Topic: [ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX] - page 955. (Read 3426921 times)

Can't quite get this working on my GTX690 on both GPU's at once.
If I use either one individually it works just fine, but as soon as I add the other GPU it drops to 1-5 khash/s

These command line options are good for 330+ khash/s on either of the GPU's seperately:
cudaminer --algo=scrypt -d 0 -H 1 -C 1 -m 1 -l K8x32  -i 1

I did have the vertcoin miner working fine on both GPU's but this is the first time trying Cudaminer from source.

Any ideas why both GPU's can't be used at the same time?

Honestly I tried almost every other combination... And yes the x86 does give better result... Any idea why my one card is getting less hps? I cant figure it out...

This is due to the new Titan kernel that the Nvidia engineer submitted. S/He stated that the best config for it will be SMX x 24. This is why this setting is preferred.
780    -l T12x24
Titan  -l T14x24
780Ti  -l T15x24

Please try all -H -C and -m settings before you write the changes off. This is KEY with this release.

I went from over to (commit 114 to release) saw a very little kh/s change but found that the core could stably OC another 50MHz improving kh/s due to more OC head room

Got a big increase in kh/s with my GTX 660TI's went from:

Nvidia non-oc'd gtx 660ti= Stays ~70c @50% fan speed

230 kh/s to 280 kh/s in new version x86 with config: -d 0 -l K28x8 -H 1 -C 1 -m 1 -i 0

Evga 660TI FTW Sig2 OC'd= Stays ~74c @ 55% fan speed

270kh/s to 315kh/s in new version x86 with config: -d 1 -C 1 -i 0 -m 1 -H 1 -l K14x20

Thank you for updating OP, your awesome!

That's like the 15x24 for some 780ti... yes me too I'd be curious to know... Is there different models (beside factory bios) ?

Thx for the 12x20, was driving me nuts haha. Good speed increase, went from 500 kh to 600 kh with a 780 going from 12-18 to newest version. Seems to draw less power to boot! Time for some more overclockiiiiiiing
And x86 kernel works a lot better now than x64 for me, earlier it was a small difference favoring x64, 10-20kh+ difference now though between them.

Also, what's up with people getting good results on their 780s using T/Z12x24 ? I get absolutely craptastic results with that!

On GTX 550 Ti I was getting 78 kh/s with old version but with new version I am getting 65 kh/s

550 kh/s on gtx780 (500@old version) and 300 kh/s on gtx670 (250@old version) (100mhz OC on both cards, can do 200mhz but don't want to stress them too much). Nice update! Any good settings for gtx570 anyone? Currently running with F15x16 with same speed as before, ie. 250 kh/s.

Not sure if the new version pumped up my hash or just the new config setting i saw (which people used to the 18/13 version) but now i'm pumping 306kh/s with GTX 760

oops my bad...  That does better the performance
but still lower then before, now about 15%, but I guess i could get the numbers up by experimenting with the -l F
a little more.  The autotune probably was better at getting the optimal F.  Lets see how it goes when -l auto is fixed.

GTX 780 Water cooled.

My old 12-18-2013 Cudaminer with this Bat.

-H 1 -i 0 -C 0 -m 1 -l T18x32 -o stratum+tcp://

Around 550-560 Khs.

New 2-2-2014 Cudaminer with this Bat...With 100+ to GTX 780 Core

-d 0 -H 1 -C 1 -m 1 -l Z12x24 -i 0 -o stratum+tcp://

Around 640-650 Khs

A nice bump. Now its so close to my 700Kh/s R9 280x rig.

Just making sure, it's -l F not -i F.

-i sets your intensity to on (1) or off (0).

I have 2.1 compute capability and with the -i F  I get a little better performance but still about 30% less then before.

thank you sir!  Appreciate your good work.

-l auto is currently broken. will post a fix in a few hours. you will want to pass -l F   -l K   or -l T  depending on what
generation your card is (Fermi, Kepler, Titan/GTX780 class)

Im getting a 30% slowdown in kh from the older build (december 2013) my settings are -H 2 -C 2 -i 1 -l auto with scrypt

Good afternoon,
I am new to this forum so please go easy on me, been lurking looking on the website for sometime and until recently have been playing about with cudaminer.
Firstly I would like to say a big thank you for Christian for all of his hard work and effort on this on going project.

I have been reading the new comments from other members about the new build and it seems a lot of people are having problems with it I have installed it today and have been having great success with it gaining a improvement of approx. 40-50 khash/s over the 2013-12-18 build

EVGA 780 ti Sc default settings averaging 690 khash/s mining litecoin & dogecoin , no overclocking no voltage increases . , this is using x86 launch . I noticed that using x64 I loose typically 30-40 khash/s. I have also had a little time to play around with overclocking using the evga software and with power target increase 106% and gpu clock offset increase I have no doubt I can reach 800 khash/s using stock voltages not run it for long as I am new to this gpu and still unsure whats a safe overclock for it anyone know what the safe limit for gpu offset for the 780 ti sc?

cudaminer.exe -d 0 -H 1 -i 0 -C 0 -l T15x24 -m 1 -o

sorry again if I come across as a newbie all these settings have been used from trail and error if there are better command line for my setup I would appreciate any comments or tips

Leafcoin, exact same settings (apart from the scrypt-jane to scrypt change ofcourse):

@kokes

how much do you get with scrypt mining?

I'm having the exact same card as you (but overclocked). I have no clue why you have a slowdown, and i get a boost :O
Also that -i 10 was a typo in my config, but too lazy to change, but it does not give a difference.

What if you try mining ultracoin, do you still only get 224 khash/s? These are my OC settings if someone is interested: