SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 47.

chimaeraa

newbie

Activity: 49

Merit: 0

Quote from: doktor83 on March 22, 2019, 01:57:37 AM

Quote from: rdome on March 21, 2019, 01:45:04 AM

I think it is v4 problem. 1.8.1 is better than 1.8.0 but still not 100% stable. It was no problem on other algos in srbminer, eth rock stable, mined beam for weaks rock stable, everything i was trowing at vegas just worked. Then it came v4 and it came problems. I just love srbminer cause it is best miner for vega cards, but when v4 cam and it was unstable, i tried other miners for v4, you have at lleast 4 of them, and they crash even faster, teamredminer crashed under 10min on same settings as srbminer. Tried to underclock gpu,mem for stability and same. Problem is only on v4, other algos are fine Wink

Something is very bad with new algo, just put loop in bat file and it will at least run again when it crashes Wink

The whole concept of how v4 works is bad on gpu's. Compiling code on every new block, then need to release that memory, but it is known that the amd cl_release api is buggy and it wont always work as it should.
For cpu's this concept of create code + compile is not a problem , but for gpu's it is.

I am trying out various things, hope for the best Grin

hi doc
v4 algor , designed for cpu not gpu.

do you test new amd driver ? it's very good. it's work for few day just few crash . Cheesy

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: rdome on March 21, 2019, 01:45:04 AM

I think it is v4 problem. 1.8.1 is better than 1.8.0 but still not 100% stable. It was no problem on other algos in srbminer, eth rock stable, mined beam for weaks rock stable, everything i was trowing at vegas just worked. Then it came v4 and it came problems. I just love srbminer cause it is best miner for vega cards, but when v4 cam and it was unstable, i tried other miners for v4, you have at lleast 4 of them, and they crash even faster, teamredminer crashed under 10min on same settings as srbminer. Tried to underclock gpu,mem for stability and same. Problem is only on v4, other algos are fine Wink

Something is very bad with new algo, just put loop in bat file and it will at least run again when it crashes Wink

The whole concept of how v4 works is bad on gpu's. Compiling code on every new block, then need to release that memory, but it is known that the amd cl_release api is buggy and it wont always work as it should.
For cpu's this concept of create code + compile is not a problem , but for gpu's it is.

I am trying out various things, hope for the best Grin

vmozara

member

Activity: 190

Merit: 59

Guys, can you please help with one small issue

7x vega 56 rig 1450mhz core 1100 mem
srbminer 1.8.1
adrenalin 18.6.1
intensity 62,

miner is always startying HAVEN at 12600H/s. But if i stop it many times and restart, it will eventually start at 13000H/s. How can i get it to always start at maximum performance?

Thanks!

Bllacky

newbie

Activity: 22

Merit: 0

Quote from: Bllacky on February 11, 2019, 04:21:53 AM

Quote from: Bllacky on January 21, 2019, 11:09:50 AM

Quote from: Bllacky on January 07, 2019, 05:54:17 AM

Quote from: doktor83 on January 04, 2019, 11:54:41 AM

Quote from: Bllacky on January 04, 2019, 03:49:08 AM

Feedback on
V. 1.7.5

Now it finally works when I have mix of 270X and 280X video cards.
Hashrates are very good for normal V8 algo:
280X 450 H/s
270X 345 H/s
290 530H/s

I do not OC.
I also sometimes get lots of STALE shares.

Good job!

Awesome, thanks for reporting.
Also stales shown in miner are not a problem because those are accepted shares.
Its just more a 'cosmetical thing'.

Feedback on
V. 1.7.6

Hashrates have gone up for normal V8 algo:
280X 455 H/s
270X 400 H/s
290 650H/s

Temperatures seem to be better on the 270X with this version, or maybe now they are read correctly.

Feedback on
V. 1.7.7

Hashrates have gone up by very little for normal V8 algo:
280X 460 H/s
270X 405 H/s
290 655H/s

Fan rpm on the 270X is now correctly reported.

Thank you!

Feedback on
V. 1.7.8

Hashrates have gone up by very little for normal V8 algo:
280X 460 H/s
270X 405 H/s
290 655H/s

Basically no change. Miner has gotten a lot more stable in the past 2 months.

Feedback on
V. 1.8.1

Hashrates are similar to normal v8 with normal V4 algo:
280X 450 H/s
270X 400 H/s
290 750H/s

Runs pretty stable for me. Ran for 1 week without any crash.
Settings are almost the same as on normal v8 except that intensities need to be adjusted for optimal performance.

Mananier

newbie

Activity: 1

Merit: 0

I have vega 56. All algorithms work fine, except CNv4. Vega resets the hashrate, the miner restarts several times within 10 to 24 hours. Then the whole system hangs up.

Before the update of the butterfly, the monero was mined normally.

GKumaran

member

Activity: 204

Merit: 10

Quote from: doktor83 on March 20, 2019, 11:55:55 AM

Quote from: RuMiner on March 20, 2019, 11:53:19 AM

Quote from: chimaeraa on March 20, 2019, 09:11:44 AM

can you see your windows event viewer and see what log save there ?

last crash (and all the previous). source: Application error, event code: 1000, category: (100), and full path to SRBMiner exe file. Nothing else near.
I'm catching error 1000 event via scheduler and run reboot.bat in case of it. Rig reboots and goes well for next 6 or more hours

Are you mining V4 ?
Im curious are the crashes happening only on V4 ?

If so, im probably not releasing some of the CL resources as i should..

Can confirm that the same happens on vega runnin non cn4/r algos.

badral07

newbie

Activity: 8

Merit: 0

No more crashes thank you, Doc.

rdome

newbie

Activity: 15

Merit: 0

I think it is v4 problem. 1.8.1 is better than 1.8.0 but still not 100% stable. It was no problem on other algos in srbminer, eth rock stable, mined beam for weaks rock stable, everything i was trowing at vegas just worked. Then it came v4 and it came problems. I just love srbminer cause it is best miner for vega cards, but when v4 cam and it was unstable, i tried other miners for v4, you have at lleast 4 of them, and they crash even faster, teamredminer crashed under 10min on same settings as srbminer. Tried to underclock gpu,mem for stability and same. Problem is only on v4, other algos are fine Wink

Something is very bad with new algo, just put loop in bat file and it will at least run again when it crashes Wink

RuMiner

member

Activity: 168

Merit: 15

Quote from: doktor83 on March 20, 2019, 11:55:55 AM

Are you mining V4 ?
Im curious are the crashes happening only on V4 ?

I was digging previous Monero algo on 1.7.x versions till income dropped below power consumption, and rig was weeks-stable. Then I switched to ETH miner for several months, it was stable also. Crashes began after launching Monero v4 on 1.8.0 and 1.8.1.

chimaeraa

newbie

Activity: 49

Merit: 0

Quote from: doktor83 on March 20, 2019, 11:55:55 AM

Quote from: RuMiner on March 20, 2019, 11:53:19 AM

Quote from: chimaeraa on March 20, 2019, 09:11:44 AM

can you see your windows event viewer and see what log save there ?

last crash (and all the previous). source: Application error, event code: 1000, category: (100), and full path to SRBMiner exe file. Nothing else near.
I'm catching error 1000 event via scheduler and run reboot.bat in case of it. Rig reboots and goes well for next 6 or more hours

Are you mining V4 ?
Im curious are the crashes happening only on V4 ?

If so, im probably not releasing some of the CL resources as i should..

not V4 with other algo i get seam problem

i see some other error come with this crash ( i use intel graphic for monitor ) it's about intel graphic services .

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: Bllacky on March 20, 2019, 09:12:11 AM

On my 280X video cards I keep getting these warnings with V4:

WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'
WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'
WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'
WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'

Apart from that, it seems to be working.

I will need to figure out the best settings next.

These warnings are harmless, nothing to worry about

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: RuMiner on March 20, 2019, 11:53:19 AM

Quote from: chimaeraa on March 20, 2019, 09:11:44 AM

can you see your windows event viewer and see what log save there ?

last crash (and all the previous). source: Application error, event code: 1000, category: (100), and full path to SRBMiner exe file. Nothing else near.
I'm catching error 1000 event via scheduler and run reboot.bat in case of it. Rig reboots and goes well for next 6 or more hours

Are you mining V4 ?
Im curious are the crashes happening only on V4 ?

If so, im probably not releasing some of the CL resources as i should..

RuMiner

member

Activity: 168

Merit: 15

Quote from: chimaeraa on March 20, 2019, 09:11:44 AM

can you see your windows event viewer and see what log save there ?

last crash (and all the previous). source: Application error, event code: 1000, category: (100), and full path to SRBMiner exe file. Nothing else near.
I'm catching error 1000 event via scheduler and run reboot.bat in case of it. Rig reboots and goes well for next 6 or more hours

rotaryfurball

newbie

Activity: 13

Merit: 0

Quote from: doktor83 on March 20, 2019, 09:11:12 AM

Quote from: RuMiner on March 20, 2019, 09:04:47 AM

Quote from: chimaeraa on March 20, 2019, 06:24:27 AM

it's page file

I don't have same event regarding swap file. also it is set to 70Gb for 8 cards, so it's not the main reason

My guess is that i have a bug somwhere in my net code, so i am fully rewriting it, it's going to take some time, but i hope it will fix these annoying random crashes Roll Eyes

Thank you for the things you do! Will be looking forward for 1.8.2

Bllacky

newbie

Activity: 22

Merit: 0

On my 280X video cards I keep getting these warnings with V4:

WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'
WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'
WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'
WARNING: Linking two modules of different data layouts!
WARNING: Linking two modules of different target triples: input.bc: 'amdil64-pc-unknown-amdopencl' and 'amdil-pc-unknown-amdopencl'

Apart from that, it seems to be working.

I will need to figure out the best settings next.

chimaeraa

newbie

Activity: 49

Merit: 0

Quote from: RuMiner on March 20, 2019, 09:04:47 AM

Quote from: chimaeraa on March 20, 2019, 06:24:27 AM

it's page file

I don't have same event regarding swap file. also it is set to 70Gb for 8 cards, so it's not the main reason

can you see your windows event viewer and see what log save there ?

doktor83

hero member

Activity: 2548

Merit: 626

Quote from: RuMiner on March 20, 2019, 09:04:47 AM

Quote from: chimaeraa on March 20, 2019, 06:24:27 AM

it's page file

I don't have same event regarding swap file. also it is set to 70Gb for 8 cards, so it's not the main reason

My guess is that i have a bug somwhere in my net code, so i am fully rewriting it, it's going to take some time, but i hope it will fix these annoying random crashes Roll Eyes

RuMiner

member

Activity: 168

Merit: 15

Quote from: chimaeraa on March 20, 2019, 06:24:27 AM

it's page file

I don't have same event regarding swap file. also it is set to 70Gb for 8 cards, so it's not the main reason

chimaeraa

newbie

Activity: 49

Merit: 0

it's page file

Quote

- System

- Provider

[ Name] Application Error

- EventID 1000

[ Qualifiers] 0

Level 2

Task 100

Keywords 0x80000000000000

- TimeCreated

[ SystemTime] 2019-03-20T10:06:13.494663500Z

EventRecordID 2069

Channel Application

Computer DESKTOP-3QESJPQ

Security

- EventData

SRBMiner-CN.exe
1.8.1.0
00000000
ntdll.dll
10.0.17763.348
ca65c822
c0000374
00000000000faf49
1cbc
01d4def2eaa5a99b
C:\Users\mahdi\Desktop\SRBMiner-CN-V1-7-0\SRBMiner-CN-V1-7-0\SRBMiner-CN.exe
C:\Windows\SYSTEM32\ntdll.dll
bfbe991a-82da-4d77-aa33-83b28f32d57c

Quote

- System

- Provider

[ Name] Application Popup
[ Guid] {47bfa2b7-bd54-4fac-b70b-29021084ca8f}

EventID 26

Version 0

Level 4

Task 0

Opcode 0

Keywords 0x8000000000000000

- TimeCreated

[ SystemTime] 2019-03-20T10:06:13.471720400Z

EventRecordID 2377

Correlation

- Execution

[ ProcessID] 536
[ ThreadID] 68

Channel System

Computer DESKTOP-3QESJPQ

- Security

[ UserID] S-1-5-18

- EventData

Caption Windows - Virtual Memory Minimum Too Low
Message Your system is low on virtual memory. Windows is increasing the size of your virtual memory paging file. During this process, memory requests for some applications may be denied. For more information, see Help.

SamAlackass

newbie

Activity: 27

Merit: 1

Hi all, I'm getting a lot of low difficulty rejects with 1.8.1 so I can't properly test. I'm back to 1.8.0 for now but it's crashing often without any indication in the logs.

Topic: SRBMiner Cryptonight AMD GPU Miner V1.9.3 - native algo switching - page 47. (Read 237247 times)