Changing the hashing algorithm
-
A:2736 R:25 HW:0 WU:4.2/m @4.44Kh/s
No errors. This was running overnight. nVidia GeForce GT 630M
defultsetx DISPLAY 0 setx GPU_MAX_ALLOC_PERCENT 100 setx GPU_USE_SYNC_OBJECTS 1 cgminer.exe --neoscrypt --worksize 256 --thread-concurrency 300 -I 8 --gpu-memclock 1000 --gpu-engine 1000 -o http://pool.ftc-c.com:19328 -u TD8nGRpSxHyrw327LR1kJxwsin21GcyDwF -p x --fix-protocol
Also, Who was this? Spiked to 33,558 h/s
-
What does WU:x mean?
-
Gigabyte Gv-R7870C-2GD 7870 PCI-E
-
What does WU:x mean?
WU: The Work Utility defined as the number of diff1 equivalent shares / minutehttp://bitcoin.stackexchange.com/questions/19260/deciphering-cgminer-interface-in-terminal
So I’m guessing the higher the number the more efficient your miner is
-
Ah nice. Cheers.
-
Which version Catalyst is best ?
my 7850 ,7970 are HW ,not work.
-
I think 13.4 is the best, but 14.4 for me was only 1% or 2% slower so not a huge difference.
I was running 8.something and that would crash cgminer.
What version are you currently using?
-
I use AMD 14.4 Catalyst + SDK 2.9 .
Can you mine with AMD 13.4 Catalyst + SDK 2.8 ?
-
I could mine with just AMD 13.4, I didn’t have SDK installed at all, although I now have installed SDK 2.9.
I have a Gigagbyte HD7870 GPU, running Windows 7 64 bit
-
There have been known mining issues with 14.4 since release, 13.12 max on windows.
-
The best choice:
AMD 13.12 Catalyst + SDK 2.9
AMD 7850 CARD =48Khash/s ;
AMD 7970 CARD =16Khash/s ;
cgminer.exe --neoscrypt --fix-protocol -I 12,12 --worksize 256,384 --thread-concurrency 8192,8192 --shaders 1024,2048 -g 1 -o http://pool.ftc-c.com:19328 -u TRsuTqxAMKZNvC7JzGNgvM49rkQNcxqjhi -p x
-
I gave it a try last night on my nVidia 550M, Win7 and I got max 8 khs. Mostly default settings, --worksize 256 -I 13, no overclock. Driver version 294.10 - a bit old
76C gpu temp -> not bad
I think it can reach up to 10 khs with some tweakin
This is what i get with -I 14
Limitation of the gpu?
And whats the logic 7850 to perform 3 times faster than 7970?
Do you guys expect the stratum implementation to raise the productivity significantly?
-
Try dropping the intensity I am using 11 on my 7870, on I 12 it drops performance by half, and on 14 I would probably only get around 8 too.
For normal Scrypt I 12 is the sweet spot for me but 11 is required for Neoscrypt
-
Is there anything special you needed to do to get cgminer working with nvidia?
-
Is there anything special you needed to do to get cgminer working with nvidia?
absolutely nothing… works right outta the box. I am curious to see what a decent desktop nvidia could do
-
i have a GTX780Ti clocked at 1Ghz i could give a bash?
Anywhere i should get all the files from? Thanks
-
ive got three 750ti cards in a rig at home I may have a try and get this miner up and running with them.
-
I have one 750Ti and get about 35khash reported by the miner with default settings, just I=10
If I set intensity to a higher value, hw errors increase.
By the way, the pool reports about 24 khash only, but no errors.Some hashes ‘lost in transit’? ;)
-
Wellenreiter, do we need external help with the stratum? Do you think it would increase productivity?
I can get my 550M to work up to I-13 without HW. -I 12 and -I 13 make little difference tho.
I hate to say it but it seems the 7850 has the edge over 750ti productivty wise while at same price… :( Tho I am not sure if lizhi had aggressive overclock or its just too hot in China :D - 86C
What temps are you getting with the 750ti ? Have you tried I-13 ?
-
I gave it a try last night on my nVidia 550M, Win7 and I got max 8 khs. Mostly default settings, --worksize 256 -I 13, no overclock. Driver version 294.10 - a bit old
76C gpu temp -> not bad
I think it can reach up to 10 khs with some tweakin
This is what i get with -I 14
Limitation of the gpu?
And whats the logic 7850 to perform 3 times faster than 7970?
Do you guys expect the stratum implementation to raise the productivity significantly?
cgminer say “Maximun buffer memory device 0 supports says 536870912”
Please reduce the thread-concurrency
