Pages:
Author

Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 62. (Read 3426944 times)

legendary
Activity: 1400
Merit: 1050
evening all. Over a thousand pages, kind of hard to figure things out.

 So., I'm new to all of this. I downloaded and setup minercontrol. I downloaded DJM's latest ccminer (M7V7app). Configured the conf file. Minercontrol launches and looks like I'm mining. But CCminer never launches. I wasn't sure which .dll files to include so I tried it with none and all, same result.
 Any idea's?
I was about to send you to miner control thread but they already send you there...
I must admit I never heard of miner control before so I can't really be of any help (just read  a few message on their thread where they tell "us"  to put our shit together (to summarize... ) may-be they should first explain what they want to do with others application rather than complaining... (especially when they never put any message here to describe what they intended to do  Grin)

ccminer is a standalone application... so you have to tell us what you want to mine otherwise we can't really help, there is no need for a config file just a command line...
newbie
Activity: 22
Merit: 0
evening all. Over a thousand pages, kind of hard to figure things out.

 So., I'm new to all of this. I downloaded and setup minercontrol. I downloaded DJM's latest ccminer (M7V7app). Configured the conf file. Minercontrol launches and looks like I'm mining. But CCminer never launches. I wasn't sure which .dll files to include so I tried it with none and all, same result.
 Any idea's?
legendary
Activity: 1400
Merit: 1050

...with MaxCoin I had 24 hours notice. This one gives us 48...

Christian


Nice to see your comments Mr. Buchner. My ears perk up when you mention MaxCoin. Those were heady times indeed. As for this new coin, it's a shame they changed their minds about the prelaunch release of source code, but as an aside, I enjoy mining the coins like these with a side benefit for science.
side benefit for science...  Grin
You want to do better than calculating prime number to help science ? Run scrypt again, open the windows to cool down the room, let the lights on, then you will make lots of entomologist happy all around the world (and you may found new viruses too...  Grin)
legendary
Activity: 3164
Merit: 1003
when mining scrypt jane n=14 on my vista computer i get 3.5 kh/s per 750ti. But on windows 8 2 different computers, i get 2.25kh/s.
I tried it with one card in so memory isnt a problem and tried different batch files. I cant get a 750ti on  windows 8 to do more then 2.25kh/s.   Tongue
Anybody have a clue way plz. thank you.
EDIT: I wonder if it was compiled on windows 7 by Christian and I need to  compiled on windows 8.

Can you share your config that gives you 3.5 kh/s? I`m getting only about 2.7 - 2.8  Thanks in advance.
what windows ?  different bat files. if windows 8 or maybe 7 i get the same hash rate as you.  Undecided   vista 3.5kh/s
full member
Activity: 170
Merit: 100
any new ccminers?
I'm getting 6200 kh/s on a gtx 970 over clocked (X11)
u guys?
ccMiner is not optimized for the 900-series yet.
They're working on it.
legendary
Activity: 3164
Merit: 1003
when mining scrypt jane n=14 on my vista computer i get 3.5 kh/s per 750ti. But on windows 8 2 different computers, i get 2.25kh/s.
I tried it with one card in so memory isnt a problem and tried different batch files. I cant get a 750ti on  windows 8 to do more then 2.25kh/s.   Tongue
Anybody have a clue way plz. thank you.
EDIT: I wonder if it was compiled on windows 7 by Christian and I need to  compiled on windows 8.

Not sure what you're doing wrong, but both of my systems running 8.1 get as much or more hash than my win7 machines.
really, the vista gets more hash on all algos.  Undecided 
sr. member
Activity: 285
Merit: 250
any new ccminers?
I'm getting 6200 kh/s on a gtx 970 over clocked (X11)
u guys?
member
Activity: 91
Merit: 10

...with MaxCoin I had 24 hours notice. This one gives us 48...

Christian


Nice to see your comments Mr. Buchner. My ears perk up when you mention MaxCoin. Those were heady times indeed. As for this new coin, it's a shame they changed their minds about the prelaunch release of source code, but as an aside, I enjoy mining the coins like these with a side benefit for science.
hero member
Activity: 756
Merit: 502
I think you chose the right time to look at prime numbers: another one here https://bitcointalksearch.org/topic/anngap-gapcoin-prime-gap-search-new-math-algo-cpu-gpu-zero-premine-822498

"Source code and binaries will be distributed about two days before launch. "

with MaxCoin I had 24 hours notice. This one gives us 48.

UPDATE: "Source code will be available before launch (excluding the PoW function)"

aww man, such party poopers.

but okay, this coin might turn out to be worthless. Like pretty much any other newly launched altcoin recently Wink

Christian
legendary
Activity: 1400
Merit: 1050
ChrisH and I, we are doing really rad 1024 bit arithmetics with warp synchronous programming.
This stuff looks exciting.

Riecoin and clones, fear us.

Also our wheel factorization and prime sieving starts to kick butt...
I think you chose the right time to look at prime numbers: another one here https://bitcointalksearch.org/topic/anngap-gapcoin-prime-gap-search-new-math-algo-cpu-gpu-zero-premine-822498
hero member
Activity: 756
Merit: 502
Does anyone think/thought about using sse/avx technology in ccminer ? sp_ ?

I know that ccminer spends a little time @CPU comparing to GPU but still there is a room for improvment.

The CPU path in cudaminer used SSE2 pretty early on for scrypt's SHA parts...
hero member
Activity: 756
Merit: 502
ChrisH and I, we are doing really rad 1024 bit arithmetics with warp synchronous programming.
This stuff looks exciting.

Riecoin and clones, fear us.

Also our wheel factorization and prime sieving starts to kick butt...
legendary
Activity: 1400
Merit: 1050
Does anyone think/thought about using sse/avx technology in ccminer ? sp_ ?

I know that ccminer spends a little time @CPU comparing to GPU but still there is a room for improvment.
there is always room for improvement... however their is practically no calculation done on cpu (apart for some initialization which are done once in a while), there is really nothing to gain for the moment (compared to what could be gained on gpu... )
newbie
Activity: 25
Merit: 0
Does anyone think/thought about using sse/avx technology in ccminer ? sp_ ?

I know that ccminer spends a little time @CPU comparing to GPU but still there is a room for improvment.
hero member
Activity: 524
Merit: 500
On Nvidia being better than AMD - I'm not so sure. If what you need is memory access, then yes, Nvidia is going to be great. Lots of 32-bit or bigger multiplications? Nvidia. Need to save power? Nvidia. Raw compute power to do operations on stuff with little memory accesses? AMD is pretty damned good.
Interthread communication without local memory - there is a hope to turn whole computing unit (or how is it called) into single hashing machine

EDIT:
Quote
__shfl()
    Direct copy from indexed lane
hero member
Activity: 524
Merit: 500
that's not bitslicing! Bitsliced keccak implementation would take at least 1600 words (8 bit ones or 256 - doesn't matter), so it's going to be pretty hard to squeeze it into AMD GPU. On the other hand, new fashioned shuffles on NVidia GPUs could be of great help here - so I've asked Smiley
You're right, I only said I did it 25-way, though, not bitsliced. Might help, though.
Thank you for sharing the code, it was quite interesting to look at CUDA way of doing things.
To be specific, I'm interested not in NVidia mining progress per se (quite sure you're getting a lot of stupid questions about it even without me:)), but in bitslicing, byteslicing, DPLL and all such algebraic games around GF(2). I continue to invent new ways to mine, but the programming side is abandoned. Well, to give these ideas a try I need a good bitslicing machine, seems that new NVidia's are better at it than AMD, so this thread is a good place to hunt for new ideas Smiley
BTW, if my old OpenCL code for Quark hash can be of any use, feel free to ask, I'll just dump it into forum.
newbie
Activity: 57
Merit: 0
when mining scrypt jane n=14 on my vista computer i get 3.5 kh/s per 750ti. But on windows 8 2 different computers, i get 2.25kh/s.
I tried it with one card in so memory isnt a problem and tried different batch files. I cant get a 750ti on  windows 8 to do more then 2.25kh/s.   Tongue
Anybody have a clue way plz. thank you.
EDIT: I wonder if it was compiled on windows 7 by Christian and I need to  compiled on windows 8.

Can you share your config that gives you 3.5 kh/s? I`m getting only about 2.7 - 2.8  Thanks in advance.
sr. member
Activity: 462
Merit: 250
www.dashpay.io
I'm looking to do a crowd funded x11 optimisation project, I think we will get really good support and I will reach out to multiple communities for donations, PM me if interested:

https://darkcointalk.org/threads/x11-mining-optimisation-project.2584/

Not sure if funding would improve the situation

here, I got a pull request for ccMiner's X11 with various optimizations

https://github.com/cbuchner1/ccminer/pull/25

but at this point I am not updating the ccMiner repo anymore. So the best way to go forward is probably going to be optimized forks of the C&C Hash factory's code.

Christian


I'm hearing there's optimisations to be had on both OpenCL and Cuda for the X algo's, it just needs time and effort.

But thanks for the tip, I will look into it...
legendary
Activity: 1400
Merit: 1050
..contuing with the 900s
On X13 970 gives about 4.8 Mh/s , 980 reaches 5.45 Mh/s
With NVminer D8...

How does gtx970 compare to 780ti in mining? The price is identical.
seriously the 780ti is at 350euros ?
legendary
Activity: 1512
Merit: 1000
quarkchain.io
..contuing with the 900s
On X13 970 gives about 4.8 Mh/s , 980 reaches 5.45 Mh/s
With NVminer D8...

a bit underwhelming vs 750ti, need opti
Fully agree
Pages:
Jump to: