|
You know your tools are getting old when
1. Your Ryzen 5 2600 shows up in HWiNFO with a 108 Degrees core temp.. like wtf
2. Your Vega 64 reference shows up in GPU-Z with 8 Gig of Micro memory. like wtf. I know it has Samsung..
3. ATI pixel clock patcher complains that the driver file is too large.
Luckily all above items were resolved by upgrading the corresponding tool...
Time to overhaul the software repo..
Goes to show you the amount of innovation going on in this field called mining..
Perhaps I should mention that I needed to upgrade OverDrivenTool to 0.2.8. beta1 after I upgraded my AMD driver to 19.4.1 and lost fan control on my RX 570s...
Then I had to upgrade my Teamviewer to 14 something after it complained that the machine I tried to connect to had 14 something....
BTW does anyone know of a replacement for TeamViewer? I am getting tired of its ways.. I can only stay connected a few minutes at a time now it suspects business use...
|
|
|
|
|
The RX 570 is the goto-card on the AMD side. I sold my RX 580s only to buy back the same number of RX 570, because they consume less power, and the hash is the same.
The 1660 looks to be a good contender to replace my 1070s...
The whattomine power figures are waay off. Looks like all cards consume 130 watts carte blanche...
My RX 570s need 110 Watts at 30.5 MH/s ETH
My GTX 1070s are at 115-117Watts mining Ravens...
|
|
|
|
@eliovp, thank you for this tool.
Do you or anybody else know if there is a tool available in Linux to count GPU memory errors? Much like HWiNFO64 in Windows?
Hey Bobben2, Yeah, it's not as obvious as on windows, that's a fact, try to check dmesg log, this gives you a lot of info as well. Because if available, it in theory would be possible to build an auto-tune type of tool to find the tightest timings that produce no memory error, for some mining algos (CN/R and ETHash for example or for Raven's X16R that might benefit from changing timings on the fly depending on which of the 16 algos currently running). Maybe such a tool could be available to miner developers as a 3rd party library that could be linked into their application... |
|
|
|
|
@eliovp, thank you for this tool.
Do you or anybody else know if there is a tool available in Linux to count GPU memory errors? Much like HWiNFO64 in Windows?
|
|
|
|
|
For some reason this miner works waay better under Linux than Windows 10.
On Xubuntu 18.04 my 2 x Sapphire Pulse RX 574 are mining Monero at 960H/s each at 1175/2000, core Volt at 850mV, consuming only 225 Watts for the whole 2-card rig.
The same settings in Win and it was above 240 Watts with core volt 900mV for stability and hash was "only" 933H/s each card..
|
|
|
|
XMR has made my Vegas profitable to run again!
...at least for the next few days.
...not yet at the ugly $0.16/KWh I'm dishing out for power, but if it keeps going in the right direction, it might just fairly soon! There's no secret though. Prices need to go up for profits to rise as well. Wonder how long it'll take for Asics (and FPGAs) to kick into gear and up the diff on CN/r. Oh btw have any other coins forked to CN/r yet, or only XMR thus far? Is it still true to say that Cryptonight algo (including the forks) is still lowest in power consumption compared to all the other algos? Yes it's the lowest power usage mining algo out there. I am actually trying to build a enclosed rig which is more living room friendly when the guests or family arrive. Seems with adequate cooling you can easily do 5-6 GPUs in a server chassis without the temps going up or without the fans going max speed. XMR however is not as easy as ETH to set up properly. The bios needs to be difficult and if you take the time to tune each GPU individually you can get decent power efficiency. So it's a good thing that it's hard to tune because most lazy miners might just stay on ETH using Claymore. Should keep the difficulty low temporarily.... until the FGPA arrive again that is. I would say XMR mining is consuming less power only by a small margin now... Here is some numbers from my test rig with 2 x Sapphire Pulse RX 574 cards. Here goes system idle 43 Watts XMR mining: 240 Watts, Hashrate: 1870H/s 1150/1975 ETH mining: 255 Watts, hashrate: 61 MH/s 1100/2000 The XMR power usage is average figure since it fluctuates between 230 and 260 Watts. I probably need to tune the GPUs more with XMR.. Even rebooting the rig with the same GPU config will yield a different hashrate/power figure. With ETH mining it is really easy... My other rigs are all mining just ETH, just for the of configuring.. |
|
|
|
HELP
Trying to run this miner with 2 x Sapphire Pulse RX 574, miner version 0.4.1. Win 10, AMD driver version 18.6.1.
Miner runs for a while, then driver gets stuck in device driver.. and system reboots..
Usually, this means too much undervolt, but I am now running both GPUs at 1200/2000 at 1V core and it still it crashes....
The previous version ran fine mining CNV8 coins.
EDIT:
Lowered core freq to 1150. Seems more stable now. But I see a lot of memory errors in HWiNFO now... and mem is only at 2000...
Hmm. Maybe these cards were not made for monero mining.. A lot better in ETH mining they are..
1V core sounds crazy high for 570s. I run 580-8s between 812-843mv w/ 1225 core (some down to 1200,) for ~1Kh/s. Mem speed is totally dependent on the mfg and your straps - 2K is pretty much the upper limit for the PBE one-click straps on samsung, so if that's what you're running, maybe dial it down a touch. Also, this algo (or maybe this implementation) is a bit funny - it almost seems like voltage requirements may be a bit quantized. For instance, on pretty much all my 580s, I couldn't get 1250 running w/o errors even at 850mv. But after dropping to 1225, I could reduce my voltage all the way down to 812mv on a number of cards before errors started appearing again. Thanks for all the advice. Yeah, 1V is uber. But power draw and temps are reasonable. I reduced core freq and mem freq to 1150/1975 with core volt 925 and rig ran stably overnight, both cards hashing at 935H/s. Still memory errors. Will work on these three parameters when I get home from work: Core freq Core volt Mem freq Looks like the mem strap is too tight for CNR, so will have to reduce mem freq further. Cards work flawlessly at 1100/2000 on ETH with core volt at 813mV. For CNR core volt needs at least 925 mV... Anyhow, the miner software looks to be stable. Thankyou goes to the developers. |
|
|
|
|
HELP
Trying to run this miner with 2 x Sapphire Pulse RX 574, miner version 0.4.1. Win 10, AMD driver version 18.6.1.
Miner runs for a while, then driver gets stuck in device driver.. and system reboots..
Usually, this means too much undervolt, but I am now running both GPUs at 1200/2000 at 1V core and it still it crashes....
The previous version ran fine mining CNV8 coins.
EDIT:
Lowered core freq to 1150. Seems more stable now. But I see a lot of memory errors in HWiNFO now... and mem is only at 2000...
Hmm. Maybe these cards were not made for monero mining.. A lot better in ETH mining they are..
|
|
|
|
Released NPlusMiner 4.7.3https://github.com/MrPlusGH/NPlusMiner/releasesLast Release:   Discord: https://discord.gg/2BCqPxe NPlusMiner with GUI, AutoUpdate, Earnings Tracker and Monitoring
Candidate for AutoUpdate = Yes - NPlusMiner checks for updates every 24 hours. - If you do not want to wait, simply restart NPlusMiner and it will pick it up on startup
ChangeLog - Added NiceHash MTP support Is there a way to avoid having to re-type the config parameters for every new version? Also, when CPU mining, it picks the least efficient version (usually sse2) even when I have Has(h)well? And for CPU mining the JayDee opt-miner is not the most efficient. Example: For M7 the specxx minerd is better |
|
|
|
|
The clue will be in the log.
|
|
|
|
|
I am using miningpoolhub for Monero (and Eth) mining for the last half year.
It is very stable and you can specify quite small payout amounts (0.05 XMR).
|
|
|
|
Some early test result with version 1.3.2 My 2 x RX 574 test rig now hashing at 913+914 H/s. And both GPUs temp/RPM are showing. So far all is well. Thank you, dev  Config: GPUs running at 1200/2000 MSI with Hynix mem. "gpu_conf" : [ { "id" : 0, "intensity" : 56, "worksize" : 8, "threads" : 2}, { "id" : 1, "intensity" : 56, "worksize" : 8, "threads" : 2} ] Algo heavy? Ah, forgot to mention its normalv7 |
|
|
|
Some early test result with version 1.3.2 My 2 x RX 574 test rig now hashing at 913+914 H/s. And both GPUs temp/RPM are showing. So far all is well. Thank you, dev Config: GPUs running at 1200/2000 MSI with Hynix mem. "gpu_conf" : [ { "id" : 0, "intensity" : 56, "worksize" : 8, "threads" : 2}, { "id" : 1, "intensity" : 56, "worksize" : 8, "threads" : 2} ] |
|
|
|
|
Often an Intel chip has a graphics unit built-in while an AMD processor may not.
The advantage of having it built in is you can connect a display to the mobo instead of the gpu, which will be mining usually..
|
|
|
|
|
Another bug?
Miner issues this
[2018-04-10 19:00:53] Warning: thread1 hashing speed is 0 (1)
[2018-04-10 19:00:53] Warning: thread2 hashing speed is 0 (1)
GPU is MSI RX 574 running at 1100/2050 and gpu_conf intensity set to 0 for auto-selection (56).
The computer freezes and reboot is required.
Happens on two separate rigs after about one day of mining..
|
|
|
|
|
Is there a bug in the reporting of GPU temp and RPM?
Always the same values for GPU 0 and 1 and 2 and 3?
Both my 4-GPU rigs exhibit this behaviour. Version 1.3.0
|
|
|
|
Now using Claymore 11.3 mining Monero with -pow7 1 and it works wonderfully. All my RX 570s hashing at 850-855H/s Claymore, you are the champ! Can you give me the detail of config, thanks ? Sure, this is my command line NsGpuCNMiner.exe -pow7 1 -h 928 -dmem 1 -xpool stratum+ssl://europe.cryptonight-hub.miningpoolhub.com:20580 -xwal bobben.%computername% -xpsw %computername% The above -h value is for 4 gig cards. If you hasve 8 gig cards, then you can set that value higher. I use OverdriveNTool to set gpu core clock (P7) to 1150 and mem (P2) to 2000. Many thanks for your feedback. Because my rigs can not run with algo -pow7 1 Update : With your config, still getting error : Error in server response : {"id":1,"jsonrpc":"2.0","error":{"code":-1,"message":"Sumohasforkedtonewalgo.YoushoulduseXMRig2.6orXMR-Stak2.4.2."}} Trying to mine Sumo? Claymore 11.3 does not support CN Heavy. Look here to find a miner that can mine you Sumokoin. https://bitcointalk.org/index.php?topic=2345911.0Actually, the pool is very good here. It suggests to you which miner to use. |
|
|
|
Now using Claymore 11.3 mining Monero with -pow7 1 and it works wonderfully. All my RX 570s hashing at 850-855H/s Claymore, you are the champ! Can you give me the detail of config, thanks ? Sure, this is my command line NsGpuCNMiner.exe -pow7 1 -h 928 -dmem 1 -xpool stratum+ssl://europe.cryptonight-hub.miningpoolhub.com:20580 -xwal bobben.%computername% -xpsw %computername% The above -h value is for 4 gig cards. If you hasve 8 gig cards, then you can set that value higher. I use OverdriveNTool to set gpu core clock (P7) to 1150 and mem (P2) to 2000. |
|
|
|
Now using Claymore 11.3 mining Monero with -pow7 1 and it works wonderfully. All my RX 570s hashing at 850-855H/s Claymore, you are the champ! |
|
|
|
|
Show Posts
