CPU mining getting nowhere with this coin now. Please make an NVIDIA miner, I have 6x 750Ti's waiting to mine this coin! Kind regards
pay me lol
it's a lot of work your demanding lol
i looked at it and it's going to be a big pain the ass and i don't feel like spending sick amounts of time screwing with it..
maybe go ask chubner on his cudaminer topic if he's working on it
anyway, i am happy to report back, i have gone 12+ hours uninterrupted at over 1mh on my 7950
In that time i had no stales rejects no hardware errors and i overclocked the shit out of my card too lol
using an older version of the driver but i'm going to try 12.6 final i just downloaded soon.
i want to point out what i think is weird behavior for the 2 different sgminer mods i tried..
normally on scrypt and scrypt-n thread-concurrency has a direct correlation to how much dedicated Dram is used on your card.
and it does NOT on marucoin !
which make me think something weird is hard coded inside and THAT is what causing crashing or weird behavior etc.
I am fairly convinced the coding done for that kernel or what ever sucks ass or could be improved if i wanted to put it nicely (or if my name was not Spoetnik) lol
Normally i can creep up my TC and watch GPU-Z as it increases the mem usage up until i get the max of my card which is 3gb's
and then when it goes over the miner won't run.. of course
Vs. marucoin mining from low to high settings it's always stuck at 523mb / 525mb dedicated dram usage. wtf ?
So i don't know if it's the reason why but when i set EXACTLY
"thread-concurrency" : "1792",
"shaders" : "0",
After trying Toooooons of settings i finally got stability and an obvious boost in speed etc.
and i stress.. i tried TONS of settings of TC like crazy for days when i put in 1792 bingo i had smooth sailing
i will post my config that i think is working really stable and tested perfect for 12 hours but i want to make a couple quick notes..
I think scan time at 1 is too short.. your cutting of the workload and asking for new work too soon never mind the fact your flooding the servers maybe with requests for work
when i changed it to
4 i was getting a lot more consecutive shares submitted before a new block was detected in the console.
i know from working on miner code that the hashing routines use the scan time setting to cut off hashing and stop and ask for work
it's a time limit.. setting it too short can do more harm than good sometimes ..based on the coin and difficulty that changes we may need to adjust it occasionally.
a key aspect to my config is i decided to optimize for two gpu-threads and i normally have always used one thread with far higher settings.
but i felt it would be more effective for X13 mining if i went with more threads.. i tried 4 threads and 2 was better i think..
BUT i also have been using Queue at 2 also as seen in my config below.
anyway here is my config i hope it helps people with tweaking or whatever
used on my PowerColor PCS+ AX7950 running a stock bios and stock voltage. w/ WIN 7 x64 (card dedicated for mining / Intel Haswell Integrated for desktop)
This is giving me a steady and clean 1.11MH/s on marucoin @ 54 degrees
Almost forgot this was running on Badman's compiled sgminer build with a difficulty mod but i was NOT using the diff modifier command-line.
Also my CPU miner is leaving two free threads and my PC ram is like,
GSkill Trident X DDR3 16gb kit ram using an XMP OC profile with my Z87 mobo at 1200mhz (DDR3-2400) 10-12-12-31 CR2 / 1.648v
no auto cgminer crap either.. it's all set manually how i want it
{
"pools" : [
{
"url" : "stratum+tcp://maru.suprnova.cc:3131",
"user" : "Spoetnik.1",
"pass" : "x"
}
]
,
"intensity" : "14",
"vectors" : "1",
"worksize" : "128",
"lookup-gap" : "2",
"thread-concurrency" : "1792",
"shaders" : "0",
"gpu-engine" : "1000",
"gpu-fan" : "50",
"gpu-memclock" : "1500",
"gpu-memdiff" : "0",
"gpu-powertune" : "0",
"gpu-vddc" : "0.000",
"temp-cutoff" : "95",
"temp-overheat" : "85",
"temp-target" : "75",
"api-mcast-port" : "4028",
"api-port" : "4028",
"no-submit-stale" : true,
"expiry" : "10",
"gpu-dyninterval" : "7",
"gpu-platform" : "0",
"gpu-threads" : "2",
"hotplug" : "5",
"log" : "5",
"no-pool-disable" : true,
"queue" : "2",
"scan-time" : "4",
"temp-hysteresis" : "3",
"shares" : "0",
"kernel-path" : "/usr/local/bin"
}
edit:
I used a hex comparison program to compare the opencl .dll's from
AMD_Catalyst_13.11_BetaV9.5 - to - 12.6 final whql
and they are identical !
