CCminer(SP-MOD) Modded GPU kernels.

[t-nelson]

ASROCK H81 PRO MB and DUAL VIDEO--

A number of thread-followers have mentioned that they run rigs on ASROCK H81 PRO motherboards.  These boards support Intel CPU-Integrated graphics as well as a wide variety of graphic cards.  Is it possible to drive a monitor with integrated Intel video, and mine with nVidia graphics cards (750ti SC)?

I will research the web before booting my new H81 rig, but I'd like to know if it has been done by any one of the readers.

Thank you in advance!       --scryptr

I achieve this on Linux with the following xorg.conf

Code:

Section "Device"
    Identifier "intel0"
    Driver "intel"
    BusID "PCI:0@0:2:0"
    Option "AccelMethod" "SNA"
EndSection

Section "Screen"
    Identifier "Screen0"
    Device "intel0"
EndSection

Section "Device"
    Identifier "nvidia0"
    Driver "nvidia"
    BusID "PCI:1@0:0:0"
    Option "ConstrainCursor" "off"
    Option "CoolBits" "28"
EndSection

Section "Screen"
    Identifier "Screen1"
    Device "nvidia0"
    Option "AllowEmptyInitialConfiguration" "on"
    Option "IgnoreDisplayDevices" "CRT"
EndSection

Section "Device"
    Identifier "nvidia1"
    Driver "nvidia"
    BusID "PCI:2@0:0:0"
    Option "ConstrainCursor" "off"
    Option "CoolBits" "28"
EndSection

Section "Screen"
    Identifier "Screen2"
    Device "nvidia1"
    Option "AllowEmptyInitialConfiguration" "on"
    Option "IgnoreDisplayDevices" "CRT"
EndSection

Section "ServerLayout"
    Identifier "layout"
    Screen 0 "Screen0"
    Screen 1 "Screen1"
    Screen 2 "Screen2"
#    Inactive "nvidia0"
#    Inactive "nvidia1"
EndSection

In the "ServerLayout" section, I have the nvidia cards assigned to screens to allow NVML to work.  It costs 18MB of VRAM to X, but I noticed no performance decrease.  It is simple enough to disable this by commenting screens 1 and 2 and uncommenting the "Inactive" lines for them.  You will lose temperature, fan speed, etc though.

EDIT:  On Ubuntu, you'll also need to disable "gpu-manager" or the dumb ass thing will overwrite your xorg.conf everytime you log in.  I just commented the entirety of /etc/init/gpu-manager.conf.  Allegedly, this also can be done by passing a kernel param in newer (than 14.04) versions.  Though I don't recall the name of said param.

[Strannik-74]

ASROCK H81 PRO MB and DUAL VIDEO--

A number of thread-followers have mentioned that they run rigs on ASROCK H81 PRO motherboards.  These boards support Intel CPU-Integrated graphics as well as a wide variety of graphic cards.  Is it possible to drive a monitor with integrated Intel video, and mine with nVidia graphics cards (750ti SC)?

I will research the web before booting my new H81 rig, but I'd like to know if it has been done by any one of the readers.

Thank you in advance!       --scryptr

Was running this setup last year, was using the iGPU of the G3220 to display.

750Ti allow to place them more compact 

[sp_]

Worked some more on the pentablake miner:

Same clocks as before.

In pentablake sp-mod 1 the rig was doing 215MHASH.
In sp-mod 2 it has increased + 20.5MHASH to 235.5MHASH

The GTX 970 is up 6MHASH
the 750ti is up 3MHASH

Around 10% increase


Sendt for free to the Pentablake Donators.

[scryptr]

ASROCK H81 PRO MB and DUAL VIDEO--

A number of thread-followers have mentioned that they run rigs on ASROCK H81 PRO motherboards.  These boards support Intel CPU-Integrated graphics as well as a wide variety of graphic cards.  Is it possible to drive a monitor with integrated Intel video, and mine with nVidia graphics cards (750ti SC)?

I will research the web before booting my new H81 rig, but I'd like to know if it has been done by any one of the readers.

Thank you in advance!       --scryptr

Was running this setup last year, was using the iGPU of the G3220 to display.

<SNIP>

750Ti allow to place them more compact 
<SNIP>

NICE RIGS! --

Are you running a Windows OS, or Linux?  I am wondering how to add the nVidia drivers without disabling the Intel internal video.       --scryptr

[tbearhere]

ASROCK H81 PRO MB and DUAL VIDEO--

A number of thread-followers have mentioned that they run rigs on ASROCK H81 PRO motherboards.  These boards support Intel CPU-Integrated graphics as well as a wide variety of graphic cards.  Is it possible to drive a monitor with integrated Intel video, and mine with nVidia graphics cards (750ti SC)?

I will research the web before booting my new H81 rig, but I'd like to know if it has been done by any one of the readers.

Thank you in advance!       --scryptr

Was running this setup last year, was using the iGPU of the G3220 to display.

You can't get a rig frame any cheaper then that.  

EDIT: What type of ram are you using?

[chrysophylax]

ASROCK H81 PRO MB and DUAL VIDEO--

A number of thread-followers have mentioned that they run rigs on ASROCK H81 PRO motherboards.  These boards support Intel CPU-Integrated graphics as well as a wide variety of graphic cards.  Is it possible to drive a monitor with integrated Intel video, and mine with nVidia graphics cards (750ti SC)?

I will research the web before booting my new H81 rig, but I'd like to know if it has been done by any one of the readers.

Thank you in advance!       --scryptr

Was running this setup last year, was using the iGPU of the G3220 to display.

You can't get a rig frame any cheaper then that. 

o i dont know about that ...

cardboard and foam cost a HUGE amount of money you know ...

nice setup ...

#crysx

[tbearhere]

ASROCK H81 PRO MB and DUAL VIDEO--

A number of thread-followers have mentioned that they run rigs on ASROCK H81 PRO motherboards.  These boards support Intel CPU-Integrated graphics as well as a wide variety of graphic cards.  Is it possible to drive a monitor with integrated Intel video, and mine with nVidia graphics cards (750ti SC)?

I will research the web before booting my new H81 rig, but I'd like to know if it has been done by any one of the readers.

Thank you in advance!       --scryptr

Was running this setup last year, was using the iGPU of the G3220 to display.

You can't get a rig frame any cheaper then that.  

o i dont know about that ...

cardboard and foam cost a HUGE amount of money you know ...

nice setup ...

#crysx

scryptr...the foam is free...comes with the asrock mb.  

[bensam1231]

I think you don't understand the license if you think you can profit off someones work as long as it's not 'public'. It still counts, its just not enforceable in anyway unless someone who receives the miner initiates it. When they talk about 'private work' that means you don't share it with anyone, not you don't share it with anyone publicly. Open licensing mainly refers to profiting off other peoples work though. That still happens even if it's not public, developers with private miners who sell them are already violating the open license, but this obviously isn't their area of forte, the only reason it's come up recently is because people are looking to use it as a excuse.

As I mentioned already, it's pretty easy to include more people with a fee schedule and this is a non-issue as this sort of stuff happens all the time in the mining community. No one is going to sue anyone else, nothing is going to happen, especially if you include all the main parties that created the work... That's mainly where I think people starting to poopoo on this. If DJM doesn't get any bit of the L2V2 fees, even though he developed most of the work for the initial miner. All of that could easily be programmed in, it's just someones address, or it could divy it up on the receiving end of the payments, which would probably be easiest.

Before continuing Your accusations, please provide GPL license text that is violated when sending binaries over private channels. You can assume that user is asking and paying developer to make better code for him from open source code. Can You make citation from GPL license that is broken with this?

Oh yeah? He's making better open source code, so he can then open source it? Using a loophole doesn't make it right. You're not supposed to be able to profit off other persons open work without their permission, that's the way it works. Just because you do it in private doesn't mean you aren't violating it. It's like doing something illegal profit still makes it illegal even if you more then likely wont get caught.

All of that could easily be programmed in, it's just someones address, or it could divy it up on the receiving end of the payments, which would probably be easiest.

easily programmed, this is where you are wrong (and I am not speaking about the logic)... it isn't easy at all in the sense you cannot just put an address and get the coins, you need to gives address/username/password for every single pool and every single coin.

For solo, it can't even work (at least not with getwork), I think it is possible to do something with gbt though...

Cause using a BTC address on nicehash is hard (or conversely for more flexibility use a proxy you can point wherever). I think you're grossly overestimating the amount of work here to try and make a point. Receiving address splits up end day, week, or whatever you want based on the fee schedule. Very easy to do and program. All it has to do is mine 2 out of 100 shares on Nicehash using the current algorithm that the user has selected.

If I can make a batfile that changes addresses for me, I assume you can do this in code.

GIGABYTE GeForce GTX 970 [GV-N970WF3OC-4GD 1.0]  
GPU 1264MHz (+150)/ mem 1753MHz (lock) / VDDC 1.2V (lock)
Rig idle 70W

     -- ccminer-1.5.69-git-spmod --
Lyra2REv2 - 9.5MH/s - 155W (Rig 225W)
X11-           8.7MH/s - 175W (Rig 245W)
Quark-        18MH/S - 220W (Rig 290W)



750Ti without power, steel risers burn through 12+ volts

Buy powered risers (they can still be the ribbon ones). Look into cutting the power cables on the risers if you have problems with power. The powered risers work fine despite what Myagui said. Unpowered ones lead to what you just saw.

Powered USB-type risers do not have the input PCIe power pins connected, they have an external power port on the riser itself, usually a SATA or MOLEX connector. This way, both riser and card get all power feeding directly from the PSU, and nothing out of the motherboard.

These risers kill 2 birds with one stone:
You're not using high power over a flat ribbon cable - which is a serious fire hazard btw -, and you're not using power off the motherboard PCIe slots, so you can use whatever motherboard you want, and not just the specialty flavors. In both aspects, it is much better (and safer) to have all power drawn straight off the PSU.

Motherboards with extra PCIe power are a waste in my mind, since you cannot fill-up all PCIe slots without using risers. And as soon as you need to use risers, go with powered ones, or go home.  

The ones I bought still tried to get power through the motherboard. I'd advise testing this as they aren't all created equal.

Any guides?

18GPU  10GPU

Pretty sure he's talking about the hardware setup. I've tried the PCI-E switches too and they don't work. Software side I've already found answers too, but on the hardware side you have to use clustering hardware which is really expensive from what I've seen.

Edit: Nevermind, he's using a realllllly expensive motherboard. Still cool though.

[djm34]

Cause using a BTC address on nicehash is hard (or conversely for more flexibility use a proxy you can point wherever). I think you're grossly overestimating the amount of work here to try and make a point. Receiving address splits up end day, week, or whatever you want based on the fee schedule. Very easy to do and program. All it has to do is mine 2 out of 100 shares on Nicehash using the current algorithm that the user has selected.

why would I do that ?
I am sorry, I don't work for nicehash, I am not even a fan of that system and have 0 reason to privilege nicehash over any other pool.
(and actually if I was doing something like, I would clearly expect a retribution from the pool who would have the monopoly over the fastest miner..   yeah I can dream...)
and we all know it would be bad for the coins.

This is the part you really fail to understand over and over again: your own interest isn't the interest of everyone (nor actually mine).
Hence if I make a fee miner, it has to work for all pool, not to force user to move to one pool which is insane.

I think your problem is that you are just greedy and don't give a shit about crypto. (sure I like to make money on it, but I try to not destroy what helps me to make money...)

[hashbrown9000]

i've never had a problem with all my unpowered ribbon cable risers on the 750ti cards. but all my cards have the extra 6 pin power connector though.

[theLosers106]

Hi,

What's the most hash per watt card right now?

Thanks
theLosers106

[bensam1231]

Cause using a BTC address on nicehash is hard (or conversely for more flexibility use a proxy you can point wherever). I think you're grossly overestimating the amount of work here to try and make a point. Receiving address splits up end day, week, or whatever you want based on the fee schedule. Very easy to do and program. All it has to do is mine 2 out of 100 shares on Nicehash using the current algorithm that the user has selected.

why would I do that ?
I am sorry, I don't work for nicehash, I am not even a fan of that system and have 0 reason to privilege nicehash over any other pool.
(and actually if I was doing something like, I would clearly expect a retribution from the pool who would have the monopoly over the fastest miner..   yeah I can dream...)
and we all know it would be bad for the coins.

This is the part you really fail to understand over and over again: your own interest isn't the interest of everyone (nor actually mine).
Hence if I make a fee miner, it has to work for all pool, not to force user to move to one pool which is insane.

I think your problem is that you are just greedy and don't give a shit about crypto. (sure I like to make money on it, but I try to not destroy what helps me to make money...)

Because Nicehash is easy and getting it going is the first step to making this work. But as I mentioned you could use a proxy and send it wherever you want and adjust it at the proxy.

This wouldn't be any worse then any other multipool, plus Nicehash can handle spurts of sudden hashing power with ease, is ddos protected, and extremely robust. It's the best solution unless someone can come up with something better.

I'm greedy because I want to pay you guys with my hash... Sorry, once again that doesn't compute, especially from the person selling private miners to select few rich individuals. THAT is exactly what cryptos aren't about. Sharing your hashrate with the developers who make things possible is the complete opposite of that.

i've never had a problem with all my unpowered ribbon cable risers on the 750ti cards. but all my cards have the extra 6 pin power connector though.

Yeah, that six pin helps. I'm guessing this may be different for the 750ti owners who don't have a aux power connector, especially those with higher wattage per card or who have unlocked their bios to provide more volts.

[AliMan]

ASROCK H81 PRO MB and DUAL VIDEO--

A number of thread-followers have mentioned that they run rigs on ASROCK H81 PRO motherboards.  These boards support Intel CPU-Integrated graphics as well as a wide variety of graphic cards.  Is it possible to drive a monitor with integrated Intel video, and mine with nVidia graphics cards (750ti SC)?

I will research the web before booting my new H81 rig, but I'd like to know if it has been done by any one of the readers.

Thank you in advance!       --scryptr

Was running this setup last year, was using the iGPU of the G3220 to display.

You can't get a rig frame any cheaper then that.  

EDIT: What type of ram are you using?

Those are Corsair Vengeance 1600Mhz, 2x4GB.

[chup]

Before continuing Your accusations, please provide GPL license text that is violated when sending binaries over private channels. You can assume that user is asking and paying developer to make better code for him from open source code. Can You make citation from GPL license that is broken with this?

Oh yeah? He's making better open source code, so he can then open source it? Using a loophole doesn't make it right. You're not supposed to be able to profit off other persons open work without their permission, that's the way it works. Just because you do it in private doesn't mean you aren't violating it. It's like doing something illegal profit still makes it illegal even if you more then likely wont get caught.

Your freewill understanding of GPL license doesn't count at all. If You can not provide citation of GPL license that is broken with before mentioned case than You should stop Your false accusations.

[theotherme]

Before continuing Your accusations, please provide GPL license text that is violated when sending binaries over private channels. You can assume that user is asking and paying developer to make better code for him from open source code. Can You make citation from GPL license that is broken with this?

Oh yeah? He's making better open source code, so he can then open source it? Using a loophole doesn't make it right. You're not supposed to be able to profit off other persons open work without their permission, that's the way it works. Just because you do it in private doesn't mean you aren't violating it. It's like doing something illegal profit still makes it illegal even if you more then likely wont get caught.

Your freewill understanding of GPL license doesn't count at all. If You can not provide citation of GPL license that is broken with before mentioned case than You should stop Your false accusations.

Once you have understood, that bensam is always right no matter how wrong he is, you start to react a lot less... 

[Slava_K]

SP_ - Lyra2REv2 - on my rig lost 1MH! on last commit... (2x980+960)

[sp_]

SP_ - Lyra2REv2 - on my rig lost 1MH! on last commit... (2x980+960)

Clean the project,get all the latest files and rebuild.

[sp_]

I am working on an optimized whirlcoin miner. But the old DJM-34 miner doesn't seem to work on the pool.

ccminer -a whirlcoin -o stratum+tcp://stratum.coinspool.cu.cc:3048 -u <joincoin adress> -p x

lots of low difficulty shares.

Anybody know how to set the --diff parameter?


Perhaps the pool is broken

[sp_]

I have added it now to github. The Whirlcoin algo is doing over 15MHASH on the 970 with overclocking. (run with -a whirl)
Does somebody have the time to solomine joincoin - whirlpool

The diff is really low at around 8.

Should do around 15MHASH on the 970 oc.


1 year ago the whirlpool algo was doing less on the gtx 980: (Hashrates before the sp-mod project)



My modded code will turn your 970 card into a 980 when mining whirlpool



And here is DJM-34's original version

[djm34]

I am working on an optimized whirlcoin miner. But the old DJM-34 miner doesn't seem to work on the pool.

ccminer -a whirlcoin -o stratum+tcp://stratum.coinspool.cu.cc:3048 -u <joincoin adress> -p x

lots of low difficulty shares.

Anybody know how to set the --diff parameter?


Perhaps the pool is broken

probably got broken when Vanilla was released... not a big deal to fix though...
(most likely not using the correct constant... or something like that...)

btw: whirlcoin is kinda dead (was never alive actually...)