Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 980. (Read 3426921 times)
January 26, 2014, 07:23:38 PM
Ooo it is definitely worth while. I have yet to sell anything I have made with cudaminer as its only 1100 yacoins so far. But when it comes to me selling, christian you have a nice chunk going your way
January 26, 2014, 07:21:23 PM
Very good for one day I must say. Over $200. And if they go up even a small amount it will make a big difference in their value
Not a small sum agreed.. But cbuchner1 is worthy of such an award - he deserved it. I think we all need to financially support such an important job.
January 26, 2014, 07:17:55 PM
Very good for one day I must say. Over $200. And if they go up even a small amount it will make a big difference in their value
January 26, 2014, 07:16:32 PM
How long did it take to get 5million?
Im guessing the block rewards are very high
Im guessing the block rewards are very high
i have many amd rigs - 5m may be 1 day
January 26, 2014, 07:14:20 PM
How long did it take to get 5million?
Im guessing the block rewards are very high
Im guessing the block rewards are very high
January 26, 2014, 07:00:34 PM
well... a Microcoin millionaire
Thanks for the donation. It is a significant amount.Remaining accelerate keccak and many of us become millionaires
January 26, 2014, 06:55:46 PM
Uh thanks. That makes me a millionaire then?
i exchange MRC to BTC rate 5m mrc / 0.22 btc
well... a Microcoin millionaire
Thanks for the donation. It is a significant amount. January 26, 2014, 06:54:22 PM
Uh thanks. That makes me a millionaire then?
i exchange MRC to BTC rate: 5m mrc / 0.22 btc
January 26, 2014, 06:51:42 PM
I downloaded a MicroCoin wallet. Donations go here please: 1Lxc4JPDpQRJB8BN4YwhmSQ3Rcu8gjj2Kd
Paid!
Great job!
Uh thanks. That makes me a millionaire then?
I got it to sync with this addnode list in microcoin.conf (from a recent forum posting)
Code:
addnode=84.125.196.167:56141
addnode=46.4.72.211:49983
addnode=188.2.174.144:61074
addnode=180.152.20.155:65409
addnode=24.18.104.42:61670
addnode=68.116.166.203:38810
addnode=62.178.107.69:63589
addnode=89.21.79.8:54693
addnode=183.171.178.83:38670
addnode=62.73.111.44:1100
addnode=80.213.13.216:62872
addnode=71.187.20.20:51552
addnode=105.236.85.49:54151
addnode=85.25.146.82:17397
addnode=188.230.132.226:3154
addnode=46.4.72.211:49983
addnode=188.2.174.144:61074
addnode=180.152.20.155:65409
addnode=24.18.104.42:61670
addnode=68.116.166.203:38810
addnode=62.178.107.69:63589
addnode=89.21.79.8:54693
addnode=183.171.178.83:38670
addnode=62.73.111.44:1100
addnode=80.213.13.216:62872
addnode=71.187.20.20:51552
addnode=105.236.85.49:54151
addnode=85.25.146.82:17397
addnode=188.230.132.226:3154
and it's already on a couple of exchanges: poloniex.com crycurex.com
January 26, 2014, 06:38:24 PM
I downloaded a MicroCoin wallet. Donations go here please: 1Lxc4JPDpQRJB8BN4YwhmSQ3Rcu8gjj2Kd
Paid!
Great job!
my microCoin wallet let's 51 connections
January 26, 2014, 06:31:31 PM
TITAN get 520k (vs AMD 290x - 2,6m)
GPU load 95%
MEMORY used 1458MB
[2014-01-27 02:10:54] GPU #1: 2303.87 khash/s with configuration T432x8
let's address your MRC - I pay 5 000 000 MRC for the excellent work!
I believe it's still spending 3/4 of the GPU time in the Keccak code. Otherwise you would
be seeing the indicated 2303.87 khash/s instead of "only" 530...
Maybe in a day or two I'vegot the speedup you need.
I downloaded a MicroCoin wallet. Donations go here please: 1Lxc4JPDpQRJB8BN4YwhmSQ3Rcu8gjj2Kd
Meh... wallet doesn't sync. initially it shows 2 connections. After a minute they drop to 0.
January 26, 2014, 06:13:40 PM
Not for YAC, but maybe people can now start to try mining somewhat lower Nfactor coins.
I believe I can still make that Keccak GPU code 50 times faster. What you see is just the first working draft.
Christian
microCoin work!
nFactor 7
TITAN get 520k (vs AMD 290x - 2,6m)
GPU load 95%
MEMORY used 1458MB
[2014-01-27 02:10:54] GPU #1: 2303.87 khash/s with configuration T432x8
[2014-01-27 02:10:54] GPU #1: using launch configuration T432x8
[2014-01-27 02:10:54] GPU #1: GeForce GTX TITAN, 300.51 khash/s
[2014-01-27 02:10:54] DEBUG: got new work in 5 ms
[2014-01-27 02:10:57] GPU #1: GeForce GTX TITAN, 520.20 khash/s
[2014-01-27 02:10:57] DEBUG: got new work in 2 ms
[2014-01-27 02:11:02] GPU #1: GeForce GTX TITAN, 522.15 khash/s
[2014-01-27 02:11:02] DEBUG: got new work in 0 ms
[2014-01-27 02:11:07] GPU #1: GeForce GTX TITAN, 521.11 khash/s
[2014-01-27 02:11:07] DEBUG: got new work in 1 ms
[2014-01-27 02:11:12] GPU #1: GeForce GTX TITAN, 523.82 khash/s
[2014-01-27 02:11:12] DEBUG: got new work in 1 ms
[2014-01-27 02:11:17] GPU #1: GeForce GTX TITAN, 526.02 khash/s
[2014-01-27 02:11:17] DEBUG: got new work in 1 ms
[2014-01-27 02:11:22] GPU #1: GeForce GTX TITAN, 529.54 khash/s
[2014-01-27 02:11:22] DEBUG: got new work in 1 ms
let's address your MRC - I pay 5 000 000 MRC for the excellent work!
January 26, 2014, 05:52:03 PM
-H 2 in scrypt-jane now puts keccak onto the GPU.
Is there any benefit to using that with YAC?
Not for YAC, but maybe people can now start to try mining somewhat lower Nfactor coins.
I believe I can still make that Keccak GPU code 50 times faster. What you see is just the first working draft.
Christian
January 26, 2014, 05:43:59 PM
Stanr010, the shutdown would allow an external script or control software to manage the failover process...
please watch the volume of quotations (> 80% in your posting)
Failover is on my TODO list. But right now I am completing the scrypt-jane feature.
please watch the volume of quotations (> 80% in your posting)
Failover is on my TODO list. But right now I am completing the scrypt-jane feature.
January 26, 2014, 05:40:49 PM
-H 2 in scrypt-jane now puts keccak onto the GPU. However the CUDA kernel is still a perfect example of how NOT to do efficient CUDA code.
Meaning without further optimizations, don't expect any miracles yet!
Christian
Is there any way to set up Cudaminer to have a failover / backup server if one goes down? Cgminer allows you to edit that in the config but there isn't a config file for cudaminer. Is there any way to set it up with the .bat?
Something like this from cgminer??
I think a good way to make this possible is with a --shutdown parameter, which would make cudaminer shutdown as soon as it has a connect failure.Something like this from cgminer??
Quote
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
C:\your\path\to\cgminer.exe --verbose --scrypt -o stratum+tcp://[mining_url] -u [workerusername] -p [workerpassword] --thread-concurrency 8192 -I 13 -g 2 -w 256 --gpu-fan 85 --lookup-gap 2 --gpu-engine 1080 --gpu-memclock 1500 --failover-only -o stratum+tcp://[failover_url] -u [workerusername] -p [workerpassword]
setx GPU_USE_SYNC_OBJECTS 1
C:\your\path\to\cgminer.exe --verbose --scrypt -o stratum+tcp://[mining_url] -u [workerusername] -p [workerpassword] --thread-concurrency 8192 -I 13 -g 2 -w 256 --gpu-fan 85 --lookup-gap 2 --gpu-engine 1080 --gpu-memclock 1500 --failover-only -o stratum+tcp://[failover_url] -u [workerusername] -p [workerpassword]
Why would I want it to shut down? I just want it to switch to another server if it disconnects and is unable to reconnect. Eventually, I'd like it to go back to the original server when it's up again.
January 26, 2014, 05:36:03 PM
-H 2 in scrypt-jane now puts keccak onto the GPU. However the CUDA kernel is still a perfect example of how NOT to do efficient CUDA code.
Meaning without further optimizations, don't expect any miracles yet!
Christian
Is there any way to set up Cudaminer to have a failover / backup server if one goes down? Cgminer allows you to edit that in the config but there isn't a config file for cudaminer. Is there any way to set it up with the .bat?
Something like this from cgminer??
I think a good way to make this possible is with a --shutdown parameter, which would make cudaminer shutdown as soon as it has a connect failure. Something like this from cgminer??
Quote
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
C:\your\path\to\cgminer.exe --verbose --scrypt -o stratum+tcp://[mining_url] -u [workerusername] -p [workerpassword] --thread-concurrency 8192 -I 13 -g 2 -w 256 --gpu-fan 85 --lookup-gap 2 --gpu-engine 1080 --gpu-memclock 1500 --failover-only -o stratum+tcp://[failover_url] -u [workerusername] -p [workerpassword]
setx GPU_USE_SYNC_OBJECTS 1
C:\your\path\to\cgminer.exe --verbose --scrypt -o stratum+tcp://[mining_url] -u [workerusername] -p [workerpassword] --thread-concurrency 8192 -I 13 -g 2 -w 256 --gpu-fan 85 --lookup-gap 2 --gpu-engine 1080 --gpu-memclock 1500 --failover-only -o stratum+tcp://[failover_url] -u [workerusername] -p [workerpassword]
January 26, 2014, 05:30:53 PM
Is there any way to set up Cudaminer to have a failover / backup server if one goes down? Cgminer allows you to edit that in the config but there isn't a config file for cudaminer. Is there any way to set it up with the .bat?
Something like this from cgminer??
Something like this from cgminer??
Quote
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
C:\your\path\to\cgminer.exe --verbose --scrypt -o stratum+tcp://[mining_url] -u [workerusername] -p [workerpassword] --thread-concurrency 8192 -I 13 -g 2 -w 256 --gpu-fan 85 --lookup-gap 2 --gpu-engine 1080 --gpu-memclock 1500 --failover-only -o stratum+tcp://[failover_url] -u [workerusername] -p [workerpassword]
setx GPU_USE_SYNC_OBJECTS 1
C:\your\path\to\cgminer.exe --verbose --scrypt -o stratum+tcp://[mining_url] -u [workerusername] -p [workerpassword] --thread-concurrency 8192 -I 13 -g 2 -w 256 --gpu-fan 85 --lookup-gap 2 --gpu-engine 1080 --gpu-memclock 1500 --failover-only -o stratum+tcp://[failover_url] -u [workerusername] -p [workerpassword]
January 26, 2014, 05:06:41 PM
-H 2 in scrypt-jane now puts keccak onto the GPU. However the CUDA kernel is still a perfect example of how NOT to do efficient CUDA code.
Meaning without further optimizations, don't expect any miracles yet!
EDIT: seems my 660Ti could do around 400 kHash/s Keccak alone (when I remove the ROMix phase of scrypt). This is about 10 times the performance I had with the CPU based Keccak (single core used). What I do not understand however is why the CPU use of that cudaminer instance still hovers around 15%... It should be idling!
Christian
Meaning without further optimizations, don't expect any miracles yet!
EDIT: seems my 660Ti could do around 400 kHash/s Keccak alone (when I remove the ROMix phase of scrypt). This is about 10 times the performance I had with the CPU based Keccak (single core used). What I do not understand however is why the CPU use of that cudaminer instance still hovers around 15%... It should be idling!
Christian
January 26, 2014, 04:11:22 PM
In most wallets you can increase it with command line options or with the wallet .conf file.
With more connections you can reach the whole network with less hops. And it might help finding less orphans. Not sure.
Also, with more connections you can download the blockchain faster (if you're not CPU and/or IO bottlenecked, which is generally the case).
The reason I asked because I solomined YAC for days and found nothing, not even an orphan block and I just recently increased the active connection limit and was not sure that it would have any connection between solomining. I guess it doesn't.
Code:
-maxconnections= "Maintain at most connections to peers (default: 125)"
-maxoutbound= "Maintain at most outbound connections to peers (default: 8)"
-maxoutbound=
With more connections you can reach the whole network with less hops. And it might help finding less orphans. Not sure.
Also, with more connections you can download the blockchain faster (if you're not CPU and/or IO bottlenecked, which is generally the case).
The reason I asked because I solomined YAC for days and found nothing, not even an orphan block and I just recently increased the active connection limit and was not sure that it would have any connection between solomining. I guess it doesn't.
Ah, so this isn't about RPC connections of miners to the wallet...
I think in all the time mining YAC (2 weeks or so) I had just a single orphan.
The YAC network doesn't seem particulary loaded with traffic, so I think 8 connections are just fine.
By the way the Windows Yacoin wallet reports 15 active connections, and my Linux wallet (which is the one I am solo mining with) reports 8 connections. Both are at v0.4.2.
Christian
January 26, 2014, 03:46:46 PM
In most wallets you can increase it with command line options or with the wallet .conf file.
With more connections you can reach the whole network with less hops. And it might help finding less orphans. Not sure.
Also, with more connections you can download the blockchain faster (if you're not CPU and/or IO bottlenecked, which is generally the case).
The reason I asked because I solomined YAC for days and found nothing, not even an orphan block and I just recently increased the active connection limit and was not sure that it would have any connection between solomining. I guess it doesn't.
As far as I saw Vertcoin is hardcoded for 8 connections, it reaches it in a couple of minutes and then (looking at the debug.log) it stops any network discovery until a connection is lost. Not sure if that is a good idea especially that the default timeout is 5000 ms for connections, so the wallet could theoritically pickup pretty shitty, high latency connections and keep them. You can decrease the default timeout delay, though.
Code:
-maxconnections= "Maintain at most connections to peers (default: 125)"
-maxoutbound= "Maintain at most outbound connections to peers (default: 8)"
-maxoutbound=
With more connections you can reach the whole network with less hops. And it might help finding less orphans. Not sure.
Also, with more connections you can download the blockchain faster (if you're not CPU and/or IO bottlenecked, which is generally the case).
The reason I asked because I solomined YAC for days and found nothing, not even an orphan block and I just recently increased the active connection limit and was not sure that it would have any connection between solomining. I guess it doesn't.
As far as I saw Vertcoin is hardcoded for 8 connections, it reaches it in a couple of minutes and then (looking at the debug.log) it stops any network discovery until a connection is lost. Not sure if that is a good idea especially that the default timeout is 5000 ms for connections, so the wallet could theoritically pickup pretty shitty, high latency connections and keep them. You can decrease the default timeout delay, though.
Jump to: