[ANN][XCN] Cryptonite | 1st mini-blockchain coin | M7 PoW | No Premine - page 122.

dboylc

full member

Activity: 153

Merit: 100

mine for future~

Quote from: catia on August 07, 2014, 10:08:11 PM

New miner up. Getting 3970kh/s on some X16 connected 750ti. Sufferage on X1 to only about 3MH. Can probably reduce memory transfer with some effort to do 4 on even X1. Gonna hold off on that until the wolf gives us run though.

compiled but cannot run minerd.
only can run minerd --help
:::::::::::::
./minerd -a m7 -o stratum+tcp://xcnpool.1gh.com:7333 -u CPtc1bSfyVLVRXna1yf8eJD6oZzxLwJBbT -p x -t 6
Try `minerd --help' for more information.

./minerd -V
cpuminer 2.4
built on Aug 8 2014
features: x86_64 SSE2 AVX AVX2 XOP
libcurl/7.35.0 OpenSSL/1.0.1f zlib/1.2.8 libidn/1.28 librtmp/2.3
libjansson 2.5

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: bitfreak! on August 07, 2014, 09:52:30 PM

At the end of the day you are right, we probably should have went with an existing PoW algorithm instead of creating a new one, but at the time we had good reasons for doing it that way and it seemed like a good idea. But it was a sub-optimal design decision, and if I could turn back time the one thing I would change is the PoW algorithm. But it's done now and we can't change it, so I don't know why you have to rant about it for 50 pages. I think we have got the point by now, you don't like new PoW algorithms. It's not like it's the end of the world, the long mining period will do much to offset any big miners in these early stages, and the private GPU miners will be made publicly available soon enough. All new PoW algorithms take time to be optimized, and complaining about it isn't going to get it done.

I think you're off base here.

I haven't ranted on it for 50 pages. Since we originally discussed it I mentioned it once or twice. In this context it had to do with the need to optimize the one you have, in fact I was encouraging the catia's efforts, and if you read between the lines of my post I was offering to donate to support that effort.

catia

full member

Activity: 288

Merit: 105

New miner up. Getting 3970kh/s on some X16 connected 750ti. Sufferage on X1 to only about 3MH. Can probably reduce memory transfer with some effort to do 4 on even X1. Gonna hold off on that until the wolf gives us run though.

bitfreak!

legendary

Activity: 1536

Merit: 1000

electronic [r]evolution

Quote from: smooth on August 07, 2014, 09:14:56 PM

Concentrated mining by people with more efficient miners is bad for the coin, especially early on. That's one reason why I slammed your use of an immature PoW instead of one that is already well optimized.

At the end of the day you are right, we probably should have went with an existing PoW algorithm instead of creating a new one, but at the time we had good reasons for doing it that way and it seemed like a good idea. But it was a sub-optimal design decision, and if I could turn back time the one thing I would change is the PoW algorithm. But it's done now and we can't change it, so I don't know why you have to rant about it for 50 pages. I think we have got the point by now, you don't like new PoW algorithms. It's not like it's the end of the world, the long mining period will do much to offset any big miners in these early stages, and the private GPU miners will be made publicly available soon enough. All new PoW algorithms take time to be optimized, and complaining about it isn't going to get it done.

Wolf0

member

Activity: 81

Merit: 1002

It was only the wind.

Quote from: dboylc on August 07, 2014, 02:39:36 AM

Quote from: Wolf0 on August 07, 2014, 02:12:56 AM

Quote from: dboylc on August 07, 2014, 02:09:28 AM

Quote from: Wolf0 on August 07, 2014, 01:58:32 AM

It seems some people are blindly going to AWS without doing the math... they're gonna lose money.

The price now is too low.

Not yet.

$0.788 for ec2.instance per hour and only get 3 coins,
even if mining @home use 750Ti, it's not profitable compare to X11 multipool.

Dude, spot instances! You can get a g2.2 for $0.1/hr!

Also, 750Ti numbers:

Hash of one 750Ti @ stock / network hash * blocks per day * XCN per block = XCN per day per 750Ti @ stock:

3100000 / 6016040021 * 1440 * 242 = ~180 XCN/day

As of this moment, 1 XCN is worth 0.0000272 BTC - making one 750Ti bring in ~0.005 BTC per day. It's not excellent, but it's not horrid, either.

PeaMine

hero member

Activity: 979

Merit: 510

Quote from: skunk on August 07, 2014, 09:10:51 PM

Quote from: catia on August 07, 2014, 08:45:13 PM

Also I'd like to gripe a little about miner donations. Let me count them for you all. 1,2,3,4, no wait, no donate, none

i've sent you 100 coins (24h worth of desktop mining)
hopefully the ones with big irons will follow the example Wink

301.02 XCN sent from my CPU mining adventures, hopefully someone releases a Windows GPU binary, as I am apparently not yet skilled enough for compiling.
Holding the rest as it seems price is actually going up! ever since the GPU was publicly released, can't explain that! Bought some more with BTC and then sold for 5% profit, so a nice spread for investors as well.
Though I tried this earlier and bought in at 7000...so...yeah...still paying for that lol.

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: catia on August 07, 2014, 08:12:37 PM

Quote from: djm34 on August 07, 2014, 08:07:51 PM

Quote from: catia on August 07, 2014, 07:56:17 PM

Quote from: mmmaybe on August 07, 2014, 07:54:03 PM

Quote from: catia on August 07, 2014, 07:27:25 PM

Catia now delivers for you 2.5MH of raw speed on 750ti. For free as per usual. Now let those complaints roll!

Great

Any chance this will compile on windows? I saw "Win32 work and 2.5MH" Huh

I've put a solid 24 hours into getting it to compile on windows. Unknown even if any real progress was made. That toolchain is so fubar.

any info on how to compile on window ?

Probably the best approach is too add support for 1gh pool to ccminer. Getting cpu-miner codes to compile under MSVC is a nightmare. Probably won't work. So far the extent of progress is getting cpu-miner.c to compile. Whether we can get the rest of the codes to compile, or eventually link with some magic libcurl and libpthread which is compiled with msvc and we don't have is a serious unknown.

for some unknown reason, I get only rejected with ccminer (both 1gh and testnet).
The share/block are compared to cpu it matches but got rejected (or stale...)

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: catia on August 07, 2014, 08:59:50 PM

Quote from: smooth on August 07, 2014, 08:53:59 PM

So you (or not necessarily you personally but anyone who is interested in seeing this PoW succeed) optimize the other hashes too.

Wolf0 doesn't have some magic secret hash code optimizing machine delivered him by aliens from the future. He's just tweaking code. He's not the only one who can tweak code.

Yeah there is one other person, me. There's like 10000 things to optimize and he will always be one step ahead. He will always stand on the sholders of the OSS miner. My only goal is to make it so he has to do huge amounts of work for small gain. If rejects were fixed he's only got 20% on us and I have some ideas to get that down.

That's what I call parity. Once you pick the low hanging fruit, Amdahl's Law kicks in and yeah there might be 10000 things to optimize but reducing each to zero only buys you 0.01%. Plus there is nothing to prevent other people who want to see this coin succeed from contributing tweaks to the open source as well.

Concentrated mining by people with more efficient miners is bad for the coin, especially early on. That's one reason why I slammed your use of an immature PoW instead of one that is already well optimized. But if you (meaning anyone who wants to see the PoW succeed) are willing to put in the effort and get to parity, you can make it work.

BTW, you can't be the only one in the world trying to implement fast open source hashing on GPUs right?

BTW #2, I can't even find where to donate for the miner work. If I can't find it I'm sure some others can't find it either. Nothing in the README for example.

skunk

sr. member

Activity: 329

Merit: 250

Quote from: catia on August 07, 2014, 08:45:13 PM

Also I'd like to gripe a little about miner donations. Let me count them for you all. 1,2,3,4, no wait, no donate, none

i've sent you 100 coins (24h worth of desktop mining)
hopefully the ones with big irons will follow the example Wink

billotronic

legendary

Activity: 1610

Merit: 1000

Crackpot Idealist

There IS something to be said for healthy competition....

It IS fucking sad that more folks are not stepping up. Catia's time would be much better spent working on the client and not a miner.

Wolf0

member

Activity: 81

Merit: 1002

It was only the wind.

Quote from: dboylc on August 07, 2014, 02:09:28 AM

Quote from: Wolf0 on August 07, 2014, 01:58:32 AM

It seems some people are blindly going to AWS without doing the math... they're gonna lose money.

The price now is too low.

Not yet.

catia

full member

Activity: 288

Merit: 105

Quote from: smooth on August 07, 2014, 08:53:59 PM

So you (or not necessarily you personally but anyone who is interested in seeing this PoW succeed) optimize the other hashes too.

Wolf0 doesn't have some magic secret hash code optimizing machine delivered him by aliens from the future. He's just tweaking code. He's not the only one who can tweak code.

Yeah there is one other person, me. There's like 10000 things to optimize and he will always be one step ahead. He will always stand on the sholders of the OSS miner. My only goal is to make it so he has to do huge amounts of work for small gain. If rejects were fixed he's only got 20% on us and I have some ideas to get that down.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: catia on August 07, 2014, 08:45:13 PM

Parity is just not possible.

Huh? Low self esteem or something?

Quote

If I develop a super ultimate sha512 for example that brings us up to same performance as Wolf. Then Wolf will just use it, along with his optimized other hashes and once again be better. Wolf will always be faster because he has my optimizations plus his own.

So you (or not necessarily you personally but anyone who is interested in seeing this PoW succeed) optimize the other hashes too.

Wolf0 doesn't have some magic secret hash code optimizing machine delivered him by aliens from the future. He's just tweaking code. He's not the only one who can tweak code.

catia

full member

Activity: 288

Merit: 105

Quote from: smooth on August 07, 2014, 08:36:01 PM

Quote from: catia on August 07, 2014, 05:56:05 PM

New gpu miner code released with 50% speed up. Unfortunately this is an arms race we can't really win. Anything I do Wolf will take to improve his miner. Best we can hope for is to make improvements he already has.

The way you "win" is by getting to parity (where neither can significantly beat the other) and there a level playing field so mining is maximally decentralized instead of concentrated with the one or two people with the best miners.

Parity is just not possible. If I develop a super ultimate sha512 for example that brings us up to same performance as Wolf. Then Wolf will just use it, along with his optimized other hashes and once again be better. Wolf will always be faster because he has my optimizations plus his own.

What would be nice is getting rid of rejects. For some reason our client is losing a lot of hash and wolf isn't. We could much faster if that were fixed. This is not cuda problem, but some issue in 1gh miner. Potentially caused by my screwing with it to try and get it to compile on windows. Windows kills even linux miner :/

Also I'd like to gripe a little about miner donations. Let me count them for you all. 1,2,3,4, no wait, no donate, none

Wolf0

member

Activity: 81

Merit: 1002

It was only the wind.

It seems some people are blindly going to AWS without doing the math... they're gonna lose money.

smooth

legendary

Activity: 2968

Merit: 1198

Quote from: catia on August 07, 2014, 05:56:05 PM

New gpu miner code released with 50% speed up. Unfortunately this is an arms race we can't really win. Anything I do Wolf will take to improve his miner. Best we can hope for is to make improvements he already has.

The way you "win" is by getting to parity (where neither can significantly beat the other) and there a level playing field so mining is maximally decentralized instead of concentrated with the one or two people with the best miners.

Wolf0

member

Activity: 81

Merit: 1002

It was only the wind.

Quote from: wyj1888 on August 07, 2014, 01:29:19 AM

Who can tell me the AMD graphics tools in that? THANKS VERY MUCH!

There are none available yet.

cubex

newbie

Activity: 50

Merit: 0

Quote from: catia on August 07, 2014, 05:56:05 PM

New gpu miner code released with 50% speed up. Unfortunately this is an arms race we can't really win. Anything I do Wolf will take to improve his miner. Best we can hope for is to make improvements he already has.

I am trying the latest gpu code. I am getting about 1.7-1.8mh (after 30mins). It is much better and we are back in the game. Thanks for your effort!

catia

full member

Activity: 288

Merit: 105

Quote from: djm34 on August 07, 2014, 08:07:51 PM

Quote from: catia on August 07, 2014, 07:56:17 PM

Quote from: mmmaybe on August 07, 2014, 07:54:03 PM

Quote from: catia on August 07, 2014, 07:27:25 PM

Catia now delivers for you 2.5MH of raw speed on 750ti. For free as per usual. Now let those complaints roll!

Great

Any chance this will compile on windows? I saw "Win32 work and 2.5MH" Huh

I've put a solid 24 hours into getting it to compile on windows. Unknown even if any real progress was made. That toolchain is so fubar.

any info on how to compile on window ?

Probably the best approach is too add support for 1gh pool to ccminer. Getting cpu-miner codes to compile under MSVC is a nightmare. Probably won't work. So far the extent of progress is getting cpu-miner.c to compile. Whether we can get the rest of the codes to compile, or eventually link with some magic libcurl and libpthread which is compiled with msvc and we don't have is a serious unknown.

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: catia on August 07, 2014, 07:56:17 PM

Quote from: mmmaybe on August 07, 2014, 07:54:03 PM

Quote from: catia on August 07, 2014, 07:27:25 PM

Catia now delivers for you 2.5MH of raw speed on 750ti. For free as per usual. Now let those complaints roll!

Great

Any chance this will compile on windows? I saw "Win32 work and 2.5MH" Huh

I've put a solid 24 hours into getting it to compile on windows. Unknown even if any real progress was made. That toolchain is so fubar.

any info on how to compile on window ?

Topic: [ANN][XCN] Cryptonite | 1st mini-blockchain coin | M7 PoW | No Premine - page 122. (Read 578501 times)