[ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX]

[wezra]

Be Interesting to play with this on my 690 when the windows binary is released, don't have the means or the brains to compile it!

Had all sorts of fun getting cudaminer running suitably well with VTC and the 690!

Does anyone have a compiled windows one with the keccak and the recent changes?

[1714219763]

1714219763

[1714219763]

1714219763

[1714219763]

1714219763

[1714219763]

1714219763

[Stanr010]

Will there be a new release of Cudaminer compiled and updated in the first post before Maxcoin launches this afternoon?

EDIT: For Windows.

[Mitchell]

Will there be a new release of Cudaminer compiled and updated in the first post before Maxcoin launches this afternoon?

EDIT: For Windows.

Yes. Half an hour before the release of the coin itself. So 7PM GMT.

EDIT: That's what OP said and I understood.

[bigjme]

any ideas about the performance on the new build christian?
im going to keep waiting for replies and just compile and test for you before release time

[cbuchner1]

Testing your builds

enable this disabled line in maxcoin.cpp (be sure to take it out again before actual mining)

Code:

        // TESTING ONLY
        ((uint32_t*)ptarget)[7] = 0x0000000f;

run cudaminer with the --debug (-D) option and watch if you get this:

Code:

[2014-02-06 15:57:01] DEBUG: hash <= target
Hash:   0000000d72a9d97d68d1e2ddb20809b4ad4623b895f0dfde715cd82b4e54d606
Target: 0000000f00000000000000000000000000000000000000000000000000000000

if you get the dreaded "does not validate on CPU" errors then your launch config or even your
build is having a problem.

This should help you figuring out if your 450 MHash/s are real or not.

EDIT: kHash/s -> MHash/s

[spur]

Will it be possible to configure a backup pool for cudaminer like you can do it in cgminer? This planned as a feature?
Lately a lot of pools are having problems, so I think this would be a nice option for pool mining.

Bumping this.

[bigjme]

christian did you see this over on reddit for maxcoin;

What do you mean by "Starting Algorithm"?

This is an issue of hardware miner resistance, such as ASICs. Keccak is the starting algorithm for MaxCoin and at this point in time no hardware miner currently exists. However, creating a Keccak ASIC is not impossible. Therefore, in order to protect against a hardware-miner future we are going to implement an "ASIC protection" feature into MaxCoin. This will work by allowing the blockchain to decide a new hashing algorithm for MaxCoin every x blocks. More specifically, the last authenticated transaction's hash is used to determine an integer and depending on this value an algorithm will be selected. This will mean hardware miners will find it difficult to create hardware in enough time to see profitable return. Purely for example, these could be:

x Algorithm 0 Keccak 1 Blake 2 Grostlx2 3 JH 4 Skein 5 Blake2 6 JH(Grostl) 7 Keccak+Blake

[cbuchner1]

christian did you see this over on reddit for maxcoin;

What do you mean by "Starting Algorithm"?

it says they intend to add a feature where they add new hashing algorithms over time, and let the blockchain switch between them according to some yet to be decided scheme.

Means I have to implement new hashing algorithms into cudaminer every few months or so, if we want to keep the advantage   Think 24-48 hours roundtrip time for me. Piece of cake.

Christian

[cbuchner1]

Will it be possible to configure a backup pool for cudaminer like you can do it in cgminer? This planned as a feature?
Lately a lot of pools are having problems, so I think this would be a nice option for pool mining.

Bumping this.

read the README, specifically the TODO section

[sgtstedanko]

Anyone compile windows binaries yet?

cheers!

[Stanr010]

Will it be possible to configure a backup pool for cudaminer like you can do it in cgminer? This planned as a feature?
Lately a lot of pools are having problems, so I think this would be a nice option for pool mining.

Bumping this.

read the README, specifically the TODO section

How about moving that from the TODO to the Changelog 

[cbuchner1]

How about moving that from the TODO to the Changelog  

I preferred to move scrypt-jane and keccak to the changelog first. For failover there are third party scripts and applications like CUDA Manager.

If you get failover integrated into cpuminer I can just merge it in with little effort.

Christian

[djm34]

Testing your builds

enable this disabled line in maxcoin.cpp (be sure to take it out again before actual mining)

Code:

        // TESTING ONLY
        ((uint32_t*)ptarget)[7] = 0x0000000f;

run cudaminer with the --debug (-D) option and watch if you get this:

Code:

[2014-02-06 15:57:01] DEBUG: hash <= target
Hash:   0000000d72a9d97d68d1e2ddb20809b4ad4623b895f0dfde715cd82b4e54d606
Target: 0000000f00000000000000000000000000000000000000000000000000000000

if you get the dreaded "does not validate on CPU" errors then your launch config or even your
build is having a problem.

This should help you figuring out if your 450 kHash/s are real or not.

hmm, I don't get the line at all...

[Stanr010]

How about moving that from the TODO to the Changelog  

I preferred to move scrypt-jane and keccak to the changelog first. For failover there are third party scripts and applications like CUDA Manager.

If you get failover integrated into cpuminer I can just merge it in with little effort.

Christian

I actually tried that and it's so bad it's not really worth using. It can't even return you back to your original server. It just keeps on going down a finite list until you get to the bottom and it just stops mining altogether.

[bigjme]

hmm, I don't get the line at all...

i ran the same on mine and i got it validated every time

[cbuchner1]

This should help you figuring out if your 450 kHash/s are real or not.

hmm, I don't get the line at all...

you're running keccak hashing with extra -D flag and no good hashes appear? Ouch...

[cbuchner1]

hmm, I don't get the line at all...

i ran the same on mine and i got it validated every time

so your ~110 MHash are real and the 450 MHash are not?

Here's one thing that bugs me. If my 660Ti can get close to 100 MHash, why can't your 780Ti exceed 200 MHash/s?

EDIT: kHash -> MHash

[ddink7]

Will this work with AMD cards as well for maxcoin or just Nvidia?

[bigjme]

i only get the 127Mhash/s and they all validate
i havent been able to push it higher. and just as a point, i get x** to 2 5which was giving me stupid number, that doesnt show anything on validation atall it just goes to the next one

just nvidia

[djm34]

This should help you figuring out if your 450 kHash/s are real or not.

hmm, I don't get the line at all...

you're running keccak hashing with extra -D flag and no good hashes appear? Ouch...

the hash doesn't show up at all
I try to recompile with maxregcount = 96