Problem - T-Rex 0.19.12 - RTX 3070

[TomARG120]

Hi, i have a few problems with my Rig

The rig have:

H110 Pro BTC+ ASRock (P1.50 02/26/2018)
CPU G3900 @ 2.80GHz AES
ATA PNY CS900 240GB 240GB
PSU Server 2450W + PSU 750W Thermaltake 80 Plus Bronze
13x RTX 3070 Nvidia

The first problem
1- When i try to mine with the 12 and 13 GPUs appear this message:
"TREX: Cant find nonce with device id=11 GPU#11 Cuda exception in synchronize 51 , unsepcified lauch failure, try to reduce overclock to stabilize GPU state…

I try changing the Riser, chage the conector of PCIe of the mother, trying with another PSU, set the stock settings

Second problem:
2- "TREX: Cant find nonce with device id=00 GPU#00 Cuda exception in set busy 85 unspecified launch failure…

Third problem:
3- cant find 999 gpu trex mining

Some one can help me?

[dukeke]

Hi, did you solve it?

I’m facing same problem but when connecting the 10th GPU.
Suddenly, T-rex shows your first msg for GPU5 and restarts. This repeats no matter what I do (raisers, cables, splitters, etc). Removing OC didn’t work.

[TomARG120]

Hi
I add more ram, i was using 4GB DDR4 then i put 4gb more and the problem was solved

[Devilz]

I have 6 rtx 3080 but have the same code error

Can’t find nonce with device [ID=4, GPU #4], cuda exception in [synchronize, 52], CUDA_ERROR_LAUNCH_FAILED, try to reduce overclock to stabilize
GPU state

Have 8 gb ddr 4

[kgoldsmith3]

Getting the same error on one of my 1080’s. Has anyone found the fix for this yet?

[goofee]

Getting the same thing on one of my 1080’s too. Using t-rex and -mt 2, if i put in -mt 0 it works again. Didn’t have this problem before.

[soepie1980]

Can’t find nonce with device [ID=3, GPU #3], cuda exception in [synchronize, 52], CUDA_ERROR_LAUNCH_FAILED, try to reduce overclock to stabilize GPU state

same problem ? tried to adjust gpu 3 nothing is working… did update nvidia drivers to 64.80 but even that didnt do anything

[goofee]

Today i enabled -mt 2 and the error is gone! I didn’t change a thing, i did the change in the IOS App. I don’t get why i had this error. I tried rebooting, updating everything etc. But it works now again.

[Sweetpayn]

@goofee

Today i enabled -mt 2

Could you please tell me what do you mean by enabaling -mt2> Blockquote

I experience the same error over and lver again… sometimes after 1h … sometimes after 2days- miner is restarting.

[miszekhive]

hello, i had the same problem. My kit is the 11x3070 (Dell) + Asrock H110 Pro BTC motherboard + 8gb RAM.

The problem with the miner was like above every one, be it t-rex, nbminer, phoenix, etc. The problem was solved by adding an additional RAM chip. Now I have 12GB of framework and 24 hours of stable work with normal OC. 61 mhs per card.

[SweetiePieFon]

Sweetpayn,

goofee, is referencing memory-tweaks. This is added in the “Extra config arguments:” under setup miner config. Depending on what mining software you are using the syntax will be slightly different. I am using T-rex right now syntax is “mt”: X. The X is the numbers from 1 to 6 depending on what memory tweak you want to apply.

Happy mining!

[alexskh]

so what we must to do to solve it ?

[rawbar]

now having this same issue with 12 x 3070s. been running absolutely fine until i blew a ZSX card. just added a 3rd PSU and replaced the card and spread out the GPUs across 3 PSU instead of 2. Now when I boot, cards mine for maybe a minute, then start throwing these errors. Note I have disabled OC on all cards, rebooted the whole system, the -mt parameter suggested above is only for 10xx cards and doesn’t do anything on 30xx cards, I have 8GB of system memory. I’m stumped and I hate that I’ve spent 8 hours working on this rig today and am going to bed with 12 cards down.

20210629 23:55:31 GPU #2(000c00): ASUS GeForce RTX 3070, intensity 22
20210629 23:55:31 GPU #1(000200): EVGA GeForce RTX 3070, intensity 22
20210629 23:55:31 GPU #0(000100): EVGA GeForce RTX 3070, intensity 22
20210629 23:55:34 TREX: Can’t find nonce with device [ID=0, GPU #0], GPU #0: not enough free memory to mine ethash at epoch 424
20210629 23:55:34 WARN: Miner is going to shutdown…
20210629 23:55:35 Main loop finished. Cleaning up resources…

[rawbar]

it APPEARS the 3rd PSU may be bad. I powered it off and the 7 remaining cards are mining absolutely fine. I had a replacement PSU next day shipped, so should be able to confirm tomorrow.

[Zdenko]

Did anybodi solved problem with this
TREX: Can’t find nonce with device [ID=1, GPU #1], cuda exception: CUDA_ERROR_LAUNCH_FAILED, try to reduce overclock to stabilize GPU state

[marcek]

I have same problem too. These are my last logs before rig goes offline:

--------------20211010 02:48:23 --------------
Mining at eu1.ethermine.org:4444, diff: 4.29 G
GPU #0: RTX 3090 - 121.03 MH/s, [T:62C, P:305W, F:85%, E:397kH/W], 147/147 R:0% I:0%
GPU #1: RTX 3080 - 95.56 MH/s, [T:59C, P:222W, F:80%, E:430kH/W], 107/107 R:0% I:0%
GPU #2: RTX 3070 - 60.75 MH/s, [T:50C, P:122W, F:74%, E:498kH/W], 81/81 R:0% I:0%
GPU #3: RTX 2080 Ti - 61.07 MH/s, [T:49C, P:156W, F:68%, E:394kH/W], 69/69 R:0% I:0%
GPU #4: RTX 3060 Ti - 60.32 MH/s, [T:49C, P:123W, F:67%, E:490kH/W], 74/74 R:0% I:0%
GPU #5: RTX 3080 - 98.29 MH/s, [T:52C, P:166W, F:75%, E:592kH/W], 125/125 R:0% I:0%
GPU #6: RTX 3070 - 60.70 MH/s, [T:43C, P:125W, F:45%, E:486kH/W], 89/89 R:0% I:0%
Hashrate: 557.72 MH/s, Shares/min: 6.477 (Avr. 7.202), Avr.P: 1218W, Avr.E: 458kH/W
Max diff share was found by GPU #5, diff: 2.33 T
Uptime: 1 hour 36 mins 15 secs | Algo: ethash | T-Rex v0.24.0

20211010 02:48:25 [ OK ] 693/693 - 557.77 MH/s, 76ms … GPU #1 | 10.88 G
20211010 02:48:27 TREX: Can’t find nonce with device [ID=1, GPU #1], cuda exception: CUDA_ERROR_LAUNCH_FAILED, try to reduce overclock to stabilize GPU state
20211010 02:48:27 WARN: Miner is going to shutdown…
20211010 02:48:28 Main loop finished. Cleaning up resources…