CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 1209.

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: chrysophylax on December 18, 2014, 10:50:43 AM

Quote from: sp_ on December 17, 2014, 05:32:11 PM

Time to make a neoscrypt port for NVIDIA. Can we beat the opencl code by WOLF0?

(nicehash)

NeoScrypt 3341 0.2021 GH/s 4.7595 BTC/GH/Day 738% 3.5289 BTC/GH/Day 5.2900 BTC/GH/Day 5.2900 BTC/GH/Day View orders

now this would be VERY nice ... Wink

i am changing ALL the cards to 750ti OC ( Gigabyte to be exact ) and will be testing it with the initial 11 cards i have currently ...

i will be more than willing to guinea pig and test neoscrypt on these cards the moment you have something out ...

#crysx

not to sound greedy, but donations are a better incentive than guinea pig Grin

chrysophylax

legendary

Activity: 2912

Merit: 1091

--- ChainWorks Industries ---

Quote from: sp_ on December 17, 2014, 05:32:11 PM

Time to make a neoscrypt port for NVIDIA. Can we beat the opencl code by WOLF0?

(nicehash)

NeoScrypt 3341 0.2021 GH/s 4.7595 BTC/GH/Day 738% 3.5289 BTC/GH/Day 5.2900 BTC/GH/Day 5.2900 BTC/GH/Day View orders

now this would be VERY nice ... Wink

i am changing ALL the cards to 750ti OC ( Gigabyte to be exact ) and will be testing it with the initial 11 cards i have currently ...

i will be more than willing to guinea pig and test neoscrypt on these cards the moment you have something out ...

#crysx

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: Epsylon3 on December 18, 2014, 10:34:23 AM

i was about to say : its in lyra2_cpu_hash_32

and i think you can even remove the line Wink

else, no impact on hashrate

indeed much better (I did saw that, but tried to remove all the (my)cudasynchronize and it didn't change anything, putting them back works) actually.

Epsylon3

legendary

Activity: 1484

Merit: 1082

ccminer/cpuminer developer

i was about to say : its in lyra2_cpu_hash_32

and i think you can even remove the line Wink

else, no impact on hashrate

Schleicher

hero member

Activity: 675

Merit: 514

Quote from: djm34 on December 18, 2014, 07:24:18 AM

did a rapid test with lyra2, the cpu usage is awfully high (I have a noisy cpu cooler...) 50% usage... against 5~7% in my version.

what did you do ? Grin

I was thinking, may-be I could use tpruvot interface, but 50% cpu usage is a big red flag for me... so please fix this...
(not the case for x11... though, but I don't use x11...)

Using cudaDeviceSynchronize() was probably not a good idea.
Fix:
https://github.com/KlausT/ccminer/commit/2a807de0c38c3e34f64bc947832a8567e411f57c

Epsylon3

legendary

Activity: 1484

Merit: 1082

ccminer/cpuminer developer

m7 has been... skipped Wink

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: Bombadil on December 18, 2014, 07:37:47 AM

Quote from: djm34 on December 18, 2014, 07:24:18 AM

did a rapid test with lyra2, the cpu usage is awfully high (I have a noisy cpu cooler...) 50% usage... against 5~7% in my version.

what did you do ? Grin

I was thinking, may-be I could use tpruvot interface, but 50% cpu usage is a big red flag for me... so please fix this...
(not the case for x11... though, but I don't use x11...)

You should merge his interface and api Tongue

I'm sure that they aren't the ones causing the CPU usage, but I've noticed it too on lyra2.

it should come from the interface I use the same code in the legacy interface and the cpu usage is normal...
Main problem for me to move in (beside cpu usage), is the fact that m7 has been ditched Roll Eyes

and I would need to rewrite everything (again) again on the interface, and that was particularly painful the first time...

Bombadil

hero member

Activity: 644

Merit: 500

Quote from: djm34 on December 18, 2014, 07:24:18 AM

did a rapid test with lyra2, the cpu usage is awfully high (I have a noisy cpu cooler...) 50% usage... against 5~7% in my version.

what did you do ? Grin

I was thinking, may-be I could use tpruvot interface, but 50% cpu usage is a big red flag for me... so please fix this...
(not the case for x11... though, but I don't use x11...)

You should merge his interface and api Tongue

I'm sure that they aren't the ones causing the CPU usage, but I've noticed it too on lyra2.

djm34

legendary

Activity: 1400

Merit: 1050

did a rapid test with lyra2, the cpu usage is awfully high (I have a noisy cpu cooler...) 50% usage... against 5~7% in my version.

what did you do ? Grin

I was thinking, may-be I could use tpruvot interface, but 50% cpu usage is a big red flag for me... so please fix this...
(not the case for x11... though, but I don't use x11...)

exemplaar

hero member

Activity: 978

Merit: 506

Quote from: sp_ on December 17, 2014, 04:47:01 PM

Quote from: exemplaar on December 17, 2014, 03:59:31 PM

Quote from: sp_ on December 17, 2014, 03:10:49 PM

Yes this version is bether on the pools when mining on high intensity. The old version only found one solution below target, this version is submitting 2 solutions. Code forked fronm 1.5.1 (tvprovet's work)

Hi SP.
I'm mining X13 on regular pool (not multipool).The hashrate showing by the pool is lower than what the ccminer shows. And there is "set stratum difficulty to x.xxxx" message appearing quite a lot.
I tried versions from 17-20. The same scene. Hashrate is lower (from 14mhs on ccminer to only 8-9mhs on pool). I then tried older version 8 and it works better (13mhs miner 12mhs pool) I have standard setting in .bat file. Do i need a manual set up of diff per rig or is there something else that I'm missing? Switching back to version 8 now...
I would appreciate any help.
Thank you.

Wich card is this? you can to change the -i parameter. (intensity) If you are not using this card to display graphic -i 20.5 used to be the best for maxwell cards (750 -980). but it can crash if you have many cards.

The diff should be set in the password of the pool. Try the password -p d=0.02

Thanks for your help.

6x 750ti Gainward rig. Intensity parameter does not do anything. 20 crashes the rig, lower brings no improvement. I will try the password tomorrow. Now back to v8 again

jpouza

legendary

Activity: 2716

Merit: 1116

Quote from: sp_ on December 17, 2014, 06:14:15 PM

Quote from: djm34 on December 17, 2014, 05:14:28 PM

Quote from: sp_ on December 17, 2014, 04:50:55 PM

Quote from: djm34 on December 17, 2014, 04:16:22 PM

from my system: 980: 9.6MH/s/9.7MH/s
750ti 3.1MH/s
780ti 6.1MH/s
I think you have missed something, I don't think there should be so much difference

780ti 6.1MH/s is pretty good. Did you try to increase the intesity to the limit? f.eks -i 19.9 or -i 20.5?
The default intesity in the miner is set to run on 1mb cards. But with compute 5.2 the default is increased. I case on the compute version and not the memory..

didn't even know there was an intensity Grin

I will check again
edit: 6.2MH at 19.9 and 6.08MH at 20.5

http://cryptomining-blog.com/2361-new-ccminer-1-0-available-with-added-support-for-x11-mining/

ccminer 1.0 did 2.8 MHASH on the 780ti. (10 may 2014)

The latest software is 221 % faster

That's awesome!

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: djm34 on December 17, 2014, 05:14:28 PM

Quote from: sp_ on December 17, 2014, 04:50:55 PM

Quote from: djm34 on December 17, 2014, 04:16:22 PM

from my system: 980: 9.6MH/s/9.7MH/s
750ti 3.1MH/s
780ti 6.1MH/s
I think you have missed something, I don't think there should be so much difference

780ti 6.1MH/s is pretty good. Did you try to increase the intesity to the limit? f.eks -i 19.9 or -i 20.5?
The default intesity in the miner is set to run on 1mb cards. But with compute 5.2 the default is increased. I case on the compute version and not the memory..

didn't even know there was an intensity Grin

I will check again
edit: 6.2MH at 19.9 and 6.08MH at 20.5

http://cryptomining-blog.com/2361-new-ccminer-1-0-available-with-added-support-for-x11-mining/

ccminer 1.0 did 2.8 MHASH on the 780ti. (10 may 2014)

The latest software is 221 % faster

myagui

legendary

Activity: 1154

Merit: 1001

Quote from: sp_ on December 17, 2014, 05:32:11 PM

Time to make a neoscrypt port for NVIDIA. Can we beat the opencl code by WOLF0?

(nicehash)

NeoScrypt 3341 0.2021 GH/s 4.7595 BTC/GH/Day 738% 3.5289 BTC/GH/Day 5.2900 BTC/GH/Day 5.2900 BTC/GH/Day View orders

Would you consider cryptonight instead? That would turn me into a more generous donator ^^
I'm guessing I'm out of luck though, if I understand correctly, tsiv's miner will be a pain to merge...

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Time to make a neoscrypt port for NVIDIA. Can we beat the opencl code by WOLF0?

(nicehash)

NeoScrypt 3341 0.2021 GH/s 4.7595 BTC/GH/Day 738% 3.5289 BTC/GH/Day 5.2900 BTC/GH/Day 5.2900 BTC/GH/Day View orders

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: djm34 on December 17, 2014, 05:14:28 PM

Quote from: sp_ on December 17, 2014, 04:50:55 PM

Quote from: djm34 on December 17, 2014, 04:16:22 PM

from my system: 980: 9.6MH/s/9.7MH/s
750ti 3.1MH/s
780ti 6.1MH/s
I think you have missed something, I don't think there should be so much difference

780ti 6.1MH/s is pretty good. Did you try to increase the intesity to the limit? f.eks -i 19.9 or -i 20.5?
The default intesity in the miner is set to run on 1mb cards. But with compute 5.2 the default is increased. I case on the compute version and not the memory..

didn't even know there was an intensity Grin

I will check again
edit: 6.2MH at 19.9 and 6.08MH at 20.5

You probobly need to finetune it.

try 18.0, 18,1,18,2.... 20.5

If you find the optimal, I will hardcode the value into the next release with compute 3.5

hehe

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: sp_ on December 17, 2014, 04:50:55 PM

Quote from: djm34 on December 17, 2014, 04:16:22 PM

from my system: 980: 9.6MH/s/9.7MH/s
750ti 3.1MH/s
780ti 6.1MH/s
I think you have missed something, I don't think there should be so much difference

780ti 6.1MH/s is pretty good. Did you try to increase the intesity to the limit? f.eks -i 19.9 or -i 20.5?
The default intesity in the miner is set to run on 1mb cards. But with compute 5.2 the default is increased. I case on the compute version and not the memory..

didn't even know there was an intensity Grin

I will check again

edit: 6.2MH at 19.9 and 6.08MH at 20.5

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: Amph on December 17, 2014, 03:33:35 PM

how much the last version, does on scrypt algo?
i've checked the source now, there isn't any scrypt algo available, you removed it?
p.s. ah silly me, scypt is in cudaminer, i always forgot that lol, mmh...old cudaminer isn't working right with 970, not optimized i supposed, it use T91

I have looked at the scrypt kernal. I think I can make it 30-50% faster on the maxwell with some work. But waste of time.

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: djm34 on December 17, 2014, 04:16:22 PM

from my system: 980: 9.6MH/s/9.7MH/s
750ti 3.1MH/s
780ti 6.1MH/s
I think you have missed something, I don't think there should be so much difference

780ti 6.1MH/s is pretty good. Did you try to increase the intesity to the limit? f.eks -i 19.9 or -i 20.5?
The default intesity in the miner is set to run on 1mb cards. But with compute 5.2 the default is increased. I case on the compute version and not the memory..

sp_

legendary

Activity: 2926

Merit: 1087

Team Black developer

Quote from: exemplaar on December 17, 2014, 03:59:31 PM

Quote from: sp_ on December 17, 2014, 03:10:49 PM

Yes this version is bether on the pools when mining on high intensity. The old version only found one solution below target, this version is submitting 2 solutions. Code forked fronm 1.5.1 (tvprovet's work)

Hi SP.
I'm mining X13 on regular pool (not multipool).The hashrate showing by the pool is lower than what the ccminer shows. And there is "set stratum difficulty to x.xxxx" message appearing quite a lot.
I tried versions from 17-20. The same scene. Hashrate is lower (from 14mhs on ccminer to only 8-9mhs on pool). I then tried older version 8 and it works better (13mhs miner 12mhs pool) I have standard setting in .bat file. Do i need a manual set up of diff per rig or is there something else that I'm missing? Switching back to version 8 now...
I would appreciate any help.
Thank you.

Wich card is this? you can to change the -i parameter. (intensity) If you are not using this card to display graphic -i 20.5 used to be the best for maxwell cards (750 -980). but it can crash if you have many cards.

The diff should be set in the password of the pool. Try the password -p d=0.02

djm34

legendary

Activity: 1400

Merit: 1050

Quote from: jpouza on December 17, 2014, 01:48:16 PM

Build20 running OK for now, nice be just one instance for mining, all in one

In X11:

3.2MH/s 750Ti@200+core/400+mems

9.6MH/s 980@150+core/0+ mems

980gtx temps at 55ºC@ EK WC.

Nice.

Cheers

ok I checked the hashrate for the 780ti, not terrible honestly compared to the other card.

from my system: 980: 9.6MH/s/9.7MH/s
750ti 3.1MH/s
780ti 6.1MH/s
I think you have missed something, I don't think there should be so much difference

Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 1209. (Read 2347597 times)