CCminer(SP-MOD) Modded GPU kernels.

[Amph]

lyra2v2 is better with 1xxx than maxwell, a single 1070 can do 41MH versus private mod of sp which is only 37MH for a 980ti

on neoscrypt also, a 980ti can do only 1MH, a 1070 can do 1150khs with +200 core and 1250khs with +400 on the core

and this without any optimization for pascal

[1715286231]

1715286231

[1715286231]

1715286231

[1715286231]

1715286231

[1715286231]

1715286231

[Epsylon3]

There is no optimization to do on pascal yet, its almost the same hardware, reduced in size

Maybe the next cuda versions will show some hidden functions, but im not confident about that

[airwalker2662]

Excellent job on your commit today for Neoscrypt, its working great!

725 kh/s on 1080 from around 197 prior
 

[airwalker2662]

So there is no advantage to compile with CUDA 8 at the moment?

[Amph]

There is no optimization to do on pascal yet, its almost the same hardware, reduced in size

Maybe the next cuda versions will show some hidden functions, but im not confident about that

it looks like you are right since benchmark shows 50-70% improvement over a 970, which is what we are getting already on all algo

Excellent job on your commit today for Neoscrypt, its working great!

725 kh/s on 1080 from around 197 prior
 

from who sp? you should get that with the djm34 version

[vaulter]

Excellent job on your commit today for Neoscrypt, its working great!

725 kh/s on 1080 from around 197 prior
 

Can you point to the version you are using - want to check 1070 - can you give the link?

[airwalker2662]

There is no optimization to do on pascal yet, its almost the same hardware, reduced in size

Maybe the next cuda versions will show some hidden functions, but im not confident about that

it looks like you are right since benchmark shows 50-70% improvement over a 970, which is what we are getting already on all algo

Excellent job on your commit today for Neoscrypt, its working great!

725 kh/s on 1080 from around 197 prior
 

from who sp? you should get that with the djm34 version

No, from tpruvot's 1.8. Isn't the latest version from djm34 from January?

[airwalker2662]

Excellent job on your commit today for Neoscrypt, its working great!

725 kh/s on 1080 from around 197 prior
 

Can you point to the version you are using - want to check 1070 - can you give the link?

https://github.com/tpruvot/ccminer

[giagge]

Hi!

I have a problem to mining a neoscrypt FTC on coinotron , i have windows 10 x64 home edition and nvidia gtx 980 with last driver .

For program use this:

http://cryptomining-blog.com/6450-new-ccminer-fork-from-djm34-with-faster-neoscrypt/

My .bat is :

ccminer-65 -a neoscrypt -o stratum+tcp://coinotron.com:3337 -u my user -p my pass

But my hashrate is too low , only 180 kh/s , but i see the hashrate with gtx 980 is 632 kh/s .

Help ? .

[airwalker2662]

Were you always getting 632? Did you check your core clock? Restart machine, etc, etc.? Need more info.

[Epsylon3]

yep, in fact i just reimported the last djm/VTC version with a lot of cleanup (but no real code changes). I didnt had the time/patience before (sp and klaus already did that)

its much better on linux, but windows results are weird... well actually i only have the pascal cards on my dev machine and its a pain to compare

[airwalker2662]

yep, in fact i just reimported the last djm/VTC version with a lot of cleanup (but no real code changes). I didnt had the time/patience before (sp and klaus already did that)

its much better on linux, but windows results are weird... well actually i only have the pascal cards on my dev machine and its a pain to compare

Ok great! I will compile with 8 and see what changes. Thanks!

[pallas]

yep, in fact i just reimported the last djm/VTC version with a lot of cleanup (but no real code changes). I didnt had the time/patience before (sp and klaus already did that)

its much better on linux, but windows results are weird... well actually i only have the pascal cards on my dev machine and its a pain to compare

980ti 190 W, linux

Djm34: 920Mh/s
Tpruvot latest: 760 Mh/s

[gghost]

Epsylon - Are you able to add blake2b support to your yiimp pool for SIA coin? Claymore's latest Dual Miner v5.0 now supports Eth+Sia mining and I'd like to continue supporting your pool as thanks for your continued open source support.

[Epsylon3]

so in the last day :

zpool some segfaults issues
decred team ask me to check gominer
mrM4D ask me quark changes

and im working on neoscrypt... not enough ?

i already checked this SIA several times... but i forgot again why i stopped...

[Amph]

There is no optimization to do on pascal yet, its almost the same hardware, reduced in size

Maybe the next cuda versions will show some hidden functions, but im not confident about that

it looks like you are right since benchmark shows 50-70% improvement over a 970, which is what we are getting already on all algo

Excellent job on your commit today for Neoscrypt, its working great!

725 kh/s on 1080 from around 197 prior
 

from who sp? you should get that with the djm34 version

No, from tpruvot's 1.8. Isn't the latest version from djm34 from January?

you can get that hash with djm34, he imported the work on his fork?

[Epsylon3]

ive almost the same values on linux 750 ti and 960

[2016-07-08 20:15:10] GPU #1: GeForce GTX 750 Ti, 221 (T= 45C F= 34% C=1306/2700)
[2016-07-08 20:15:10] accepted: 6/6 (100.00%), 597.97 kH/s yes!
[2016-07-08 20:15:21] GPU #0: GeForce GTX 960, 403 (T= 56C F= 20% C=1341/3004)

me
[2016-07-08 18:20:36] GPU #1: Gigabyte GTX 750 Ti, 226.73 kH/s
[2016-07-08 18:20:36] accepted: 23/23 (diff 0.001), 621.69 kH/s yes!
[2016-07-08 18:20:44] GPU #0: Gigabyte GTX 960, 395.18 kH/s

[gghost]

so in the last day :

zpool some segfaults issues
decred team ask me to check gominer
mrM4D ask me quark changes

and im working on neoscrypt... not enough ?

i already checked this SIA several times... but i forgot again why i stopped...

I just wanted to bring it to your attention. You can implement if it makes sense at your leisure. I thought you might stand to gain some revenue from a bunch of people jumping on the SIA bandwagon with the dual miner.

[Amph]

tpruvot's 1.8 is out, thread still report 1.7.6? so i can get rid of that "theft" from cryptominingblog for 11evo

[airwalker2662]

I've been asked to compile for someone, might as well share it. It's Tpruvots 1.8 CUDA 8 version from today : https://1drv.ms/f/s!Aq23cheCBvi_bfyZ1qVCgcKiASA