6344khash?
How fast is djm34`s official lyra2v2 miner on the gtx 960?
Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 866. (Read 2347659 times)
nVIDIA DRIVER 3.58.91--
I woke, read the latest on this thread, and linked up to hashpower.co using the auto-switch batch file. My GTX 960 SSC rotated to Lyra2v2 and began hashing with no problems at about 344kh/s. This hash rate is about 15-20kh/s better than the hash rate with NiceHash.com for the same algo.
I suspect a poolside problem with NiceHash, now. The driver is working OK mining at hashpower.co However, my Linux rigs are mining on NiceHash with no problem.
--scryptr
I woke, read the latest on this thread, and linked up to hashpower.co using the auto-switch batch file. My GTX 960 SSC rotated to Lyra2v2 and began hashing with no problems at about 344kh/s. This hash rate is about 15-20kh/s better than the hash rate with NiceHash.com for the same algo.
I suspect a poolside problem with NiceHash, now. The driver is working OK mining at hashpower.co However, my Linux rigs are mining on NiceHash with no problem.
--scryptr
54% higher payouts is not a little bit..
Somone is renting your rigs cheap from nicehash and make 50% on the mulitpool!
Somone is renting your rigs cheap from nicehash and make 50% on the mulitpool!
You can use the nicehashminer or just switch to another pool like this one:
http://hashpower.co/site/mining
It pays bether than Nichash in most algos. but 2% fee and 2% more with BTC payouts.
This time only Lyra2v2 a bit more... Other is chiper. http://hashpower.co/site/mining
It pays bether than Nichash in most algos. but 2% fee and 2% more with BTC payouts.
You can use the nicehashminer or just switch to another pool like this one:
http://hashpower.co/site/mining
It pays bether than Nichash in most algos. but 2% fee and 2% more with BTC payouts.
hashpower.co lyra2v2 the last 24HOURS 0.3915BTC / GHASH
Nichash lyra2v2 the last 24HOURS 0.2530BTC/ GHASH
http://hashpower.co/site/mining
It pays bether than Nichash in most algos. but 2% fee and 2% more with BTC payouts.
hashpower.co lyra2v2 the last 24HOURS 0.3915BTC / GHASH
Nichash lyra2v2 the last 24HOURS 0.2530BTC/ GHASH
nVidia Driver 358.91, Win 7 x64--
I upgraded the driver on my Win 7 box today, and all I get from NiceHash is "retry in 30 seconds" using release dot 73. Not just Quark, but Lyra2v2 and Neoscrypt also. I rebooted, no better performance. I then played a game, as my GTX 960 on this box is usually the first to try something new. My rigs are Linux boxes.
I'll maybe roll the driver back tomorrow. --scryptr
It is a pool side problem. I use miner control and after 3 retries switching to another stratum.I upgraded the driver on my Win 7 box today, and all I get from NiceHash is "retry in 30 seconds" using release dot 73. Not just Quark, but Lyra2v2 and Neoscrypt also. I rebooted, no better performance. I then played a game, as my GTX 960 on this box is usually the first to try something new. My rigs are Linux boxes.
I'll maybe roll the driver back tomorrow. --scryptr
Well it's not likely a pool side problem if he is is the only one with the problem.
I have seen two similar problems on nicehash for which I have no explanation.
All of my rigs fail to connect to nicehash on the first atempt but suceed on the retry. There is no
error message other than the retry message. This has been happening for a few months and is
consistent. No one else seems to have reported this problem.
I have also seen repeated connection failures on one machine only. It would retry forever and always fail.
Rebottong the machine would fix it temporarilly. This problem ocurred for a while then vanished.
I'm not sure what made it go away, maybe a driver upgrade.
These problems have never occurred on any other pool, just nicehash, or more specifically westhash usa.
what is the default intensity for Lyra2rev2 on a 980 ti??
The default intensity is -X 18. You can test with -X 30. Might be faster?No overclocking. The intensity parameter adjusts the size of the number of hashes that are stored in memory.
You can verify this by running gpu-z and observe the allocated memory.
The -X intensity is linear growth of the memory buffer. The -i is exponential.
setting the intensity is a matter of using the gpu resources better.
overclocking works on top of it.
If I am not mistaking - I believe it goes like this - -X 18,18,15,20,18 (where the placement of the intensity number refers to your GPU number) So GPU 0,1,2,3,4 in a (5) card system? Correct me if I am wrong please.
So if I use the -X and -i on 3 cards it would look like this.
-d 0,1,2 -i 16,0,0 -X 18,30,30
Device 0 is the one I use for the graphics on this computer and the other 2, on this computer not using for graphics , I like to get the best hashrate.
Is this correct and since -X 18 is default?
thx
EDIT: I tried it on my 980ti that's in a p2 state and the -X 18 works and -X 20 works but -X 22 ran out of memory. So I'll not use the -X parameter at all and no real improvment.
nVidia Driver 358.91, Win 7 x64--
I upgraded the driver on my Win 7 box today, and all I get from NiceHash is "retry in 30 seconds" using release dot 73. Not just Quark, but Lyra2v2 and Neoscrypt also. I rebooted, no better performance. I then played a game, as my GTX 960 on this box is usually the first to try something new. My rigs are Linux boxes.
I'll maybe roll the driver back tomorrow. --scryptr
It is a pool side problem. I use miner control and after 3 retries switching to another stratum. I upgraded the driver on my Win 7 box today, and all I get from NiceHash is "retry in 30 seconds" using release dot 73. Not just Quark, but Lyra2v2 and Neoscrypt also. I rebooted, no better performance. I then played a game, as my GTX 960 on this box is usually the first to try something new. My rigs are Linux boxes.
I'll maybe roll the driver back tomorrow. --scryptr
nVidia Driver 358.91, Win 7 x64--
I upgraded the driver on my Win 7 box today, and all I get from NiceHash is "retry in 30 seconds" using release dot 73. Not just Quark, but Lyra2v2 and Neoscrypt also. I rebooted, no better performance. I then played a game, as my GTX 960 on this box is usually the first to try something new. My rigs are Linux boxes.
I'll maybe roll the driver back tomorrow. --scryptr
I upgraded the driver on my Win 7 box today, and all I get from NiceHash is "retry in 30 seconds" using release dot 73. Not just Quark, but Lyra2v2 and Neoscrypt also. I rebooted, no better performance. I then played a game, as my GTX 960 on this box is usually the first to try something new. My rigs are Linux boxes.
I'll maybe roll the driver back tomorrow. --scryptr
My poor superclocked 950 can barely get 5400 KH while my overclocked 750 TI can get north of 5200 KH using much less power- is there any more data I can give you to help with dialing in the 950? Intensity for both seems to top out at 20.
There is nothing wrong with your 950 or your config. All 9-series maxwell gpus are poor in lyra2v2 performance compared to 7-series maxwell hash/w wise. 950/60 really suck because of their crippled 128-bit memory lane. Better mine some other algo with them.
980 TIs arent too bad- but I get what youre saying. I think there is some room for optimization on the 950 however- Its should be able to hang and outpace the 750 TI by a wider margin-
Usually I mine Quark- but the payrate has tanked as of late...
The 750ti will go down as "The little engine that could". For what they are capable of power wise they can't be beat. If the other Maxwell cards could give the same performance (hash per wattage per price).
thefarm is a complete gigabyte 750ti oc lp powerhouse ...
the amount of power expended for the hashrate - has been the best i have had in a VERY long time ...
i cant give you static figures - just bills ( which i will on refer to - not specify ) - but the power bills in comparison to when thefarm was completely running gigabyte 7970 / 20x oc cards is a HUGE difference ...
even though the hashrate is much lower in comparison to the amd cards - nvidia has about half of the hashrate at about one third of the electricity bills ...
and from where im standing ( and forking out my pocket from ) thats a damn good deal ...
the private kernels would make it that much more efficient also - if anyone was willing to make a deal on them ...
#crysx
My poor superclocked 950 can barely get 5400 KH while my overclocked 750 TI can get north of 5200 KH using much less power- is there any more data I can give you to help with dialing in the 950? Intensity for both seems to top out at 20.
There is nothing wrong with your 950 or your config. All 9-series maxwell gpus are poor in lyra2v2 performance compared to 7-series maxwell hash/w wise. 950/60 really suck because of their crippled 128-bit memory lane. Better mine some other algo with them.Everything can be improved with bether code. Most of the developers have worked with the 750ti. But fast code on the 750ti isn't always fast on the other maxwell's.
My poor superclocked 950 can barely get 5400 KH while my overclocked 750 TI can get north of 5200 KH using much less power- is there any more data I can give you to help with dialing in the 950? Intensity for both seems to top out at 20.
There is nothing wrong with your 950 or your config. All 9-series maxwell gpus are poor in lyra2v2 performance compared to 7-series maxwell hash/w wise. 950/60 really suck because of their crippled 128-bit memory lane. Better mine some other algo with them.
980 TIs arent too bad- but I get what youre saying. I think there is some room for optimization on the 950 however- Its should be able to hang and outpace the 750 TI by a wider margin-
Usually I mine Quark- but the payrate has tanked as of late...
The 750ti will go down as "The little engine that could". For what they are capable of power wise they can't be beat. If the other Maxwell cards could give the same performance (hash per wattage per price).
Fully agree- especially when you consider the price point I picked it up at-
My poor superclocked 950 can barely get 5400 KH while my overclocked 750 TI can get north of 5200 KH using much less power- is there any more data I can give you to help with dialing in the 950? Intensity for both seems to top out at 20.
There is nothing wrong with your 950 or your config. All 9-series maxwell gpus are poor in lyra2v2 performance compared to 7-series maxwell hash/w wise. 950/60 really suck because of their crippled 128-bit memory lane. Better mine some other algo with them.
980 TIs arent too bad- but I get what youre saying. I think there is some room for optimization on the 950 however- Its should be able to hang and outpace the 750 TI by a wider margin-
Usually I mine Quark- but the payrate has tanked as of late...
The 750ti will go down as "The little engine that could". For what they are capable of power wise they can't be beat. If the other Maxwell cards could give the same performance (hash per wattage per price).
My poor superclocked 950 can barely get 5400 KH while my overclocked 750 TI can get north of 5200 KH using much less power- is there any more data I can give you to help with dialing in the 950? Intensity for both seems to top out at 20.
There is nothing wrong with your 950 or your config. All 9-series maxwell gpus are poor in lyra2v2 performance compared to 7-series maxwell hash/w wise. 950/60 really suck because of their crippled 128-bit memory lane. Better mine some other algo with them.
980 TIs arent too bad- but I get what youre saying. I think there is some room for optimization on the 950 however- Its should be able to hang and outpace the 750 TI by a wider margin-
Usually I mine Quark- but the payrate has tanked as of late...
My poor superclocked 950 can barely get 5400 KH while my overclocked 750 TI can get north of 5200 KH using much less power- is there any more data I can give you to help with dialing in the 950? Intensity for both seems to top out at 20.
There is nothing wrong with your 950 or your config. All 9-series maxwell gpus are poor in lyra2v2 performance compared to 7-series maxwell hash/w wise. 950/60 really suck because of their crippled 128-bit memory lane. Better mine some other algo with them.
what is the default intensity for Lyra2rev2 on a 980 ti??
The default intensity is -X 18. You can test with -X 30. Might be faster?
Thanks-
I have been setting intensity by device since I have a 950 in the mix. 20 seems to work well with the non-monitor 980 ti-
My poor superclocked 950 can barely get 5400 KH while my overclocked 750 TI can get north of 5200 KH using much less power- is there any more data I can give you to help with dialing in the 950? Intensity for both seems to top out at 20.
what is the default intensity for Lyra2rev2 on a 980 ti??
The default intensity is -X 18. You can test with -X 30. Might be faster?No overclocking. The intensity parameter adjusts the size of the number of hashes that are stored in memory.
You can verify this by running gpu-z and observe the allocated memory.
The -X intensity is linear growth of the memory buffer. The -i is exponential.
setting the intensity is a matter of using the gpu resources better.
overclocking works on top of it.
If I am not mistaking - I believe it goes like this - -X 18,18,15,20,18 (where the placement of the intensity number refers to your GPU number) So GPU 0,1,2,3,4 in a (5) card system? Correct me if I am wrong please.
what is the default intensity for Lyra2rev2 on a 980 ti??
The default intensity is -X 18. You can test with -X 30. Might be faster?No overclocking. The intensity parameter adjusts the size of the number of hashes that are stored in memory.
You can verify this by running gpu-z and observe the allocated memory.
The -X intensity is linear growth of the memory buffer. The -i is exponential.
setting the intensity is a matter of using the gpu resources better.
overclocking works on top of it.
what is the default intensity for Lyra2rev2 on a 980 ti??
The default intensity is -X 18. You can test with -X 30. Might be faster?No overclocking. The intensity parameter adjusts the size of the number of hashes that are stored in memory.
You can verify this by running gpu-z and observe the allocated memory.
The -X intensity is linear growth of the memory buffer. The -i is exponential.
setting the intensity is a matter of using the gpu resources better.
overclocking works on top of it.
Jump to: