 Show Posts
|
|
Pages: [1] 2 »
|
Hi
I'm having problem with the latest release (srbminer v.2.3.6) together with Radeon VII cards and Dynex. (SRB v.2.3.5 work's like a charm)
Error message:
GPU0 couldn't create program for ctx->algorithm
Are there a work around?
I have same error with VII + 2.3.6, rolling back to 2.3.5 |
|
|
|
How to get rid of this Red Error shown in screenshot attached!
Upgrade to v1.23 its ver 1.23 |
|
|
|
Dear SP_, How to get rid of this Red Error shown in screenshot attached!  |
|
|
|
Hi SP__ Please Check the Screenshot Attached, I cant have the miner Stable on my 6x3090 Rig. you advise highly appreciated  2 of your cards are running on a lower memclock. 2021 watt is alot. Is your PSU strong enough for that? Yes I manually adjusted these 2 cards, because on higher clock they produces invalid shares. For PSU, I Have 3 x1000 watt Psu, each 2 cards connected to one PSU. Why I cant run the miner Stable ? |
|
|
|
Hi SP__ Please Check the Screenshot Attached, I cant have the miner Stable on my 6x3090 Rig. you advise highly appreciated |
|
|
|
What happens if you use --lock-cclock [[1200,1200] ? for me its same hash rate but way less power used as 1500/1600, and temps less for sure since its less power consuming
Higher clocks need more power. When you tell the device that you don't want higher than 1200 clock it will undervolt automaticly. For the laptop gpu's this is giving a nice boost in the Hashrate. Less Power, no trottleing, run cooler, more hash.., For Laptop it works well with 1200, what about Desktop's GPU! do you think 3070 and 3060Ti needs more than 1200 Core Clock ? Daggerhashimoto needs more Clocks on Memory as I know. what do you suggest? |
|
|
|
@KillrBee
I have a rig of 3090s (6 cards) running on HiveOS. I couldn't reach 11326 mem clock (+3000 in Hive) max could reach was +2900 and after several minutes miner crashes. Cards cooled up using AC direct cold air to them.
I think you are using Windows ?!
one more thing, I used +1200 on Core, didn't try +1600 since Ethhash a memory hungry algorithm
Yeah, Linux and Windows are different beasts when it comes to the GPU clocks. +1200 would probably be just fine -- essentially I'm running at .75 - .80 Volts. I was very lucky with how well my card overclocks, I can get that memory up to 11400 for a while, but eventually the card throws an error (even though temps are well controlled). Try pushing the intensity up and keeping the memory overclock to +2600 (I'm running XInt = 3108 right now and I can keep the memory clock down to a level that the miner doesn't drop or gpu hang for 24hr+ {which is 11326ish for my card}). It's taken A LOT of experimenting and testing to find a stable and fast hashrate/share rate, and it continues to be refined every day and every release. I am probably spending way too much time tweaking vs. doing my day job, but I'm a performance freak and always have to find the useable limits of everything. :-) What happens if you use --lock-cclock [[1200,1200] ? for me its same hash rate but way less power used as 1500/1600, and temps less for sure since its less power consuming |
|
|
|
|
--lock-cclock [[1500,1500],[1500,1500],[1500,1500],[1500,1500],[1500,1500],[1500,1500]]
Is This Good to use with RTX 3060Ti 3070 and 3090 ?? or better use 1200 ?
I can confirm that 1200 uses less power and same hash-rate as 1500! but in previous posts I saw many likes 1500 !!!
|
|
|
|
|
@KillrBee
I have a rig of 3090s (6 cards) running on HiveOS. I couldn't reach 11326 mem clock (+3000 in Hive) max could reach was +2900 and after several minutes miner crashes. Cards cooled up using AC direct cold air to them.
I think you are using Windows ?!
one more thing, I used +1200 on Core, didn't try +1600 since Ethhash a memory hungry algorithm
|
|
|
|
2014:  My work was in the first opensource ethereum cuda miner for nvidia. Good Job man  |
|
|
|
I have "No devices are set" on RaveOS. Adding "-d" didn't help. What should I do?
try --api-port 64068 |
|
|
|
Fixed The issue, - Newer Driver installed 470.82.00
- Added --api-port 64068
Thank you guys, Happy mining ⚒ everyone  |
|
|
|
Others have managed to run 1.20 without any problems. 1.20 was released a few hours ago, why don't you wait until hiveos add it.
wanted to test --dagintensity |
|
|
|
Have you tested this driver? I have RTX 3xxx Cards, I need to make sure not miss with rig, its located far away from me. |
|
|
|
I have Latest Drivers installed : N 470.74 Drivers The only thing done, is that I manually replaced 1.19 to 1.20 by using commands in bash. may be something went wrong or I miss to do something while replacing 1.19 ? add this line in miner.config --api-port 64068 Didnt help, still nothing. let me try replace 1.19 Back and check the results, seems issue in 1.20 !! |
|
|
|
perhaps you need to give execute permissions.
$ chmod a+rwx tbminer
did that command, didnt help. FYI, I only cp TBminer from tar file to 1.19 directory. shall I copy also miner.conf cacert.pem ? |
|
|
|
I have Latest Drivers installed : N 470.74 Drivers The only thing done, is that I manually replaced 1.19 to 1.20 by using commands in bash. may be something went wrong or I miss to do something while replacing 1.19 ?
470.74 is not the latest. this driver only support cuda 11.3
this is the latest available on HiveOS, used : nvidia-driver-update --list and it was latest one in the list [moderator's note: consecutive posts merged] |
|
|
|
any resolution for this issue ?
There is a bug in 1.20 that if you use a worker on ethermine with high case letters, it wont show as reported hashrate. this has been fixed in v1.21. So rename all your workers to lowercase is the workaround. I am not mining to ethermine, I am mining @ 2miners pool. please check screenshot attached  |
|
|
|
Why don't you try v1.20 and see if you can reproduce it.
I have manually updated HiveOS miner to 1.20, replaced miner in ver 1.19 Folder with the new one. The issue I cant see now stats in HiveOS portal page, only see it in Shell (ssh), seems like HiveOS not getting data from miner (1.20) any resolution for this issue ? |
|
|
|
use another program to set memclocks.
msi afterburner is good
I am on HiveOS (Linux) no afterburner there  HiveOS over/underclocks your GPUs independently of the miner. Do the overclocking for each card in the HiveOS worker screen not the miner. What are the best settings ( C-Clock and Mem-Clock) that works good with this miner for: RTX 3060Ti ( non LHR) RTX 3070 ( non LHR) RTX 3090? Thank you in advance |
|
|
|
|
