[ANN] TeamRedMiner v0.10.10 - Ironfish/Kaspa/ZIL/Kawpow/Etchash and More - page 145.

MoneroCrusher

jr. member

Activity: 88

Merit: 1

Quote from: kerney666 on November 01, 2018, 02:34:04 PM

Quote from: MoneroCrusher on November 01, 2018, 02:24:05 PM

RX 550, 2 GB Polaris12 (not Baffin!)
Team Red Miner version 0.3.5
[2018-11-01 20:21:52] Auto-detected AMD OpenCL platform 0
[2018-11-01 20:21:53] Failed to initialize device number 0.
[2018-11-01 20:21:53] Failed to initialize device number 1.
[2018-11-01 20:21:53] Failed to initialize device number 2.
[2018-11-01 20:21:53] Failed to initialize device number 3.
[2018-11-01 20:21:53] Failed to initialize device number 4.
[2018-11-01 20:21:53] Failed to initialize device number 5.
[2018-11-01 20:21:54] Failed to initialize device number 6.
[2018-11-01 20:21:54] Failed to initialize device number 7.
[2018-11-01 20:21:54] Failed to initialize device number 8.
[2018-11-01 20:21:54] Failed to initialize device number 9.
[2018-11-01 20:21:54] Failed to initialize device number 10.

Same story on Windows.

Yep, we'll be getting a new version with 550 support out soon, current last version doesn't recognize the cards. We delayed it do finish some stability tests for large Vega rigs to include those fixes as well. I got a report from an early tester that it still failed for his 550s under 18.5.1 though, zzz...

MoneroCrusher, if you have a mins, can you run "clinfo" from a command prompt and find this section for one of your 550s? I'd like to verify you're seeing "gfx804" as the device name.

... Platform ID: 00007FFF3BE5FFD0 Name: gfx804 Vendor: Advanced Micro Devices, Inc. Device OpenCL C version: OpenCL C 2.0 Driver version: 2580.6 Profile: FULL_PROFILE Version: OpenCL 2.0 AMD-APP (2580.6) ...

Yup it's gfx804.
You have Discord or Telegram? Can I also test it on Linux?

misterkit

newbie

Activity: 47

Merit: 0

Quote from: kerney666 on November 01, 2018, 01:36:50 PM

Quote from: misterkit on November 01, 2018, 12:43:02 PM

I tried this and quit. The power usage was about the same but the 'printed' hashrate was higher.

I did not care for all of the connections being made in the background. http://gratified.approvalbureau.com/, inpervious.choppedgreen.com, etc.

I dont mind paying a dev fee for something.....

Sorry to hear it. I know we are as clean as can be, so not sure what you are 'insinuating' really? We also tried hard to let non-affiliated forum members verify poolside hashrate before releasing just to not have your kind of posts in the thread. The hashrate displayed is legit.

There is a single outbound connection for the dev fee to a single IP. It will be active at all times and should not change. There is no dev fee switching. For gratified.approvalbureau.com I can't even make a successful public DNS lookup using Google (8.8.8.8, 8.8.4.4) for it so I'm guessing you have a typo somewhere. The other URL resolves to a 207.148.248.143 which is not affiliated with our miner in any way.

Using standard Win10 tools, open the Resource Monitor, go to the Network tab, sort the TCP Connections on Image name, find teamredminer.exe in the sorted list, verify you have two outbound tcp/ip connections, one to your pool and another for our dev fee. On linux, use netstat -nltp | egrep teamred while running the miner. If you're running our binary downloaded from github, there will be nothing else there to find unless that account has been hacked, which I doubt.

I'm not insinuating anything. The rates displayed on my screen were higher than I have seen with any other miner. Of course that has me a little suspicious but open minded. Closed source software can print anything they want. If they #s are legit, then kudos to you guys for finding a better way to get these kinds of results. You've done something that the other miners out there havent been able to. I would have kept mining with this but when I saw what was happening in the background it started giving me some concerns.

The software I downloaded from the first page in this thread. The github link.

As soon as I started up the process for the delivered .bat file it started spawning additional threads to a ip.googleusercontent site and an additional thread to the various addresses I mentioned below. As soon as I killed them another would spawn (sometimes with a different name like shipload.choppedgreen) or the miner would generate a message about the dev pool/reduced rates.

None of these showed up when I use another miner (like srb or xmrstak). Process Hacker 2 shows these threads coming from your software.

I'm on in Win 10.

edit- wrong process hacker version.

kerney666

member

Activity: 658

Merit: 86

Quote from: MoneroCrusher on November 01, 2018, 02:24:05 PM

RX 550, 2 GB Polaris12 (not Baffin!)
Team Red Miner version 0.3.5
[2018-11-01 20:21:52] Auto-detected AMD OpenCL platform 0
[2018-11-01 20:21:53] Failed to initialize device number 0.
[2018-11-01 20:21:53] Failed to initialize device number 1.
[2018-11-01 20:21:53] Failed to initialize device number 2.
[2018-11-01 20:21:53] Failed to initialize device number 3.
[2018-11-01 20:21:53] Failed to initialize device number 4.
[2018-11-01 20:21:53] Failed to initialize device number 5.
[2018-11-01 20:21:54] Failed to initialize device number 6.
[2018-11-01 20:21:54] Failed to initialize device number 7.
[2018-11-01 20:21:54] Failed to initialize device number 8.
[2018-11-01 20:21:54] Failed to initialize device number 9.
[2018-11-01 20:21:54] Failed to initialize device number 10.

Same story on Windows.

Yep, we'll be getting a new version with 550 support out soon, current last version doesn't recognize the cards. We delayed it do finish some stability tests for large Vega rigs to include those fixes as well. I got a report from an early tester that it still failed for his 550s under 18.5.1 though, zzz...

MoneroCrusher, if you have a mins, can you run "clinfo" from a command prompt and find this section for one of your 550s? I'd like to verify you're seeing "gfx804" as the device name.

... Platform ID: 00007FFF3BE5FFD0 Name: gfx804 Vendor: Advanced Micro Devices, Inc. Device OpenCL C version: OpenCL C 2.0 Driver version: 2580.6 Profile: FULL_PROFILE Version: OpenCL 2.0 AMD-APP (2580.6) ...

MoneroCrusher

jr. member

Activity: 88

Merit: 1

RX 550, 2 GB Polaris12 (not Baffin!)
Team Red Miner version 0.3.5
[2018-11-01 20:21:52] Auto-detected AMD OpenCL platform 0
[2018-11-01 20:21:53] Failed to initialize device number 0.
[2018-11-01 20:21:53] Failed to initialize device number 1.
[2018-11-01 20:21:53] Failed to initialize device number 2.
[2018-11-01 20:21:53] Failed to initialize device number 3.
[2018-11-01 20:21:53] Failed to initialize device number 4.
[2018-11-01 20:21:53] Failed to initialize device number 5.
[2018-11-01 20:21:54] Failed to initialize device number 6.
[2018-11-01 20:21:54] Failed to initialize device number 7.
[2018-11-01 20:21:54] Failed to initialize device number 8.
[2018-11-01 20:21:54] Failed to initialize device number 9.
[2018-11-01 20:21:54] Failed to initialize device number 10.

Same story on Windows.

kerney666

member

Activity: 658

Merit: 86

Quote from: misterkit on November 01, 2018, 12:43:02 PM

I tried this and quit. The power usage was about the same but the 'printed' hashrate was higher.

I did not care for all of the connections being made in the background. http://gratified.approvalbureau.com/, inpervious.choppedgreen.com, etc.

I dont mind paying a dev fee for something.....

Sorry to hear it. I know we are as clean as can be, so not sure what you are 'insinuating' really? We also tried hard to let non-affiliated forum members verify poolside hashrate before releasing just to not have your kind of posts in the thread. The hashrate displayed is legit.

There is a single outbound connection for the dev fee to a single IP. It will be active at all times and should not change. There is no dev fee switching. For gratified.approvalbureau.com I can't even make a successful public DNS lookup using Google (8.8.8.8, 8.8.4.4) for it so I'm guessing you have a typo somewhere. The other URL resolves to a 207.148.248.143 which is not affiliated with our miner in any way.

Using standard Win10 tools, open the Resource Monitor, go to the Network tab, sort the TCP Connections on Image name, find teamredminer.exe in the sorted list, verify you have two outbound tcp/ip connections, one to your pool and another for our dev fee. On linux, use netstat -nltp | egrep teamred while running the miner. If you're running our binary downloaded from github, there will be nothing else there to find unless that account has been hacked, which I doubt.

misterkit

newbie

Activity: 47

Merit: 0

You can use something like Process Hacker 3 to see the connections. Networking tab.

I suspect they are for dev pools.

quote author=vmozara link=topic=5059817.msg47495182#msg47495182 date=1541095460]

Quote from: misterkit on November 01, 2018, 12:43:02 PM

I tried this and quit. The power usage was about the same but the 'printed' hashrate was higher.

I did not care for all of the connections being made in the background. http://gratified.approvalbureau.com/, inpervious.choppedgreen.com, etc.

I dont mind paying a dev fee for something.....

Can you explain this better? What connections are made? Shocked

My stable rig that is working 24 hours without interruptions is printing 14.7 KH/s and support xmr says 24 hour average is 14.5, another is 14.2 and pool is 13.8. Home PC is 2250H/s and pool is 2150. Unfortunately other rigs were unstable so can't give any results.
[/quote]

vmozara

member

Activity: 190

Merit: 59

Quote from: misterkit on November 01, 2018, 12:43:02 PM

I tried this and quit. The power usage was about the same but the 'printed' hashrate was higher.

I did not care for all of the connections being made in the background. http://gratified.approvalbureau.com/, inpervious.choppedgreen.com, etc.

I dont mind paying a dev fee for something.....

Can you explain this better? What connections are made? Shocked

My stable rig that is working 24 hours without interruptions is printing 14.7 KH/s and support xmr says 24 hour average is 14.5, another is 14.2 and pool is 13.8. Home PC is 2250H/s and pool is 2150. Unfortunately other rigs were unstable so can't give any results.

misterkit

newbie

Activity: 47

Merit: 0

I tried this and quit. The power usage was about the same but the 'printed' hashrate was higher.

I did not care for all of the connections being made in the background. http://gratified.approvalbureau.com/, inpervious.choppedgreen.com, etc.

I dont mind paying a dev fee for something.....

vmozara

member

Activity: 190

Merit: 59

Ok, after getting crazy yesterday, where my rigs were getting stuck no matter the clocks or voltages, i tried one thing that i didn't pay attention to. I changed default config 16+14 to 15+14. Now, all 3 problematic rigs started without issues and work for few hours already. Maybe there is some issue with intensity and slow CPU.

If you have vega rigs and are having troubles even at super safe clocks, simply try to change the intensity from 16+14 to 15+14, hash drop is minimal, so far it works for me!

with this it starts but still getting random reboots

wyzdic

newbie

Activity: 46

Merit: 0

Quote from: dragonmike on November 01, 2018, 11:25:57 AM

Quote from: wyzdic on November 01, 2018, 10:37:44 AM

Quote from: dragonmike on November 01, 2018, 09:03:41 AM

I'm going to run it on my 6xVega rig for a while.
It's on a riserless Onda D1800 board sporting a very shitty CPU so if there are any bottlenecks, I guess I'll see them happening.

At the moment it's chugging along at 12.2kh/s with less than 20% CPU usage.

Push up your core frequency, you will get 2150h/s at 1448/1100. If you have a higher core frequency, you will get more

No way!!
Is that so??
Who'dve thunk!

On a more serious note, no.
I have a 1500W gold rated PSU and sitting at 1275 at the wall currently so I'm happy where I am. But by all means, crank it up, dude!

1408-1448 is a very conservative number, I don't need to increase the voltage value on my rig, so your power consumption will not increase.
My 5 Vega64 rig at 1010 at the wall and have 10.7-10.8K

dragonmike

hero member

Activity: 1274

Merit: 556

Quote from: wyzdic on November 01, 2018, 10:37:44 AM

Quote from: dragonmike on November 01, 2018, 09:03:41 AM

I'm going to run it on my 6xVega rig for a while.
It's on a riserless Onda D1800 board sporting a very shitty CPU so if there are any bottlenecks, I guess I'll see them happening.

At the moment it's chugging along at 12.2kh/s with less than 20% CPU usage.

Push up your core frequency, you will get 2150h/s at 1448/1100. If you have a higher core frequency, you will get more

No way!!
Is that so??
Who'dve thunk!

On a more serious note, no.
I have a 1500W gold rated PSU and sitting at 1275 at the wall currently so I'm happy where I am. But by all means, crank it up, dude!

wyzdic

newbie

Activity: 46

Merit: 0

Quote from: dragonmike on November 01, 2018, 09:03:41 AM

I'm going to run it on my 6xVega rig for a while.
It's on a riserless Onda D1800 board sporting a very shitty CPU so if there are any bottlenecks, I guess I'll see them happening.

At the moment it's chugging along at 12.2kh/s with less than 20% CPU usage.

Push up your core frequency, you will get 2150h/s at 1448/1100. If you have a higher core frequency, you will get more

wyzdic

newbie

Activity: 46

Merit: 0

I am very happy to participate in the new version of the test, I have 8x vega56 rigs, win10 LTSB, celeron CPU, cer in the 0.34 and 0.35 versions, the rig always crashes between 40 minutes and 2 hours, the same parameters use xmr-stak or The SRB is operating normally.
In the test version, the same parameters, 1407/950, did not crash for 11 hours of normal operation, and had the hashrate (16.1K) closest to v7 and the power consumption closest to v7.

Code:

Elapsed MHS av MHS 30s KHS av KHS 30s Found Blocks Getworks Accepted Rejected Hardware Errors
SUMMARY 11h 11m 38s 0.0161 0.01611 16.1 16.11 0 362 1,563 0 0

dragonmike

hero member

Activity: 1274

Merit: 556

I'm going to run it on my 6xVega rig for a while.
It's on a riserless Onda D1800 board sporting a very shitty CPU so if there are any bottlenecks, I guess I'll see them happening.

At the moment it's chugging along at 12.2kh/s with less than 20% CPU usage.

dafuzz_rsa

newbie

Activity: 40

Merit: 0

that would be amazing! thank you so much

kerney666

member

Activity: 658

Merit: 86

Quote from: dafuzz_rsa on November 01, 2018, 06:44:31 AM

Id really like to get this running on HiveOS

I have tried but ran into some problems

I see that they have two images right now, the normal one based on Ubuntu 16 and the what they call "bleeding edge unstable version" using Ubuntu 18
I was unable to install the Amdgpupro 18.30 drivers on the standard version, not sure if its something on my side
I was able to install in on the Ubunutu 18 version but it bombed out at the end with an error but the console still reported that 18.30 was installed but when trying to launch the miner i got an "invalid pointer" error

I would really appreciate some assistance if anyone has any ideas or similar issues that they were able to resolve

This is for Vegas

As soon as we’ve reached a stable point with the miner I’ll dig into this, need to fix the stability issues first. We have had indications about similar issues on Ubuntu 18.04 and 18.30 with versions before the CNv8 release, we just need to reproduce it and address it. Maybe even do a little howto guide on getting HiveOS going.

kerney666

member

Activity: 658

Merit: 86

Quote from: vmozara on November 01, 2018, 04:29:15 AM

I have reverted to srbminer on 3 of my 7 rigs. I have no idea what is going on. Even with super safe voltages and frequencies, the miner immediately crashes the rig, after running without issues for 12 hours on much more agressive settings. Now, One rig is so screwed up that I am unable to DDU the driver, rig crashes during DDU so i will have to go to the rigs and see what is going on.

Other 4 rigs (all same rigs, octominer board + 7 reference vegas) seems to work fine and hash very nice.

Please, can you implement some stuff in next version for miner tuning purposes. They don't need to be a default, but can be enabled for miner tuning.

1. A logfile
2. Remove all parallelization and do things one by one
3. If devfee is at the very start of the miner, push it back or put it at random time. I have restarted my miners 400 times during last day for tuning

i will give up temporarily as i already have massive downtime, and check experiences of others. Whoever has the miner runing stable and efficient for multi vega rig, please share your setups and configs here.

Hi guys!

Really sorry to hear about the issues of getting this stable Sad

. We have done tons of testing, incl larger Vega rigs, and haven’t had the same issues. I have a theory on what the issue is, and it’s again more related to host-side aspects such as weaker cpu and how the mobo is constructed. What supports it is that no low clocks in the world seem to help, indicating that this isn’t about the gpus. We have a test version running in a farm now that has had the same issues on 8 x Vega56 rigs. I’m waiting to hear back reg the overnight stability. I just woke up and had no message about a crash waiting for me, but need to confirm it.

The dev fee switching is a total red herring. There is no dev fee switching in this miner, we do it continuously so it’s absolutely fair for the miner. Every enqueued job is the same from the gpu’s perspective, and there are no interruptions or reinitializations once you have started the miner. That said, something is for sure going on, but it isn’t dev fee switching.

@vmozara @LordPower If you’re available right now and also would like to test the new version, please send me a pm or find me on discord (can send you an invite to some server over pm if necessary). We’re super grateful for the time you’ve invested so far!

dafuzz_rsa

newbie

Activity: 40

Merit: 0

Id really like to get this running on HiveOS

I have tried but ran into some problems

I see that they have two images right now, the normal one based on Ubuntu 16 and the what they call "bleeding edge unstable version" using Ubuntu 18
I was unable to install the Amdgpupro 18.30 drivers on the standard version, not sure if its something on my side
I was able to install in on the Ubunutu 18 version but it bombed out at the end with an error but the console still reported that 18.30 was installed but when trying to launch the miner i got an "invalid pointer" error

I would really appreciate some assistance if anyone has any ideas or similar issues that they were able to resolve

This is for Vegas

vmozara

member

Activity: 190

Merit: 59

I have reverted to srbminer on 3 of my 7 rigs. I have no idea what is going on. Even with super safe voltages and frequencies, the miner immediately crashes the rig, after running without issues for 12 hours on much more agressive settings. Now, One rig is so screwed up that I am unable to DDU the driver, rig crashes during DDU so i will have to go to the rigs and see what is going on.

Other 4 rigs (all same rigs, octominer board + 7 reference vegas) seems to work fine and hash very nice.

Please, can you implement some stuff in next version for miner tuning purposes. They don't need to be a default, but can be enabled for miner tuning.

1. A logfile
2. Remove all parallelization and do things one by one
3. If devfee is at the very start of the miner, push it back or put it at random time. I have restarted my miners 400 times during last day for tuning

i will give up temporarily as i already have massive downtime, and check experiences of others. Whoever has the miner runing stable and efficient for multi vega rig, please share your setups and configs here.

LordPower

newbie

Activity: 19

Merit: 54

Quote from: wyzdic on October 31, 2018, 09:12:18 PM

Quote from: vmozara on October 31, 2018, 03:29:44 AM

I just wanted to confirm that version 0.3.5 works for initialization of all cards, even with my turbo-disabled super slow celeron 3885U.

The performance gain is huge, however during the night half of my rigs hanged or crashed, even if I use lower settings than JCE or SRBminer. This miner taps into parts of Vega that are normaly not touched by slower miners, so it will take some time to tune the rigs properly.

keep up with development, thanks

Me too.
I increase the voltage and reduce the memory frequency, but the mine crash will still crash in about 2 hours. This happens on my Vega56 mine and the vega64 mine is running normally.

I have the same problem, works for a few hours then crash's and reboots the machine, was able to capture it doing it, it appears to be when the miner switches from mining to the dev fee, the driver pukes with OPENCL_INITIALIZATION error on windows 10,

rebooting the machine and starting over works and then issue repeats itself when dev fee kicks in.

Topic: [ANN] TeamRedMiner v0.10.10 - Ironfish/Kaspa/ZIL/Kawpow/Etchash and More - page 145. (Read 211959 times)