a bit slower but noticeable (2.5% - 3% slower)
but yes, it seems more stable. Im running latest drivers (384.94)

Rig Detail
OS      : Windows 10
CPU    : Intel Core i3-6100@3.70 GHz
VGAs  : 6 x ASUS Dual GTX 1060 3gb
Driver : 382.53

AfterBurner settings
PL       : 80%
Core    : +125
Mem   : -500

With x86 and -i 24, CWIgm gives me 102.43MH/s (avg 17.07MH/s per card)
With x64 and -i 24, CWIgm gives me 102.59MH/s (avg 17.09MH/s per card)

For me, there is no significant difference between x86 and x64.
I tried -i 25 and it crashed, just like sp-version

PS. skunksp-mod3 and -i 24 gave me 104MH/s for the same rig (17.33MH/s per card)

Rig Detail
OS      : Windows 10 X64
GPU's  : 3 x GTX 1070
Driver  : 382.05

Afterburner Settings
Core clock      :  +120
Memory clock :  -140
Power            :   85%

Performance
Intensity      : 25
Hashrate      : 26 mh/s to 27 mh/s  (per card)
Power Draw  : 115 W to 130 W   (per card)


I was getting the same hashrates and power draw with spmod's ccminer.

tested on 750Ti and K2200
it doesnt work, there are some unsupported instructions because it complains about "Ilegal instruction"

I'd like to be a tester too. Having 6x1060 & 2xRX580 running

My rig has 6 pcs of 1080 , hash rate about 221 MH/s.

It alway connect to port 6000, that is for hi-diff.

Do I need to change to port 2000 , lo-diff? and How?

Your miner is clearly bit better (2-4%) than the one im using at the moment (ccminer1.0-skunk) but im bit annoyed by the pool restriction. Could you just add the current pool fee .7% to the miner and lose the pool restriction?

our entire network was disconnected for most of the night ...

tanx to the dumbass isp we have ( which 'should' be changing to a better one in two days time ) - our networks ( yes - ALL of them ) continuously disconnect due to dynamic ip changes ... imbeciles set us up on this type of internet connections even AFTER we specifically asked for a set of points to be strictly adhered to on our connection ... seriously deficient these blasted telcos ...

ok - bitch is over ... we are back ...

changed the stratum diff for the lodiff port 2000 - to max out target at 2.5 - it should go no higher than that ... the hidiff port 6000 is actually not in use with this current version CWI-gm-0.9.7 due to its locking on port 2000 ... sending the stratum into a seesaw rollercoast in diff ...

some issues should be fixed on that end the scale ...

hope you will come back soon - as we have a few changes in the next version ( for port adjustment and a couple of other parameters - to name a few ) ...

#crysx

hi wolf0 ...

well - sp is sp ... and does what he does for his own greedy benefit ... we all know that ...

much respect to you mate - after all you have for the community AND us ... if you decide you wish to come on board and help with an amd miner - we would luv to do it ... just like old times ...

we could always cut a deal for a bin that is tied to this test ( and pool ) if it can be done mate ... but as always - a faster sgminer bin is always welcome anyway - even though we dont really focus on amd much anymore - we would luv to help the community in that way also ... maybe bring some amd miners to the pool also ...

let me know mate ...

#crysx

hi all ...

we have been surprised and absolutely ecstatic about the response for the beta test from this awesome community of miners ...

we appreciate it very very much ...

we are working diligently to rectify a few issues that have risen - as we were not expecting an influx of testers land interest as you have shown ...

there will be a small update to the miner that will be released in the coming hours - which will focus on improving the usability and functionality of the miner ... stability is still the biggest factor here and as such - default intensities and a couple of other tweaks have been made to the miner ...

we will also release a specific compute50 version of the miner - to test those older maxwell cards ... we had huge numbers of the gigabyte 750ti oc lp in our theFARM before it got blown to bits - so these compute50 cards are an interesting test in itself ...

we are pleasantly surprised at some of the results and stability that you are getting from the results and settings you have posted - and are equally intrigued by the crashes and lockups that seems to occur with the few ... please keep posting these figures - as these stats are VERY important for the growth of CWIgm ...

the amount of pm we have had to join the beta test is massive ... and again - we are humbled by the interest you have all shown ... which is also the reason we have been holding back a little with posting the link to the current miner ...

CWIgm-0.9.7.1 will be released with some of the changes ive mentioned in the coming hours ... but we really do need a different way of pushing this miner out to the current and future beta testers ...

till then - we will continue to go through each and every pm we ( i ) get ...

tanx ...

#crysx