[ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 21.

djm34

legendary

Activity: 1400

Merit: 1050

for yescrypt and compute 5.0 and higher cards (750ti and newer) use sgminer ( Grin

)

github/djm34/sgminer

with these parameters: (tested on gtx980, 750ti)
sgminer.exe --no-submit-stale --kernel yescrypt-multi -o stratum+tcp://mine2.bsty.nonce-pool.com:4095 -u djm34.1 -p password --gpu-platform 0 -w 4 -g 2

(may-be it is due to the fact that nvidia now support opencl 1.2... don't know or cuda doesn't like at all sha256... at least not how I write it Grin

)

it is significantly faster for the 900's 1.5kh/s versus 1.2kh/s
Don't use it for compute 3.5 cards, works like shit... Grin

ccminer works a lot better for the older cards...

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: CPUcoins on May 02, 2015, 02:55:06 AM

Hello. I am trying this new GPU instance to mine Yescrypt. I am new to mining with a GPU and was hoping I could get some assistance on setting it up.

https://github.com/djm34/ccminer-tpsp/tree/windows

I have an NVIDIA GeForce GTX 750 Ti. Driver Version 9.18.13.4752

Here is the post about the release.

https://bitcointalksearch.org/topic/m.11252741

Edit: Here is the string I am using for my bat file. I also tried the below with -a yescrypt, and I get the same as below.

Code:

ccminer.exe --algo=yescrypt -d 1 -o stratum+tcp://stratum.coin-miners.info:3575 -u user.1 -p pass -i 0

I also have ccminer.exe in the folder C:\ccminer-tpsp-windows

Not sure what I am doing wrong.

doesn't look like the version I released... but rather a version released by sp

edit: the link is correct, but the screenshot is from sp version
here the windows binaries: https://mega.co.nz/#!dR1mwTIb!22mH7em1FX2F81-C5vl9Wbb7o4RG57e82bNVfvFwInE

CPUcoins

newbie

Activity: 11

Merit: 0

Hello. I am trying this new GPU instance to mine Yescrypt. I am new to mining with a GPU and was hoping I could get some assistance on setting it up.

https://github.com/djm34/ccminer-tpsp/tree/windows

I have an NVIDIA GeForce GTX 750 Ti. Driver Version 9.18.13.4752

Here is the post about the release.

https://bitcointalksearch.org/topic/m.11252741

Edit: Here is the string I am using for my bat file. I also tried the below with -a yescrypt, and I get the same as below.

Code:

ccminer.exe --algo=yescrypt -d 1 -o stratum+tcp://stratum.coin-miners.info:3575 -u user.1 -p pass -i 0

I also have ccminer.exe in the folder C:\ccminer-tpsp-windows

Not sure what I am doing wrong.

https://i.imgur.com/76yAhNO.png

tbearhere

legendary

Activity: 3164

Merit: 1003

Quote from: o00o on May 01, 2015, 01:52:20 PM

It would be much appreciated if someone can confirm as to which algorithms/miners are supported by my Compute Capability 1.1 GPUs such as 9800GTX+/GTS250. I've tried mining x11 w/ Cudaminer& ccminer on 340.52, the latest driver w/ support for my GPUs but unfortunately, GPU utilization is ~ 1%, making Scrypt my only viable yet unprofitable option.

Thanks to anyone in advance for offering their help!

Goto OP and download
cudaminer-2014-02-18.zip [32+64bit version] (7.1 MB) also runs on Maxwell
And read the readme file it will show some coins that are mineable.
And goto this thread for ccminer https://bitcointalksearch.org/topic/ccminersp-mod-modded-nvidia-maxwell-pascal-kernels-826901
But I don't think ccminer will work on Compute 1.1 not sure .. ask there.

o00o

member

Activity: 93

Merit: 11

It would be much appreciated if someone can confirm as to which algorithms/miners are supported by my Compute Capability 1.1 GPUs such as 9800GTX+/GTS250. I've tried mining x11 w/ Cudaminer& ccminer on 340.52, the latest driver w/ support for my GPUs but unfortunately, GPU utilization is ~ 1%, making Scrypt my only viable yet unprofitable option.

Thanks to anyone in advance for offering their help!

myagui

legendary

Activity: 1154

Merit: 1001

Quote from: Houndirno on April 29, 2015, 05:40:12 PM

I'm still getting now where with 52 version ;c

You have a better shot at someone helping you, if you post:
- What coin it is that you are mining,
- What your miner launch command, (omit password if you care, but otherwise be accurate)
- What ccminer release you are using (2 recommendations below)

by tpruvot
https://github.com/tpruvot/ccminer/releases

by SP_
https://github.com/sp-hash/ccminer/releases/

Good Luck!

Houndirno

sr. member

Activity: 474

Merit: 263

Quote from: Amph on April 29, 2015, 02:07:18 PM

Quote from: Houndirno on April 29, 2015, 12:13:44 PM

Can someone help me, I'm just testing a GTX 750 no ti card and I keep getting either accepted 0/soso or when startum has found a new block

cudaminer is for scrypt, while ccminer for the rest, search for the sp_ version, also you need to run a ccminer_52 version

I'm still getting now where with 52 version ;c

Amph

legendary

Activity: 3248

Merit: 1070

Quote from: Houndirno on April 29, 2015, 12:13:44 PM

Can someone help me, I'm just testing a GTX 750 no ti card and I keep getting either accepted 0/soso or when startum has found a new block

cudaminer is for scrypt, while ccminer for the rest, search for the sp_ version, also you need to run a ccminer_52 version

Houndirno

sr. member

Activity: 474

Merit: 263

Can someone help me, I'm just testing a GTX 750 no ti card and I keep getting either accepted 0/soso or when startum has found a new block

Epsylon3

legendary

Activity: 1484

Merit: 1082

ccminer/cpuminer developer

New gen can also reduce ports size... actually the socket 2011 cpus have more pcie/usb3 lanes than 115x.

I think next ones will have more than ever... even the sata/sata express ports tend to move to M2 pci express system

Amph

legendary

Activity: 3248

Merit: 1070

Quote from: djm34 on April 16, 2015, 08:45:56 AM

Quote from: sp_ on April 16, 2015, 07:37:55 AM

Quote from: smolen on April 16, 2015, 07:10:52 AM

Quote from: sp_ on April 16, 2015, 04:32:25 AM

What if you have 4 gpu's in your rig and each thread is executed on a seperate gpu. x11 is then reduced to x2+.

Hmm, PCIe 2.0 x16 bandwidth is 8 GB/s, 100 MH/s (desired speed) * 64 bytes (internal state size) = 6.4 GB/s

x11 with the latest ccminer is currenly hashing at around 3MH/s (on the 750ti)

PCIe 4.0×16 has a 31.5 GB/s bandwidth. (3*4(GPU)) above 12 MH/s (desired speed) * 64 bytes (internal state size) = 0,576 GB/s

with 4 gpu's and 4 threads you need to copy the data 3 times. (the last gpu pass will only need to return the winning nounce.)

The copy process can be done in parallell 0,576 GB/s by dividing the pci-e bus timeline

-----------TimeLine--------------->
GPU1: --Copy--Calc--calc--Calc
GPU2: --Calc--Copy--Calc--Calc
GPU3: --Calc--Calc--Copy--Calc
GPU4: --Calc--Calc--Calc--Result

A PCIe 2.0 x16 bandwidth is 8 GB/s and this meens that at least 7.2% of the time is wasted with moving memory. So your Kernals needs to run at least 8% faster.

seems slow. Could work for a cryptonite miner since the desired speed is 500h-1KH/s

PCIe 4.0 ? (planned to be released late 2016, by that time there will be for sure x11 asic Grin

)
also you never have more than 2 gpu's at x16 (unless you go to the utterly expensive x99 setup, not mining friendly)

oh pcie 4 is planned for 2016, did not know that, interesting, but isn't pci-e 3.0 still not completely saturated?

uplodex

jr. member

Activity: 40

Merit: 13

Hi,

there is not a solution of CPU mining "windows 7" for CACHeCoin?

reorder

sr. member

Activity: 462

Merit: 250

Quote from: Bitice on April 23, 2015, 07:15:11 AM

Apologies if this has been covered before but what is the status on mining Blake algo with Nvidia vs AMD? AMD still much faster?

Thank you.

Titan X might beat 290X on this one.

Bitice

member

Activity: 66

Merit: 10

Apologies if this has been covered before but what is the status on mining Blake algo with Nvidia vs AMD? AMD still much faster?

Thank you.

mach123coin

newbie

Activity: 29

Merit: 0

Quote from: cbuchner1 on April 13, 2015, 02:55:17 AM

Quote from: Nubminer on April 12, 2015, 10:08:59 PM

I am trying to load the cudaminer.sln in visual studio Commmunity 2013.. will this work.
it fails every time so I am guessing not.. or maybe I missed a key step ?

the precompiled dependencies in the OP will only work for VS 2010.

You would have to build all the dependencies with VS 2013 if you want to use this newer IDE version.

Christian

Still seems to work with VS2010. Donated .01 LTC - transaction fee for what I mined for 5 days. Cudaminer on a EVGA GTX570/Q6600. Wish I had the faster cards. Best of luck and keep optimizing the code!

smolen

hero member

Activity: 524

Merit: 500

Quote from: sp_ on April 16, 2015, 07:37:55 AM

A PCIe 2.0 x16 bandwidth is 8 GB/s and this meens that at least 7.2% of the time is wasted with moving memory.

IIRC, calculations and bus I/O are independent and can go in parallel. PCIe bus is bidirectional, so the full bandwidth (up+down) should be twice as large.
But this chimerical AMD/NVidia monster going to be quite unstable

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: sp_ on April 16, 2015, 07:37:55 AM

Quote from: smolen on April 16, 2015, 07:10:52 AM

Quote from: sp_ on April 16, 2015, 04:32:25 AM

What if you have 4 gpu's in your rig and each thread is executed on a seperate gpu. x11 is then reduced to x2+.

Hmm, PCIe 2.0 x16 bandwidth is 8 GB/s, 100 MH/s (desired speed) * 64 bytes (internal state size) = 6.4 GB/s

x11 with the latest ccminer is currenly hashing at around 3MH/s (on the 750ti)

PCIe 4.0×16 has a 31.5 GB/s bandwidth. (3*4(GPU)) above 12 MH/s (desired speed) * 64 bytes (internal state size) = 0,576 GB/s

with 4 gpu's and 4 threads you need to copy the data 3 times. (the last gpu pass will only need to return the winning nounce.)

The copy process can be done in parallell 0,576 GB/s by dividing the pci-e bus timeline

-----------TimeLine--------------->
GPU1: --Copy--Calc--calc--Calc
GPU2: --Calc--Copy--Calc--Calc
GPU3: --Calc--Calc--Copy--Calc
GPU4: --Calc--Calc--Calc--Result

A PCIe 2.0 x16 bandwidth is 8 GB/s and this meens that at least 7.2% of the time is wasted with moving memory. So your Kernals needs to run at least 8% faster.

seems slow. Could work for a cryptonite miner since the desired speed is 500h-1KH/s

PCIe 4.0 ? (planned to be released late 2016, by that time there will be for sure x11 asic Grin

)
also you never have more than 2 gpu's at x16 (unless you go to the utterly expensive x99 setup, not mining friendly)

sp_

legendary

Activity: 2954

Merit: 1087

Team Black developer

Quote from: smolen on April 16, 2015, 07:10:52 AM

Quote from: sp_ on April 16, 2015, 04:32:25 AM

What if you have 4 gpu's in your rig and each thread is executed on a seperate gpu. x11 is then reduced to x2+.

Hmm, PCIe 2.0 x16 bandwidth is 8 GB/s, 100 MH/s (desired speed) * 64 bytes (internal state size) = 6.4 GB/s

x11 with the latest ccminer is currenly hashing at around 3MH/s (on the 750ti)

PCIe 4.0×16 has a 31.5 GB/s bandwidth. (3*4(GPU)) above 12 MH/s (desired speed) * 64 bytes (internal state size) = 0,576 GB/s

with 4 gpu's and 4 threads you need to copy the data 3 times. (the last gpu pass will only need to return the winning nounce.)

The copy process can be done in parallell 0,576 GB/s by dividing the pci-e bus timeline

-----------TimeLine--------------->
GPU1: --Copy--Calc--calc--Calc
GPU2: --Calc--Copy--Calc--Calc
GPU3: --Calc--Calc--Copy--Calc
GPU4: --Calc--Calc--Calc--Result

A PCIe 2.0 x16 bandwidth is 8 GB/s and this meens that at least 7.2% of the time is wasted with moving memory. So your Kernals needs to run at least 8% faster.

seems slow. Could work for a cryptonite miner since the desired speed is 500h-1KH/s

smolen

hero member

Activity: 524

Merit: 500

Quote from: sp_ on April 16, 2015, 04:32:25 AM

What if you have 4 gpu's in your rig and each thread is executed on a seperate gpu. x11 is then reduced to x2+.

Hmm, PCIe 2.0 x16 bandwidth is 8 GB/s, 100 MH/s (desired speed) * 64 bytes (internal state size) = 6.4 GB/s

sp_

legendary

Activity: 2954

Merit: 1087

Team Black developer

Quote from: ?? on ??

Quote from: sp_ on April 16, 2015, 04:32:25 AM

What if you have 4 gpu's in your rig and each thread is executed on a seperate gpu. x11 is then reduced to x2+.

advantages:
-Smaller kernals, bether register usage, less memory needed, more cache hits, more paralell threads
-Hybrid mining is possible. (run AES algos on the AMD, and the rest on NVIDIA)

disadvangtages:

-throughput must be passed from gpu to gpu trough the pci-E to memory and back.
-You need 4 gpu's (but the algorithm can be scalable to support x gpu's)

Terrible idea - implies a device-to-host copy followed by a host-to-device copy, or at least a device-to-device copy, if I'm understanding you. Terribly slow.

You are probobly right.

Perhaps faster with a crossfire cable and a 2 gpu setup

Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 21. (Read 3426976 times)