Pages:
Author

Topic: Claymore's Dual Ethereum AMD+NVIDIA GPU Miner v15.0 (Windows/Linux) - page 19. (Read 6590718 times)

member
Activity: 220
Merit: 12
Hello Guys, I have a 8x RX580 8GB Rig stable for last 2 months but from last 3 days it's having "cannot write buffer for DAG" issue.
I'm using nicehash, It is also happening on Betterhash, and I even tried using Claymore standalone mining on a demo address, it still gives " cannot write buffer for DAG ".

I've experimented a lot and observed that as soon as the eth epoch #379 starts to load up, the DAG size becomes 3.96GB and the DAG error comes up, But if it's still working on epoch #368 (size 3.78GB) then it keeps working fine.

I have tested on Windows 10 Build 2004 and also on Windows 10 build 1708 (just to make sure it's not windows update related) and have also tried 6 different Radeon Drivers ranging from Dec 2019 to Sep 2020 but all are having the same issue. On the Phoenix, it seems to work fine and Team Red Miner also works fine.
I have my Virtual memory set to 250GB (yes TwoHundredFifty GB) and I have upgraded my 4GB to 8GB system RAM and it still gives the same error. I've tried using the Gen 3 link speed on the PCI-E bus, the same error. This issue is only in Claymore, seems to me that somehow the claymore is unable to access more than 4GB of VRAM on 8GB RX580 cards. Please, if anyone found a working solution, post it here. Thanks for reading all of it.
Update the drivers to the newest ones.
newbie
Activity: 7
Merit: 1
Hello Guys, I have a 8x RX580 8GB Rig stable for last 2 months but from last 3 days it's having "cannot write buffer for DAG" issue.
I'm using nicehash, It is also happening on Betterhash, and I even tried using Claymore standalone mining on a demo address, it still gives " cannot write buffer for DAG ".

I've experimented a lot and observed that as soon as the eth epoch #379 starts to load up, the DAG size becomes 3.96GB and the DAG error comes up, But if it's still working on epoch #368 (size 3.78GB) then it keeps working fine.

I have tested on Windows 10 Build 2004 and also on Windows 10 build 1708 (just to make sure it's not windows update related) and have also tried 6 different Radeon Drivers ranging from Dec 2019 to Sep 2020 but all are having the same issue. On the Phoenix, it seems to work fine and Team Red Miner also works fine.
I have my Virtual memory set to 250GB (yes TwoHundredFifty GB) and I have upgraded my 4GB to 8GB system RAM and it still gives the same error. I've tried using the Gen 3 link speed on the PCI-E bus, the same error. This issue is only in Claymore, seems to me that somehow the claymore is unable to access more than 4GB of VRAM on 8GB RX580 cards. Please, if anyone found a working solution, post it here. Thanks for reading all of it.
newbie
Activity: 2
Merit: 1
Hello Everyone   Smiley
first post!!

i am running 2 rigs, all cards are RX 580 8g armors, Windows 10 with claymore v15, both rigs stopped working when the ETC epoch hit #379
All GPU's  "not enough GPU memory to place DAG, you cannot mine this coin with this GPU"
"OpenCL error -61 - cannot allocate big buffer for DAG"
"OpenCl error -38 - cannot write buffer for DAG"
server: accept failed with error: 10004

my virtual mem is at 55000
AMD Drivers 19.9.2
im using AMD software for tuning GPU's "GPU 1165MHz, 853mV" "Memory 2080MHz, 853mV"

Now when i update drivers to either latest 20.9.2 or drivers back to near beginning of the year i can mine again but now the hashrate is 19.2
Without tuning gpus i get 22.4 hash

Before #379 my hashrate was 31.3

In my bat file i have "-eres 0 -epsw x -mode 1 -r 1 -strap 1 -rxboost 1"

looking forward to solution
thanks
hero member
Activity: 1036
Merit: 500
What should I do?
I tried mining using 4 gpu 1050ti 4gb, but I found that there was mine, there was an explanation of 0 mh/s and another 15mh/s.

Does anyone have a solution?
please helpme
I think there is something wrong in the VGA bios or take a look in the device manager, does the VGA driver have an exclamation mark (!)?
If the VGA driver shows a (!) Sign then I suggest restarting your mining rig
sr. member
Activity: 1498
Merit: 443
What should I do?
I tried mining using 4 gpu 1050ti 4gb, but I found that there was mine, there was an explanation of 0 mh/s and another 15mh/s.

Does anyone have a solution?
please helpme
legendary
Activity: 3808
Merit: 1723
Claymore may be placing both ETH and ETC DAG to each of your Cards.
No that is not what the -allcoins 1 command does. What it means is that during every single dev mining session you will need to load new DAG to mine for Claymore and after the dev miner is done, it will reload the older DAGs to continue your regular mining.
Isn't that for: -allcoins -1 not for -allcoins 1? TIA

It didn't worked anyways, so the issue is somewhere else.

I think what the above poster should try is to increase the virtual memory, only issue I can think of that causes this since all his gpus are 8GB. Phoenix works but maybe it allocates a slightly different amount of virtual memory and hence might not run into this issue.

Regarding "allcoins" command, I used it many times in the past.

Basically the issue is that if you mine some new altcoin with a Epoch of 10 and later switch to an Epoch of 300 for ETH, it will sometimes crash when it tries to switch to Epoch 300, so thats why there is the -allcoins -1 which mines dev fee first so loads Epoch 300 and later Epoch 10, this resolves some crashes however the issue is the constant writing of the DAG on the GPUs every 15 minutes, eventually one or more GPUs will crash.

This is why the only way to mine alts with Claymore was with the -devfree 0 at a reduced speed. It mined slower and not optimized but at least it never crashed because it never switched to dev fee mode.
legendary
Activity: 2534
Merit: 6080
Self-proclaimed Genius
Claymore may be placing both ETH and ETC DAG to each of your Cards.
No that is not what the -allcoins 1 command does. What it means is that during every single dev mining session you will need to load new DAG to mine for Claymore and after the dev miner is done, it will reload the older DAGs to continue your regular mining.
Isn't that for: -allcoins -1 not for -allcoins 1? TIA

It didn't worked anyways, so the issue is somewhere else.
legendary
Activity: 3808
Merit: 1723

Claymore may be placing both ETH and ETC DAG to each of your Cards.

No that is not what the -allcoins 1 command does. What it means is that during every single dev mining session you will need to load new DAG to mine for Claymore and after the dev miner is done, it will reload the older DAGs to continue your regular mining.

I know this for a fact because it was very difficult to mine small dagger alts with Claymore. Basically if you got an undervolted GPU, usually after the 5th or 10th DAG loading, one of the GPUs would hang.

So its definately not loading both ETH and ETC.
legendary
Activity: 2534
Merit: 6080
Self-proclaimed Genius
-snip-
Claymore may be placing both ETH and ETC DAG to each of your Cards.
I have the same problem... I can mine eth but not etc ( 8gb rx470 & rx480 ) -allcoins etc/-1 didn't work.
Then it may be a bug on ETC on Claymore's side, I can't see any other possible options to fix that other than using phoenix miner.
newbie
Activity: 2
Merit: 0
hey guys and miners
when i mine ethereum classic i keep get error with openCL error - cannot allocate big buffer for DAG & not enough gpu memory to place dag.
I use 8g rx580 & rx570 and claymore v15 and windows 10
i tried -eres 0 command line and allocate more virtual memory(60gb) but not get resloved!!
what should i do if want to mine ETC??
Hey. I have the same problem. Claymore stopped with error, launched on Phoenix, only the hashrate is lower



on Phoenix,miner(.bat file) keep exiting and not running but for eth didn't cause any problem.
newbie
Activity: 13
Merit: 0
Hi

I have asus mining expert and win 10 pro. I have 14 different nvidia card. Mining was succesful many years but now i have problems. I updates my rig(Fresh win10,drivers,mining software etc). All cards show ok in win10. Claymore get cuda error  if i enabled more than 12 gpus when mining. I tested all cards and all cards mining but only 12 same time. All card passes benchmark fine one by one. Multiple benchmark never work:) Is claymore limit mining gpus to only 12?


PetriJuhani

newbie
Activity: 1
Merit: 0
when i mine ethereum classic i keep get error with openCL error - cannot allocate big buffer for DAG & not enough gpu memory to place dag.
I use 8g rx580 & rx570 and claymore v15 and windows 10
That's odd for an 8GB card, try to add:
Code:
-allcoins etc
or
Code:
-allcoins -1
Claymore may be placing both ETH and ETC DAG to each of your Cards.
I have the same problem... I can mine eth but not etc ( 8gb rx470 & rx480 ) -allcoins etc/-1 didn't work.
newbie
Activity: 29
Merit: 0
hey guys and miners
when i mine ethereum classic i keep get error with openCL error - cannot allocate big buffer for DAG & not enough gpu memory to place dag.
I use 8g rx580 & rx570 and claymore v15 and windows 10
i tried -eres 0 command line and allocate more virtual memory(60gb) but not get resloved!!
what should i do if want to mine ETC??
Hey. I have the same problem. Claymore stopped with error, launched on Phoenix, only the hashrate is lower
legendary
Activity: 3808
Merit: 1723
Hello, ive a problem with Claymore (with Phonenix without problem)

ive 20 GB in virtual memory.  and ive 5 cards (the first main 8 gb, the rest 4 gb, all RX 570))

but in claymore only 1 start (the main, 8 gb) . SO i have 28 Mh/s 0, 0, 0 0 ....

With phoenix i dont have problems.

what can i do?

Win 10, Adrenaline 17.5 in compute mode

I had the exact same issue but with Ubuntu, for some reason Claymore 15 doesn't work anymore with my 4GB cards. With Windows it stopped working even earlier than that. Which is a shame because I liked Claymore with the remote manager and stability and all. However Phoenix pretty much seems to do everything Claymore can, the commands might be slightly different but the features are the same.

The dev is also more active on the forum there and he can do any bug fixes that you are experiencing. Claymore hasn't been around for almost a year and I am pretty sure he won't be doing anymore updates or bug fixes for his miners.
jr. member
Activity: 60
Merit: 2
Hello, ive a problem with Claymore (with Phonenix without problem)

ive 20 GB in virtual memory.  and ive 5 cards (the first main 8 gb, the rest 4 gb, all RX 570))

but in claymore only 1 start (the main, 8 gb) . SO i have 28 Mh/s 0, 0, 0 0 ....

With phoenix i dont have problems.

what can i do?

Win 10, Adrenaline 17.5 in compute mode
DrG
legendary
Activity: 2086
Merit: 1035
Hi guys, hope you are doing well today.

Wanted to know, any chance to ask Claymore to update his drvfee address? As Dwarfpool is not working anymore his mining configuration may be affecting the mining process and dropping the hashrate more than it should. I don't know if that's just me, but in my case the hashrate drops significantly everytime the rig wants to connect to his Dwarfpool mining address as the pool is not working anymore. Anyone can let him know? please.

We asked him to fix this many times in the past. I also had issues with the Dwarfpool dev fee issues. I also had other issues where it fails to mine on the dev fee pool and eventually stops mining all together because the software thinks you are trying to cheat it. Very annoying because it causes the rigs to do a full hang and require a cold restart.

Last time he was active in this thread was Nov 2019, he was usually very active in the past, even when he didn't make updates. Now I think he might be retired from developement. He made crazy gains since his miner was the only decent software during the 2017 market boom and probably saved up tons of ETH and just living on an island somewhere.

I supposed something like that. So, we are using an out of date software that is not being maintained and is causing problems due to the developer lack of commitment. Is he even alive? I can't believe someone could just let the golden eggs chicken and just leave. Anyway, hope he fixes the problem.

Thanks for letting me know...

There are plenty of other capable miners that mining as well if not better than Claymore (at least on the pool reported hashes end). Just because a miner says it's giving a high rate doesn't neccessarily mean it is performing the best. A lot of lazy miners don't bother looking to make sure the pool side is reflecting what the terminal side says. If dual mining was still viable this miner would be decent but those days have for the most part passed.

Hi, thanks for your response. In my case it is not about lazy it's about knowledge. I will search for new miners then. It is a shame Claymore has this problems, it is a very good software. Any chance you could point the name of some better software you are aware of?

Thanks a lot.

Sorry I didn't see your response earlier. Phoenixminer seems to be a good alternative for mining under Windows environment (at least for Dagger Hashimoto). TeamRedMiner is also well supported by many pools. There are other older miners but you can expect some issues with various pools and perhaps not the best performance on newer hardware using those older miners. To find them you simply need to scroll through pages here on the forums.

Be careful if searching for miners on google. That's a good way to introduce virus/trojans onto your system/network. They prey on new users who don't pay attention to unsafe links they click on.
legendary
Activity: 2534
Merit: 6080
Self-proclaimed Genius
when i mine ethereum classic i keep get error with openCL error - cannot allocate big buffer for DAG & not enough gpu memory to place dag.
I use 8g rx580 & rx570 and claymore v15 and windows 10
That's odd for an 8GB card, try to add:
Code:
-allcoins etc
or
Code:
-allcoins -1
Claymore may be placing both ETH and ETC DAG to each of your Cards.
legendary
Activity: 1163
Merit: 1005
hey guys and miners
when i mine ethereum classic i keep get error with openCL error - cannot allocate big buffer for DAG & not enough gpu memory to place dag.
I use 8g rx580 & rx570 and claymore v15 and windows 10
i tried -eres 0 command line and allocate more virtual memory(60gb) but not get resloved!!
what should i do if want to mine ETC??
Have you set the VGA bios to mining bios?
If you have changed the bios to mining bios, I think you have to update the bios or drivers.
newbie
Activity: 2
Merit: 0
hey guys and miners
when i mine ethereum classic i keep get error with openCL error - cannot allocate big buffer for DAG & not enough gpu memory to place dag.
I use 8g rx580 & rx570 and claymore v15 and windows 10
i tried -eres 0 command line and allocate more virtual memory(60gb) but not get resloved!!
what should i do if want to mine ETC??
DrG
legendary
Activity: 2086
Merit: 1035
you all having doubts on what to mine with 4GB or even 3GB cards, the news is that ETC will re-calibrate the DAG file size making it possible again

for more info -> https://twitter.com/eth_classic/status/1314106504513871872?s=20


ETC has a network hashrate of 5TH, ETH has a network hashrate of 250TH.
It is assumed 50% of that 250 is 4GB or less (Gpus or Asics).
ETC mining profitability with drop like bricks if they drop the DAG and all that 125TH hashrate shifts to it - could even be higher if the already offline 3GB gpus join in.
Network hashrate will go from 5TH to ~130TH - thats a 2600% increase and will see a proportional difficulty increase, tanking profitability , so no, its not really an alternative.

The ones running on expensive power who came back to ETH mining from dusty closets will go back to dusty closets or give up mining. The ones running on cheap/free power will switch over to ETC. ETC profitability will of course tank, but the network will also become much more secure (hopefully not 100TH coming from a giant ASIC farm in China). The market will typically value ETC higher and the price will rise slightly - but obviously not enough to compensate for a 25x increase in difficulty. Perhaps this will breathe new life into ETC - it's happened to other coins before.
Pages:
Jump to: