Amph
Legendary
Offline
Activity: 3248
Merit: 1070
|
|
June 14, 2014, 07:43:21 AM |
|
it seems that x13 isn't working with this version?
edit: nevermind, command line was fucked....
so now the consumption is higher too? because i noticed that temps went up
I didnt think x13 was implemented in this version? I tried putting in -a x13 and that was a no go. What parameters did you use? yeah you are right, it doesn't work, i confused the miner...
|
|
|
|
PVmining
|
|
June 14, 2014, 08:02:48 AM Last edit: June 14, 2014, 08:32:59 AM by PVmining |
|
I'm getting the following errors compiling ccminer v1.1 on linux. I'm using cuda6 libraries. Is it possible that the __shfl operation isn't in there? groestl_functions_quad.cu(258): error: identifier "__shfl" is undefined groestl_functions_quad.cu(263): error: identifier "__shfl" is undefined groestl_functions_quad.cu(268): error: identifier "__shfl" is undefined bitslice_transformations_quad.cu( : error: identifier "__shfl" is undefined bitslice_transformations_quad.cu(414): error: identifier "__shfl" is undefined I have zero experience with KopiemTu 1.3, but it was up and running JPC for me with the included ccminer. I tried to do all the compile commands on there and saw a bunch of these streaming by, as well as a bunch that said there was 'something defined but never called'. I tried to compile it and do it over and over with the same results... It would finish, and wouldn't complain at the end, but I couldn't get the cards to actually mine, they just sat there at 0kh/s. If I copied back over the old files and re-started it would go back to mining not using the latest version. If anyone gets it to work please let me know if you had to do anything special. Wanted to report back that it didn't work for me... Here were the steps that I did to compile (as provided to me by PVMining in a previous post). cd /opt/miners/ mv /opt/miners/ccminer /opt/miners/ccminer-[date] git clone https://github.com/cbuchner1/ccminer cd ccminer ./autogen.sh ./configure make ...it seems the "mine start" & "monitor" command does not work anymore with the new ccminer. jk_14 gave the hint we should re-use the old util.c because he modified it. but that doesn't help. buuuut - if you compile it the right way you could start it via command right out of the dir were the new ccminer is located: screen (you need this if you logged in via ssh, and like to close the console afterwards)cd /opt/miners/ccminer (location of the new miner)./ccminer -a algo -o stratum+tcp://pool -u x -p xmake sure you stopped the monitor & the kopiemtu mining command (mine stop / monitor stop) before. edit. one little hint. the previous ccminer was located ccminer-2014-05-20. the dir "ccminer" is/was only a softlink. you could check it with "ls -l /opt/miners/ccminer" and see were it points to. you could download (our just move) the new miner into "ccminer-2014-06-13" and make a softlink with this command "ln -s /opt/miners/ccminer-2014-06-13 /opt/miners/ccminer". before you have to delete the old "ccminer" softlink.
|
|
|
|
cbuchner1 (OP)
|
|
June 14, 2014, 08:48:27 AM |
|
Would it have remained faster by leaving it out and still just doing what you were doing before?
Why the change?
"It's faster in the end - this is what counts"." let that sink in.
|
|
|
|
PVmining
|
|
June 14, 2014, 08:51:46 AM |
|
Would it have remained faster by leaving it out and still just doing what you were doing before?
Why the change?
"It's faster in the end - this is what counts"." let that sink in. btw. many thanks! the numbers of the new ccminer let me smile... never thought I could reach that much mhashes without a huge hardware upgrade. wow!
|
|
|
|
cbuchner1 (OP)
|
|
June 14, 2014, 08:54:28 AM |
|
I'm getting the following errors compiling ccminer v1.1 on linux. I'm using cuda6 libraries. Is it possible that the __shfl operation isn't in there?
__shfl is there beginning with compute 3.0 capability
|
|
|
|
jack80
|
|
June 14, 2014, 08:57:54 AM Last edit: June 14, 2014, 09:18:56 AM by jack80 |
|
I with 1 nvidia 750ti can not do 8000 kh / s but maximum 5500 kh / s I overclocked GPU +135 +300 mem use your program with this configuration: ccminer35-50.exe -jackpot -o stratum+tcp://jpc.hashatme.com:3333 -u user -p pass Help. Ok sorry im understand , you use another algorythm ( nist5 ) and dont Jackpot algorythm . 5500 kh/s with jackpot on ccminer v1.1 its correct . I have understood correctly ? .
|
|
|
|
Amph
Legendary
Offline
Activity: 3248
Merit: 1070
|
|
June 14, 2014, 09:19:55 AM |
|
so i guess, now my rig just consume more electricity nice...lol
|
|
|
|
AizenSou
|
|
June 14, 2014, 10:42:56 AM |
|
I wake up late today because of Saturday. Check TAC and wonder how does the diff increase 30%. So it turns out that Christian is the culprit again. With this again we leave AMD users in the dusts. Feel so bad for them.
|
|
|
|
Parja
Newbie
Offline
Activity: 36
Merit: 0
|
|
June 14, 2014, 10:44:12 AM |
|
so i guess, now my rig just consume more electricity nice...lol
10% more electricity for 60% more hashes? I'll take that ratio. I wake up late today because of Saturday. Check TAC and wonder how does the diff increase 30%. So it turns out that Christian is the culprit again. With this again we leave AMD users in the dusts. Feel so bad for them. AMD also got a significant boost in nist5 with the 14.6beta drivers as well. Not as significant as ccminer v1.1, but still pretty good.
|
|
|
|
jpouza
Legendary
Offline
Activity: 2842
Merit: 1122
|
|
June 14, 2014, 10:56:49 AM |
|
wow new release kicking ass on quarkcoin, #1 pool statistics x11 getting 2.7MH/s per card
Tks!
|
|
|
|
cayars
|
|
June 14, 2014, 11:12:04 AM |
|
I wake up late today because of Saturday. Check TAC and wonder how does the diff increase 30%. So it turns out that Christian is the culprit again. With this again we leave AMD users in the dusts. Feel so bad for them. Don't feel to bad for the AMD users. They have a couple profitable algos that we don't have yet for nVidia like CryptoNight.
|
|
|
|
Amph
Legendary
Offline
Activity: 3248
Merit: 1070
|
|
June 14, 2014, 11:14:32 AM |
|
so i guess, now my rig just consume more electricity nice...lol
10% more electricity for 60% more hashes? I'll take that ratio. yes but everyone else has it now, so it's pointless, you just consume 10% more
|
|
|
|
AizenSou
|
|
June 14, 2014, 11:17:45 AM |
|
AMD also got a significant boost in nist5 with the 14.6beta drivers as well. Not as significant as ccminer v1.1, but still pretty good.
That info is around for a while but not a lot of people are willing to upgrade, because with 14.6 everything else is broken, except groestl and nist5 (some minor algos more which I don't remember). Don't feel to bad for the AMD users. They have a couple profitable algos that we don't have yet for nVidia like CryptoNight.
CryptoNight for GPU is not profitable as others algos. The only profitable cryptonight miner now is in hand of Christian, which he doesn't publish yet. And I don't think of many profitable algos for AMD which we don't have. Could you elaborate ? yes but everyone else has it now, so it's pointless, you just consume 10% more
No I don't think so. The majority of TAC miners are still AMD users. I know some monsters at the pools (over 600Mhs NIST5) and they are all AMDs. Those monsters are made up 30% of total nethash already.
|
|
|
|
djm34
Legendary
Offline
Activity: 1400
Merit: 1050
|
|
June 14, 2014, 11:36:28 AM |
|
AMD also got a significant boost in nist5 with the 14.6beta drivers as well. Not as significant as ccminer v1.1, but still pretty good.
That info is around for a while but not a lot of people are willing to upgrade, because with 14.6 everything else is broken, except groestl and nist5 (some minor algos more which I don't remember). Don't feel to bad for the AMD users. They have a couple profitable algos that we don't have yet for nVidia like CryptoNight.
CryptoNight for GPU is not profitable as others algos. The only profitable cryptonight miner now is in hand of Christian, which he doesn't publish yet. And I don't think of many profitable algos for AMD which we don't have. Could you elaborate ? yes but everyone else has it now, so it's pointless, you just consume 10% more
No I don't think so. The majority of TAC miners are still AMD users. I know some monsters at the pools (over 600Mhs NIST5) and they are all AMDs. Those monsters are made up 30% of total nethash already. actually I am pretty sure with some tweaking the amd could get to the level of nvidia with the new driver. (as it is the case for groestl) But as long as cg/sgminer/claymore dev will continue playing with old versions of the driver (and say the new one is broken) it won't happen
|
djm34 facebook pageBTC: 1NENYmxwZGHsKFmyjTc5WferTn5VTFb7Ze Pledge for neoscrypt ccminer to that address: 16UoC4DmTz2pvhFvcfTQrzkPTrXkWijzXw
|
|
|
Madychoux
|
|
June 14, 2014, 12:14:18 PM |
|
Would it have remained faster by leaving it out and still just doing what you were doing before?
Why the change?
"It's faster in the end - this is what counts"." let that sink in. Thanks again for your work and contribution to NVIDIA miners !
|
|
|
|
Perl++
Newbie
Offline
Activity: 28
Merit: 0
|
|
June 14, 2014, 12:36:57 PM |
|
Hi Christian. How many H/s (cryptonight) you got using your NVIDIA miner ?
|
|
|
|
cbuchner1 (OP)
|
|
June 14, 2014, 01:14:04 PM |
|
How many H/s (cryptonight) you got using your NVIDIA miner ?
trade secret wow. much hash.
|
|
|
|
AizenSou
|
|
June 14, 2014, 01:20:27 PM |
|
How many H/s (cryptonight) you got using your NVIDIA miner ?
trade secret wow. much hash. People Christian need to earn money in order to pay for the Munich Biergarten. It's ***king expensive there. So everyone who mentions about Cryptonight now will be banned. Me including of course. Sorry Christian I split the bean and now people whining you about the new miner. They should be happy with X13 and killer groestl.
|
|
|
|
Samuray
|
|
June 14, 2014, 01:46:43 PM |
|
How many H/s (cryptonight) you got using your NVIDIA miner ?
trade secret wow. much hash. I thank you very much for what you are doing for this comunity !!! Wunderbar!!!
|
|
|
|
leicas
Newbie
Offline
Activity: 9
Merit: 0
|
|
June 14, 2014, 02:01:37 PM |
|
So, does this seem correct for one 750ti?: -X11:2.6MH/s -Jackpot:5.5MH/s -NIST5: 8.4MH/s -Quark: 4.7MH/s -Groestl: 8MH/s (!!!!!! was 3.1MH/s before !!!!!) I used these numbers to generate following list for one 750ti: http://pastebin.com/xcFqtX9EStill old hashrates for X13. Want to see Christian's magic first Again, this list is only a snapshot for this moment, with current prices and difficulties, so lower your own estimates. Is there any way to get the source for your profit calculator ? I'm using a 770 and a 780ti so the number for a 750ti aren't quite relevant. (got ~8mh/s for nist5 with the new CCminer version and 3mh/s for X11 on my 770gtx, quite surprised it's doing worse than a 750ti by the way...) Also small tip for those having hashrate loss while doing cpu mining, on windows setting a "Real time" priority on ccminer.exe in Task manager helps a lot.
|
|
|
|
|