Quote
Transaction ID: d3088e0d25037723d9d540fe4c8e7b94fe6813786c63258ac5e1fc7b69b7fb12-000
Thanks:)
Your welcome.
Got r78....testing now with lyra2v2. With latest drivers.
Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 823. (Read 2347664 times)
Thanks:)
Your welcome.
Got r78....testing now with lyra2v2. With latest drivers.
Quote
Transaction ID: d3088e0d25037723d9d540fe4c8e7b94fe6813786c63258ac5e1fc7b69b7fb12-000
Thanks:)
Try the build from the cryptomining blog
http://cryptomining-blog.com/6365-updated-windows-binary-of-the-ccminer-1-5-78-git-fork-by-sp-for-maxwell/
sp.....I can't get r78 zip to open.
And here is a couple of beers.
Transaction ID: d3088e0d25037723d9d540fe4c8e7b94fe6813786c63258ac5e1fc7b69b7fb12-000
And here is a couple of beers.
Transaction ID: d3088e0d25037723d9d540fe4c8e7b94fe6813786c63258ac5e1fc7b69b7fb12-000
ccminer.exe --algo=neoscrypt -o stratum+tcp://hashpower.co:4233 -O x:y
Is it normal?
Use @ CCMiner v1.5.74 for neoscrypt.Is it normal?
More work is needed on neoscrypt. The compiler has changed. Alot of retuning / rewriting is needed.
In release 78 I added 50% from release 77.
If you compile DJM34's offical version with cuda 7.5 my version is 50% faster.:-)
I've tried compiling the current neoscrypt kernel (single invocation which seems to be slightly better on cuda 7.5) with cuda 6.5. Interestingly enough, it is as slow as with 7.5. Thus I'm working on the old version and managed to improve it a couple percent. Maybe I should make a new fork with neoscrypt for cuda 6.5.
No gains with p0 state on 960, 970 and 980. Only ETH gains some % with memory oc.
Anyway I managed to put all the cards to p0 state also in windows. It is the same "trick" as in linux using nvidia-smi tool.
I'm not responsible if You kill Your GPU. The risk is Yours.
The procedure is:
1) open cmd.exe with the admin privileges
2) cd C:\Program Files\NVIDIA Corporation\NVSMI
3) nvidia-smi -q -d SUPPORTED_CLOCKS -i 0
4) nvidia-smi -ac 3505,1455 -i 0
-i 0 means display adapter 0. -i 4 would mean display adapter 4.
Step 3 is used to show the possible clock settings combination for each power state. The numbers will change if You overclock Your gpu before running step 3.
The result after typing step 3 would be looking like that:
Example:
nvidia-smi -ac 3505,1446 -i 0
The nvidia inspector shows Your card in p0 and You can overclock also the memory with msi afterburner or any other oc tool.
If You type nvidia-smi -ac 3505,1440 -i 0 it won't be accepted because 1440 is not in a list of matching Mhz for memory clock 3505.
I'm not responsible if You kill Your GPU. The risk is Yours.
For a 980ti, using Quark algo for a reference, I think if one could get the memory clock to p0 7ghz, one should get about 2 mh/s more.Anyway I managed to put all the cards to p0 state also in windows. It is the same "trick" as in linux using nvidia-smi tool.
I'm not responsible if You kill Your GPU. The risk is Yours.
The procedure is:
1) open cmd.exe with the admin privileges
2) cd C:\Program Files\NVIDIA Corporation\NVSMI
3) nvidia-smi -q -d SUPPORTED_CLOCKS -i 0
4) nvidia-smi -ac 3505,1455 -i 0
-i 0 means display adapter 0. -i 4 would mean display adapter 4.
Step 3 is used to show the possible clock settings combination for each power state. The numbers will change if You overclock Your gpu before running step 3.
The result after typing step 3 would be looking like that:
Code:
C:\Windows\system32>cd C:\Program Files\NVIDIA Corporation\NVSMI
C:\Program Files\NVIDIA Corporation\NVSMI>nvidia-smi -q -d SUPPORTED_CLOCKS -i 0
==============NVSMI LOG==============
Timestamp : Sat Jan 02 19:09:48 2016
Driver Version : 359.06
Attached GPUs : 5
GPU 0000:01:00.0
Supported Clocks
Memory : 3505 MHz
Graphics : 1635 MHz
Graphics : 1623 MHz
Graphics : 1610 MHz
Graphics : 1598 MHz
Graphics : 1585 MHz
Graphics : 1572 MHz
Graphics : 1560 MHz
Graphics : 1547 MHz
Graphics : 1534 MHz
Graphics : 1522 MHz
Graphics : 1509 MHz
Graphics : 1496 MHz
Graphics : 1484 MHz
Graphics : 1471 MHz
Graphics : 1458 MHz
Graphics : 1446 MHz
Graphics : 1433 MHz
Graphics : 1420 MHz
Graphics : 1408 MHz
Graphics : 1395 MHz
It always starts with the clocks of the p0 state. In my example it shows memory 3505 after that You can find listed all the possible graphics settings. You must chose one and type it exactly the same number in the step 4.C:\Program Files\NVIDIA Corporation\NVSMI>nvidia-smi -q -d SUPPORTED_CLOCKS -i 0
==============NVSMI LOG==============
Timestamp : Sat Jan 02 19:09:48 2016
Driver Version : 359.06
Attached GPUs : 5
GPU 0000:01:00.0
Supported Clocks
Memory : 3505 MHz
Graphics : 1635 MHz
Graphics : 1623 MHz
Graphics : 1610 MHz
Graphics : 1598 MHz
Graphics : 1585 MHz
Graphics : 1572 MHz
Graphics : 1560 MHz
Graphics : 1547 MHz
Graphics : 1534 MHz
Graphics : 1522 MHz
Graphics : 1509 MHz
Graphics : 1496 MHz
Graphics : 1484 MHz
Graphics : 1471 MHz
Graphics : 1458 MHz
Graphics : 1446 MHz
Graphics : 1433 MHz
Graphics : 1420 MHz
Graphics : 1408 MHz
Graphics : 1395 MHz
Example:
nvidia-smi -ac 3505,1446 -i 0
The nvidia inspector shows Your card in p0 and You can overclock also the memory with msi afterburner or any other oc tool.
If You type nvidia-smi -ac 3505,1440 -i 0 it won't be accepted because 1440 is not in a list of matching Mhz for memory clock 3505.
I'm not responsible if You kill Your GPU. The risk is Yours.
I was trying memory oc on my 750ti for a reference.
But being a noob at your programming layout to do this, I can't try this.
If there is a way sp or djm ect could program this into their programs, that would be great.
Or is this a one step..and it will stay that way even on reboot?
Thx
ps we still have cuda miner for special algo's.
ccminer.exe --algo=neoscrypt -o stratum+tcp://hashpower.co:4233 -O x:y
Is it normal?
Use @ CCMiner v1.5.74 for neoscrypt.Is it normal?
More work is needed on neoscrypt. The compiler has changed. Alot of retuning / rewriting is needed.
In release 78 I added 50% from release 77.
If you compile DJM34's offical version with cuda 7.5 my version is more than 50% faster.:-)
I don't know. What I remember of .74/6.5 might be wrong. Don't lose any sleep over it. I know you're selling souped up kernels for a premium and that's fine. I appreciate your work.
If you use cuda 7.5 use the latest version @github.
if you use cuda 6.5 compile release 74.
1.5.74(sp-MOD)
@sp-hash sp-hash released this on 15 Nov 2015 · 41 commits to windows since this release
Disregard my last post. Looks like I had a partially corrupt hybrid CUDA 6.5/7.5. I reverted back to nouveau drivers, ununstalled CUDA completely, started again from scratch and voi-la! It compiles using 7.5. Didn't see any hash rate improvements, though. 
What cards do you have.
On the 750ti release 78: x11,quark, lyra2v2 and qubit should be faster. (than release 74 (cuda 6.5))
On gtx 970 only quark and lyra2v2 is faster
I don't know. What I remember of .74/6.5 might be wrong. Don't lose any sleep over it. I know you're selling souped up kernels for a premium and that's fine. I appreciate your work.
Z
ccminer.exe --algo=neoscrypt -o stratum+tcp://hashpower.co:4233 -O x:y
Hardware
@ CCMiner v1.5.77
@ CCMiner v1.5.78
Is it normal?
Use @ CCMiner v1.5.74 for neoscrypt. Hardware
@ CCMiner v1.5.77
@ CCMiner v1.5.78
Is it normal?
Disregard my last post. Looks like I had a partially corrupt hybrid CUDA 6.5/7.5. I reverted back to nouveau drivers, ununstalled CUDA completely, started again from scratch and voi-la! It compiles using 7.5. Didn't see any hash rate improvements, though.
What cards do you have.
On the 750ti release 78: x11,quark, lyra2v2 and qubit should be faster. (than release 74 (cuda 6.5))
On gtx 970 only quark and lyra2v2 is faster
sp_ how much do you want for your quark private miner? or DJM? or even x11 and Lyra2REv2
Not for sale. But if djm34 sell his kernal I might sell it.
Just to summarize a bit the way I do things, for me for a sell to be profitable it has to reflect, somehow,
the time I spent in coding (regardless of the coin value ) hence if the coin value is low,
it wouldn't be profitable for a miner to buy from me or it wouldn't be profitable for me to sell below my time/money equivalent
(and would be bad business practices as well since when I work for a coin dev I charge along that time/money equivalent).
so for the moment, it isn't for sale...
also as already said multiple times: trust is important for me... (private miners are sold in priority to people I can trust)
sp_ how much do you want for your quark private miner? or DJM? or even x11 and Lyra2REv2
Not for sale. But if djm34 sell his kernal I might sell it.
sp_ how much do you want for your quark private miner? or DJM? or even x11 and Lyra2REv2
I am not for sell ESL
I'm having trouble compiling with CUDA 7.5. I installed 7.5.18 from the nVidia web site but when I run ccminer it still says compiled with CUDA 6.5? I'm stumped.
Do you have the latest version of the sourcecode?
release 78 has this in the heading:
Compiled with Visual C++ 18 using Nvidia CUDA Toolkit 7.5
Some kernals are faster. Some are slower.
sp_ how much do you want for your quark private miner? or DJM? or even x11 and Lyra2REv2
I am not for sell
sp_ how much do you want for your quark private miner? or DJM? or even x11 and Lyra2REv2
I'm having trouble compiling with CUDA 7.5. I installed 7.5.18 from the nVidia web site but when I run ccminer it still says compiled with CUDA 6.5? I'm stumped.
Do you have the latest version of the sourcecode?
release 78 has this in the heading:
Compiled with Visual C++ 18 using Nvidia CUDA Toolkit 7.5
I am failing to p0 state on my gtx 960 and gtx 970.
how can I do it with nvidia inspector?
how can I do it with nvidia inspector?
You don't. With nvidia inspector your 9xx cards will always be stuck with P2 (which is just shy of the top memory bins). There's other ways to get there as someone else posted though.
Yep, you are right, no p0 state with inspector. But you can overclock your memory to p0 levels at p2 if you want.
I am failing to p0 state on my gtx 960 and gtx 970.
how can I do it with nvidia inspector?
how can I do it with nvidia inspector?
You don't. With nvidia inspector your 9xx cards will always be stuck with P2 (which is just shy of the top memory bins). There's other ways to get there as someone else posted though.
Jump to: