Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 1132. (Read 2347601 times)
@chrysophylax: I am considering changing over from kopiemtu, as the developer seems to have left the project, and compiling with each new ccminer release was very difficult and wrought with errors. I am considering using VNC or Teamviewer and throwing a boot HDD onto each of my two Maxwell rigs, so I can still run them headless.
Thanks alot. i have added scrypt-n to my fork now. Copied it from the tvpruvot branch, so no modding yet
@ sp
0.3 btc coming your way
Transaction ID: e5079c35cc8255fe67100d0a9f1ed23b565c85ca4449ee1466e9453472e5de33-000
You did some great work although I hardly ever used your x algo code so far.
I'm looking forward to your scrypt-n implementation and of course your BCR implementation
I hope some others follow up and donate some as well. It seems like the last few weeks were slightly more profitable for all of us and sp_ definitely deserves some donations for his great work.
0.3 btc coming your way
Transaction ID: e5079c35cc8255fe67100d0a9f1ed23b565c85ca4449ee1466e9453472e5de33-000
You did some great work although I hardly ever used your x algo code so far.
I'm looking forward to your scrypt-n implementation and of course your BCR implementation
I hope some others follow up and donate some as well. It seems like the last few weeks were slightly more profitable for all of us and sp_ definitely deserves some donations for his great work.
Looks like the kernal code from cudaminer has beeen copied cleaned and adapted to ccminer standards. The performance should be equal to cudaminer.
I see some possible optimizing here. I might add it to my fork tonight.
I see some possible optimizing here. I might add it to my fork tonight.
sp - i thought you would like to know that v45 crashed last night on the farm in quark algo ...
no parameters were set ... just stock standard ... v44 under x11 runs continuously - though i havent tried it out under quark ...
tanx again mate ...
#crysx
Looks like the kernal code from cudaminer has beeen copied cleaned and adapted to ccminer standards. The performance should be equal to cudaminer.
I see some possible optimizing here. I might add it to my fork tonight.
I see some possible optimizing here. I might add it to my fork tonight.
Looks like scrypt-n has been added in the other branch by tpruvot
ccMiner release 1.6.2-tpruvot (May 2015) - "Scrypt algos (dev)" (github)
ccMiner release 1.6.2-tpruvot (May 2015) - "Scrypt algos (dev)" (github)
Does it perform better on higher vram cards?
are there any linux guys running kopiemtu here?
run it once and discarded it ...
it IS a good package but it does not 'fit in' with what we do with our farm ...
we use our own setup and distro - so can help a little with more generic linux questions if you have some ...
didnt help much there - but just so you know
...#crysx
Looks like scrypt-n has been added in the other branch by tpruvot
ccMiner release 1.6.2-tpruvot (May 2015) - "Scrypt algos (dev)" (github)
ccMiner release 1.6.2-tpruvot (May 2015) - "Scrypt algos (dev)" (github)
are there any linux guys running kopiemtu here?
Im getting : does not valid on the cpu error !!
im using the last version ccminer-1.5.45-git-spmod (ive tried last one too)
is there a way to reduce those or repair the problem (lower intensity or idk) and what does it mean, my cpu is like at 3% runnin ??
Gtx 970 X2
16 bg ram
ccminer -i 18,18.5 -o stratum+tcp://us-east.multipool.us:11111 -p x -u xxx.xxx
Thanks for all you guyz doin a real good job !!
Windows 7 Édition Intégrale 64-bit SP1
AMD FX-8150 4.2ghz 32 °C
Technologie Zambezi 32nm
16,0 Go Canal-Double DDR3 @ 802 MHz
ASUSTeK COMPUTER INC. SABERTOOTH 990FX R2.0 (Socket 942) 25 °C
ASUS VG278HR (1920x1080@120Hz)
4095 MoNVIDIA GeForce GTX 970 (EVGA) 62 °C
4095 MoNVIDIA GeForce GTX 970 (EVGA) 56 °C
Version ForceWare: 350.12
111GB INTEL SS DSC2CT120A3 SATA Disk Device (SSD)
2794GB Seagate ST3000DM 001-9YN166 SATA Disk Device (SATA) 25 °C
DTSOFT Virtual CdRom Device
ASUS DRW-24B1ST c SATA CdRom Device
Realtek High Definition Audio
im using the last version ccminer-1.5.45-git-spmod (ive tried last one too)
is there a way to reduce those or repair the problem (lower intensity or idk) and what does it mean, my cpu is like at 3% runnin ??
Gtx 970 X2
16 bg ram
ccminer -i 18,18.5 -o stratum+tcp://us-east.multipool.us:11111 -p x -u xxx.xxx
Thanks for all you guyz doin a real good job !!
Windows 7 Édition Intégrale 64-bit SP1
AMD FX-8150 4.2ghz 32 °C
Technologie Zambezi 32nm
16,0 Go Canal-Double DDR3 @ 802 MHz
ASUSTeK COMPUTER INC. SABERTOOTH 990FX R2.0 (Socket 942) 25 °C
ASUS VG278HR (1920x1080@120Hz)
4095 MoNVIDIA GeForce GTX 970 (EVGA) 62 °C
4095 MoNVIDIA GeForce GTX 970 (EVGA) 56 °C
Version ForceWare: 350.12
111GB INTEL SS DSC2CT120A3 SATA Disk Device (SSD)
2794GB Seagate ST3000DM 001-9YN166 SATA Disk Device (SATA) 25 °C
DTSOFT Virtual CdRom Device
ASUS DRW-24B1ST c SATA CdRom Device
Realtek High Definition Audio
For whirlpoolx,
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
You try my bat?
Yea man, I was missing the --cpu-priority parameter, but with that, it runs the cards in two's, like the first and fourth, or first and third and so on.
Quark algo is doing fantastic - between 31550 and 31600 with 5 x 750Ti.
Anyone know what coin it is: block 1544717?
Anyone know what coin it is: block 1544717?
For whirlpoolx,
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
You try my bat?
Yea man, I was missing the --cpu-priority parameter, but with that, it runs the cards in two's, like the first and fourth, or first and third and so on.
For whirlpoolx,
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
You try my bat?
For whirlpoolx,
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
The '-d' parameter is broken when using alongside with '-g 2'. So that speed increase only works when using all the cards together, but it can't run individual cards with '-g 2'. And running all the cards together gives a lot lower performance because of overall lower shares.
Quote from: scryptr
WHIRLPOOLX --
The Whirlpoolx algo runs at 464.6Mh/s without the gputhreads switch, and over 510Mh/s with "-g 2". Before, with v44, my six 750ti rig was getting 448Mh/s.
my 3x960STIX_OC befor 410Mh, after 770Mh... The Whirlpoolx algo runs at 464.6Mh/s without the gputhreads switch, and over 510Mh/s with "-g 2". Before, with v44, my six 750ti rig was getting 448Mh/s.
sp - ive reverted back to just using no other parameters as it seems to be quite unstable and crashes a great deal when using combinations ( and singularly ) the -g and -i parameters ...
if you look at the westhash test that is being performed ... it is stable and pretty constant ( apart from the dips from the different coins that get mined by renters ) ...
ill play with the settings a little later tomorrow when i get back into the office ...
tanx for the great work though ...
#crysx
if you look at the westhash test that is being performed ... it is stable and pretty constant ( apart from the dips from the different coins that get mined by renters ) ...
ill play with the settings a little later tomorrow when i get back into the office ...
tanx for the great work though ...
#crysx
Look on WhirpoolX whith -g 2... +80%
WHIRLPOOLX --
The Whirlpoolx algo runs at 464.6Mh/s without the gputhreads switch, and over 510Mh/s with "-g 2". Before, with v44, my six 750ti rig was getting 448Mh/s. --scryptr
Thanks for sharing SP, I tried with remote from home but ccminer crashed a lot, will try when at office..
Look on WhirpoolX whith -g 2... +80%
Jump to: