forzendiablo
Legendary
Offline
Activity: 1526
Merit: 1000
the grandpa of cryptos
|
|
November 10, 2013, 03:16:22 PM |
|
guys whe ni run scrypt.exe + cudminer at http://localhost:porteveryhing is fine. but when i try to run without scrypt.exe just straight from miner with command: cudaminer.exe -o stratum+tcp://stratum2.sever.com:333 -O workername:password then it runs and closes right away sayign to read --help. any ideas?
|
yolo
|
|
|
cocorickoo
Newbie
Offline
Activity: 12
Merit: 0
|
|
November 10, 2013, 05:53:02 PM |
|
GTX650 Ti OC : ~80khash/s with cudaminer. was 30 kh/s with cgminer. cgminer is almost instantaneous when you execute it, it stats mining. i found that cudaminer takes a bit more time to start when using stratum. i tried again using the http path to my pool and it started faster. also, i see that my GPU is running around 90% instead of 99% and CPU is at 7% instead of 25%. Is there any chances that you make a GUI for cudaminer ? it could be really cool... thanks a lot for your hard work. Good job Christian Buchner !
|
|
|
|
forzendiablo
Legendary
Offline
Activity: 1526
Merit: 1000
the grandpa of cryptos
|
|
November 11, 2013, 04:29:23 PM |
|
possible to mien other scrypt coins with it? i cant find a way.
|
yolo
|
|
|
cbuchner1 (OP)
|
|
November 12, 2013, 02:02:31 PM |
|
GTX650 Ti OC : ~80khash/s with cudaminer. was 30 kh/s with cgminer. also, i see that my GPU is running around 90% instead of 99% and CPU is at 7% instead of 25%. -i 0 flag for more speed, but possibly less desktop responsiveness
|
|
|
|
cbuchner1 (OP)
|
|
November 12, 2013, 02:03:40 PM |
|
Those running on two SLI'ed cards and two separate cudaminer instances might want to consider giving the -H 1 flag, otherwise both cudaminer instances will bind to the first CPU core only, competing for whatever CPU resources there are available on just one core.
|
|
|
|
hax0red
Member
Offline
Activity: 79
Merit: 10
|
|
November 13, 2013, 12:44:45 AM Last edit: November 13, 2013, 01:30:42 AM by hax0red |
|
Getting 220 kh/s with my GTX 670 consistently. Any numbers for 780 Ti?
|
|
|
|
sgk
Legendary
Offline
Activity: 1470
Merit: 1002
!! HODL !!
|
|
November 13, 2013, 05:12:01 PM |
|
Hello, I am fairly new to mining, just started. I have a PC with Intel Core i7 2600K; and NVIDIA GeForce GTX 550Ti graphics card; and 8GB of RAM. I am running Windows 7 Professional 64-bit I wanted to start mining LTC. I read that CudaMiner is best for NVIDIA GPUs, so I thought to start with that. I joined WeMineLTC.com and setup my workers. But since I didn't know what options to put in the .bat file, I had a look at the README.txt and tried to follow the pattern. My current .bat file looks like this: cudaminer.exe -H 1 -d 0,1,2 -i 1,0,0 -l auto,F27x3,K28x4 -C 0,2,1 -o stratum+tcp://stratum2.wemineltc.com:3334 -O btcsgk.home1:x When I run the .bat file, CudaMiner starts and makes the following progress: But then it stops: So I thought of using GUI-Miner, which also has a CudaMiner included. The options I set are as below: But as you can see, there is no mining progress. On the bottom Right corner, it just keeps showing "Starting" but never actually starts mining. Can anyone please guide me on how I should go about this? I am OK if you help me get through any one of the two methods above. Thanks in advance!
|
|
|
|
Eli0t
|
|
November 13, 2013, 05:24:46 PM |
|
that line is for using 3 devices to mine, start with this to find your auto tune settings and add the result after -l cudaminer.exe -H 1 -i 0 -C 2 -l -m 1 -o stratum+tcp://stratum2.wemineltc.com:3334 -O btcsgk.home1:x or try cudaminer.exe -H 1 -i 0 -C 2 -l 16x8 -m 1 -o stratum+tcp://stratum2.wemineltc.com:3334 -O btcsgk.home1:x some jealous soul has deleted all the settings from the google doc you need version 305 drivers or higher
|
LTC: LKpJf3uk7KsHU73kxq8iFJrP1AAKN7Yni7 DGC: DKXGvEbj3Rwgrm2QQbRyNPDDZDYoq4Y44d XPM: AWV5AKfLFyoBaMjg9C77rGUBhuFxz5DGGL
|
|
|
sgk
Legendary
Offline
Activity: 1470
Merit: 1002
!! HODL !!
|
|
November 13, 2013, 06:17:54 PM |
|
that line is for using 3 devices to mine, start with this to find your auto tune settings and add the result after -l cudaminer.exe -H 1 -i 0 -C 2 -l -m 1 -o stratum+tcp://stratum2.wemineltc.com:3334 -O btcsgk.home1:x or try cudaminer.exe -H 1 -i 0 -C 2 -l 16x8 -m 1 -o stratum+tcp://stratum2.wemineltc.com:3334 -O btcsgk.home1:x some jealous soul has deleted all the settings from the google doc you need version 305 drivers or higher Thanks Eli0t for your prompt response! When I try the first code you provided (to find auto tune settings), the command prompt opens and closes so fast that I can't even see what happened. When I try the second code, I receive the same error I posted in my post above ("cudaminer.exe has stopped working") I also waited on the GUI Miner to see if it ever connects to pool. After 10 minutes or so, it reported an error (on the bottom Left corner): "json_rpc_call failed" Edit: Below is a screenshot of my current driver. Do I need to upgrade?
|
|
|
|
Eli0t
|
|
November 13, 2013, 07:12:58 PM |
|
yes you need to update your drivers, looks like your using 301. for the version number right click the desktop and select Nvidia Control Panel and the home icon will tell you the driver version
i recommend 314.22 but you can try the latest ones if you want (if you get system lockups use 314.22)
|
LTC: LKpJf3uk7KsHU73kxq8iFJrP1AAKN7Yni7 DGC: DKXGvEbj3Rwgrm2QQbRyNPDDZDYoq4Y44d XPM: AWV5AKfLFyoBaMjg9C77rGUBhuFxz5DGGL
|
|
|
elbandi
|
|
November 14, 2013, 02:52:16 PM |
|
Hi, I want to test cudaMiner on ubuntu saucy (319.60 nvidia driver ver). When i start cudaminer, it do notting: [2013-11-14 14:47:02] Starting Stratum on stratum+tcp://stratum.hashfaster.com:3333 [2013-11-14 14:47:02] 1 miner threads started, using 'scrypt' algorithm. [2013-11-14 14:47:02] Stratum detected new block [2013-11-14 14:47:15] Stratum detected new block [2013-11-14 14:51:36] Stratum detected new block
And the cpu usage is 100%. Start params are: -H 1 -d 0 -i 1 -l auto -C 0 What sould i do?
|
|
|
|
sgk
Legendary
Offline
Activity: 1470
Merit: 1002
!! HODL !!
|
|
November 14, 2013, 04:54:19 PM |
|
yes you need to update your drivers, looks like your using 301. for the version number right click the desktop and select Nvidia Control Panel and the home icon will tell you the driver version
i recommend 314.22 but you can try the latest ones if you want (if you get system lockups use 314.22)
Thank you very much Eli0t for your help! I updated my driver to 314.22 and ran CudaMiner on auto mode. It is giving me 25 khash/s. I guess this is normal rate for my card?
|
|
|
|
Eli0t
|
|
November 14, 2013, 07:28:11 PM |
|
yes you need to update your drivers, looks like your using 301. for the version number right click the desktop and select Nvidia Control Panel and the home icon will tell you the driver version
i recommend 314.22 but you can try the latest ones if you want (if you get system lockups use 314.22)
Thank you very much Eli0t for your help! I updated my driver to 314.22 and ran CudaMiner on auto mode. It is giving me 25 khash/s. I guess this is normal rate for my card? would expect a little more than that but now you have it working you can start tweaking the -l setting. the google docs page has been resurrected https://docs.google.com/spreadsheet/lv?key=0AjMqJzI7_dCvdG9fZFN1Vjd0WkFOZmtlejltd0JXbmc&f=true&noheader=true&gid=0so you can check out other peoples settings, try -l F16x8 Hi, I want to test cudaMiner on ubuntu saucy (319.60 nvidia driver ver). When i start cudaminer, it do notting: [2013-11-14 14:47:02] Starting Stratum on stratum+tcp://stratum.hashfaster.com:3333 [2013-11-14 14:47:02] 1 miner threads started, using 'scrypt' algorithm. [2013-11-14 14:47:02] Stratum detected new block [2013-11-14 14:47:15] Stratum detected new block [2013-11-14 14:51:36] Stratum detected new block
And the cpu usage is 100%. Start params are: -H 1 -d 0 -i 1 -l auto -C 0 What sould i do? your using the worker username and pass from your account on that pool? cpu usage should not be 100% maybe not detecting your card for some reason
|
LTC: LKpJf3uk7KsHU73kxq8iFJrP1AAKN7Yni7 DGC: DKXGvEbj3Rwgrm2QQbRyNPDDZDYoq4Y44d XPM: AWV5AKfLFyoBaMjg9C77rGUBhuFxz5DGGL
|
|
|
cbuchner1 (OP)
|
|
November 14, 2013, 09:03:25 PM |
|
cudaminer 2013-11-14.zip was just posted on your favourite download site MEGA (chuckle) - jump to first posting in thread!
10-15% improvements over the previous best "Kepler" launch configuration may exist for Compute 3.5 enabled cards (anyone else can skip this update).
Geforce 780, 780 Ti, Titan, GT 640 with GK208 chip.
Be sure to run the "Titan" kernel this time! autotune to find the optimal config. And you don't need to give a -C flag here, as I can make use of the texture cache without actually defining textures (a new feature of Compute 3.5 devices).
Post some numbers for the aforementioned GPUs please. I am curious. Getting close to 64khash/s on GT 640, lol. beat this.
Will we beat the 400 kHash mark on the 780 Ti?
Christian
|
|
|
|
elbandi
|
|
November 14, 2013, 09:56:31 PM |
|
Hi, I want to test cudaMiner on ubuntu saucy (319.60 nvidia driver ver). When i start cudaminer, it do notting: [2013-11-14 14:47:02] Starting Stratum on stratum+tcp://stratum.hashfaster.com:3333 [2013-11-14 14:47:02] 1 miner threads started, using 'scrypt' algorithm. [2013-11-14 14:47:02] Stratum detected new block [2013-11-14 14:47:15] Stratum detected new block [2013-11-14 14:51:36] Stratum detected new block
And the cpu usage is 100%. Start params are: -H 1 -d 0 -i 1 -l auto -C 0 What sould i do? your using the worker username and pass from your account on that pool? cpu usage should not be 100% maybe not detecting your card for some reason i think there is something with my card. i replaced with another type of nvidia card, and with that, works. maybe my card is broken...
|
|
|
|
robot_meh
Newbie
Offline
Activity: 8
Merit: 0
|
|
November 15, 2013, 01:32:03 PM |
|
Hey, I am unable to compile the code on Xubuntu. I get the following error: ./configure.sh configure: error: cannot run /bin/bash ./config.sub
I have the libtool and automake libraries installed. I was unable to exectute the ./autogen.sh, however I was able to fix that using autoreconf --install. Could someone who managed to compile the source on Linux give me a hand please or at least point out the possible cause. Thanks a lot!
|
|
|
|
cbuchner1 (OP)
|
|
November 15, 2013, 01:40:34 PM |
|
code taken from github or from .zip archive?
|
|
|
|
cbuchner1 (OP)
|
|
November 15, 2013, 01:42:15 PM Last edit: November 15, 2013, 03:42:46 PM by cbuchner1 |
|
I "found" 8 more kHash in the Kepler kernel (my 660Ti (slightly OC) 200->208 khash). will publish another update soon.
EDIT: now at +10 kHash, which is a 5% improvement.
|
|
|
|
robot_meh
Newbie
Offline
Activity: 8
Merit: 0
|
|
November 15, 2013, 01:53:08 PM Last edit: November 15, 2013, 03:28:03 PM by robot_meh |
|
code taken from github or from .zip archive?
From zip from Mega. I didn't see any github links.Edit: Got it from github. Had to get a few tools and almost there. However having CUDA library installed its not included in the PATH variable and so far its being rather difficult exporting the $LD_LIBRARY_PATH to the correct location. Got to love Linux...
|
|
|
|
cbuchner1 (OP)
|
|
November 15, 2013, 11:06:37 PM |
|
Posted another cudaminer release, this time with performance improvements across the board (Fermi, Kepler, Titan)
Most notable probably for Kepler cards where you can get 5% more (went from 200 to 210 kHash on an OC'ed 660 Ti). This is mostly due to enabling the 8 byte shared memory bank mode and using ulonglong2 memory transfers where possible (dual vectors of 64 bit integers - looks very nice in the PTX code).
My 660 Ti launch config.
-C1 -l K14x16 -i 0
The card is an OC edition and I run it at 105% power limit with +100 MHz core, +300 MHz mem clock. Gets 210 kHash with this.
Now some of my cards run into thermal throttling... meh.
Christian
|
|
|
|
|