Author

Topic: [ANN][YAC] YACoin ongoing development - page 150. (Read 379983 times)

sr. member
Activity: 425
Merit: 262
June 13, 2013, 06:18:44 AM
the cpu ratio is closer to 2:1 in terms of kh/s for ltc vs yac. for cpu its more profitable to mine yac and for gpu its even at current prices.

In the long run and if everybody did the same GPU mining LTC and trading for YACs would end being more profitable, as price would rise instead of increasing the supply.

Don't forget that YAC block reward decreases with difficulty going up. More hashrate means YAC price must go up else mining becomes unprofitable.

[BOUNTY]

Anyway, is there someone skilled enough to add coin control to YACoin client? More info here > https://bitcointalksearch.org/topic/yet-another-coin-control-release-closed-144331
I'll pay 1 BTC for the job which must include source code and working Win32 and 64 bit binaries. Others are welcomed to add some coins to bounty.

Hah, this is attractive. I'll see if I have spare time after tomorrow. I will take it if there's no one else accept this request. ;-)

But what's used for? I don't know why it's worth for 1BTC bounty.

OK, I got it. But now I'm wondering why not put this coin control directly in the yacoin-qt client?
sr. member
Activity: 425
Merit: 262
June 13, 2013, 04:30:53 AM
maybe time for a YAC pool with stratum now the gpu miner is out, getting quite a few stales

I was trying to modify eloipool to work with scrypt-jane but I've got segmentation fault errors and I don't know how to debug it (I am not skilled at all in python). In my opinion problem with high stale rate is because for some reason (scrypt-jane dependant calculations?) both pushpool and yacoind are very hungry for resources - CPU usage in average ~10 times higher than other coins. Even if my server is 8 core with 24 GB RAM it creates some delays internally which leads to stales...

I find each time it hashes, it will allocate the memory buffer (call scrypt_alloc) and scrypt_free at the end of hash. Will it impact the performance when there're so many threads processing? How about make this memory buffer becomes thread-local to reduce these memory re-allocation?

Otherwise, I can't find more difference of it between the yacoin & other scrypt coins.

If you are asking me - I have no idea how to do this Smiley but I can easily test it if you can provide example.

I made the change of code by using boost's thread-local facility. Here is the patch against the current git: https://github.com/yacoin/yacoin
http://sourceforge.net/projects/hnindev/files/yacoin_mem_opt_patch.zip

I've locally built the client, it works no error.
legendary
Activity: 1197
Merit: 1000
June 13, 2013, 01:19:00 AM
maybe time for a YAC pool with stratum now the gpu miner is out, getting quite a few stales

I was trying to modify eloipool to work with scrypt-jane but I've got segmentation fault errors and I don't know how to debug it (I am not skilled at all in python). In my opinion problem with high stale rate is because for some reason (scrypt-jane dependant calculations?) both pushpool and yacoind are very hungry for resources - CPU usage in average ~10 times higher than other coins. Even if my server is 8 core with 24 GB RAM it creates some delays internally which leads to stales...

I find each time it hashes, it will allocate the memory buffer (call scrypt_alloc) and scrypt_free at the end of hash. Will it impact the performance when there're so many threads processing? How about make this memory buffer becomes thread-local to reduce these memory re-allocation?

Otherwise, I can't find more difference of it between the yacoin & other scrypt coins.

If you are asking me - I have no idea how to do this Smiley but I can easily test it if you can provide example.
sr. member
Activity: 425
Merit: 262
June 13, 2013, 01:10:48 AM
maybe time for a YAC pool with stratum now the gpu miner is out, getting quite a few stales

I was trying to modify eloipool to work with scrypt-jane but I've got segmentation fault errors and I don't know how to debug it (I am not skilled at all in python). In my opinion problem with high stale rate is because for some reason (scrypt-jane dependant calculations?) both pushpool and yacoind are very hungry for resources - CPU usage in average ~10 times higher than other coins. Even if my server is 8 core with 24 GB RAM it creates some delays internally which leads to stales...

I find each time it hashes, it will allocate the memory buffer (call scrypt_alloc) and scrypt_free at the end of hash. Will it impact the performance when there're so many threads processing? How about make this memory buffer becomes thread-local to reduce these memory re-allocation?

Otherwise, I can't find more difference of it between the yacoin & other scrypt coins.
hero member
Activity: 802
Merit: 1003
GCVMMWH
June 13, 2013, 12:28:01 AM
What I think we need next TBH is a proper mining program with stratum support. That way we can capitalize on the hashrate we still have left and get better feedback and less stale shares.

I looked at this very briefly a few days ago and it looks like it could be adapted to YACoin if someone has time to spend a few hours on it    https://github.com/CryptoManiac/stratum-mining
legendary
Activity: 1197
Merit: 1000
June 13, 2013, 12:26:38 AM
maybe time for a YAC pool with stratum now the gpu miner is out, getting quite a few stales

I was trying to modify eloipool to work with scrypt-jane but I've got segmentation fault errors and I don't know how to debug it (I am not skilled at all in python). In my opinion problem with high stale rate is because for some reason (scrypt-jane dependant calculations?) both pushpool and yacoind are very hungry for resources - CPU usage in average ~10 times higher than other coins. Even if my server is 8 core with 24 GB RAM it creates some delays internally which leads to stales...
sr. member
Activity: 252
Merit: 250
June 12, 2013, 07:37:22 PM
maybe time for a YAC pool with stratum now the gpu miner is out, getting quite a few stales
member
Activity: 81
Merit: 10
June 12, 2013, 06:33:07 PM
GPU miner released, heh..
I'm not so hyped about it.. YAC isn't special anymore, right?

Well, hanzac's GPU miner doesn't perform especially well compared with mtrlt's normal scrypt+salsa kernel for mining other scrypt coins on the same hardware.  Hanzac's OpenCL kernel was already mostly released way way back in this thread, he just needed to fix a few bugs.

I tested it long enough to mine about 5 blocks of YAC then switched my GPU farm back to mining LTC.
GPU farm? You have a CPU farm too, right?
What's your job?  Grin

Perhaps farming?  Smiley
sr. member
Activity: 406
Merit: 250
The cryptocoin watcher
June 12, 2013, 05:12:34 PM
the cpu ratio is closer to 2:1 in terms of kh/s for ltc vs yac. for cpu its more profitable to mine yac and for gpu its even at current prices.

In the long run and if everybody did the same GPU mining LTC and trading for YACs would end being more profitable, as price would rise instead of increasing the supply.
member
Activity: 75
Merit: 10
June 12, 2013, 04:45:07 PM
I can do 300kh/s for ltc and 90 kh/s for yac with Hanzac's code.

thats .5 LTC versus 27 yacoin at the current difficulty of 0.76. it's about even price-wise.

the cpu ratio is closer to 2:1 in terms of kh/s for ltc vs yac. for cpu its more profitable to mine yac and for gpu its even at current prices.
newbie
Activity: 56
Merit: 0
June 12, 2013, 01:51:34 PM
A public GPU miner is only going to add miners to YAC, which, in my opinion is a good thing. I have a few GPU rigs and ALL of them have the absolute worst, yet most power efficient CPU's - so, not viable for YAC mining. Now that I can point some of them at YACoin, I can get back in the game. I'm guessing it's the same for many. Also, as WM already mentioned, the performance difference isn't like any other scrypt implementations - so for those that have good CPU's, mining is also still viable for them.  

I agree with this, especially since N is at a high enough level now so that GPU miners don't have THAT much of an advantage over CPU miners. I'm still solo-mining on my i7-2700k just fine.

Agree and disagree - I know there were people who were able to do their own OpenCL implementation and get them running ahead of one being publicly released.... However, the GPU in my laptop (7850M) is running at roughly 4x the rate of the CPU (i7 3610M) - that's quite a bit of an advantage!

In comparing YAC on GPU to BTC on GPU, It's looking to me that the newer cards (7xxx) do great at this algorithm, while older cards (5870, I'm looking at you), aren't so good, even though they may crush in BTC hashrate...

7850 -> 140 KH/sec YAC, 310 MH/sec BTC
5870 -> 100 KH/sec YAC, 400 MH/sec BTC



Are you running Techinfernos Throttlestop its great for disabling the boost throttling on laptops

cause as soon as you load it up the boost gets throttled instantly down as you may well know already. But i get your point too i just had to add that in there, that program is the bomb, it really brought a m11x r2 to life that i had
hero member
Activity: 693
Merit: 500
June 12, 2013, 01:34:03 PM
A public GPU miner is only going to add miners to YAC, which, in my opinion is a good thing. I have a few GPU rigs and ALL of them have the absolute worst, yet most power efficient CPU's - so, not viable for YAC mining. Now that I can point some of them at YACoin, I can get back in the game. I'm guessing it's the same for many. Also, as WM already mentioned, the performance difference isn't like any other scrypt implementations - so for those that have good CPU's, mining is also still viable for them.  

I agree with this, especially since N is at a high enough level now so that GPU miners don't have THAT much of an advantage over CPU miners. I'm still solo-mining on my i7-2700k just fine.

Agree and disagree - I know there were people who were able to do their own OpenCL implementation and get them running ahead of one being publicly released.... However, the GPU in my laptop (7850M) is running at roughly 4x the rate of the CPU (i7 3610M) - that's quite a bit of an advantage!

In comparing YAC on GPU to BTC on GPU, It's looking to me that the newer cards (7xxx) do great at this algorithm, while older cards (5870, I'm looking at you), aren't so good, even though they may crush in BTC hashrate...

7850 -> 140 KH/sec YAC, 310 MH/sec BTC
5870 -> 100 KH/sec YAC, 400 MH/sec BTC
newbie
Activity: 56
Merit: 0
June 12, 2013, 01:03:26 PM
i have a problem sending funds

when i do
Code:
yacoind sendtoaddress xxxxxxx xx
i get
Code:
error: {"code":-4,"message":"Error: Transaction creation failed  "}

thanks in advanced

Ill try and help although ive never used yacoind to send funds

Firstly make sure your yacoind is synced first before you try sending funds type yacoind getinfo to check your balence to make sure you have funds

then yacoind getmininginfo to see what block your daemon is synced to as of right now mine says  88886

if its not synced then it wont let you send funds if it is synced let me know if that helps

i understand your running your own p2pool node i have one also but havent sent funds from it yet

and im not sure if you are able to send while its running with python as ive never tried
yes it is synced up, and yes i am runing my own p2pool node.

i turned it off and tried to send payment. no good.

try restarting it with the -rescan command it could be a number of things any chance you have looked at the debug log
maybe there are too many inputs also are you accounting for the transaction fee its hard to help if we i cant see any logs
because that error would come up for a number of reasons you can try running it with the command  yacoind -rescan -keypool 1000

Break the amount you want to send into few smaller ones. It is very likely that you are attempting to send so much inputs that YACoin client
refuses to create such transaction due to transaction size and / or fee limits.

As far as I know, -keypool value works only when making new wallet. It specifies the number of keys wallet is to keep as reserve at all times
and it is useful only if person is not making wallet backups frequently but regulary and frequently creates many new addresses for receiving coins.
If both criteria are fulfilled but person uses default keypool of 100, it can happen that when rolling back to old backup some coins will be missing
due to keypool of old wallet being exhausted, e.g. newer wallet "knows" more used keys than older one.

Thats correct its for creating reserve addresses I was trying to get him to throw random ideas at the problem notice how i didnt format -keypool 1000 properly
that was also a work of deductive arts Smiley Sherlock would be so proud, but thanks mate
hero member
Activity: 588
Merit: 500
June 12, 2013, 12:07:11 PM
A public GPU miner is only going to add miners to YAC, which, in my opinion is a good thing. I have a few GPU rigs and ALL of them have the absolute worst, yet most power efficient CPU's - so, not viable for YAC mining. Now that I can point some of them at YACoin, I can get back in the game. I'm guessing it's the same for many. Also, as WM already mentioned, the performance difference isn't like any other scrypt implementations - so for those that have good CPU's, mining is also still viable for them.  

I agree with this, especially since N is at a high enough level now so that GPU miners don't have THAT much of an advantage over CPU miners. I'm still solo-mining on my i7-2700k just fine.
sr. member
Activity: 406
Merit: 250
One does not simply mine Bitcoins
June 12, 2013, 10:15:10 AM
Also, who has java enabled in their browser these days?

Oh, everyone? At least if you live in a country like I do, where all public, governmental and online banking sites are secured with a centralized java applet. Disable java and you have no way to interact with the government or your bank.

- Michael

Hah, gave me a good laugh, thanks! Cheesy
hero member
Activity: 802
Merit: 1003
GCVMMWH
June 12, 2013, 08:43:38 AM
A public GPU miner is only going to add miners to YAC, which, in my opinion is a good thing. I have a few GPU rigs and ALL of them have the absolute worst, yet most power efficient CPU's - so, not viable for YAC mining. Now that I can point some of them at YACoin, I can get back in the game. I'm guessing it's the same for many. Also, as WM already mentioned, the performance difference isn't like any other scrypt implementations - so for those that have good CPU's, mining is also still viable for them.  
member
Activity: 117
Merit: 10
June 12, 2013, 08:37:33 AM
Also, who has java enabled in their browser these days?

Oh, everyone? At least if you live in a country like I do, where all public, governmental and online banking sites are secured with a centralized java applet. Disable java and you have no way to interact with the government or your bank.

- Michael
newbie
Activity: 56
Merit: 0
June 12, 2013, 06:32:20 AM
Dont make me be the first person to post here >>>http://yacointalk.com/index.php/topic,273.0.html
hero member
Activity: 637
Merit: 500
June 12, 2013, 06:11:32 AM
GPU farm? You have a CPU farm too, right?
What's your job?  Grin
OMG don't ask obvious questions, if he has a farm, he is obviously a farmer  Grin
sr. member
Activity: 686
Merit: 259
June 12, 2013, 05:23:40 AM
GPU miner released, heh..
I'm not so hyped about it.. YAC isn't special anymore, right?

Well, hanzac's GPU miner doesn't perform especially well compared with mtrlt's normal scrypt+salsa kernel for mining other scrypt coins on the same hardware.  Hanzac's OpenCL kernel was already mostly released way way back in this thread, he just needed to fix a few bugs.

I tested it long enough to mine about 5 blocks of YAC then switched my GPU farm back to mining LTC.
GPU farm? You have a CPU farm too, right?
What's your job?  Grin
Jump to: