[ANN] cudaMiner & ccMiner CUDA based mining applications [Windows/Linux/MacOSX]

[djm34]

so it seems the psu i am using on my mining cards is dieing, the noise is horrible, time to order a new psu

the 2 months old corsair RM ?

[bigjme]

so it seems the psu i am using on my mining cards is dieing, the noise is horrible, time to order a new psu

the 2 months old corsair RM ?

no, i replaced my old xfx psu with the RM1000, used the xfx to power my GPU's
the fan on the xfx has started doing the typical "im about to die" noise

[DougB62]

donations for ccMiner

in JPC: JYFBypVDkk583yKWY4M46TG5vXG8hfgD2U
in MNC MShgNUSYwybEbXLvJUtdNg1a7rUeiNgooK

consider these the official donation addresses

100K JPC Your way!

[THeZoiD]

ok now doing 8.3mh/s with gtx770
 

Are you using x86 or x64 ? What are your clocks ? I can't get past 4.7 Mh/s with mine

yeah sorry my misstake...the pool was showing 8.3 but my miner was showing 4.8

o got +109 gpu offset and +361 mem offset

my card is evga gtx770 oc acx

[bitcoinvideos]

v 0.7 from a few days ago compiles fine on Ubuntu...I've ran autogen and then compile then make with the same results.

okay, last lines in Makefile.am should say this instead.

Code:

# ABI requiring code modules
# this module doesn't compile with Compute 2.0 unfortunately
JHA/cuda_jha_compactionTest.o: JHA/cuda_jha_compactionTest.cu        
        $(NVCC) -I . -I cudpp-2.1/include @CFLAGS@ -Xptxas "-abi=yes -v" -gencode=arch=comp

it was a copy&paste from our internal SVN repo where we use a different filename, sorry.

Ok same errors (also don't you dare apologize you guys f-ing rock!):

 -c JHA/cuda_jha_compactionTest.cu
JHA/cuda_jha_compactionTest.cu(117): error: identifier "__shfl_up" is undefined

JHA/cuda_jha_compactionTest.cu(144): error: identifier "__shfl_up" is undefined

2 errors detected in the compilation of "/tmp/tmpxft_00004cc1_00000000-13_cuda_jha_compaction                     Test.compute_20.cpp1.ii".
make[2]: *** [JHA/cuda_jha_compactionTest.o] Error 2
make[2]: Leaving directory `/home/nick/ccminer'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/nick/ccminer'
make: *** [all] Error 2

[cbuchner1]

Ok same errors (also don't you dare apologize you guys f-ing rock!):

ain't that bizarre. Compiled on my Ubuntu 12.04 after making said change and running ./autogen.sh

[bitcoinvideos]

Sorry I figured out the problem...I removed the entire last line instead of just replacing the few changes you listed...

[AizenSou]

donations for ccMiner

in JPC: JYFBypVDkk583yKWY4M46TG5vXG8hfgD2U
in MNC MShgNUSYwybEbXLvJUtdNg1a7rUeiNgooK

consider these the official donation addresses

10% increase for Titan
30% for Maxwell.

Not bad at all Christian. But to be honest I did expect more . Anyway we have another cuda coin again.

Send 1mil JPC to your address Tell me if you don't get it.

Thanks again for your hard-work.

Regards,

[cbuchner1]

10% increase for Titan
30% for Maxwell.

Not bad at all Christian. But to be honest I did expect more . Anyway we have another cuda coin again.

Send 1mil JPC to your address Tell me if you don't get it.

yes got it. Made my total go from 11 mil to 12 mil. Thanks.

did you adjust the MAXWELL_OR_FERMI flag in the quark/cuda_groestl512.cu module? The Titan wants the flag at 0, Maxwell wants it at 1. Groestl is still
part of Jackpot, but in our implementation only in the 3rd round

in my super secret lab my 3 780 Ti cards currently pull this performance

Code:

[2014-05-05 23:56:28] 3 miner threads started, using 'jackpot' algorithm.
[2014-05-06 00:23:18] accepted: 1/1 (100.00%), 56563 khash/s (yay!!!)

[Steve23J]

please help ... only version working for my old gtx 580 is v0.4 ... getting around 7 mhs on groestl ... is this normal?

[Bearclaw]

on it now. if anyone can find it, can you post the change i need to make for groestl that christian posted? im too lazy xD but i will write it down this time, i think compute 3.0 was setting it to 0
i will post a reply to this will all the versions, so we want compute 3.5, 3.0, 2.0, anything else?

I know it's a bit late but any chance for a compute 5.0 version?  Or am I completely blind and miss it in a post somewhere...

[bitcoinvideos]

10% increase for Titan
30% for Maxwell.

Not bad at all Christian. But to be honest I did expect more . Anyway we have another cuda coin again.

Send 1mil JPC to your address Tell me if you don't get it.

yes got it. Made my total go from 11 mil to 12 mil. Thanks.

did you adjust the MAXWELL_OR_FERMI flag in the quark/cuda_groestl512.cu module? The Titan wants the flag at 0, Maxwell wants it at 1. Groestl is still
part of Jackpot, but in our implementation only in the 3rd round

in my super secret lab my 3 780 Ti cards currently pull this performance

Code:

[2014-05-05 23:56:28] 3 miner threads started, using 'jackpot' algorithm.
[2014-05-06 00:23:18] accepted: 1/1 (100.00%), 56563 khash/s (yay!!!)

Holy smoke balls!

[AizenSou]

please help ... only version working for my old gtx 580 is v0.4 ... getting around 7 mhs on groestl ... is this normal?

You got better than me. Version 5 gives me only 5.8Mhs. Maybe I will try ccminer0.4 too. Thanks for this tips

yes got it. Made my total go from 11 mil to 12 mil. Thanks.

did you adjust the MAXWELL_OR_FERMI flag properly? The Titan won't need it, Maxwell will. Groestl is still
part of Jackpot, but in our implementation only in the 3rd round

in my super secret lab my 3 780 Ti cards currently pull this performance

Code:

[2014-05-05 23:56:28] 3 miner threads started, using 'jackpot' algorithm.
[2014-05-06 00:23:18] accepted: 1/1 (100.00%), 56563 khash/s (yay!!!)

Sorry I didn't see that. I only deleted the HEAVY part in parse_config and parse_cmdline because that bug causes ccminer can't read from config file. I will check the flag later then. My 780 gives me only 8Mhs. And jackpot still doesn't work with Fermi, right?

Btw, Ich hoffe dass du damit eine Kiste von König Pilsener holen kannst, oder?

[cbuchner1]

Btw, Ich hoffe dass du damit eine Kiste von König Pilsener holen kannst, oder?

In Munich we prefer local beers like Augustiner, Franziskaner, Paulaner

[bigjme]

on it now. if anyone can find it, can you post the change i need to make for groestl that christian posted? im too lazy xD but i will write it down this time, i think compute 3.0 was setting it to 0
i will post a reply to this will all the versions, so we want compute 3.5, 3.0, 2.0, anything else?

I know it's a bit late but any chance for a compute 5.0 version?  Or am I completely blind and miss it in a post somewhere...

the compute 3.5 version will work on compute 5.0

[cbuchner1]

the compute 3.5 version will work on compute 5.0

before compiling, set all occurences of #define MAXWELL_OR_FERMI

to 1 then it runs a good chunk faster on Maxwell.

[bigjme]

before compiling, set all occurences of #define MAXWELL_OR_FERMI

to 1 then it runs a good chunk faster on Maxwell.

it comes with it set to 1 as default so i just compile it compute 3.0 and 2.0 i change it to 0

[AizenSou]

Btw, Ich hoffe dass du damit eine Kiste von König Pilsener holen kannst, oder?

In Munich we prefer local beers like Augustiner, Franziskaner, Paulaner

Achso, so du kommst aus Muenchen Ich mag Augustiner auch.

Ok time to report. With MAXWELL_OR_FERMI = 0 flag my 780 stays at 7.2Mhs. A tiny increase from 7.1Mhs, but it's difficult to tell because the hashrate still fluctuates so much.
I will try to optimize tomorrow then.
G9.

[Spoetnik]

Any chance you could compile it for compute 3.0 ? Thanks

on it now. if anyone can find it, can you post the change i need to make for groestl that christian posted? im too lazy xD but i will write it down this time, i think compute 3.0 was setting it to 0
i will post a reply to this will all the versions, so we want compute 3.5, 3.0, 2.0, anything else?

you know you were mighty quick to call me arrogant before

little note bud.. __shfl_up does not exist on Fermi class GPU's (compute 2.0 devices)
so no amount of screwing around compiling will ever mean anything.. it won't work it's as simple as that.
chubner knows this VERY well too.. you can't find any docs online that talk about shuffle with a lot of warnings saying pre-Kepler cards can't do it.

game over for guys running various 5xx cards when they put v7 of ccminer out they put a note on GitHub saying the dropped Fermi and 64bit support.
So we have to find something else to do with our cards now.. hopefully the same procedure down the line does not keep happening too. (device support dumping)
be kind off gay to be dropping support for older cards in favor of a faster speed on the newest ones only..

I'm out of my league on Cuda compared to these guys though i have not invested the time and energy learning it like these guys have.
you also need a propensity for math i think and i sux at math

All i know is it's a drag hearing about guys running Nvidia cards doing 20mh's or what ever when mine won't hash at all

[dga]

little note bud.. __shfl_up does not exist on Fermi class GPU's (compute 2.0 devices)
so no amount of screwing around compiling will ever mean anything.. it won't work it's as simple as that.
chubner knows this VERY well too.. you can't find any docs online that talk about shuffle with a lot of warnings saying pre-Kepler cards can't do it.

game over for guys running various 5xx cards when they put v7 of ccminer out they put a note on GitHub saying the dropped Fermi and 64bit support.
So we have to find something else to do with our cards now.. hopefully the same procedure down the line does not keep happening too. (device support dumping)
be kind off gay to be dropping support for older cards in favor of a faster speed on the newest ones only..

I'm out of my league on Cuda compared to these guys though i have not invested the time and energy learning it like these guys have.
you also need a propensity for math i think and i sux at math

All i know is it's a drag hearing about guys running Nvidia cards doing 20mh's or what ever when mine won't hash at all

To save C&C the time of responding, since I haven't done much for the nvidia crowd lately:

__shfl is fundamental to some of the optimizations that are being done for newer cards, and it makes it possible to take very different design approaches than is possible if you have to go through shared memory.

There aren't a lot of other API changes that are as important for crypto.  I doubt any of us will stop supporting Kepler/Maxwell any time soon, but dropping Fermi is probably the right thing to do for the long term.