Author

Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 1199. (Read 2347597 times)

hero member
Activity: 644
Merit: 500
Fixed slow cpu. X11,x13,x14,x15,x17, nist, quark. Lyra is still a bit slow but faster than build 22


The 23nd public beta(windows executable) is available here: (29-12-2014)

http://www.filedropper.com/ccminer23

The sourcecode is available here:

https://github.com/sp-hash/ccminer

Post your stats here. Card name/gpu clock/memclock


Same hashrate on X11 but only 12% cpu usage now ^^ It was 8% before build22 for 2x 750TIs. Thx.


(Started build23 after the last red area/line, ignore the others Tongue)

Now let's hope you'll keep the API so I can keep on monitoring.
I'm nearing the first release of it, and it would be a real shame if all was for nothing.
Source of almost-100%-working ccmonitor is already out there: https://github.com/KBomba/ccmonitor
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
Fixed slow cpu. X11,x13,x14,x15,x17, nist, quark. Lyra is still a bit slow but faster than build 22


The 23nd public beta(windows executable) is available here: (29-12-2014)

http://www.filedropper.com/ccminer23

The sourcecode is available here:

https://github.com/sp-hash/ccminer

Post your stats here. Card name/gpu clock/memclock
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
I have checked in a fix now on gihub. The cpu usage is reduced to 1/3 in x11. Hashrate is the same as in 22.

building 23 now.
legendary
Activity: 1797
Merit: 1028
Version 22 unstable? --

I also dropped v22, and returned to tpruvot's 1.5.2-git work.  I hope that your improved hashing routines make it into his version soon; there was a real increase in performance.  I'll probably try v23 as soon as it is up.

My slower rig has a Sempron 145, 8 gigs ram.  It would work well for 12 hours or so, then the hash rate would drop to a crawl and the rig would crash.  The faster rig is holding its own and I will let it be with v22.  It has a faster AMD FX CPU.  Maybe a memory leak?       --scryptr
legendary
Activity: 3164
Merit: 1003
Nice graph Smiley The new version seems to have more stable hashrate. I am working to remove some cpu load for the next build
With 22 I got improvement some but a lot more power usage. My team-viewer crashes sometimes with 22. I'm back to 21 release, waiting for 23.
Yes, less cpu usage would be great.  Smiley
sp_
legendary
Activity: 2926
Merit: 1087
Team Black developer
The problem is that when after I merged my kernals into 1.5.0 and above, the rate has gone down on some pools. Release 8 or 9 is displaying the full hash(1.4.9-mod). The only way to test for sure is to merge kernals down to the
djm34 release. This is some hours work. I will merge the kernals and then release a test build. I plan to fix the bug in the latest version, but the merging is needed to find the bug.

10 accepted in 1 minute is a bit high. You should increase the difficulty or use vardiff.
legendary
Activity: 2912
Merit: 1091
--- ChainWorks Industries ---
Good job as always SP! Lyra2 hash is a bit better.

However, same issue as with x11: Pool does not report the full hashrate. I think this is a ccminer 1.5.x related problem. For the original ccminer-Lyra2RE, pools show the correct hashrate as it is displayed in the miner. Would you mind to include the Lyra2 improvements into djm34's ccminer-Lyra2RE release?

I have done some fixes in x11 , is it still underrepporting in release 21?

Lyra is currently only finding one solution and the throughput is big. I will add 2 solutions soon.

I am considering to change the fork to the DJM34 fork. Clean it and replace all the kernals.

1. Less crashes.
2. Fix the problem on some pools.

But will lack some features like the API etc..


Hi, hope you guys had a good Christmas! As promised earlier, I have tested X11 with release21, and the hashrate is still not fully recognized by (some?) pools. Here's the graph for 16h of mining.



from what i can gather with all the mining thats been done - there is NO pool that will give a totally accurate hashrate ...

this is due to the collation of the previous accepted shares and the hashrate is calculated on THOSE accepted shares only for the period of time that the pool takes the sample shares ... most pools take this period of time to be 10 minutes - but some longer ...

this means that no matter what the miner itself says its doing - the pool will not show the same ...

this is why we can only get an _estimate_ of what the hashrate may be from the pools - and a more _accurate estimate_ from the miner itself ...

the hashrates from the miner are usually around the same figure - with a small splay percentage ...

the hashrates from the pools are usually much more splayed and have a higher degree of variance ...

result? ... the most accurate representation of hashrate is from the miner itself ...

thats my take on it anyway ... can anyone shed more light on this situation? ...

#crysx
member
Activity: 61
Merit: 10
Good job as always SP! Lyra2 hash is a bit better.

However, same issue as with x11: Pool does not report the full hashrate. I think this is a ccminer 1.5.x related problem. For the original ccminer-Lyra2RE, pools show the correct hashrate as it is displayed in the miner. Would you mind to include the Lyra2 improvements into djm34's ccminer-Lyra2RE release?

I have done some fixes in x11 , is it still underrepporting in release 21?

Lyra is currently only finding one solution and the throughput is big. I will add 2 solutions soon.

I am considering to change the fork to the DJM34 fork. Clean it and replace all the kernals.

1. Less crashes.
2. Fix the problem on some pools.

But will lack some features like the API etc..


Hi, hope you guys had a good Christmas! As promised earlier, I have tested X11 with release21, and the hashrate is still not fully recognized by (some?) pools. Here's the graph for 16h of mining.

legendary
Activity: 1151
Merit: 1001
Is this build supporting cryptonight?!
legendary
Activity: 2912
Merit: 1091
--- ChainWorks Industries ---
I chose and stuck with teamviewer for its cross platform simplicity. That, and not having to manage all the port forwarding across the 10 rigs i have in my little farm. Teamviewer lets me connect easily from my iphone, from which I do most of my work while I am away from home. I have been considering switching to linux, however my farm is 4500km away on the other side of the country, and I already burden my brother with enough hardware troubleshooting as it is!

hence the reason to go with linux ...

the control over the system via console is unparalleled ... and im in a similar situation as you are with the systems when im not here in the office ...

a simple boot and ssh daemon activation on the main system ( or pxe boot with scripts to auto install with your settings ) or even a prebuilt mining package ( which i dont use as i want my own setup - hence why i use fedora on the miners ) and viola - done in a matter of a few minutes ... o - and its free ... Wink

anyway - teamviewer is a great little package and if it serves you well - then all good i say ...

need to run some errands - then will be back to build another system with nvidia cards ... testing with one first though - and the latest compile of ccminer ... so see you in about two hours ... Smiley

#crysx
HoY
newbie
Activity: 26
Merit: 0
I chose and stuck with teamviewer for its cross platform simplicity. That, and not having to manage all the port forwarding across the 10 rigs i have in my little farm. Teamviewer lets me connect easily from my iphone, from which I do most of my work while I am away from home. I have been considering switching to linux, however my farm is 4500km away on the other side of the country, and I already burden my brother with enough hardware troubleshooting as it is!
legendary
Activity: 2912
Merit: 1091
--- ChainWorks Industries ---
Yes, +180 +180 with factory overclocked MSI cards. the hashrate is usually slightly higher on 2 of those cards but when I'm connected on teamviewer to that rig they both slow down 100-200kh/s each. been pretty consistent and I haven't found a simple solution to that


nice ... so the assumption is that its a windows system then ...

if so - try the rdp connection thats built in to windows instead of teamviewer ... im a teamviewer advocate also - but in this instance - i would find a way to keep track with as little overhead as possible ...

or switch to linux ...

the only thing im missing knowledge ( and not many can answer this without actually modding the card itself ) is the method of overclocking within linux itself ...

sgminer is quite granular and can change settings via commandline ... ccminer has no such thing that i know of - apart from the intensity setting ... and that doesnt seem to be doing much at all at the moment ...

still - nice hashrates ...

#crysx
HoY
newbie
Activity: 26
Merit: 0
Yes, +180 +180 with factory overclocked MSI cards. the hashrate is usually slightly higher on 2 of those cards but when I'm connected on teamviewer to that rig they both slow down 100-200kh/s each. been pretty consistent and I haven't found a simple solution to that
legendary
Activity: 2912
Merit: 1091
--- ChainWorks Industries ---
I just checked the CPU usage on this rig, 99% for ccminer.exe lol. its just a dual core intel g2 3220

on the plus side, it doesnt do anything but mine so I can deal with the extra lag when I log in to manage it Smiley


they are nice hashrates HoY ...

overclocked? ...

#crysx
legendary
Activity: 2912
Merit: 1091
--- ChainWorks Industries ---
not liking the higher cpu usage on the new release

I would revert to release 21 until 23 comes out with its lower CPU usage. It's working wonderfully on my rig (changed motherboard and all of my issues went away. Could have been a riser on the 7th card or simply a driver issue I couldn't fix)

yep I reverted.  7 card rig?  nice!  I'm just running some extra 750ti's on my kids computers and trying to get some extra coin.

I'm wondering how others are doing with crashes?  I run on 2 win8 machines and a win7 machine and they all seem to crash in less than 24 hours.  Seems like I might have better luck with the win 8 machine or with the win7 and win8 machine that are only running one gpu.  The 3 gpu system I have seems to crash in less than 24 hours constantly. 

Last thing I did to try to improve was set compatibility to win8 and run in admin mode... not sure if that will help any.

Anyone else crashing consistently before 24 hours?

no downgrade ... just compiled from the git ...

fedora 19 x64 - 99% cpu load with approx 2700-2750KH on gigabyte 750ti cards - both powered and non-powered ( low profile ) cards ...

no crashes as yet ... been running for a most of today ... 6 card rig - 5 card rig ... with another 10 x 6card rigs over the next few weeks being setup with EXACT specs ...

a huge testbed soon enough Wink

cpu load is the only issue at the moment it seems ...

#crysx
HoY
newbie
Activity: 26
Merit: 0
I just checked the CPU usage on this rig, 99% for ccminer.exe lol. its just a dual core intel g2 3220

on the plus side, it doesnt do anything but mine so I can deal with the extra lag when I log in to manage it Smiley
https://i.imgur.com/WItJZmh.png?1
legendary
Activity: 2296
Merit: 1031
not liking the higher cpu usage on the new release

I would revert to release 21 until 23 comes out with its lower CPU usage. It's working wonderfully on my rig (changed motherboard and all of my issues went away. Could have been a riser on the 7th card or simply a driver issue I couldn't fix)

yep I reverted.  7 card rig?  nice!  I'm just running some extra 750ti's on my kids computers and trying to get some extra coin.

I'm wondering how others are doing with crashes?  I run on 2 win8 machines and a win7 machine and they all seem to crash in less than 24 hours.  Seems like I might have better luck with the win 8 machine or with the win7 and win8 machine that are only running one gpu.  The 3 gpu system I have seems to crash in less than 24 hours constantly. 

Last thing I did to try to improve was set compatibility to win8 and run in admin mode... not sure if that will help any.

Anyone else crashing consistently before 24 hours?
HoY
newbie
Activity: 26
Merit: 0
not liking the higher cpu usage on the new release

I would revert to release 21 until 23 comes out with its lower CPU usage. It's working wonderfully on my rig (changed motherboard and all of my issues went away. Could have been a riser on the 7th card or simply a driver issue I couldn't fix)
legendary
Activity: 2296
Merit: 1031
not liking the higher cpu usage on the new release
legendary
Activity: 1797
Merit: 1028
Nice graph Smiley The new version seems to have more stable hashrate. I am working to remove some cpu load for the next build
Indeed, heavy CPU usage, my 5930k is at 4,45GHz 44% usage  Shocked

I have fixed it now. 1/3 of the cpu and only 3KHASH lost (750ti), but needs some more work and getting late here.

Good News!   Grin  I look forward to compiling and running it tomorrow.

I just finished compiling and running the new build 22.  First, since about v1.5.0, there is a very steady hash rate on my rigs, where it varies within single digits.  On build 21, I was running about 18405kh/s, and 17810kh/s for x11 on my faster and slower rigs.  On build 22, my faster rig now runs at 18425kh/s and my slower rig saw a larger improvement to 17945kh/s.  This was an increase of 130kh/s for the slower rig.  My slower rig has no 6-pin connectors on the cards.

One thing I noticed was that the hash rate took a long time to build to maximum, about 100 accepts.       --scryptr

 
Jump to: