Author

Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 602. (Read 3426922 times)

sr. member
Activity: 434
Merit: 250
XPM miner for Nvidia GPU IS COMING .
full member
Activity: 126
Merit: 100
Hi all!

I'm a bit new to nVidia mining, so I've some little questions:

I've seen that the card on the 1x-16x riser cable has less performance, I read that
this issue is already under investigation, is there any improvement in progress?

Is it more effective to mine non-scrypt coins to get around this bottleneck?



Running the miner twice solves the issue. So just run your bat file twice and leave both running
i have a 5 card evga sc 750 7i  hvc with 2 instances  -d,0,1,2   and d,3,4  getting 56mh/s 1gh hashe is a bit slow, but super stable.

For my 5* 750 ti mining HVC I run 1 instance for gpu 0. 1 instance for gpu 1,2. 1 instance for gpu 3,4. I get 58 Mh/s with +160 OC on the evga non SC. Mining on Zhpool.

If you are mining HVC please give some thought to your vote. A low vote now will extend the duration of the mint phase. During the limit and sustain phases you will be forced to adjust your vote. Your GPU's want you to vote 1.
hero member
Activity: 812
Merit: 1000
Sounds like a scammer...don't give him anything lol

Hey,wait a minute! You have a license from the primegpu guy for some of your cards,don't you?

Yup, why?

It's bound to a hardware ID, which is unique per rig (you swap components and you need a new one).
sr. member
Activity: 476
Merit: 250
Hi all!

I'm a bit new to nVidia mining, so I've some little questions:

I've seen that the card on the 1x-16x riser cable has less performance, I read that
this issue is already under investigation, is there any improvement in progress?

Is it more effective to mine non-scrypt coins to get around this bottleneck?



Running the miner twice solves the issue. So just run your bat file twice and leave both running

He's talking for regular scrypt and that does not solve the problem.

From what I can tell, in HVC you can run 2 instances because it's using less pci bandwidth and isn't running the card as hard.  When the cards are fully stressed to 99% it sounds like pcie 2.0 times 3+ cards starts to exhibit problems unles your mobo+processor play nice together(if you're on intel) or it comes down to the board on AMDs.

My socket 1150 Core i3 powered ASrock h81 rig runs fine and all cards get 295-300 after overclocks.  With how high they're overclocked I would assume I'd be getting the 315khs some people report, but I'm starting to think that's a myth and they just read that number from the dashboard on a pool or something.

In comparison, my socket 1155 board with a celeron only gives the main card pcie 2.0 and the rest show up as pcie 1.1 and have decreased hashrates. Even though the board supports pcie 3.0, the cpu doesn't, so they're a the mercy of intel for that one. luckily all socket 1150 i-series have pcie3.0 support this time around, unlike 1155.
sr. member
Activity: 350
Merit: 250
Yeh 1gh is good. I have dwarfpool on one of my miners and their stratum is very funny
hero member
Activity: 773
Merit: 508
Bitcore (BTX) - The Future is Now
Hi all!

I'm a bit new to nVidia mining, so I've some little questions:

I've seen that the card on the 1x-16x riser cable has less performance, I read that
this issue is already under investigation, is there any improvement in progress?

Is it more effective to mine non-scrypt coins to get around this bottleneck?



Running the miner twice solves the issue. So just run your bat file twice and leave both running
i have a 5 card evga sc 750 7i  hvc with 2 instances  -d,0,1,2   and d,3,4  getting 56mh/s 1gh hashe is a bit slow, but super stable.
full member
Activity: 126
Merit: 100
Anyone wanna help me figure out why ccMiner isn't hashing?

I have a 6 x 750TI rig, and I'm wanting to mine HVC.  My .bat looks like this:

ccminer35.exe -a heavy -o stratum+tcp://<
>  -u <> -p <> -v 512

It connects to the stratum and starts detecting new blocks however this is the only output on the window.

Binding thread 0 to cpu 0
Binding thread 1 to cpu 1
Binding thread 2 to cpu 2
Binding thread 3 to cpu 3
Binding thread 4 to cpu 4
Binding thread 5 to cpu 5

And that's it.  It never actually starts hashing, and yet the stratum keeps detecting new blocks.

Also via Hardware Monitor I can see that none of my cards are under any load at all.

Thoughts?

Still looking for someone to help out here, please.

What pool are you trying to connect to?
sr. member
Activity: 308
Merit: 250
Sounds like a scammer...don't give him anything lol

Hey,wait a minute! You have a license from the primegpu guy for some of your cards,don't you?
sr. member
Activity: 308
Merit: 250
What a bad day it is. First the new primecoin GPU dev asks for 50 BTC for linux-only code. Then my power-supply fails on me for the second time. Now all my coins at Ypool get stolen. I don't want my crypto experiment ending like this Cry
you mean the nvidia one ?

Yeah,that one. Though I believe it can be cross-compiled for windows,no one is going to give a 50 BTC bounty.
Is only asking for 12500xpm now... for a code written in opencl and "optimized" for gtx570 (without any giving proof he has anything...)

Heh,no one's gonna pay that much for unproven code. Meanwhile,the AMD miner dev is asking for 35 XPM for one license. This is suddenly becoming very closed-source. Defeats the whole "open-structure" of cryptos. Next thing you know,AMD or Nvidia is making cards with custom software at a price. God forbid that ever happens.
hero member
Activity: 812
Merit: 1000
Sounds like a scammer...don't give him anything lol
legendary
Activity: 1400
Merit: 1050
What a bad day it is. First the new primecoin GPU dev asks for 50 BTC for linux-only code. Then my power-supply fails on me for the second time. Now all my coins at Ypool get stolen. I don't want my crypto experiment ending like this Cry
you mean the nvidia one ?

Yeah,that one. Though I believe it can be cross-compiled for windows,no one is going to give a 50 BTC bounty.
Is only asking for 12500xpm now... for a code written in opencl and "optimized" for gtx570 (without any giving proof he has anything...)
sr. member
Activity: 308
Merit: 250
What a bad day it is. First the new primecoin GPU dev asks for 50 BTC for linux-only code. Then my power-supply fails on me for the second time. Now all my coins at Ypool get stolen. I don't want my crypto experiment ending like this Cry
you mean the nvidia one ?

Yeah,that one. Though I believe it can be cross-compiled for windows,no one is going to give a 50 BTC bounty.
legendary
Activity: 1400
Merit: 1050
What a bad day it is. First the new primecoin GPU dev asks for 50 BTC for linux-only code. Then my power-supply fails on me for the second time. Now all my coins at Ypool get stolen. I don't want my crypto experiment ending like this Cry
you mean the nvidia one ?
sr. member
Activity: 350
Merit: 250
What a bad day it is. First the new primecoin GPU dev asks for 50 BTC for linux-only code. Then my power-supply fails on me for the second time. Now all my coins at Ypool get stolen. I don't want my crypto experiment ending like this Cry

Very very bad luck
sr. member
Activity: 308
Merit: 250
What a bad day it is. First the new primecoin GPU dev asks for 50 BTC for linux-only code. Then my power-supply fails on me for the second time. Now all my coins at Ypool get stolen. I don't want my crypto experiment ending like this Cry
sr. member
Activity: 350
Merit: 250
So far my miner runs at ~3300 khash/s. GPU overclock +91 Mem overclock +650.
Is there any benefit to lowering Mem clock like for HVC?
Do these numbers sound about right for GroestlCoin on a 750ti?

i just moved all 3 of my 750Ti's over to groestl, i have my memory clocks set to -502
right now i am getting;

3.2 mh/s
2.8 mh/s
2.9 mh/s

but they don't submit shares very often. over 1 minute per share

any idea's on how to distribute the new groestl code closed source, christian?
it does seem to be the most profitable now, today hvc dropped to the point of it not being profitable to mine at all with my 3 750Ti's
sr. member
Activity: 350
Merit: 250
It may be providing too much because of the extra molex. At that your better off using risers and still get the same performance.

As for the power reading, the cards themselves do pull around 39W but the efficiency of your psu will effect that. So an 80% psu will add an extra 8w ontop

Older psu's can be less then 70%. And even at that it would add an extra 10W or so on. Making them to the 50w area.
hero member
Activity: 676
Merit: 500
Moded to 45w. But even 65w will not make that much. It is more like 75-80w

defective Watt meter maybe? it may be overreporting certain kinds of load. And the load posed by a switching power supply is certainly a challenging one to measure correctly.


I got the stock bios back and with it the mashine uses 510 w overclocked.
What i saw is that if i put only one card , stock bios and start cudaminer the system from 50w goes up to 115w , so the card with or without moding the bios is pulling 65w (this is on the 16x pci-e). When it is on riser on 1x pci-e it pulls about 50 w. I don`t know , but it is not definetly 38w , try yourself , may be the board is giving max power to card at load , i don`t know. Huh Huh Huh
sr. member
Activity: 350
Merit: 250
Is it possible to add API to Cudaminer? Kind of liked that i could have all stats on one computer for all my rigs with CGRemote...

There are a few of us after this already. Im unsure how easy it is to do, or if its just not important right now compared to what else he is doing
newbie
Activity: 48
Merit: 0
I started mining GroestlCoin today on a single 750ti.

I am wondering what kind of hash rate others have seen mining with ccminer35 v0.5?

So far my miner runs at ~3300 khash/s. GPU overclock +91 Mem overclock +650.
Is there any benefit to lowering Mem clock like for HVC?
Do these numbers sound about right for GroestlCoin on a 750ti?

Thanks!
Jump to: