CCminer(SP-MOD) Modded GPU kernels.

[tbearhere]

750 ti
About 6150kh/s  100kh/s +  quark
X11 3,235kh/s on one of my cards.

[tbearhere]

I expect couple of testing samples of GTX960 next week. Would be interesting what could they do

That card is what I have been waiting for. Where can I buy one please.

[go6ooo1212]

I expect couple of testing samples of GTX960 next week. Would be interesting what could they do

That card is what I have been waiting for. Where can I buy one please.

You couldt yet, because the official announce is at the end of January 😊

[tbearhere]

I expect couple of testing samples of GTX960 next week. Would be interesting what could they do

That card is what I have been waiting for. Where can I buy one please.

You couldt yet, because the official announce is at the end of January 😊

Thanks...I missed the part where he said next week.  :p

[sp_]

-Faster bmw,keccak, echo and jh512

Not much though.

The 29th public beta(windows executable) is available here: (11-jan-2015)

http://www.filedropper.com/release29

The sourcecode is available here:

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

[scryptr]

-Faster bmw,keccak, echo and jh512

Not much though.

The 29th public beta(windows executable) is available here: (11-jan-2015)

http://www.filedropper.com/release28

The sourcecode is available here:

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

SP_ v29 RESULTS--

Initially, quark is +100kh/s on my faster rig, not much change on my slower rig.  Its Sempron 145 compiles noticably more slowly than the AMD FX on my faster rig.  If the CPU processing load is a factor in hashing speed, it may be a bottleneck on my slower rig. I now get 36.3-4mh/s on my faster rig, and 35.1-2mh/s on my slower rig.  Version 28 pushed the slow rig to 35.1mh/s, and the fast rig to 36.2mh/s.       --scryptr

[sp_]

I corrected the link, it was pointing to the 28 release binary.
In release 22 and up I increased the cpu usage abit. In benchmark mode, my rig with 4 cards does 600KHASH  bether than mining on a pool I think it is connected to the increased cpu usage.

[chrysophylax]

-Faster bmw,keccak, echo and jh512

Not much though.

The 29th public beta(windows executable) is available here: (11-jan-2015)

http://www.filedropper.com/release28

The sourcecode is available here:

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

SP_ v29 RESULTS--

Initially, quark is +100kh/s on my faster rig, not much change on my slower rig.  Its Sempron 145 compiles noticably more slowly than the AMD FX on my faster rig.  If the CPU processing load is a factor in hashing speed, it may be a bottleneck on my slower rig. I now get 36.3-4mh/s on my faster rig, and 35.1-2mh/s on my slower rig.  Version 28 pushed the slow rig to 35.1mh/s, and the fast rig to 36.2mh/s.       --scryptr

why dont you make the sempron machine faster by changing the cpu scryptr? ...

or are you not in a position to get that done? ...

i have most of my machines upgraded to the 4300 cpu for the same reason ... now i compile once and distribute it - rather than compile on every machine due to the different cpu's that were in them ...

although now im changing over to the h81-pro-btc boards and intel chips ... couldnt go any other way with those boards ...

quark is running more stable though i havent seen a lift in hashrate ...

maybe need to recompile with the latest 29 ... do that after i build the next machine ...

#crysx

[sp_]

The latest (29) should give a small boost in x11,x13,x14,x15

x14 has been more profitable than quark the last 24 hours. 0.4 BTC for 1GHASH.
x11 on nicehash has paid 0.3816 BTC/GH/Day the last 24hours.

All my rigs are using profit switching software. Miner control (1.6). Windows only.

[sp_]

On windows ccminer is faster when compiled and run in 32bit mode(x86).  Linux rigs should probobly run with 32 bit installation/drivers for a boost. Has anyone tried?

[jpouza]

The latest (29) should give a small boost in x11,x13,x14,x15

x14 has been more profitable than quark the last 24 hours. 0.4 BTC for 1GHASH.
x11 on nicehash has paid 0.3816 BTC/GH/Day the last 24hours.

All my rigs are using profit switching software. Miner control (1.6). Windows only.

The only true profitable coin at the moment is SpreadCoin, it uses SpreadX11 algo and no pools are allowed, only solo mining, I think you know that...is there a way to improve tsiv´s miner to increase Maxwell hash rates in SpreadX11?

At the moment I get 5.5MH/s per 980 Oced card and 1.7MH/s per 750Ti Oced.


Cheers

[scryptr]

-Faster bmw,keccak, echo and jh512

Not much though.

The 29th public beta(windows executable) is available here: (11-jan-2015)

http://www.filedropper.com/release28

The sourcecode is available here:

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

SP_ v29 RESULTS--

Initially, quark is +100kh/s on my faster rig, not much change on my slower rig.  Its Sempron 145 compiles noticably more slowly than the AMD FX on my faster rig.  If the CPU processing load is a factor in hashing speed, it may be a bottleneck on my slower rig. I now get 36.3-4mh/s on my faster rig, and 35.1-2mh/s on my slower rig.  Version 28 pushed the slow rig to 35.1mh/s, and the fast rig to 36.2mh/s.       --scryptr

why dont you make the sempron machine faster by changing the cpu scryptr? ...

or are you not in a position to get that done? ...

i have most of my machines upgraded to the 4300 cpu for the same reason ... now i compile once and distribute it - rather than compile on every machine due to the different cpu's that were in them ...

although now im changing over to the h81-pro-btc boards and intel chips ... couldnt go any other way with those boards ...

quark is running more stable though i havent seen a lift in hashrate ...

maybe need to recompile with the latest 29 ... do that after i build the next machine ...

#crysx

CryptoBadger--

CryptoBadger touted the Sempron 145 as as good enough to mine, and cheap.  I followed his advice, but also had some better AMD CPUs around when building my rigs.  However, last night I remembered some conflicting notes online as to whether the Sempron 145 was a 1- or 2-core CPU and did some research.

I found that the Sempron 145 was an "unlockable" CPU, that frequently could be enabled as a 2-core CPU via BIOS.  My motherboards had this built-in capability (they are identical), and I cycled through my rigs.  In less than 10 minutes, I had an "AMD II X2 4450e" dual-core capable of 2800MHz, rather than my Sempron 145 single-core.  The other rigs were already at max cores.  The stock AMD II X2 4450e chips were clocked at 2300MHz, by the way.  I had unlocked my Sempron!  

The rig is running steady for several hours now.  It runs at least 50kh/s faster than before being unlocked.       --scryptr

[chrysophylax]

-Faster bmw,keccak, echo and jh512

Not much though.

The 29th public beta(windows executable) is available here: (11-jan-2015)

http://www.filedropper.com/release28

The sourcecode is available here:

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

SP_ v29 RESULTS--

Initially, quark is +100kh/s on my faster rig, not much change on my slower rig.  Its Sempron 145 compiles noticably more slowly than the AMD FX on my faster rig.  If the CPU processing load is a factor in hashing speed, it may be a bottleneck on my slower rig. I now get 36.3-4mh/s on my faster rig, and 35.1-2mh/s on my slower rig.  Version 28 pushed the slow rig to 35.1mh/s, and the fast rig to 36.2mh/s.       --scryptr

why dont you make the sempron machine faster by changing the cpu scryptr? ...

or are you not in a position to get that done? ...

i have most of my machines upgraded to the 4300 cpu for the same reason ... now i compile once and distribute it - rather than compile on every machine due to the different cpu's that were in them ...

although now im changing over to the h81-pro-btc boards and intel chips ... couldnt go any other way with those boards ...

quark is running more stable though i havent seen a lift in hashrate ...

maybe need to recompile with the latest 29 ... do that after i build the next machine ...

#crysx

CryptoBadger--

CryptoBadger touted the Sempron 145 as as good enough to mine, and cheap.  I followed his advice, but also had some better AMD CPUs around when building my rigs.  However, last night I remembered some conflicting notes online as to whether the Sempron 145 was a 1- or 2-core CPU and did some research.

I found that the Sempron 145 was an "unlockable" CPU, that frequently could be enabled as a 2-core CPU via BIOS.  My motherboards had this built-in capability (they are identical), and I cycled through my rigs.  In less than 10 minutes, I had an "AMD II X2 4450e" dual-core capable of 2800MHz, rather than my Sempron 145 single-core.  The other rigs were already at max cores.  The stock AMD II X2 4450e chips were clocked at 2300MHz, by the way.  I had unlocked my Sempron!  

The rig is running steady for several hours now.  It runs at least 50kh/s faster than before being unlocked.       --scryptr

awesome ...

it seems that there are ways around getting things like that for free ...

good work mate ...

#crysx

[sp_]

The latest (29) should give a small boost in x11,x13,x14,x15
x14 has been more profitable than quark the last 24 hours. 0.4 BTC for 1GHASH.
x11 on nicehash has paid 0.3816 BTC/GH/Day the last 24hours.
All my rigs are using profit switching software. Miner control (1.6). Windows only.

The only true profitable coin at the moment is SpreadCoin, it uses SpreadX11 algo and no pools are allowed, only solo mining, I think you know that...is there a way to improve tsiv´s miner to increase Maxwell hash rates in SpreadX11?
At the moment I get 5.5MH/s per 980 Oced card and 1.7MH/s per 750Ti Oced.
Cheers

I am working on a spreadcoin miner. Fork from tsivs work.

x11 is profitiable now. The payout is 300% higher today than it was last week.

[chrysophylax]

On windows ccminer is faster when compiled and run in 32bit mode(x86).  Linux rigs should probobly run with 32 bit installation/drivers for a boost. Has anyone tried?

i only compile in x64 fedora 19 ( for the moment ) ...

tomorrow i am building a x32 intel based fedora 21 system ...

ill install all the libraries and compile 29 on that to see if there is indeed any boost ...

i have an extra 750ti from the shipment due to a dodgy pcie port on one of the older motherboards - so ill mine with that and let you know ...

that is - if there is enough time tomorrow ... otherwise it will be the day after ...

#crysx

[chrysophylax]

The latest (29) should give a small boost in x11,x13,x14,x15
x14 has been more profitable than quark the last 24 hours. 0.4 BTC for 1GHASH.
x11 on nicehash has paid 0.3816 BTC/GH/Day the last 24hours.
All my rigs are using profit switching software. Miner control (1.6). Windows only.

The only true profitable coin at the moment is SpreadCoin, it uses SpreadX11 algo and no pools are allowed, only solo mining, I think you know that...is there a way to improve tsiv´s miner to increase Maxwell hash rates in SpreadX11?
At the moment I get 5.5MH/s per 980 Oced card and 1.7MH/s per 750Ti Oced.
Cheers

I am working on a spreadcoin miner. Fork from tsivs work.

x11 is profitiable now. The payout is 300% higher today than it was last week.

can the algorithm be implemented within ccminer - or does it need a new miner altogether?

#crsyx

[sp_]

It wil be a seperate miner that can only mine spreadcoin.
I am to lazy to integrate it in the other fork of ccminer.

[chrysophylax]

It wil be a seperate miner that can only mine spreadcoin.
I am to lazy to integrate it in the other fork of ccminer.

hehehe ... kool ...

i appreciate that it maybe a bit of work involved ...

would be very nice to have one miner mine the majority of algorithms - if not all ...

tanx sp ...

#crysx

[bluejam00]

Hi

On my stock Geforce GTX 980 - in PrecisionX I OC's 110% power,  +150MHz GPU, Mem +116MHZ, + 6mV - I then get 10MH+ for a few seconds - then it falls over with this error:

[2015-01-12 23:21:03] GPU #0: result for 01f4ebc0 does not validate on CPU!
[2015-01-12 23:21:03] GPU #0: GeForce GTX 980, 158879 kH/s
Cuda error in func 'quark_blake512_cpu_setBlock_80' at line 482 : the launch tim
ed out and was terminated.

Not sure what this means - i was getting temps around 65% and 10MH+ so keen to see why this is occuring? I'm using SP release 29.

[sp_]

Seems like you have clocked the card to high. Reduce the gpuclock and increase the memory clock. Did the driver crash as well? then you might need to reboot the machine.

You might also want to experiment with the Intensity parameter.  start with -i 18 and then continue to find the best setting for your card

 -i 20.5 might work good for your overvolted card.