Author

Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 457. (Read 6590565 times)

sr. member
Activity: 1008
Merit: 297
Grow with community
Is it possible to delay all mining until DAG generation completes?  I have a GPU that's pretty touchy and if the mining starts before it's DAG is generated the GPU will never come up.  This causes the app to restart and sometimes crash.   I've used the serial options and tried to slow it down but that doesn't prevent the mining from starting before DAG generation on all cards.  Just a thought.

Thanks!


see the documentation:

-gser   this setting can improve stability on multi-GPU systems if miner hangs during startup. It serializes GPUs initalization routines. Use "-gser 1" to serailize some of routines and "-gser 2" to serialize all routines.
   Using values higher than 2 allows you also to set custom delay between DAG generation on GPUs, for example, "-gser 5" means same as "-gser 2" and also adds 3sec delay between DAG generation (can be useful for buggy drivers and/or weak PSU).
   Default value is "0" (no serialization, fast initialization).

 i would recomment starting with -gser 3 to serialize it with 3 sec. delay.

if that doesnt help you can also try it with -lidag 2, which sets the intensity lower when DAG creation

Thanks for the response.   I've already tried the queuing options for serializing the startup including 'gser' and 'lidag'.  It's not that the card crashes during the DAG generation, but when mining starts before the DAG generation is completed.    The 'gser' option works to serialize the start, but it doesn't stop mining from starting before DAG is complete. 

Any other thoughts?


did you set your Virtual mem to at least 16GB, I was also thinking a hardware issue, are you using risers? if you could post your HW specs then its more appreciated.

newbie
Activity: 6
Merit: 0
Is it possible to delay all mining until DAG generation completes?  I have a GPU that's pretty touchy and if the mining starts before it's DAG is generated the GPU will never come up.  This causes the app to restart and sometimes crash.   I've used the serial options and tried to slow it down but that doesn't prevent the mining from starting before DAG generation on all cards.  Just a thought.

Thanks!


see the documentation:

-gser   this setting can improve stability on multi-GPU systems if miner hangs during startup. It serializes GPUs initalization routines. Use "-gser 1" to serailize some of routines and "-gser 2" to serialize all routines.
   Using values higher than 2 allows you also to set custom delay between DAG generation on GPUs, for example, "-gser 5" means same as "-gser 2" and also adds 3sec delay between DAG generation (can be useful for buggy drivers and/or weak PSU).
   Default value is "0" (no serialization, fast initialization).

 i would recomment starting with -gser 3 to serialize it with 3 sec. delay.

if that doesnt help you can also try it with -lidag 2, which sets the intensity lower when DAG creation

Thanks for the response.   I've already tried the queuing options for serializing the startup including 'gser' and 'lidag'.  It's not that the card crashes during the DAG generation, but when mining starts before the DAG generation is completed.    The 'gser' option works to serialize the start, but it doesn't stop mining from starting before DAG is complete. 

Any other thoughts?
full member
Activity: 209
Merit: 100
So, new dual miner algo's
Keccak
TimeTravel
m7
BlakeR8, i'm aware of the ASIC, however, they are merged miner with at least four coins, BLC, LIT, PHO, UMO.  This algo is best known for Verge at the moment and is likely to draw attention only there.
newbie
Activity: 43
Merit: 0
my prediction is that now dual mining is dead, claymore will lose a lot of clients unless he finds some other coins to dual mine

Some things just can't happen. I found a reddit post that seems legit, which explains that:

Ethereum is memory intensive and mining is bottlenecked by the memory speed, while the compute units are waiting on the memory it uses those cycles to mine the second currency

That "second currency" was based on Blake algorithm -which now has ASICs available-.
What you imagine as "some other coins to dual mine" would be of another algorithm, lets say CryptoNight for example. As far as I know, CryptoNight is memory intensive (just like Ethhash) and there is only room for one memory intensive coin per card.

All of the above is my personal understanding, do not take it for granted.

That is right. We cannot mine the DCR and SIA soon.

That depends on your electricity cost. In my case (I have very expensive electricity, 0.3$/kwH) I have already stopped dual mining

Maybe you misread, i never said we can mine cryptonight and ethash together.

This is why i said he must find out other coins to be mined.
newbie
Activity: 1
Merit: 0
Hi Guys,

Not sure if I am in the right thread but google just dropped me here.

I have a Sapphire TriXX R9 290x where I've just replaced the crappy stock fans with aftermarket one's that are NOT using the Fan Connector of the GPU.
They run full speed all the time and all seems to be fine but the card is not detected by Claymore.
However it is showing up in Afterburner and Windows detected it as well.

The card was detected correctly with the Stock Fans but they died and I replaced them with 100mm ones that connect to a Fan Management.

Is there some settings in Claymore that prevent the card from showing up if no Fan is connected?

Cheers and hoping for some help as I do not want to buy original replacement fans from China that need ages to arrive.
full member
Activity: 209
Merit: 100
Noob here and kind of late in the game but still want to see how things work.

I have a PC with Windows 10, Intel based with eVGA 980 TI video card.  I just downloaded 10.6 version of  Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v10.6 (Windows/Linux).

I am getting around 10.70+ Mh/s as shown on the Windows console running "ETHEREUM-ONLY MINING MODE ENABLED"

Is this normal?  If not, what can I do get the yield higher?

If things go well, I am thinking of building a rig with 4 GPUs but we will see.

Thanks,

Rick..
Use Linux even with the workaround in windows the speed is still slow on Nviidia cards
member
Activity: 119
Merit: 10
my prediction is that now dual mining is dead, claymore will lose a lot of clients unless he finds some other coins to dual mine

Some things just can't happen. I found a reddit post that seems legit, which explains that:

Ethereum is memory intensive and mining is bottlenecked by the memory speed, while the compute units are waiting on the memory it uses those cycles to mine the second currency

That "second currency" was based on Blake algorithm -which now has ASICs available-.
What you imagine as "some other coins to dual mine" would be of another algorithm, lets say CryptoNight for example. As far as I know, CryptoNight is memory intensive (just like Ethhash) and there is only room for one memory intensive coin per card.

All of the above is my personal understanding, do not take it for granted.

That is right. We cannot mine the DCR and SIA soon.

That depends on your electricity cost. In my case (I have very expensive electricity, 0.3$/kwH) I have already stopped dual mining
newbie
Activity: 56
Merit: 0
my prediction is that now dual mining is dead, claymore will lose a lot of clients unless he finds some other coins to dual mine

That is right. We cannot mine the DCR and SIA soon.
newbie
Activity: 43
Merit: 0
my prediction is that now dual mining is dead, claymore will lose a lot of clients unless he finds some other coins to dual mine
newbie
Activity: 7
Merit: 0
Anyone got an idea how to reach my payout limit on Siacoins at nanopool - now ASIC mining is destroying GPU mining of Siacoins and it seems impossible to get any coins with claymore.
I am only short of 5 coins to reach the payout limit...

   How about you keep on mining? a single AMD 570/580 Gpu can still do ~2.75 siacoins a day (and this is severely downclocked/undervolted). If you got multiple GPUs 5 coins could be produced in less than a day. If you got Nvidia, a 1070 should be about 60% of that when dual mining. 5 siacoins is far away form being considered impossible atm, even with single GPUs. P.S: I am no longer dual mining, but just checked whattomine with my card hashrates to see how many coins one would be getting at current difficulty. 
Horrible idea. Compete with ASIC that is 1000x faster? DCR is still good dual mining, but it is one of option on Baikal Giant B, so it may go down at some point.

See profit here...

http://xeridea.us/charts

xeridea, i think you are missing the point. area511 is asking how he can reach his payout threshold to get his balance out of the pool, he is only 5 siacoins short, so I am telling him that mining those last 5 coins would not require that much time as he believes, even with one GPU. I am not suggesting he keeps on mining after that. the longer he waits to mine those last coins the harder it will get.  
newbie
Activity: 2
Merit: 0
Hi, Claymore

I have a problem with my 6*rx470.

When I try dual mining, the clock core could not be locked, it changed from 600Mhz to 1120Mhz,
So the hashrate was changing at dual mining mode.
It almost locked when only mining eth at 1120Mhz.

claymore: 10.5
jr. member
Activity: 73
Merit: 3

Thank you both for your comments. I am still trying to figure out what can be happening.


This is my current situation, after trying every power option in my windows 8.1. You can see that after exactly 2 minutes being AFK, mi hashrate drops from 30.5 Mh/s to 22 Mh/s, and returns to 30.5 Mh/s exactly when i move my mouse.

Also I've been mesuring power comsuption while in 30.5 Mh/s and while in 22 Mh/s and both are the same. It doesn't save any power...

I wasn't able to figure out how to change power level options in my GPU (Asus GTX 1070) so, maybe I still have light at the end of the tunnel.

You know, the funny thing is ... your temp doesn't drop.  It looks like the card is doing the same work but is being miscounted for some reason.  When mine slows, I usually see a couple of degrees drop off of the temp.

I finally found it. I was killing every process and service in windows until I reached the one... CCleaner.

I don't know how or what it was doing (I will investigate further), but being in background it was slowing my mining by 25-30%.

Now it is fine. Thanks to everyone that helped or tried to. I hope this can help somebody else.

That's good news.  Mine seemed to be related to the Hynix memory timings in PBE 1.6.7.  I changed them out and am now stable at 31.1MH/s for 24 hours running Claymore 10.5.  I'm going to leave it alone for a while.
newbie
Activity: 1
Merit: 0
Noob here and kind of late in the game but still want to see how things work.

I have a PC with Windows 10, Intel based with eVGA 980 TI video card.  I just downloaded 10.6 version of  Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v10.6 (Windows/Linux).

I am getting around 10.70+ Mh/s as shown on the Windows console running "ETHEREUM-ONLY MINING MODE ENABLED"

Is this normal?  If not, what can I do get the yield higher?

If things go well, I am thinking of building a rig with 4 GPUs but we will see.

Thanks,

Rick..
sr. member
Activity: 449
Merit: 251
Anyone got an idea how to reach my payout limit on Siacoins at nanopool - now ASIC mining is destroying GPU mining of Siacoins and it seems impossible to get any coins with claymore.
I am only short of 5 coins to reach the payout limit...

   How about you keep on mining? a single AMD 570/580 Gpu can still do ~2.75 siacoins a day (and this is severely downclocked/undervolted). If you got multiple GPUs 5 coins could be produced in less than a day. If you got Nvidia, a 1070 should be about 60% of that when dual mining. 5 siacoins is far away form being considered impossible atm, even with single GPUs. P.S: I am no longer dual mining, but just checked whattomine with my card hashrates to see how many coins one would be getting at current difficulty. 
Horrible idea. Compete with ASIC that is 1000x faster? DCR is still good dual mining, but it is one of option on Baikal Giant B, so it may go down at some point.

See profit here...

http://xeridea.us/charts
sr. member
Activity: 449
Merit: 251
Would it be possible to do ETH + XMR on an AMD graphics card?
ETH with the RAM, XMR with the processor.
Yes you could. Cards will be mine etherium and cpu independent from cards will mine monero.
I meant XMR with the GPU. The processor of the graphics card.
No, as has been stated 9000 times already. XMR is memory intensive. It barely uses any power when mining. Stock 570 (Hynix) does ~600H/s, with memory mod and OC can do 850+. Samsung mem 970+. Needs ~1240 core for Samsung mem, but power draw is super low.

XMR can be reasonably mined on CPU (~500H/s on Ryzen 1700X, stock) because it need about 2MB per thread, which can fit in the L3 cache on CPU, which of course has a high amount of bandwidth (Ryzen 7 has 16MB L3 cache).

XMR on Nvidia cards is just terrible. 500H/s on 1060, 700 on 1070, 950 on 1080 TI.
newbie
Activity: 7
Merit: 0
Anyone got an idea how to reach my payout limit on Siacoins at nanopool - now ASIC mining is destroying GPU mining of Siacoins and it seems impossible to get any coins with claymore.
I am only short of 5 coins to reach the payout limit...

   How about you keep on mining? a single AMD 570/580 Gpu can still do ~2.75 siacoins a day (and this is severely downclocked/undervolted). If you got multiple GPUs 5 coins could be produced in less than a day. If you got Nvidia, a 1070 should be about 60% of that when dual mining. 5 siacoins is far away form being considered impossible atm, even with single GPUs. P.S: I am no longer dual mining, but just checked whattomine with my card hashrates to see how many coins one would be getting at current difficulty. 
newbie
Activity: 1
Merit: 0
Hi Guys,

I'm trying to get the miner up and working on two pools, one in eth, one in dcr/lbc etc.
In Windows, eth connects fine, neither dcr nor lbc will work.
With the default start file, the eth connects, but everything else times out.

In Linux, same behavior, eth is fine (primary), dcr or lbc (secondary) do not. Both dcr and lbc are using suprnova.cc pools.

Router has put machine in DMZ. Firewall off in W10, no difference. Tried connection through two different ISP's, no difference.

Again to recap, secondary connection does connect according to the miner, and netstat shows established. However, only eth shows authorized.

The vid card is an NVidia 1080.

Have tried OP's drivers, latest drivers and mix of in betweens, incase it was cuda related.

terminal:
 LBC - Total Speed: 0.000 Mh/s, Total Shares: 0, Rejected: 0
 LBC: GPU0 0.000 Mh/s
 LBC: Stratum - connecting to 'lbry.suprnova.cc' <193.70.45.182> port 6256
 LBC: Stratum - Connected (lbry.suprnova.cc:6256)
ETH: 01/25/18-19:50:49 - New job from us1.ethermine.org:14444
.
ETH: 01/25/18-19:50:52 - New job from us1.ethermine.org:14444
.
ETH: Share accepted (63 ms)!
.
ETH: 01/25/18-19:51:02 - New job from us1.ethermine.org:14444
 LBC - Total Speed: 0.000 Mh/s, Total Shares: 0, Rejected: 0
 LBC: GPU0 0.000 Mh/s
Socket was closed remotely (by pool)
 LBC: Job timeout, disconnect, retry in 20 sec...

Thanks!

EDIT: Got it. Asus router/Merlin Firmware, Trend Micro settings. While all my A/V software was set correct, the routers A/V caught and killed the connections of the secondary connection created by the miner.
full member
Activity: 298
Merit: 100
21:06:07:808   1c6c   NVML: cannot get current temperature, error 999 (an internal driver error occurred)
21:06:07:808   1c6c   NVML: cannot get fan speed, error 999 (an internal driver error occurred)

Help?
newbie
Activity: 182
Merit: 0
Would it be possible to do ETH + XMR on an AMD graphics card?
ETH with the RAM, XMR with the processor.
Yes you could. Cards will be mine etherium and cpu independent from cards will mine monero.
I meant XMR with the GPU. The processor of the graphics card.
newbie
Activity: 17
Merit: 0
Is it possible to delay all mining until DAG generation completes?  I have a GPU that's pretty touchy and if the mining starts before it's DAG is generated the GPU will never come up.  This causes the app to restart and sometimes crash.   I've used the serial options and tried to slow it down but that doesn't prevent the mining from starting before DAG generation on all cards.  Just a thought.

Thanks!


see the documentation:

-gser   this setting can improve stability on multi-GPU systems if miner hangs during startup. It serializes GPUs initalization routines. Use "-gser 1" to serailize some of routines and "-gser 2" to serialize all routines.
   Using values higher than 2 allows you also to set custom delay between DAG generation on GPUs, for example, "-gser 5" means same as "-gser 2" and also adds 3sec delay between DAG generation (can be useful for buggy drivers and/or weak PSU).
   Default value is "0" (no serialization, fast initialization).

 i would recomment starting with -gser 3 to serialize it with 3 sec. delay.

if that doesnt help you can also try it with -lidag 2, which sets the intensity lower when DAG creation
Jump to: