GMiner v3.24 Ergo+KASPA+ZIL/KAWPOW/Equihash/Cortex - page 72.

kbs1

jr. member

Activity: 107

Merit: 1

Hi guys, my windows 10 1903 AMD rig is now showing this weird behavior:

6x RX580 8GB
driver 17.x
all cards set to 'compute' mode
gminer 1.50
attemping to mine grin C29
HDMI dongle connected to one of the GPUs
using only RDP to launch the miner

It 'just works'. For 12 hours. Or for 4 days. Sometimes 3. There are no exact times. When it 'stops working', I just see on the pool the miner disconnected.

I connect via RDP and violla - it starts up instantly!

Then I disconnect. Then it dies. In console, there's the following:

21:21:31 Total Speed: 11.01 G/s Fidelity: 0.993 Shares Accepted: 3973 Rejected: 5 Power: 725W 0.02 G/W
21:21:31 Uptime: 0d 18:14:37 Electricity: 13.222kWh $1.98
21:21:34 Connection Error: Write timed out
21:21:50 Connect to eu-west-stratum.grinmint.com:4416 Failed: No such host is known
21:21:52 Stopped Mining on GPU0
21:21:52 Stopped Mining on GPU1
21:21:52 Stopped Mining on GPU2
21:21:52 Stopped Mining on GPU3
21:21:52 Stopped Mining on GPU4
21:21:52 Stopped Mining on GPU5
21:22:06 Connect to eu-west-stratum.grinmint.com:4416 Failed: No such host is known
21:22:06 All user mining pools are unavailable
21:22:09 Miner terminated, watchdog will restart process after 10 seconds <--- miner keeps on restarting every 10 seconds without any further error message. It never starts.

Then I connect to RDP and boom, it instantly starts mining again. I don't have to restart the miner process by hand. Power profile is set to high performance. There are no sleeps, hibernates, power-downs, no "energy saving" features or anything like that.

It's not a connection issue, as I can always connect back to the machine via RDP. So the connection isn't 'asleep'. It's a ethernet connection, not a wifi.

Has anyone ever seen this behavior? It is a first time for me. It might be a bug in GMiner, but it's very very weird that it 'magically' comes back alive INSTANTLY when I connect via RDP when dead. Maybe some windows 'this proccess is using a lot of resources, so I'll pause it' feature? Or something like that? But as I said, I've never encountered anything like that.

Any help is greatly appreciated. Thank you!

coinzzzman

newbie

Activity: 13

Merit: 0

GRIN coin is updating its POW algorithm at block 262080 in next 48 hours.

Is GMINER coming out with an update before then?

jebidia

jr. member

Activity: 161

Merit: 1

Quote from: TBOEMC on July 12, 2019, 01:58:57 AM

Quote from: Zminer777 on July 10, 2019, 07:42:36 AM

Quote from: trucobit on July 10, 2019, 05:14:57 AM

Gminer already has several versions that does not work well in my rigti 1080ti and RTX 2080. It starts, detects the cards and randomly or stays stuck there for hours without undermining, or mine well. The only thing is reboot al rig. I am a user of Awesome miner and we recommend not using Gminer to the thousands of users of this program.

It occurs in the cucka and 150.5 and in some more, and it is not the OC, because it reboots and usually works with the same parameters.

Very much 2% Fee for such an unstable miner. I hope they can solve it because in the end the ones that lose the most are them, the less people use Gminer, the less they earn for the FEE. In its place for the Cucka I have to use nbminer that yields less but never fails.

Have you tried versions 1.48+? I hope we have fixed this issue.
Btw nbminer have same comission for grin - 2%, https://github.com/NebuTech/NBMiner

I have these same issues in rig with RTX 2070
Rig after fresh start works well, but randomly - sometimes after 2hours, sometimes half a day suddenly Gminer cannot start and there is only his black empty window, which is blinking cause it start and ends. The only possible thing is reset a whole rig. It not dependis on the version of Gminer (since 1.44 to 1.50) - i have this issue only with RTX. With 1070 and 1080 is it good. RTX cards did this creepy thing with standard settings, slight OC (+50 +300) and with higher OC (+100 +900). Nothing helped.

Same issue here...Switched over to miniz 14o for beam for now...1080ti's work fine...It's definitely with the 2XXX cards...

soda37

jr. member

Activity: 63

Merit: 1

Quote from: Dr_Victor on July 14, 2019, 10:44:47 AM

Quote from: soda37 on July 14, 2019, 09:05:26 AM

I've just tried Gminer today on my test rig and strange enough on one of 1060 6G it says "not enough VRAM", but only for one of them.

Win 10? Does monitor plug to this card?

That's right, it was HDMI dongle, but today I upgraded drivers to 431.36 and seems Grin mining works fine, although with Beam I'm still getting "Illegal memory access was encountered". I can't get rid of that HDMI dongle and can't lower intensity like in other miners to alleviate over that problem Sad

Update: removing HDMI didn't help as well as lowering my OC settings, different cards give "Illegal memory access was encountered" with beam, while Grin works more or less fine.

Dr_Victor

sr. member

Activity: 954

Merit: 250

Quote from: soda37 on July 14, 2019, 09:05:26 AM

I've just tried Gminer today on my test rig and strange enough on one of 1060 6G it says "not enougth VRAM", but only for one of them.

Win 10? Does monitor plug to this card?

soda37

jr. member

Activity: 63

Merit: 1

I've just tried Gminer today on my test rig and strange enough on one of 1060 6G it says "not enougth VRAM", but only for one of them.

cryptoyes

member

Activity: 297

Merit: 10

HF is in less than 3 days and there are multiple floonet (testnet) pools available to test the new hashing algo. On Telegram you said a few days ago that you already have the new miner up and running but you refuse to offer a public beta version for others to test. That's shortsighted and everyone loses. Keeping it secret until fork time is bound to lead to a miriad of issues at fork time, issues which could be solved nicely beforehand. Bminer hasn't released any beta either. It shows how immature this space (mining) continues to be ...

jasta more

member

Activity: 224

Merit: 18

devs will your miner be ready for upcoming grin fork from start?

2h4u

newbie

Activity: 55

Merit: 0

Quote from: qexing on July 12, 2019, 12:06:02 AM

Quote from: 2h4u on July 05, 2019, 09:09:03 AM

Quote from: qexing on July 05, 2019, 02:08:05 AM

Code:

resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000d0000-0x000d3fff window]
caller _nv001126rm+0xe3/0x1d0 [nvidia] mapping multiple BARs
NVRM: Xid (PCI:0000:01:00): 13, Graphics SM Warp Exception on (GPC 3, TPC 0): Out Of Range Address
NVRM: Xid (PCI:0000:01:00): 13, Graphics Exception: ESR 0x51c648=0x3000e 0x51c650=0x20 0x51c644=0xd3eff2 0x51c64c=0x17f
NVRM: Xid (PCI:0000:01:00): 43, Ch 00000008, engmask 00000101

Xids 13 and 43 do indicate the GPU memory error (not the core error like them 61/62/79). Cause is the o/c, mostly.

@2h4u @Zminer777

I have tried mining BEAM and ZelCash with miniZ miner, both works fine in the same GTX 1080 hardware and Linux system,
never get 'Out Of Range Address' error when mining with miniZ.

It's completely normal though. Different miners, different kernels. Every other miner humps a GPU in its own unique way. As I told you, 13/43 are solely memory overclock errors.

TBOEMC

jr. member

Activity: 41

Merit: 1

Quote from: Zminer777 on July 10, 2019, 07:42:36 AM

Quote from: trucobit on July 10, 2019, 05:14:57 AM

Gminer already has several versions that does not work well in my rigti 1080ti and RTX 2080. It starts, detects the cards and randomly or stays stuck there for hours without undermining, or mine well. The only thing is reboot al rig. I am a user of Awesome miner and we recommend not using Gminer to the thousands of users of this program.

It occurs in the cucka and 150.5 and in some more, and it is not the OC, because it reboots and usually works with the same parameters.

Very much 2% Fee for such an unstable miner. I hope they can solve it because in the end the ones that lose the most are them, the less people use Gminer, the less they earn for the FEE. In its place for the Cucka I have to use nbminer that yields less but never fails.

Have you tried versions 1.48+? I hope we have fixed this issue.
Btw nbminer have same comission for grin - 2%, https://github.com/NebuTech/NBMiner

I have these same issues in rig with RTX 2070
Rig after fresh start works well, but randomly - sometimes after 2hours, sometimes half a day suddenly Gminer cannot start and there is only his black empty window, which is blinking cause it start and ends. The only possible thing is reset a whole rig. It not dependis on the version of Gminer (since 1.44 to 1.50) - i have this issue only with RTX. With 1070 and 1080 is it good. RTX cards did this creepy thing with standard settings, slight OC (+50 +300) and with higher OC (+100 +900). Nothing helped.

qexing

newbie

Activity: 7

Merit: 0

Quote from: 2h4u on July 05, 2019, 09:09:03 AM

Quote from: qexing on July 05, 2019, 02:08:05 AM

Code:

resource sanity check: requesting [mem 0x000c0000-0x000fffff], which spans more than PCI Bus 0000:00 [mem 0x000d0000-0x000d3fff window]
caller _nv001126rm+0xe3/0x1d0 [nvidia] mapping multiple BARs
NVRM: Xid (PCI:0000:01:00): 13, Graphics SM Warp Exception on (GPC 3, TPC 0): Out Of Range Address
NVRM: Xid (PCI:0000:01:00): 13, Graphics Exception: ESR 0x51c648=0x3000e 0x51c650=0x20 0x51c644=0xd3eff2 0x51c64c=0x17f
NVRM: Xid (PCI:0000:01:00): 43, Ch 00000008, engmask 00000101

Xids 13 and 43 do indicate the GPU memory error (not the core error like them 61/62/79). Cause is the o/c, mostly.

@2h4u @Zminer777

I have tried mining BEAM and ZelCash with miniZ miner, both works fine in the same GTX 1080 hardware and Linux system,
never get 'Out Of Range Address' error when mining with miniZ.

HFG5521

newbie

Activity: 26

Merit: 0

Needs an update with new Nvidia drivers. Performance regression and instability is back on them.

sikke1

jr. member

Activity: 44

Merit: 18

Quote from: cryptoyes on July 02, 2019, 06:07:50 AM

Quote from: 2h4u on July 01, 2019, 11:11:41 PM

Quote from: cryptoyes on July 01, 2019, 08:42:33 PM

Importantly, I have 12 GPUs in every node, with an i3-7100 cpu. The nodes with gtx1070ti keep the CPU at 100% utilisation (on both 1.45 and 1.46/1.47) and the GPUs cannot handle anything else (even nvidia-smi fails, takes 5 minutes to show 1 line).

Try --watchdog 0 --pec 0 options to temper gminer's thirst for NVML calls. And, I guess, you have old nvidia drivers like 3xx.something. Try 4xx branch, it's kinda optimized for multi-gpu environment--less LA, more responsiveness.

p.s. @Zminer777, it would be nice to add smth like --no-nvml option, it'll help a lot.

As I said above, I'm using 415.29 drivers, cuda10.0. I'm also already using --pec 0 and --watchdog 0. None of these are an issue though, since the problem is only on gtx1070ti and not on gtx1070. I've tried even newer drivers (no luck).

p.s. this problem of 100% cpu on gtx1070ti (but not on gtx1070) exists for all other algos with Gminer, but not with other mining software, suggesting the problem is Gminer, not drivers or hardware.

I have same issues with multiple rigs with 13mining cards. All running ubuntu linux GMiner takes 100% CPU usage with Core i3, Core i5 and Core i7. It just kills it. I can mine GMiner with under 10GPUs but anything over that will cause rejected shares and 100% CPU usage. I am getting bit better results with 1.41. It still hits 100% CPU time but will mine with less rejected.

This is 19GPU with Core i5 7500.
https://imgur.com/a/JxZywCI

trucobit

jr. member

Activity: 756

Merit: 2

Quote from: sefirotius on July 11, 2019, 04:30:41 AM

today it is the best miner for mining; all other miners are worse to keep

nbminer
ttminer
gminer
t-rex
cryptodredge

Zminer777

member

Activity: 488

Merit: 37

Quote from: loop3 on July 11, 2019, 09:02:07 AM

Yes this is on beam nodes, apparently it is a Gminer formatting issue, other miners report correct difficulty.

what's miners do you mean?
gminer only print integer from stratum (you can check it via wireshark for example), does not format it.

loop3

newbie

Activity: 3

Merit: 0

Quote from: Zminer777 on July 11, 2019, 08:13:16 AM

Quote from: loop3 on July 11, 2019, 07:56:36 AM

Could you please explain how does console difficulty reporting work?
Solo mining I am getting Difficulty: 463806055 reported in console, when it should be around 219,255,616 according to the official block explorer.

Do you mean beam mining?
Miner just forward this information from node, you can ask your question in beam community.

Yes this is on beam nodes, apparently it is a Gminer formatting issue, other miners report correct difficulty.

Zminer777

member

Activity: 488

Merit: 37

Quote from: loop3 on July 11, 2019, 07:56:36 AM

Could you please explain how does console difficulty reporting work?
Solo mining I am getting Difficulty: 463806055 reported in console, when it should be around 219,255,616 according to the official block explorer.

Do you mean beam mining?
Miner just forward this information from node, you can ask your question in beam community.

loop3

newbie

Activity: 3

Merit: 0

Could you please explain how does console difficulty reporting work?
Solo mining I am getting Difficulty: 463806055 reported in console, when it should be around 219,255,616 according to the official block explorer.

sefirotius

full member

Activity: 658

Merit: 102

IDENA.IO - Proof-Of-Person Blockchain

today it is the best miner for mining; all other miners are worse to keep

Marvell2

full member

Activity: 1148

Merit: 132

=how to stop the miner from adding stratum:// to ever pool even if i already have stratum+tcp

Topic: GMiner v3.24 Ergo+KASPA+ZIL/KAWPOW/Equihash/Cortex - page 72. (Read 307024 times)