miniZ v2.0a Equihash 144,5 125,4 210,9 150,5 192,7 BeamHash3 ProgPoW Ethash CFX - page 57.

DanGB

newbie

Activity: 25

Merit: 0

Quote from: miniZ on April 28, 2019, 04:43:22 PM

Quote from: DanGB on April 28, 2019, 04:17:56 PM

miniZ - The miner on my one rig crashed. It ran for 7.5 hours. This is the error in the log:

GPU[2]: CUDA error 'unspecified launch failure' in func 'eq_a[[[ac>5;:c_26dEbB, SM, PS2>::solve' line 957
GPU[0]: CUDA error 'unspecified launch failure' in func 'eq_a[[[ac>5;:c_26dEbB, SM, PS2>::solve' line 957

This rig has an i5 CPU and 4GB ram running Win10x64. NVIDIA driver version is 417.35.

Hi DanGB,
Thank you for the test.
That Cuda error is not related to the other issue.
This may be related to your GPUs OCs. Remind us which GPUs and OCs are you using.

Nevertheles, we have updated the testing file with few more small corrections.
miniZ v1.3n++ @ cuda v10.0, Apr 28 2019 20:05:51

Thanks.
Cheers

Those two specific cards are as follows (settings done with Afterburner):
GPU0 - PNY 1070 XLR8 Power 66%, temp limit 65deg, core +160, mem +0
GPU2 - EVGA 1070ti FTW2 Power 66%, temp limit 70deg, core +200, mem +0

These are the settings I had been using on gminer since January.

Thanks

miniZ

member

Activity: 690

Merit: 17

Quote from: impynick on April 28, 2019, 04:20:37 PM

Quote from: miniZ on April 28, 2019, 08:28:24 AM

Quote from: impynick on April 28, 2019, 05:45:05 AM

MiniZ is faster on my 1070s vs Gminer; however Gminer is faster on my 1080ti.

I noticed f11=1 (stock setting I think) is the best one with oc2. I’ve teied “2” and “3”3 and “0” for f11 values and it’s slower on these other values

Hi impynick,
Thanks for reporting this.
Could you share your hashrate and OC values for those GPUs?
Cheers

1070 and 1080ti on same settings = 100%PL ; +100 core +500 mem

1070s putting down 70 sols and 1080ti (FE) at 90-92

On Gminer I’m getting 65 sols on my 1070s and 97 sols on the 1080ti.

Hi impynick,
thank you for the information.
We will look into performance as soon as we manage to solve this stability issue.
Cheers

miniZ

member

Activity: 690

Merit: 17

Quote from: DanGB on April 28, 2019, 04:17:56 PM

miniZ - The miner on my one rig crashed. It ran for 7.5 hours. This is the error in the log:

GPU[2]: CUDA error 'unspecified launch failure' in func 'eq_a[[[ac>5;:c_26dEbB, SM, PS2>::solve' line 957
GPU[0]: CUDA error 'unspecified launch failure' in func 'eq_a[[[ac>5;:c_26dEbB, SM, PS2>::solve' line 957

This rig has an i5 CPU and 4GB ram running Win10x64. NVIDIA driver version is 417.35.

Hi DanGB,
Thank you for the test.
That Cuda error is not related to the other issue.
This may be related to your GPUs OCs. Remind us which GPUs and OCs are you using.

Nevertheles, we have updated the testing file with few more small corrections.
miniZ v1.3n++ @ cuda v10.0, Apr 28 2019 20:05:51

Thanks.
Cheers

miniZ

member

Activity: 690

Merit: 17

Quote from: $@X on April 28, 2019, 03:26:03 PM

Hi miniZ,

Thank you so much for the continued efforts to improve mining Beam. W

What version of Nvidia drivers is your team testing on?

Also, do you have a SHA256 checksum for the latest testing version?

Thanks!

Hi $@X,
Thank you for your message.
We test mostly on NVIDIA 390.48, in linux. But, on HiveOS 418.43.
(*edit*) On Windows 10, 418.91.

We added the miniZ.exe sha to the website.
We have updated the file with few more small corrections.
  miniZ v1.3n++ @ cuda v10.0, Apr 28 2019 20:05:51

Cheers

impynick

jr. member

Activity: 77

Merit: 6

Quote from: miniZ on April 28, 2019, 08:28:24 AM

Quote from: impynick on April 28, 2019, 05:45:05 AM

MiniZ is faster on my 1070s vs Gminer; however Gminer is faster on my 1080ti.

I noticed f11=1 (stock setting I think) is the best one with oc2. I’ve teied “2” and “3”3 and “0” for f11 values and it’s slower on these other values

Hi impynick,
Thanks for reporting this.
Could you share your hashrate and OC values for those GPUs?
Cheers

1070 and 1080ti on same settings = 100%PL ; +100 core +500 mem

1070s putting down 70 sols and 1080ti (FE) at 90-92

On Gminer I’m getting 65 sols on my 1070s and 97 sols on the 1080ti.

DanGB

newbie

Activity: 25

Merit: 0

miniZ - The miner on my one rig crashed. It ran for 7.5 hours. This is the error in the log:

GPU[2]: CUDA error 'unspecified launch failure' in func 'eq_a[[[ac>5;:c_26dEbB, SM, PS2>::solve' line 957
GPU[0]: CUDA error 'unspecified launch failure' in func 'eq_a[[[ac>5;:c_26dEbB, SM, PS2>::solve' line 957

This rig has an i5 CPU and 4GB ram running Win10x64. NVIDIA driver version is 417.35.

$@X

newbie

Activity: 19

Merit: 0

Hi miniZ,

Thank you so much for the continued efforts to improve mining Beam. W

What version of Nvidia drivers is your team testing on?

Also, do you have a SHA256 checksum for the latest testing version?

Thanks!

DanGB

newbie

Activity: 25

Merit: 0

Quote from: ARTRN on April 28, 2019, 01:14:24 PM

Quote from: miniZ on April 28, 2019, 08:15:29 AM

Quote from: ARTRN on April 28, 2019, 06:51:34 AM

FYI, v1.3n++ still have the same 'losing miner' on pool. Cheers.

Hi ARTRN,
Thanks a lot for your quick feedback.
It looks like you can replicate faster the problem than we do!

Now, we know what it is (the communication with the pool gets into some weird state and the miner
was getting stuck trying to recover). We fixed it, but it still requires testing...
This stage will be much faster with some extra help Wink

The miniZ v1.3n++ testing version (Win 10, Cuda10) was updated. You can download from our Download page https://miniz.ch/download/
We did not change the version name, so check you have the right build with miniZ.exe --version
It should be Apr 28 2019 12:28:20

Cheers,
miniZ

Hi, miniZ v1.3n++ testing version (Win 10, Cuda10) Apr 28 2019 12:28:20 so far so good.
It's been running close to 4 hours without fail.
Cheers.

miniZ - I agree. I am at 5 hours now with no issues, so I am hopeful. CPU utilization is holding steady so far. I need to get to 12 hours with no issues before I will be pretty confident that it is resolved.

ARTRN

newbie

Activity: 63

Merit: 0

Quote from: miniZ on April 28, 2019, 08:15:29 AM

Quote from: ARTRN on April 28, 2019, 06:51:34 AM

FYI, v1.3n++ still have the same 'losing miner' on pool. Cheers.

Hi ARTRN,
Thanks a lot for your quick feedback.
It looks like you can replicate faster the problem than we do!

Now, we know what it is (the communication with the pool gets into some weird state and the miner
was getting stuck trying to recover). We fixed it, but it still requires testing...
This stage will be much faster with some extra help Wink

The miniZ v1.3n++ testing version (Win 10, Cuda10) was updated. You can download from our Download page https://miniz.ch/download/
We did not change the version name, so check you have the right build with miniZ.exe --version
It should be Apr 28 2019 12:28:20

Cheers,
miniZ

Hi, miniZ v1.3n++ testing version (Win 10, Cuda10) Apr 28 2019 12:28:20 so far so good.
It's been running close to 4 hours without fail.
Cheers.

hogwash.89m

newbie

Activity: 32

Merit: 0

https://i.imgur.com/OMGwSrh.png

v1.3n+ and v1.3n++ (beta build, Apr 28 2019 12:28:20) CUDA 10 versions doesn't work on sparkpool, see picture above.
Sparkpool is displaying that there are too many rejected shares, but it's not reporting it back to miner, so in miner it looks good, but on poolside >95% of shares are rejected (or stales).

Then I left it to mine on leafpool and it keeps mining fine for more than 8 hours.
Only thing I noticed is that at some point it hits 100% CPU usage and stays like that, I need to investigate this more.

CPU: Celeron G3930 (@2.90 GHzm 2C/2T)
RAM: 4 GB
Virtual Mem.: 64 GB
GPUs: 1x 1080ti MSI Gaming
5x 1080 inno3d twin X2

OS: Windows Server 2016 (version 1607, OS Build 14393.2273)
Nvidia driver: 419.17 (CUDA 10.10)

MSI Afterburner (for OC): 4.6.0 -> noting this as it can be cause of 100% CPU usage, because on 1080ti fanspeed readings in miniZ console becomes 0%, maybe there are conflict between querying Nvidia-SMI stats from miner and Afterburner, idk.

MineOrDie

jr. member

Activity: 75

Merit: 1

Quote from: miniZ on April 28, 2019, 08:25:50 AM

Quote from: MineOrDie on April 28, 2019, 04:02:46 AM

With 1.3n+ I also have a problem that my worker just disappear from the pool but it keeps hashing on my rig. I will try 1.3++ now.
Also on Windows 10, 6x GTX1060 6GB and a G3930 CPU with version 1.3n+ is still 100% CPU usage.

EDIT: With 1.3++ my LA on Hive OS rig is 5,6+ again.

Hi MineOrDie,
Thanks for keeping us posted.
We did not update the Hive OS package yet.
We are currently testing and looking into it. (we think we got the offline issue already)
Cheers

Well I added one more + to the link, and it downloaded something that is apparently the ++ version, so yeah...

miniZ

member

Activity: 690

Merit: 17

Quote from: sharmanov on April 28, 2019, 09:25:36 AM

I and 2 of my mining buddies can confirm the 1.3n+ did not resolve the "OFFLINE WORKER" issue as it is still happening. I really hope you guys get to the bottom of that issue soon. Thanks

Quote from: BIOLERUP4LIFE on April 28, 2019, 09:33:59 AM

Just fired up the new version now. Ill let you know how it goes. it was consistently failing every 6-10 hours. If you don't mind me asking what did you find going on?

Hi sharmanov, BIOLERUP4LIFE,
Thanks for all the feedback.

We think we got it now, sometimes (not often) the communication with the pool gets into some weird state and the miner
was getting stuck in a loop trying to recover. We fixed it, but it still requires testing...

For now, we made available a miniZ v1.3n++ testing version (Win 10, Cuda10).
You can download from our Download page https://miniz.ch/download/

To check if you have the right build run miniZ.exe --version
It should be Apr 28 2019 12:28:20

Thanks again.
Cheers

BIOLERUP4LIFE

newbie

Activity: 10

Merit: 0

Quote from: miniZ on April 28, 2019, 08:15:29 AM

Quote from: ARTRN on April 28, 2019, 06:51:34 AM

FYI, v1.3n++ still have the same 'losing miner' on pool. Cheers.

Hi ARTRN,
Thanks a lot for your quick feedback.
It looks like you can replicate faster the problem than we do!

Now, we know what it is (the communication with the pool gets into some weird state and the miner
was getting stuck trying to recover). We fixed it, but it still requires testing...
This stage will be much faster with some extra help Wink

The miniZ v1.3n++ testing version (Win 10, Cuda10) was updated. You can download from our Download page https://miniz.ch/download/
We did not change the version name, so check you have the right build with miniZ.exe --version
It should be Apr 28 2019 12:28:20

Cheers,
miniZ

Just fired up the new version now. Ill let you know how it goes. it was consistently failing every 6-10 hours. If you don't mind me asking what did you find going on?

sharmanov

newbie

Activity: 54

Merit: 0

Quote from: miniZ on April 27, 2019, 09:48:49 AM

Hi everyone,

Again, tanks a lot for your feedback!

There is a new miniZ version v1.3n+ for you to download.

* Reduced CPU usage to background levels.
* Better invalid share detection.
* Ignore pers auto for 150,5.
* Fixing issues with worker going offline.

I and 2 of my mining buddies can confirm the 1.3n+ did not resolve the "OFFLINE WORKER" issue as it is still happening. I really hope you guys get to the bottom of that issue soon. Thanks

DanGB

newbie

Activity: 25

Merit: 0

Quote from: miniZ on April 28, 2019, 08:15:29 AM

The miniZ v1.3n++ testing version (Win 10, Cuda10) was updated. You can download from our Download page https://miniz.ch/download/
We did not change the version name, so check you have the right build with miniZ.exe --version
It should be Apr 28 2019 12:28:20

Cheers,
miniZ

miniZ - I have this latest version running on my 3 rigs right now. The CPU utilization is a little better on my two Celeron rigs (74% and 89% instead of 100%). I'll let you know how it goes.

miniZ

member

Activity: 690

Merit: 17

Quote from: impynick on April 28, 2019, 05:45:05 AM

MiniZ is faster on my 1070s vs Gminer; however Gminer is faster on my 1080ti.

I noticed f11=1 (stock setting I think) is the best one with oc2. I’ve teied “2” and “3”3 and “0” for f11 values and it’s slower on these other values

Hi impynick,
Thanks for reporting this.
Could you share your hashrate and OC values for those GPUs?
Cheers

miniZ

member

Activity: 690

Merit: 17

Quote from: MineOrDie on April 28, 2019, 04:02:46 AM

With 1.3n+ I also have a problem that my worker just disappear from the pool but it keeps hashing on my rig. I will try 1.3++ now.
Also on Windows 10, 6x GTX1060 6GB and a G3930 CPU with version 1.3n+ is still 100% CPU usage.

EDIT: With 1.3++ my LA on Hive OS rig is 5,6+ again.

Hi MineOrDie,
Thanks for keeping us posted.
We did not update the Hive OS package yet.
We are currently testing and looking into it. (we think we got the offline issue already)
Cheers

miniZ

member

Activity: 690

Merit: 17

Quote from: rgsnedds on April 27, 2019, 09:22:31 PM

no worries thanks for the response on it. it was the only thing i could think of but if it had been the case then it would have stuck on the 4minute initiation of the devfee, not after 5 hrs . good luck finding the issue.

Hi rgsnedds,
Thanks a lot for the tips, they are much appreciated.
We think we got this one. Testing it still.
Cheers

miniZ

member

Activity: 690

Merit: 17

Quote from: ARTRN on April 28, 2019, 06:51:34 AM

FYI, v1.3n++ still have the same 'losing miner' on pool. Cheers.

Hi ARTRN,
Thanks a lot for your quick feedback.
It looks like you can replicate faster the problem than we do!

Now, we know what it is (the communication with the pool gets into some weird state and the miner
was getting stuck trying to recover). We fixed it, but it still requires testing...
This stage will be much faster with some extra help Wink

The miniZ v1.3n++ testing version (Win 10, Cuda10) was updated. You can download from our Download page https://miniz.ch/download/
We did not change the version name, so check you have the right build with miniZ.exe --version
It should be Apr 28 2019 12:28:20

Cheers,
miniZ

ARTRN

newbie

Activity: 63

Merit: 0

FYI, v1.3n++ still have the same 'losing miner' on pool. Cheers.

Topic: miniZ v2.0a Equihash 144,5 125,4 210,9 150,5 192,7 BeamHash3 ProgPoW Ethash CFX - page 57. (Read 60004 times)