://IP1:8085/").
-define(Peer, "http://IP2:8085/").
-define(Peer, "http://IP3:8085/").
Z
01:01
Zack
In reply to this message
true. It would just use the last peer defined, or it would throw an error. I don't know which.
01:01
In reply to this message
false.
01:01
In reply to this message
That is my new pool I just set up. It seems to not crash so far.
P|
01:01
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
Should I try to DDOS it? πŸ€‘
Z
01:02
Zack
In reply to this message
I would rather you ddos my old one that isn't making me money
AS
01:02
Aizen Sou
In reply to this message
how to do it then, Zack ?
Z
01:02
Zack
In reply to this message
Are you using my C miner? you can only connect to 1 peer.
P|
01:03
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
I'm on Mandel's pool, so I have all the incentive to DDOS your running pool 😀
Z
01:03
Zack
In reply to this message
or you could join my pool
P|
01:03
Paul N. | BAND (πŸ…‘) - I'll never DM you
ask @Iridescence to open source his superior miner lol
OK
01:03
O K
I've tried
P|
01:04
Paul N. | BAND (πŸ…‘) - I'll never DM you
if he does, it'd jump to Zack pool in a beat
OK
01:04
O K
That's what he's afraid of
P|
01:04
Paul N. | BAND (πŸ…‘) - I'll never DM you
if I could install cuda on my ubuntu lol
OK
01:04
O K
I think he should work something out with Mandel to get a % of fees from Mandel's pool, he'll get more blocks that way
01:05
I can help you Paul
P|
01:05
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
He (@Iridescence) can create a new pool protocol that's not compatible with Zack's, and force us to use his
01:06
it's what Mandel did
OK
01:06
O K
That too, I've suggested many alternatives to his current model
01:06
I'm not going to purchase something I can't see working with my own eyes
P|
01:06
Paul N. | BAND (πŸ…‘) - I'll never DM you
I like how the shares are spit though
01:06
yeah
OK
01:06
O K
I would rather pay a fee
P|
01:06
Paul N. | BAND (πŸ…‘) - I'll never DM you
me too
I
01:07
Iridescence
I understand
01:07
I'm currently facing bugs with the mining pool anyway
01:07
So my miner isn't doing much
P|
01:07
Paul N. | BAND (πŸ…‘) - I'll never DM you
Also Zack do you know if it's possible for someone to mine on your pool, but simultaneously submitting work for Mandel's pool?
01:07
I think it's a good strategy
OK
01:07
O K
If you could work out a deal with Mandel, it would benefit both you and everyone on Mandel's pool
P|
01:07
Paul N. | BAND (πŸ…‘) - I'll never DM you
especially if that person can dominate hashing power
Z
01:08
Zack
In reply to this message
not possible.
P|
01:09
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
Why? Because of some string provided by the pools, that miner needs to include in their nouces?
Z
01:09
Zack
there is no step of the process that is paralelizable
P|
01:09
Paul N. | BAND (πŸ…‘) - I'll never DM you
Sry for the dumb question, i'm not very knowledgable in this field
01:09
No
OK
01:10
O K
Also if we can get our community hashrate up, it makes mining even less profitable for kassel, and maybe he will wind down a bit
P|
01:12
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
Oh right, because the pubkey of the pool is included in the input of hashing function. My bad.
Z
01:12
Zack
In reply to this message
you understand. great
OK
01:14
O K
These notes were also helpful, but not directly the solution https://www.linkedin.com/pulse/installing-nvidia-cuda-80-ubuntu-1604-linux-gpu-new-victor
P|
01:15
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
I think Kassel is mining at approx. 5x my hash power, given the same cost. The key to his hash power should come from

1) Access to V100 GPU β€” That's 3x more cost effective than the closest competitor (P100)
2) Access to better miner β€” That's 1.5-2x more hash rate
01:16
And on the V100 GPU, there's deep learning computing module, which, from my understanding, can do 4x-10x faster in matrix operations.
01:16
that's probably why he's so superior in pumping out hash
OK
01:17
O K
In reply to this message
I was basing my assumptions off this
P|
01:17
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
This can be tricky, depending on the cloud provider.
01:20
Zack there'll be difficulty adjustment at block #6000?
Z
01:23
Zack
In reply to this message
Yes
I
01:23
Iridescence
In reply to this message
To nitpick, hashing does not use matrix operations, and so it does not benefit from those features
01:24
But the V100 is the latest and greatest, and so it had superior hashpower
P|
01:26
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
fair enough
Kyle Spencer joined group by link from Group
Deleted joined group by link from Group
P|
02:26
Paul N. | BAND (πŸ…‘) - I'll never DM you
I'm buying 100 Veo for 0.3 BTC
02:27
PM me if you're interested
I
02:56
Iridescence
I wonder what the trading volume of VEO is so far
MF
03:08
Mr Flintstone
I think ~400
03:08
that would be my guess
G
03:40
Gonzalo
So finally i did it... new AMOVEO POOL at 51.15.75.100:8085
No fees for the next 10 blocks, then only 2% fee. 😁
03:40
let's have some fun
03:48
Hey Kassel, BATman... over here! πŸ‘‹
03:48
πŸ˜‚
BB
05:04
Brian Brian
Only rewarding block finders?
MF
05:11
Mr Flintstone
Is there a new light wallet link?
OK
05:12
O K
Correct
G
05:12
Gonzalo
In reply to this message
Yes, not shared
Z
05:13
Zack
The old light wallet site should still work too.
OK
05:13
O K
It's been down
M
05:13
Mike
How do I go about generating a vanity address
Z
05:13
Zack
Oh. Thanks for telling me.
OK
05:13
O K
I would have told you sooner, but I misunderstood and thought it was just a transfer
G
05:14
Gonzalo
In reply to this message
OK
05:14
O K
Neat, whose site is that
05:15
I made on in erlang, and it comes with the full node.
OK
05:15
O K
Eventually we should probably get away from generating private keys online, someone is going to get got someday soon
MF
05:16
Mr Flintstone
@O K
Z
05:16
Zack
It is a little more optimized too. I only generate a random number once for your private key, and keep incrementing it to find your vanity address.
MF
05:16
Mr Flintstone
you can right click the light wallet white space
G
05:16
Gonzalo
In reply to this message
I did it, just made a few mods on Zacks wallet
MF
05:16
Mr Flintstone
save as
Z
05:16
Zack
Instead of generating a new private key for each try
MF
05:16
Mr Flintstone
and then open it locally
05:16
that way you can be more sure everything is client side
05:16
then, you can input whatever entropy you want
OK
05:16
O K
πŸ‘
05:17
I just mean as a community, foundationally, the light wallets were convenient but will eventually be used to scam someone
BB
05:17
Brian Brian
I think we need to stop calling nodes pools. They aren’t pools they are just full nodes people are using to not have to run one themselves
G
05:17
Gonzalo
In reply to this message
Way better πŸ‘
MF
05:17
Mr Flintstone
this way, you can get yours before you get got doe
Z
05:18
Zack
In reply to this message
Yes.
MF
05:18
Mr Flintstone
I’m honestly surprised we haven’t seen people try to phish the light wallet yet
Z
05:18
Zack
In reply to this message
The full node software and the mining pool software come from different github repositories. What would you call the mining pool repository instead?
OK
05:18
O K
Plus the NSA has all our private keys πŸ€”
MF
05:18
Mr Flintstone
everyone please be extra mindful of revealing your private key, though that should go without saying
Z
05:19
Zack
In reply to this message
If you use a light wallet served by your own full node, I don't think the NSA can take your keys. The key is only stored locally, never transmitted.
OK
05:20
O K
Even if your full node is on a VPS?
Z
05:20
Zack
Yes
OK
05:20
O K
It's generated locally and not shipped via TCP?
05:20
Good to know, thanks
Z
05:20
Zack
It is a light node. It is verifying headers on your local machine. You use the private key to sin transactions
05:20
Verifying headers in the browser
OK
05:21
O K
The "Save private keys to file"
Z
05:21
Zack
The private key is only stored in your own browser. Never sent to anyone
G
05:22
Gonzalo
In reply to this message
You should always check the code. @ amoveo.online the "generator" is like the wallet, is just JavaScript and It runs on your browser, nothing is sent back.
OK
05:22
O K
I trust you Gonzalo but hear me now, someone, someday is going to bait and switch
Luke B joined group by link from Group
Z
05:22
Zack
Of course, it is trivial to make a scam version of the light wallet that does steal your private key.
05:23
I like the light wallet because I can have Veo on my phone
05:23
It is convenient for small amounts
BB
05:28
Brian Brian
In reply to this message
Miner node maybe? I dunno I’ll think on it but as far as I can tell we only have 1 real pool so far and it’s Mandels
J M joined group by link from Group
OK
05:42
O K
Are you working on 253 Zack? It's still down
Z
05:43
Zack
no
05:45
I did discover that the same error is happening on the bigger node. but the bigger node doesn't crash, it just starts wasting resources heavily for 10 minutes.
So hopefully I can figure out what is breaking, and then Amoveo wont crash on lighter hardware.
OK
05:47
O K
Interesting, I just ssh'd into my node and it's barely responsive
Z
05:49
Zack
What does "log into your node" and "really crawling" mean?
OK
05:49
O K
It doesn't look that bad looking at top though
Z
05:49
Zack
is the light node syncing headers slowly?
OK
05:49
O K
edited
G
05:50
Gonzalo
In reply to this message
Yes you're right
Z
05:50
Zack
you are probably experiencing the problem I am working on.
OK
05:51
O K
No one uses my 'mining node'
05:51
as far as I know
05:51
Wouldn't it show up in free -m or in top as high usage
Deleted joined group by link from Group
Z
05:52
Zack
they are both called "beam" in top.
05:52
I dont' know about free -m
OK
05:53
O K
Beam is using 7.6%MEM of a gig
05:59
I don't seem to be having the same problem, whatever it was it resolved. Probably my connection
M
06:02
Mike
No AMD miner yet?
Z
06:16
Zack
There was a nice log file this time, so now I know it is crashing because potential_block gets stuck in a non-functional state.
The miner keeps querying an api that call potential_block, it fills up too much ram.

potential block is not written very well. I guess I will rewrite it more carefully, and see if that fixes it.
OK
06:17
O K
Mike No public AMD miner to my knowledge, do you have AMD cards ready to work?
MF
06:19
Mr Flintstone
In reply to this message
nice
M
06:19
Mike
In reply to this message
Yeah over 20
MF
06:20
Mr Flintstone
I thought Zack had an opencl miner?
06:20
not sure if it ever worked
Z
06:21
Zack
I never got it working for gpu.
OK
06:25
O K
Mike I have some people looking into it, I'll keep you posted
06:26
I have some sitting idle too
M
06:26
Mike
In reply to this message
Sounds good
MF
06:28
Mr Flintstone
me as well, though probably not on the same scale as you guys
AS
06:39
Aizen Sou
anyone knows who owns BKyLqCvF5vt7KBiPLId7xlW64Bb/uV+icPnvn5mgEKncqlCKiaXUTkRcZv7iYK3GKxSspNh0leIhc4YA+DJAY/s= ?
Z
06:39
Zack
that is Mandel's pool address
06:39
I think
AS
06:41
Aizen Sou
ur 120.84 pool is down again, Zack ?
Z
06:41
Zack
In reply to this message
yes, thanks for telling me.
06:41
I think I found a fix to our problem, I am doing some final testing on it.
AS
06:42
Aizen Sou
no need to thank. just bad luck because everytime ineed to check balance it went down πŸ˜…
06:42
awesome, thanks Zack
Z
06:42
Zack
I found a race condition in block_absorber where if your miner asked for a potential_block:read at exactly the right moment, you would end up deleting some parts of the merkel tree.
06:43
but it is fixed now.
Z
06:59
Zack
I pushed the fix, and updated both my nodes. Lets see if it works.
OK
07:03
O K
πŸ‘
Z
07:07
Zack
the small node still crashed. a mystery.
07:08
I notice that Mandel's pool mines faster than 10 minutes per block when my pool is off.
07:08
I guess the difficulty is going to go up a lot for the next 2 retargets at least.
07:14
if you rent a server and keep money on it, do the people hosting your hardware ever steal the tokens?
OK
07:18
O K
I couldn't find news related to that, I'd like to think that people keep all kinds of sensitive data on VPS and professional providers should be trustworthy. Seems like a good rule of thumb to clear the account regularly though
07:18
Could set up a simple honeypot
Z
07:19
Zack
I want to launch a lightning network provider, so I need to keep a fairly large amount of Veo on a hub.
07:19
profit is in proportion to how much Veo is on the hub.
07:20
I guess I will start small, and see how it goes
OK
07:21
O K
Like the guy that put $100 in LTC on private keys that he used as passwords for various sites to see if they were salting passwords
Z
07:21
Zack
cool idea
07:30
maybe the first server has a botnet pointed at it.
Maybe someone opened the web miner in like 100 browsers.
And the excessive noise is making it crash.

And maybe the second server crashed because of that race condition I solved.

I left the first server running without the mining pool, the second server still has it's mining pool.
K
07:35
Kenny
@Mandelhoff looks like your pool is down
OK
07:40
O K
The site loaded for me eventually
07:40
I wonder if he's getting DDOS
07:41
Miners still offline
Z
07:42
Zack
It does say his hashrate is nearly 400 gh now
07:42
He is probably having the same problem I was.
OK
07:43
O K
oh no 😨
07:43
Where do you check hashrate statistics?
Z
07:43
Zack
It is about 1.1 TH for the network.
07:43
http://amoveopool.com/pool
this is mandel's pool info.
OK
07:44
O K
How do you get network statistics?
Z
07:44
Zack
block:hashrate_estimate(). tells you the estimate for the network.
07:44
it estimates based on the 10 most recent blocks.
OK
07:45
O K
How does that 1.1 figure compare to yesterday?
K
07:46
Kenny
yeah. the miner is just getting getwork failed so im pretty sure any hashrate going there is effectively 0 unless they move to zack. its been so smooth lately too. oh well. also looks like he hasnt been here since midnight last night.
OK
07:46
O K
0.73 blocks per minute since launch, slowing down a bit
K
07:46
Kenny
zack do you have another way to contact him? you might try to see if you can get a way to notify him when its down other than just telegram.
Z
07:49
Zack
L = lists:map(fun(N) -> round(block:hashrate_estimate(N)) end, lists:seq(100, block:height(), 100)).
rp(L).
07:50
[2,2,2,3,1,1,1,1,1,2,1,2,2,2,2,2,1,2,2,1,3,21,70,91,114,108,96,108,114,96,87,91,96,96,108,114,131,131,96,96,335,566,415,432,476,804,1219,410,773,1058,1277,796,307,376,638,320,659,354]
MH
07:51
Mandel Hoff
Try to fix now. Yes, pool has crashed many times today. Syncing from 0. Must have corrupted cause couldn't sync after simple restart
K
07:52
Kenny
thx. keep up the good work!!!!
Z
07:52
Zack
In reply to this message
maybe the update will help
MH
07:52
Mandel Hoff
Sounds good. I did a git pull so should have the latest now.
Z
07:52
Zack
In reply to this message
πŸ‘
M
07:57
Mike
In reply to this message
Everything’s back working thanks
OK
07:58
O K
In reply to this message
what is lists:seq(x,y,z).
Z
07:58
Zack
I think we will go above 100 terahash per block after height 6000.
08:00
It give the estimated hash rate every 100 blocks in gh/s
OK
08:00
O K
aw man I almost figured it out
08:00
wow very interesting
MF
08:01
Mr Flintstone
In reply to this message
why do you think so?
Z
08:01
Zack
because we are at 20 th/block now, and I think the average block period during the last 2000 blocks has been below 2 minutes.
08:02
it is about 90 seconds now
08:02
104 seconds
OK
08:02
O K
So the last 700 blocks have been under 1 TH?
J
08:02
Jim
At some point, min payout at amoveopool.com and tx fee should be lowered. Prob after next difficulty adjustment.
Z
08:03
Zack
In reply to this message
no. each of those blocks is 20 TH.
The rate of mining was below 1 terahash per second.
OK
08:04
O K
Sorry, yes
Z
08:05
Zack
oh, it is actually 80 th/ block right now by my estimate.
Mandel's estimate is 20th/block. I wonder which of us is more accurate.
MH
08:08
Mandel Hoff
In reply to this message
Tx fee can not be lowered. It's a governance value.
08:09
In reply to this message
I'm sure you are closer. My pool always estimates high for an unknown reason
Z
08:10
Zack
I am adding a command block:hashes_per_block(). to calculate that.
Also useful are block:hashrate_estimate(). for gh/s and block:period_estimate(). for seconds per block.
OK
08:10
O K
Is there a consolidated document of all commands?
Z
08:10
Zack
In reply to this message
there is a secondary fee controlled in your config file that can be lowered. but I think it is like 20 satoshi right now
OK
08:11
O K
I've been through those
08:12
Deleted Account
@Mandelhoff ive had an upaid balance in the pool for over 24 hours
08:12
how long does it take to pay out?
MH
08:13
Mandel Hoff
In reply to this message
Payout triggers at 1.0 veo.
08:14
Deleted Account
In reply to this message
Got it.
08:14
Thanks!
BB
08:16
Brian Brian
all these people trying to get their unpaid balance and i can't get a fraction of a veo with my measily 220 MH/s since I don't have windows
OK
08:17
O K
There must be a bad link somewhere, because somehow I can start at https://github.com/zack-bitcoin and end up on a "blob" that has old information on it
08:17
Then you link me the same thing and it's up to date
08:17
I'll try to help you find it
Z
08:18
Zack
In reply to this message
cool, thank you for noticing this
08:18
yes, grep quickly found a lot of examples of this problem
08:19
are you sure "blob" is a problem?
OK
08:19
O K
I'm not sure
08:19
I'll be browsing the github on a different computer and end up on an old page
08:19
maybe I got there by searching
G
08:21
Gonzalo
potential block died!false 3
08:21
sorry to long
08:21
any idea Zack? nodes are not syncing
Z
08:22
Zack
In reply to this message
are you running a mining pool?
OK
08:22
O K
I see, I remembered there being an advanced guide, so I searched for "advanced" -- took me to quick_start_developer_guide.md from which I went up a level to docs and by then I'm on an old update
G
08:22
Gonzalo
yes, but I copy pasted another xterm windows, sorry
Z
08:22
Zack
ive been having similar problems, we are trying to find the cause.
It seems to me that the mining pool is wasting a bunch of ram occasionally, and it can make you crash.
G
08:23
Gonzalo
To be sure I've hired a nice server, 8 cores, 32GB, SSD... a bit oversized haha
08:24
htop mem is always around 275mb
Z
08:24
Zack
In reply to this message
I am having trouble repeating this problem. Why do you think it is an old version?
08:25
In reply to this message
Did you get that recent update a few hours ago? I fixed a race condition that can make potential block died! happen.
OK
08:27
O K
https://github.com/zack-bitcoin/amoveo/search?utf8=%E2%9C%93&q=advanced&type= click that guide, - getting started - turn it on - commands - mining
G
08:28
Gonzalo
I did a pull like 5 hours ago and right now again, this is what I got:
apps/amoveo_core/src/consensus/chain/block_absorber.erl  |  5 ++--- 
apps/amoveo_core/src/consensus/chain/potential_block.erl | 26 +++++++++++++-------------
OK
08:28
O K
It's probably just me using github poorly, but that's how I ended up on an old fork inadvertently
Z
08:29
Zack
In reply to this message
great, I reproduced it. thank you for the help
08:29
In reply to this message
The update in question only chanced like 4 lines.
08:29
that is a different update.
08:29
and 5 hours is too long ago
G
08:30
Gonzalo
In reply to this message
here in "Amoveo world" it really is
Z
08:31
Zack
In reply to this message
I think this is a limitation of github search.
But if you simply change to the "master" branch after doing a search, I think it fixes it.
08:31
yes, it was only 2 hours ago
LB
08:32
Luke B
so is this guy in Kassel winning every block?
OK
08:32
O K
Okay, I will be more careful
Z
08:32
Zack
In reply to this message
yes. almost all of the blocks.
G
08:33
Gonzalo
@Zack are your nodes updated? I mean block height?
LB
08:33
Luke B
I have tesla GPUs running and cant win a block
Z
08:33
Zack
5868, both mine and Mandels
G
08:33
Gonzalo
ok
OK
08:33
O K
Don't solo mine Christopher
Z
08:33
Zack
In reply to this message
are you in my mining pool?
LB
08:33
Luke B
no
Z
08:34
Zack
I recommend joining my mining pool to see if that helps
LB
08:34
Luke B
^is that the IP and port?
Z
08:34
Zack
there is issues with block propagation.
08:34
oh, no
08:34
the port is 8085
08:35
There are issues with block propagation when we are finding blocks this fast
LB
08:35
Luke B
Would you give me some VEO to add CSS to the block explorer page? lol
Z
08:36
Zack
We could set up a smart contract to potentially pay you to do this.
08:36
In reply to this message
I think that is what Mandel did http://amoveopool.com/pool
LB
08:37
Luke B
yeah his page looks really nice
G
08:42
Gonzalo
syncing with this peer now [-7,[-7,159,65,120,84],8080]
trade peers
get their top header
get blocks from them.
another get_blocks thread
block organizer add
Z
08:43
Zack
In reply to this message
looks good.
G
08:43
Gonzalo
... but block height is stuck
08:49
I had to clean up and resync
08:50
sometimes the first peer you start downloading from decides to stop talking to you, and then you need to find a different peer to get the rest of the blocks from.
G
08:50
Gonzalo
thanks is ok now
Z
08:51
Zack
especially since an update I did recently.
It used to be you would download blocks from a hard-coded peer. now you download the list of peers first, and then randomly select someone to download blocks from.

This reduced bandwidth requirements for the nodes on the hardcoded list.
08:51
but it came with the cost that syncing is less reliable.
G
08:53
Gonzalo
that's why I asked before if you were synced, to force my node syncing with you
08:54
i think this is the first time in more than 24 hours that the node loses sync
Z
08:54
Zack
you are running a mining pool?
G
08:55
Gonzalo
08:55
but nobody is coming to my grand opening
Z
08:56
Zack
oh right, we decided it was probably the race condition from block_absorber that I fixed 2 hours ago
G
08:56
Gonzalo
0% fees the next 10 blocks 😁
08:57
are there any usefull commands on the pool side for the pool op?
Z
08:59
Zack
https://github.com/zack-bitcoin/amoveo-mining-pool/blob/master/apps/amoveo_mining_pool/src/http_handler.erl#L25
it has 2 commands worth calling. the 3rd command just turns it on, and you don't want to turn it on more than once.
MF
08:59
Mr Flintstone
oh man 6000 is soon
Z
09:00
Zack
yes, but I think it wont help much. The block period will still be too short.
G
09:01
Gonzalo
thanks Zack
Z
09:01
Zack
maybe once we reach 8000 it will behave more normal.
09:01
one command is for requesting a problem to work on. the other command is for submitting a solution.
09:02
70 seconds per block now.
09:03
Deleted Account
anyone know why in the pool stats the hash rate on last share bounces so much?
09:03
on the amoveo pool that is
Z
09:03
Zack
In reply to this message
It is closed source. I am not certain how he calculates those numbers.
09:04
I guess I should put some of this info into the block explorer page
G
09:06
Gonzalo
come on KasselMAN, BATman!! Who wants some 0% blocks?? Grand opening tonight! Free drinks, pool dancers all around 😁
OK
09:06
O K
8085
G
09:06
Gonzalo
In reply to this message
thanks, Im a bit drunk
09:06
party you know
Z
09:07
Zack
less than 150 minutes till 6000 it looks like.
MF
09:10
Mr Flintstone
it’s roughly a minute per block, so 100 minutes
09:10
?
09:10
unless my block height is off
09:11
5898
G
09:11
Gonzalo
5899 now
MF
09:11
Mr Flintstone
In reply to this message
my guess is calculating expected hash rate based on time between last two valid shares and their difficulty
09:12
maybe he uses more than just the last two but it is rly volatile so irk
09:12
Idk
Z
09:15
Zack
In reply to this message
you are right.
09:16
In reply to this message
don't people submit partial pow solutions, and he gives rewards? I think he could use those to get a more accurate estimate.
M
09:20
Mike
In reply to this message
This would be a dominant assurance contract?
Z
09:20
Zack
yes
M
09:21
Mike
I’m confused about how I get interest if I contribute but it isn’t built
MH
09:23
Mandel Hoff
In reply to this message
Yes, the pool uses your partial pow solution difficulty and how long it takes you to give that to estimate your hashrate. That bounces around very similar to how block times range from 20 seconds to several minutes even though they have the same difficulty. The pool is not doing any averaging on your partial pow to make a "smooth hashrate".
MF
09:27
Mr Flintstone
In reply to this message
ya that’s what I mean
09:28
you submit a solution that may be less difficult than the block difficulty and you are awarded a share of some kind
Z
09:28
Zack
In reply to this message
the entrepenuer is taking on risk.
M
09:33
Mike
In reply to this message
Ok I get it now. Thanks
Z
09:35
Zack
The only smart contracts I have programmed so far are lightning payments and binary derivatives.

We can use binary derivatives and a modified api to approximate a dominant assurance contract. But I think we might be able to do even better by writing a new specialized smart contract for dominant assurance contracts.
09:36
Another important smart contract to make is the 2x2 conditional contract. It measures the correlation between the likelihood of different future events coming to pass.
KR
09:42
Kevin Roberts BWD
Any Ideas?
Z
09:43
Zack
error code 44 I am guessing?
I
09:43
Iridescence
That's my miner
KR
09:43
Kevin Roberts BWD
yep
I
09:43
Iridescence
Did the code compile correctly?
KR
09:43
Kevin Roberts BWD
yes
I
09:44
Iridescence
What gpu?
KR
09:44
Kevin Roberts BWD
let me check (aws)
I
09:45
Iridescence
nvidia-smi command will tell you as well
MF
09:50
Mr Flintstone
In reply to this message
these are on your github?
Z
09:54
Zack
they are in amoveo. you can use them from the light node.
09:54
but I don't think anyone has a server set up for it yet
09:55
for lightning payments and channels, the server needs a positive balance, and unlocked keys.
the size of the server's balance is a limit on how big the channel can be, since the server's money gets locked up in the channel.
09:59
Mandel's node looks like it is finding a block every 10 minutes now
MF
10:03
Mr Flintstone
In reply to this message
do you have the code on github?
Z
10:04
Zack
yes. the entire amoveo full node in erlang is open source, and so is the entire amoveo light node in javascript.
MF
10:05
Mr Flintstone
I mean the smart contract code
10:06
the market one is split into two pages, because the two parts are more reusable this way.
10:07
this is the main repository for the smart contract system: https://github.com/zack-bitcoin/chalang
10:08
it has compilers and a virtual machine and documentation and example code
MF
10:34
Mr Flintstone
thank you
10:35
Deleted Account
Hi, I finally succeeded to compile Mandel's cuda GPU miner on ubuntu.
Z
10:36
Zack
In reply to this message
great
I
10:36
Iridescence
but have you succeeded in compiling my ubuntu cuda GPU miner on Windows? 😁
10:36
Deleted Account
Oh not yet
10:36
I do not have windows lol
I
10:36
Iridescence
haha
10:36
such is life
N
10:36
NM$L
@Mandelhoff when i get paid balance?
10:36
I have 0.4+ unpaid balance
Z
10:37
Zack
at 1.0 I think.
10:38
Feel free to use if you want
N
10:38
NM$L
wow
G
10:38
Gonzalo
In reply to this message
πŸ‘πŸ‘πŸ‘
N
10:38
NM$L
Thank you very much
Z
10:38
Zack
isn't it risky to share software that way?
10:38
why not github?
MF
10:39
Mr Flintstone
ya be careful
Z
10:39
Zack
maybe it will just take all your amoveo tokens
10:39
Deleted Account
It includes binary and sources. You can compile by yourself
10:41
but github is better of course. sorry I'm a bit in hurry
P|
10:50
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
Use trusted provide like AWS or Google Cloud
M
10:55
Mike
In reply to this message
Is there any reading on this type of contract?
Phil Bonello joined group by link from Group
Z
11:13
Zack
Paul Sztorc wrote great stuff.
11:14
This is one of my favorites
11:14
but that entire website is great.
M
11:15
Mike
In reply to this message
Lol I had gotten to page 6 last week and stopped. The answer is on page 8 :p
11:20
11:20
Perfect explanation. Thanks Zach
G
11:24
Gonzalo
@Mandelhoff I think your pool is stuck at height 5944. Networks is at 5968
MH
11:26
Mandel Hoff
Thanks - checking. Probably something wrong.
11:27
Yes, lots of different died! errors on node. Fixing
G
11:27
Gonzalo
I had to make prod-clean on my nodes
11:27
before
11:27
both where stuck at diff heights
MH
11:30
Mandel Hoff
Yes, restart just went in to more died errors. Doing a clean sync now.
G
11:30
Gonzalo
πŸ‘
11:31
are you planing any update on your miners? better h/s?
MH
11:33
Mandel Hoff
Yes, probably have new cuda tomorrow. I rewrote a lot. It's a lot faster on my 1060 from 250 to 380mhs. Like 107 to 125 mhs on my 750ti.
G
11:36
Gonzalo
wow great
11:38
Deleted Account
Great!!!!
11:39
how much is an veo?
MF
12:02
Mr Flintstone
10 more blocks
K
12:19
Kenny
whats happening..... i keep getting new work from mandels pool with the old difficulty, but zacks pool is stuck at 5999
MF
12:20
Mr Flintstone
I see 6000 in light wallet
K
12:21
Kenny
what address?
MH
12:21
Mandel Hoff
New Diff: 84,937,273,245,696
K
12:21
Kenny
nvm i see it as well
MF
12:21
Mr Flintstone
12:21
this is connected to one of zacks nodes I believe
MH
12:21
Mandel Hoff
Up from 24.4 T
K
12:21
Kenny
wooooo
J
12:21
Jim
Less than 100T? Woot
K
12:22
Kenny
any guesses on the next block time?
MF
12:22
Mr Flintstone
so we should see ~5 minute blocks
K
12:22
Kenny
any bets?
12:22
lol
MF
12:22
Mr Flintstone
If hash rate doesn’t change
12:22
but it’ll probs go up
12:35
no blocks yet
K
12:36
Kenny
yup 16 minutes and going.....
MF
12:36
Mr Flintstone
if anyone feels like they want to sell their veo pm me
12:37
I am happy to provide liquidity
K
12:37
Kenny
lol
MF
12:37
Mr Flintstone
if Kassel dipped what is expected block time
12:37
what % of blocks was he? 90%?
12:38
if so, then I think like 45 or 50 minute blocks
K
12:42
Kenny
i would guess 1.5 to 3 hrs
12:43
2000 blocks might be a little bit too high of a number to retarget difficulty
12:44
an elegant algo that does it every block would be ideal.... not saying it would be easy, but it would be nice
MF
12:46
Mr Flintstone
it’s been about 25 minutes
K
12:49
Kenny
it has been a while
12:49
this cant be bad luck
12:49
or it could be
MF
12:49
Mr Flintstone
maybe he is playing games with us
12:49
lmao
K
12:50
Kenny
fucking batman
MH
12:51
Mandel Hoff
Good night. Forgive me if the pool goes down. It may be about 95% of the hash power now. πŸ™ˆ
MF
12:51
Mr Flintstone
lol
K
12:51
Kenny
noooooooo
12:51
have a good night
12:51
i cant believe you dont wanna see 6001
12:52
ill mine to ya.... ride or die!!!!!!!!!!!
MH
12:52
Mandel Hoff
We may still be at 6k when I wake up if my pool goes down. πŸ‘
K
12:52
Kenny
ahahahhahahahahahhahhahahahah
k
12:52
kdmlll
Zack, iri pool also down?
K
12:54
Kenny
i think zacks pool is up
12:54
his node is anyways, which usually means the pool is as well
I
12:56
Iridescence
my pool is not currently up
BB
13:02
Brian Brian
6001!
K
13:02
Kenny
wooooooooo
13:02
amoveopool got it!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
13:04
Deleted Account
kassel out of mining?
T
13:04
TownAlwaysGet
Whoaaaa
N
13:04
NM$L
what?
T
13:05
TownAlwaysGet
Why im a getting about 500 mh/s on my cpu?
13:05
What sorcery is this?
13:06
Im on mandels pool
N
13:06
NM$L
THat's not correct
13:07
6003
T
13:09
TownAlwaysGet
Ofcourse its not, but why is it showing that? Probably something wrong with the estimation
MF
13:10
Mr Flintstone
6004
13:34
This version is much faster than the previous version. Gtx1060 goes from 250mhs to 357mhs.

SEED input is now a command line argument that is optional. Use this if you run multiple miners on the same address. Pick a different seed per miner process to avoid nonce collisions.

NUMBLOCKS default is now 256. The 192 numblocks is no longer better for most cards. The K80 is better with 512, but that's the only one that got better with a value other than 256.
BB
13:37
Brian Brian
do you include the optimization that was posted on reddit the other day that was deleted? had to do with hashing the nonce after hashing the rest of the data or something
MH
13:40
Mandel Hoff
I had made that "reuse context" fix in previous versions, but this version is a big refinement with the get/submit work put on separate threads to let the cuda kernel not have to wait.
BB
13:41
Brian Brian
hm so it won't gain any efficiency reusing the ctx?
MH
13:42
Mandel Hoff
Yes, it is already doing that. I implemented a crude version of that the night before that redit post, but it has been refined a few times. This one is turbo.
K
13:43
Kenny
@ mandel.... just fyi.... on .0.0.3 with 0 192 after my address i would get 397mh.............. with .0.0.4 default i get ......394mh............with 0.0.4 with params .... 0 192 after the address i get 490mh!!!!!!!!!!!!!!!!!!!!!!!!!!
13:44
1080gtx
MH
13:45
Mandel Hoff
πŸ‘
K
13:45
Kenny
hell yea!!!!!!!
13:58
Deleted Account
@Mandelhoff Great work. Do you plan to release linux version of this? As I posted before, I managed to compile older version of your miner in Ubuntu but my work is a bit dirty, because I'm not so skilled at C++/C yet.
14:05
If you will do linux version yourself I think it's far better.
I
14:08
Iridescence
Who is BPWND?
14:08
is that Zack's other pool?
D
14:15
D
In reply to this message
Yes
Deleted joined group by link from Group
14:38
Deleted Account
my node seems to be stuck on block 6011... any suggestions? πŸ™‚
I
15:11
Iridescence
@penfold try make prod-restart
15:25
Deleted Account
thanks, seems to have caught up this time
BB
15:44
Brian Brian
@Mandelhoff catweed I managed to get the newest version compiled on ubuntu following catweeds instructions and fiddling for a couple hours
Deleted joined group by link from Group
I
16:04
Iridescence
I will give a bounty to the person that is able that get my ubuntu cuda miner working on windows lol
16:54
Deleted Account
@Mandelhoff how to specify device ID with the new miner? cant get it to work. It just hangs
G
16:57
Greg
In reply to this message
you just add 0 or 1 after your address, same as before. the SEED seems to be the second last argument
16:58
Deleted Account
In reply to this message
Thanks! Yeah I tried this.. Error Getwork Failed
G
16:58
Greg
wait I will copy paste my command
17:00
for a 1070: AmoveoMinerGpuCuda.exe XXXXXX 0 128 65536 phFn7VUFW9CMX5Ne http://amoveopool.com/work
17:01
Deleted Account
woah thats nothing like what i have lol
G
17:01
Greg
sometimes there is an Error Getwork Failed but its temporary instability
17:01
Deleted Account
0 128 65536 <-- whats this
17:01
?
17:01
phFn7VUFW9CMX5Ne <-- is this your seed?
G
17:03
Greg
In reply to this message
replace XXXX by your address, 0 by device ID, 128 is great for 1070, 65536 is NumBlocks, phFn7VUFW9CMX5Ne is a random seed
17:03
Deleted Account
ok awesome, thank u!
17:03
what would u use for 1080s?
G
17:03
Greg
set different seeds on diffrent devices
17:04
slight improvement on 1070s, more boost on 1060 with this version
17:04
Deleted Account
interesting
17:05
Deleted Account
It seems Mandel's pool is down
17:05
Deleted Account
seems that way yeah
17:06
In reply to this message
what does the 128 mean?
17:06
In reply to this message
yeah
17:06
Deleted Account
oh thanks
17:06
didnt see that
G
17:07
Greg
@Mandelhoff we get 'SubmitWork Exception', web is down too
Z
17:23
Zack
In reply to this message
It would be easy. I have written it before.
2000 blocks is for security reasons. There are lots of ways to attack the timing system.
K
17:57
Kenny
Zack i get security.....its a hard problem https://twitter.com/VladZamfir
17:58
i only copy vlad because i think very highly of his reseach
18:01
would be nice if his ideas were merged into amoveo
18:12
@Mandelhoff looks like your pool has been down for a bit.....just fyi
G
18:22
Greg
In reply to this message
he's spleeping 😁
Z
18:22
Zack
2 am for Mandel
My pool is up
MH
18:54
Mandel Hoff
Zack No one to sync with error? Why please?
18:54
18:55
I'm at block height 0. If I try to sync with a single peer, it says "nothing to sync"
18:55
sync:start([{{159,89,106,253}, 8080}]).
nothing to sync
19:00
peers:all shows only my 52.x address
Z
19:00
Zack
In reply to this message
You probably don't have headers. api:height ().
MH
19:01
Mandel Hoff
Ok, I did clea and restart. Can't find peers to sync
19:01
How do I fix that?
Z
19:02
Zack
In reply to this message
I should probably add my new mining pool to the default list of peers.
MH
19:03
Mandel Hoff
sync:start([{{159,65,120,84}, 8080}]). to your new address also says nothing to sync. I'm at height 0 from doing a clean and restart.
Z
19:05
Zack
I added the new server to the default list of peers.
19:05
In reply to this message
that is concerning. I will try it out
MH
19:06
Mandel Hoff
I crashed in to this state - first time I've seen this.
Z
19:06
Zack
looks like the first peer I attempted to sync with worked fine. they gave me blocks.
,159,65,120,84
MH
19:10
Mandel Hoff
Just did a rm -rf and rebotoed the computer. Fresh clone. Build and go. same problem.
19:11
19:15
Deleted Account
Has this been sorted?
MH
19:15
Mandel Hoff
No
Z
19:16
Zack
looks like every node in the network died.
I kept a copy of the blocks, and am reverifying them from my hard drive.
MH
19:16
Mandel Hoff
πŸ‘
19:17
Thank you for the update. I''ll try to stand by. I have to run soon for a while though, so I'm only around another 15 min or so. Early jobs today
Z
19:20
Zack
Having more nodes in the default list would make this less likely to happen.
Even more importantly, we should solve whatever is making them crash.
19:21
I got one of my nodes up, the second is syncing.
OK
19:22
O K
Mine crashed overnight too
19:23
Syncing
19:23
Are there any log files that would be helpful?
Z
19:24
Zack
possibly _build/prod/rel/amoveo_core/log/crash.log
MH
19:24
Mandel Hoff
What node is up? I'm still getting no peers and nothing to sync when I give both your node addresses
Z
19:25
Zack
In reply to this message
both my nodes are up and synced.
Maybe reboot to repopulate your peers list.
MH
19:25
Mandel Hoff
Reboot computer or make prod-restart?
Z
19:26
Zack
make prod-restart
19:26
I don't think I have ever had a situation where rebooting the computer would help
MH
19:26
Mandel Hoff
Peers all still shows only me
OK
19:27
O K
I had to do
Z
19:27
Zack
I have had to do ./scripts/kill_all_erlang.sh
OK
19:27
O K
19:27
I did the sync:get_headers and api:txs and another sync:start
19:27
That got me moving
MH
19:28
Mandel Hoff
Yes, api off, halt, kill all erlang, git pull, make prod clean and restart, no peers
19:30
OK
19:30
O K
My node is synced at 159,65,173,9
Z
19:30
Zack
did you follow the direction in sync.md?
It sounds like you skipped the step where you verify that headers are downloaded.
MH
19:30
Mandel Hoff
don't have a parent for this header - is that a thing?
MH
19:30
Mandel Hoff
Ok, got it syncing. I got angry and did kill all erlang like 20 times and that worked... after a restart
19:34
Deleted Account
Is amoveopool.com down until all is synced?
OK
19:34
O K
Thanks for cutting it so close mandel
Z
19:35
Zack
looks like Mandel synced the blocks
OK
19:36
O K
Miners still offline
Z
19:36
Zack
my pool is working
19:37
I will be gone for the next hour or two. good luck.
OK
19:37
O K
Thank you
19:37
Deleted Account
In reply to this message
Do you have an executable?
OK
19:39
O K
Thanks Mandel!!
MH
19:40
Mandel Hoff
Ok, my pool is up
OK
19:41
O K
The new miner has dupe nonce issue
19:41
For later
19:41
Doesn't ask for random string
MH
19:41
Mandel Hoff
Set the seed on the command line now. It's an argument instead of input
OK
19:41
O K
Ooh
19:41
Thanks again
MH
19:41
Mandel Hoff
It was requested for easier scripting. πŸ‘
OK
19:43
O K
It doesn't say where the seed input belongs on the GitHub though
19:44
In the advanced usage template
MH
19:47
Mandel Hoff
Right. It's after the block size. Run the tool with no args for the template. It shows there.
19:47
Sorry finished the new miner just a few hours ago and was really tired. Forgot to update docs.
OK
19:50
O K
Thank you :)
19:57
Deleted Account
In reply to this message
Thank you for ur extraordinary job on maintain your pool. I finally got one veo today!bb
M
20:08
Minieep21
Anyone having issues with Mandel's new CUDA miner?
20:12
Nevermind, seems to be working
Z
20:22
Zack
Blocks propagate between nodes pretty fast now.

The bottleneck is that it takes around 80 seconds for the mining pool to switch to working on the new problem.

I will work on this soon.
MF
20:38
Mr Flintstone
In reply to this message
has the 80 seconds always been an issue, or did this just pop up?
MH
20:39
Mandel Hoff
Yes that's pretty significant barrier to competition with the block times already very low.
MF
20:43
Mr Flintstone
In reply to this message
so I assume this is new
20:43
In reply to this message
Block times are already low again?
MH
20:43
Mandel Hoff
I don't think it's new.
MF
20:45
Mr Flintstone
lol it’s probably ev positive to manually see what pool found a block, then switch hash rate to that pool...
Z
21:02
Zack
In reply to this message
It didn't matter before because it wasn't the bottleneck. It has always been an issue.
Ryuu joined group by link from Group
MH
21:42
Mandel Hoff
Yes blocks times are still almost never over 5 min
21:42
Pretty fast. Many are still about 1min.
Z
21:43
Zack
I think that the network would be doing a lot more work if it wasn't for this bottleneck.
Evans Pan joined group by link from Group
MH
21:59
Mandel Hoff
Pool news:
I have 31 Veo in payments being auto-reprocessed. Txs are still getting lost by the network. Many have been sent 9 or 10 times. They will continue to be resent until they get confirmed.
21:59
22:00
Deleted Account
any Chinese there?
G
22:02
Gonzalo
my pool is up and sync'd: 51.15.75.100:8085
22:03
@Mandelhoff great improvement on your GPU miner, but I still need the 192 param to see more h/s
MH
22:04
Mandel Hoff
Ok, 256 was much better on my 750ti, 1050 and 1060. My K80 tests were better with 512.
22:05
Yes, do try them all and find the best fit. Yes, I'm very happy with this speed boost.
G
22:06
Gonzalo
not on the 1080ti Im testing:
v1003 with 192 = 550mhs
v1004 with 256 = 550mhs
v1004 with 192 = 775 - 810 mhs
22:07
really nice @Mandelhoff , well done! πŸ‘
MH
22:07
Mandel Hoff
Thanks. I'm very happy with it even if it is "a little too late".
G
22:09
Gonzalo
I dont care, Im here 30% for the money, 70% for fun 😁
MH
22:09
Mandel Hoff
I'm probably just hooked on the adrenaline rush at this point. Pure Addiction.
22:09
Deleted Account
Is there a way to mine veo for Chinese people ? I want to join the mining
RL
22:10
R L
No very different for Chinese people
22:10
Deleted Account
But my English is so poor....
MH
22:10
Mandel Hoff
RL
22:10
R L
I am just kidding
22:12
Deleted Account
In reply to this message
Thanks for your help!
MH
22:14
Mandel Hoff
In reply to this message
Wow - your 1080ti is double my 1060. Incredible. I wonder what this would do on the P100 and V100s. I don't have access to those though. Would love to get some numbers.
22:21
Deleted Account
In reply to this message
the 1.0.04 doesn't work fast than 1.0.0.3 with my 1050ti
MH
22:24
Mandel Hoff
Maybe try setting Numblocks to 192. Default is now 256, which was better for nearly all my tests. Gonz reports 192 is better on his 1080ti, so maybe it's a "TI" model thing.
G
22:24
Gonzalo
In reply to this message
have you tested numblocks 192? v1003 default was 192, v1004 default is 256 again I think
MH
22:24
Mandel Hoff
πŸ‘
G
22:24
Gonzalo
:)
OK
22:32
O K
192 is still better for me on 1080, 1070
MH
22:34
Mandel Hoff
πŸ‘ Thank you for checking. I think leaving that setting configurable is probably the best solution now. I'll try mine again, but I'm pretty sure 256 was better.
OK
22:34
O K
Which cards are you using again Mandel?
22:34
I don't have Ti
MH
22:36
Mandel Hoff
I have a gtx1060 6gb, a gtx1050 (not TI), a 750TI, and I test K80s on Azure - but I don't run them long. They are expensive junk for mining, but fair to test. I've been in ML for a long time, so I'm comfortable with the K80s. Their 12GB memory is great for ML on large data.
OK
22:37
O K
Cool
WL
22:38
Wilson Lau
256 runs better on my 1060, 345mh vs 365mh
MH
22:38
Mandel Hoff
πŸ‘
OK
22:40
O K
my cards vary way too much to recognize a difference of 20mh/s
22:40
over time
G
22:43
Gonzalo
I dont understand why nobody wants to test my mining pool, with 0% fees for the first 10 blocks and then 2% fees. Maybe is because Kassel is the only one mining at Zacks pool, and so he's the only one mining in a non shared pool?.
MH
22:43
Mandel Hoff
Also interesting - I tested the miner without the hash2integer function that scores solutions. I thought it was likely "slow", but it's not slow. "Not scoring hashes" gives like a 1% speedup or less. I was very surprised by that.
OK
22:44
O K
I use Mandel's pool because it's a true shared pool, and he releases his code for free to the public. Gonzalo you basically run a mining node. Whose miner would we use to connect to your pool?
22:45
The only other GPU miner available is Iri's, which is harder to get working, it's deliberately down throttled, and he charges a premium to unthrottle it, only so we can connect it to an unshared pool
MH
22:45
Mandel Hoff
Browser Miner πŸ˜†
OK
22:45
O K
😐
22:45
lol
G
22:45
Gonzalo
In reply to this message
right, that's what I believe... so why is Kassel mining on Zacks pool at 20% fees??
OK
22:46
O K
I don't think Zack's fees are that high, are they?
MH
22:46
Mandel Hoff
Yes, he's at 20%
P|
22:46
Paul N. | BAND (πŸ…‘) - I'll never DM you
In reply to this message
Yes
OK
22:46
O K
To some extent he is probably scratching Zack's back
22:46
I don't mean that negatively
22:46
but it's a strategy
G
22:46
Gonzalo
yes, thats default fees on zack pool
OK
22:46
O K
plus Zack's node historically has been up to date
22:47
and online
22:47
He's had the most hashpower during various periods, so no stale blocks there
G
22:47
Gonzalo
In reply to this message
yes, probably, but thats because hes the first one getting the updates ;)
OK
22:48
O K
I meant in sync
G
22:49
Gonzalo
maybe its not a 100% fair situation ??
OK
22:49
O K
Can you give me an example of a 100% fair situation?
G
22:49
Gonzalo
In reply to this message
NO
OK
22:49
O K
lol
G
22:49
Gonzalo
haha
Z
22:49
Zack
In reply to this message
Thanks for sharing. I guess it is too insignificant to bother fixing.
MH
22:56
Mandel Hoff
Another funny thing is that when I "optimize" my cuda kernal such that is breaks, it makes my hash rates look really fast but it's not doing valid work. The kernel just fails to launch and exits without error or crashing. I run in to this often, but my pool validates the partial work and tells me when I'm making bad solutions.
Z
22:56
Zack
In reply to this message
Yes. I got confused by this same thing.
I
23:03
Iridescence
In reply to this message
I would really like to get a shared pool up and running unfortunately I only have so c much spare time for Amoveo
Z
23:05
Zack
me too. I already dedicated all my spare time to other parts of Amoveo.
I will make a shared pool eventually, if no one else does first.
For now I just want to stop my pool from blocking everyone else from mining.
23:09
hashrate looks like 2.4 th/s
23:17
ok, I pushed a fix. Hopefully the mining pool is working on the most recent block now.
23:18
it passed some tests, but it is hard to know how it will behave when live
23:18
if Mandel's pool starts getting significantly more blocks, then that means it worked.
MH
23:24
Mandel Hoff
πŸ‘
23:24
Just a pool git pull then? Any node changes?
Z
23:25
Zack
In reply to this message
I only modified the mining pool, not amoveo.
23:25
I think you don't use the mining pool I wrote.
23:25
And yours is closed source, so I can not check if you have the same problem.
MH
23:25
Mandel Hoff
Yes, I do call it's endpoints just like a mining tool would.
23:26
Updated.
Z
23:26
Zack
my mining pool had a problem with how it was storing data. the api for talking to the full node is unchanged.
MH
23:27
Mandel Hoff
I see "mining data block hash is...." much faster on my node now. I think I'm observing the updates.
OK
23:27
O K
If the mining node software is the issue, couldn't someone maliciously use bad mining node software to attack the network/ other miners?
Z
23:27
Zack
so my mining pool is embedded in yours somehow?
23:28
In reply to this message
It only works if one miner has >40% or so. I don't know the exact number, but I heard it is less than 51%.
23:29
I suspect in practice it is 51%, because we can talk to each other and react to problems.