Pages:
Author

Topic: SILENTARMY v5: Zcash miner, 115 sol/s on R9 Nano, 70 sol/s on GTX 1070 - page 40. (Read 209314 times)

sr. member
Activity: 652
Merit: 266
0.3.3] Fix VM_CONTEXT1_PROTECTION_FAULT_ADDR.
Confirmed.Errors gone now.
Will test for a few hours to see the difference, but looks promising 540S/s against 380S/s for FOSS miner.

EDIT:
After 12 minutes one of the cards went gone...
Quote
[12836.439466] amdgpu 0000:03:00.0: GPU fault detected: 147 0x0c908402
[12836.439485] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013F592
[12836.439501] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A084002
[12836.439517] VM fault (0x02, vmid 5) at page 1308050, read from 'TC7' (0x54433700) (132)
[12964.285030] amdgpu 0000:03:00.0: GPU fault detected: 147 0x0c908402
[12964.285047] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013F592
[12964.285063] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A084002
[12964.285080] VM fault (0x02, vmid 5) at page 1308050, read from 'TC7' (0x54433700) (132)
[12976.758284] amdgpu 0000:03:00.0: GPU fault detected: 147 0x00004802
[12976.758303] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x00200000
[12976.758320] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A048002
[12976.758337] VM fault (0x02, vmid 5) at page 2097152, read from 'TC4' (0x54433400) (72)
[12976.766850] amdgpu 0000:03:00.0: GPU fault detected: 147 0x0c908402
[12976.766869] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013F592
[12976.766885] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A084002
[12976.766901] VM fault (0x02, vmid 5) at page 1308050, read from 'TC7' (0x54433700) (132)
[13065.948618] amdgpu 0000:03:00.0: GPU fault detected: 147 0x0c908402
[13065.949169] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013F592
[13065.949713] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A084002
[13065.950283] VM fault (0x02, vmid 5) at page 1308050, read from 'TC7' (0x54433700) (132)
[13461.645347] amdgpu 0000:03:00.0: GPU fault detected: 147 0x0c908402
[13461.645898] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013F592
[13461.646437] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A084002
[13461.646977] VM fault (0x02, vmid 5) at page 1308050, read from 'TC7' (0x54433700) (132)
[13474.869840] amdgpu 0000:03:00.0: GPU fault detected: 147 0x0c908402
[13474.870393] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0013F592
[13474.870931] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0A084002
[13474.871471] VM fault (0x02, vmid 5) at page 1308050, read from 'TC7' (0x54433700) (132)

Still issues with this miner.
hero member
Activity: 1246
Merit: 708
0.3.3] Fix VM_CONTEXT1_PROTECTION_FAULT_ADDR.
legendary
Activity: 1470
Merit: 1024
Did you try 3.3. Error is gone for me in this upgrade.
sr. member
Activity: 652
Merit: 266
Yeah, I think it's totally worth it to put up a serious open source competition against Sir. Claymore  Wink
We shall be back.
I'll try get Claymore's v5/optiminer speed in 2-3 days.

p.s. he have a more motivation because his model (miner with fee) works perfectly - opposite to my (opensource miner and own pool), miners don't want support coinsforall.io Sad

I personally think optiminer's way is far better. The dev fee is reducted from the hashrate, without us knowing. I hate seeing devfee on the miner and trying to reconnect for ages. Maybe you could do that?

Oh btw, its optiminer's speed you have to compete with atm, minus the def fees I'm having 130 stable hashrate with an rx 470 which was never possible or stable with claymore.
You can't see that on windows, but I have enabled PCIe debug mode on my kernel and with optiminer I'm seeing a lot of VM faults which can possibly damage your card...
With silentarmy I don't see any and is 99.9% stable.

Quote
2016-11-15 16:12:06,081 INFO  [default] Optiminer/Zcash 0.3.2 (C) Optiminer 2016
2016-11-15 16:12:06,176 INFO  [GPU0] Using device 'Ellesmere
2016-11-15 16:12:06,207 INFO  [stratum] Extranonce is dcfa8fc46b
2016-11-15 16:12:06,266 INFO  [stratum] Connected to eu1-zcash.flypool.org:3333
2016-11-15 16:12:06,267 INFO  [stratum] Mining target is 004189374bc6a7ef9db22d0e5604189374bc6a7ef9db22d0e5604189374bc6a7
2016-11-15 16:12:10,839 INFO  [GPU1] Using device 'Ellesmere
2016-11-15 16:12:15,545 INFO  [GPU2] Using device 'Ellesmere
2016-11-15 16:12:16,086 INFO  [Total]  94.0 I/s 180.0 S/s (1s)
2016-11-15 16:12:16,116 INFO  [GPU0]  67.0 I/s 129.0 S/s (1s)
2016-11-15 16:12:20,316 INFO  [GPU3] Using device 'Ellesmere
2016-11-15 16:12:20,788 INFO  [GPU1]  66.0 I/s 120.0 S/s (1s)
2016-11-15 16:12:25,480 INFO  [GPU2]  67.0 I/s 118.0 S/s (1s)
2016-11-15 16:12:26,088 INFO  [Total]  258.0 I/s 508.0 S/s (1s)
2016-11-15 16:12:26,128 INFO  [GPU0]  67.0 I/s 145.0 S/s (1s)
2016-11-15 16:12:26,160 INFO  [GPU3] Share accepted! (1/1)
2016-11-15 16:12:27,475 INFO  [GPU0] Share accepted! (2/2)
2016-11-15 16:12:27,917 INFO  [GPU3] Share accepted! (3/3)

Quote
[12124.304150] amdgpu 0000:01:00.0: GPU fault detected: 147 0x09f8c401
[12124.304177] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x04102F3F
[12124.304194] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x040C4001
[12124.304210] VM fault (0x01, vmid 2) at page 68169535, read from 'TC3' (0x54433300) (196)
[12130.680841] amdgpu 0000:03:00.0: GPU fault detected: 147 0x09f8c401
[12130.680861] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x04102F3F
[12130.680878] amdgpu 0000:03:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0C0C4001
[12130.680894] VM fault (0x01, vmid 6) at page 68169535, read from 'TC3' (0x54433300) (196)
[12134.350936] amdgpu 0000:02:00.0: GPU fault detected: 147 0x09f8c401
[12134.350957] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x04102F3F
[12134.350973] amdgpu 0000:02:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x060C4001
[12134.350990] VM fault (0x01, vmid 3) at page 68169535, read from 'TC3' (0x54433300) (196)
[12135.382376] amdgpu 0000:01:00.0: GPU fault detected: 147 0x09f8c401
[12135.382395] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x04102F3F
[12135.382411] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x040C4001
[12135.382428] VM fault (0x01, vmid 2) at page 68169535, read from 'TC3' (0x54433300) (196)
[12135.382716] amdgpu 0000:01:00.0: GPU fault detected: 147 0x09f8c401
[12135.382745] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x04102F3F
[12135.382774] amdgpu 0000:01:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x040C4001
[12135.382803] VM fault (0x01, vmid 2) at page 68169535, read from 'TC3' (0x54433300) (196)
[12147.365389] amdgpu 0000:04:00.0: GPU fault detected: 147 0x09f8c401
[12147.365408] amdgpu 0000:04:00.0:   VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x04102F3F
[12147.365424] amdgpu 0000:04:00.0:   VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x0E0C4001
[12147.365440] VM fault (0x01, vmid 7) at page 68169535, read from 'TC3' (0x54433300) (196)

legendary
Activity: 1470
Merit: 1024
Yeah, I think it's totally worth it to put up a serious open source competition against Sir. Claymore  Wink
We shall be back.
I'll try get Claymore's v5/optiminer speed in 2-3 days.

p.s. he have a more motivation because his model (miner with fee) works perfectly - opposite to my (opensource miner and own pool), miners don't want support coinsforall.io Sad

I personally think optiminer's way is far better. The dev fee is reducted from the hashrate, without us knowing. I hate seeing devfee on the miner and trying to reconnect for ages. Maybe you could do that?

Oh btw, its optiminer's speed you have to compete with atm, minus the def fees I'm having 130 stable hashrate with an rx 470 which was never possible or stable with claymore.
sr. member
Activity: 652
Merit: 266
I merged the low CPU Nvidia patch into my Windows branch. Don't have the hardware to verify correct functionality.

Also submitted a tiny (potential) speed bump for all cards that don't have 36 compute units.

link? i've found only the linux version on your github i want to try it
https://github.com/Genoil/silentarmy/tree/windows
hero member
Activity: 1246
Merit: 708
As I cannot find Optiminer thread I decided to post here:
Quote
ERROR [default] Failed to send message!
2016-11-15 15:12:29,802 ERROR [stratum] IO/Error communicating to zec.coinmine.pl

Earlier I had an issue where screen with miner disappeared few times  Huh
legendary
Activity: 3248
Merit: 1070
I merged the low CPU Nvidia patch into my Windows branch. Don't have the hardware to verify correct functionality.

Also submitted a tiny (potential) speed bump for all cards that don't have 36 compute units.

link? i've found only the linux version on your github i want to try it
full member
Activity: 142
Merit: 100
I merged the low CPU Nvidia patch into my Windows branch. Don't have the hardware to verify correct functionality.

Also submitted a tiny (potential) speed bump for all cards that don't have 36 compute units.

Nice, will test later  Smiley
sr. member
Activity: 438
Merit: 250
I merged the low CPU Nvidia patch into my Windows branch. Don't have the hardware to verify correct functionality.

Also submitted a tiny (potential) speed bump for all cards that don't have 36 compute units.
sr. member
Activity: 652
Merit: 266
Yeah, I think it's totally worth it to put up a serious open source competition against Sir. Claymore  Wink
We shall be back.
I'll try get Claymore's v5/optiminer speed in 2-3 days.

p.s. he have a more motivation because his model (miner with fee) works perfectly - opposite to my (opensource miner and own pool), miners don't want support coinsforall.io Sad

you could not do a normal windows package, instead of installing python and all the linux stuff? not all people are able to use this miner Roll Eyes Roll Eyes Roll Eyes Roll Eyes

That's the point...instead of doing UI and ease-of-use enhancements devs are targeting stability and hash speed... I think speed/stability are high priority now.


P.S. Modified fan.sh to match only AMD cards due to script crash when having IntelHD card enabled.
https://drive.google.com/drive/folders/0B72yKpOokCMcOWxNTHBPRm4xRms
Go to scripts/fan.sh
Set THERMOSTAT=~60; temperature as it suits you and add to crontab
*/1 * * * * /path-to-fan.sh  > /dev/null 2>&1
hero member
Activity: 798
Merit: 1000
Yeah, I think it's totally worth it to put up a serious open source competition against Sir. Claymore  Wink
We shall be back.
I'll try get Claymore's v5/optiminer speed in 2-3 days.

p.s. he have a more motivation because his model (miner with fee) works perfectly - opposite to my (opensource miner and own pool), miners don't want support coinsforall.io Sad

you could not do a normal windows package, instead of installing python and all the linux stuff? not all people are able to use this miner Roll Eyes Roll Eyes Roll Eyes Roll Eyes
legendary
Activity: 1500
Merit: 1002
Mine Mine Mine
the army marches on ! go go go . . .
legendary
Activity: 1274
Merit: 1000
The more people port the better the testing is..
I like to test if you don't mind...

Some details about your port?

I wasn't sure if was going to post... but have a look.

I am not a developer, this was my own personal proof of concept, nothing special, I know nothing about coding.

I suspect it is easy to do, in reality, not for me  Smiley

I used sa-solver built by Genoil, V5 sources and Python34 with cx_Freeze.

I take no credit for this work.

https://mega.nz/#F!KsIlmLQC!1m8SdCwcpLZmSFkOaW4R9w

I got it working following what you posted but seemed the same as zawawa speed wise if that's you Smiley .


Your exe version works great .

Thank you for the compliment, speed is the same as zawawa, but does not rely on full Cygwin install.

Your Welcome i would rather use a exe file over a lot of other programs but the plus side for me i got a fast crash course on how to compile in windows without using Linux or using UBUNTU for windows 10 or use a Linux PC or USB drive  . i all ways used a full Linux version.
sr. member
Activity: 574
Merit: 250
Fighting mob law and inquisition in this forum
Well its working..but when you close it ..it is freezing the system the problem I always had with genoils builds :-D
full member
Activity: 142
Merit: 100
The more people port the better the testing is..
I like to test if you don't mind...

Some details about your port?

I wasn't sure if was going to post... but have a look.

I am not a developer, this was my own personal proof of concept, nothing special, I know nothing about coding.

I suspect it is easy to do, in reality, not for me  Smiley

I used sa-solver built by Genoil, V5 sources and Python34 with cx_Freeze.

I take no credit for this work.

https://mega.nz/#F!KsIlmLQC!1m8SdCwcpLZmSFkOaW4R9w

I got it working following what you posted but seemed the same as zawawa speed wise if that's you Smiley .


Your exe version works great .

Thank you for the compliment, speed is the same as zawawa, but does not rely on full Cygwin install.
sr. member
Activity: 728
Merit: 304
Miner Developer
i decided to dump Cygwin because it was giving me too much headache.
I will apply all the latest patches to Genoil's port and remove the Python dependency from it.
legendary
Activity: 1274
Merit: 1000
The more people port the better the testing is..
I like to test if you don't mind...

Some details about your port?

I wasn't sure if was going to post... but have a look.

I am not a developer, this was my own personal proof of concept, nothing special, I know nothing about coding.

I suspect it is easy to do, in reality, not for me  Smiley

I used sa-solver built by Genoil, V5 sources and Python34 with cx_Freeze.

I take no credit for this work.

https://mega.nz/#F!KsIlmLQC!1m8SdCwcpLZmSFkOaW4R9w

I got it working following what you posted but seemed the same as zawawa speed wise if that's you Smiley .


Your exe version works great .
full member
Activity: 142
Merit: 100
The more people port the better the testing is..
I like to test if you don't mind...

Some details about your port?

I wasn't sure if was going to post... but have a look.

I am not a developer, this was my own personal proof of concept, nothing special, I know nothing about coding.

I suspect it is easy to do, in reality, not for me  Smiley

I used sa-solver built by Genoil, V5 sources and Python34 with cx_Freeze.

I take no credit for this work.

https://mega.nz/#F!KsIlmLQC!1m8SdCwcpLZmSFkOaW4R9w
sr. member
Activity: 574
Merit: 250
Fighting mob law and inquisition in this forum
The more people port the better the testing is..
I like to test if you don't mind...

Some details about your port?
Pages:
Jump to: