Claymore's ZCash/BTG AMD GPU Miner v12.6 (Windows/Linux) - page 152.

IOTUSA

sr. member

Activity: 504

Merit: 250

Quote from: PandaPanda! on February 17, 2017, 11:44:16 AM

Quote from: IOTUSA on February 17, 2017, 11:00:24 AM

Question: I have a few Bios modded Rx 470/480 that I'm trying to get running under Windows 10 ver. 1607. On install they show as 'problem detected, code 43' in device manager. Do I have to use modded driver files and disable signature check etc. to solve the issue? I tried 15.12 and 16.3 and I get the issue with both.

Anyone of you scholars that know?

Use 16.11.5 - will work perfectly. 16.12 and up won't work with modded BIOSes as far as i know.

Ah, OK. Damn I hate AMD drivers.

PandaPanda!

newbie

Activity: 35

Merit: 0

Quote from: IOTUSA on February 17, 2017, 11:01:31 AM

Quote from: PandaPanda! on February 17, 2017, 10:47:12 AM

My farm is a mess with v12.0 - i'm moving back to v11.1 until claymore sorts out all the issues with crashing GPUs etc.

The strange part is that it seems v12.0 actually messed up my Fury cards somehow so now when I revert to 11.x they still crash. Could be some kind of odd coincidence but it seems unlikely.

You know, its funny you say that - i've just noticed the stability I had when I was originally using 11.1 has now gone after downgrading from 12.0 back to 11.1. This is strange.. However, the instability seems to be coming from only my 2xx/3xx cards.. RX4xx series cards seem to be running okay.

PandaPanda!

newbie

Activity: 35

Merit: 0

Quote from: IOTUSA on February 17, 2017, 11:00:24 AM

Question: I have a few Bios modded Rx 470/480 that I'm trying to get running under Windows 10 ver. 1607. On install they show as 'problem detected, code 43' in device manager. Do I have to use modded driver files and disable signature check etc. to solve the issue? I tried 15.12 and 16.3 and I get the issue with both.

Anyone of you scholars that know?

Use 16.11.5 - will work perfectly. 16.12 and up won't work with modded BIOSes as far as i know.

spyshagg

full member

Activity: 176

Merit: 100

suprnova shows my hashrate but not my " SOL " hashrate nor my " SOL " sharerate. Anyone else?

Redhotdog

newbie

Activity: 4

Merit: 0

Quote from: fr4nkthetank on February 17, 2017, 08:42:16 AM

for the incorrect data thingy, I had the same problem with 280x but at i -2 it seems to be good. One card is good at -3. So lower i it should work.

Tried to reduce intensity and remove undervoltage = recieve definetly less "incorrect data" messages. Will see.

IOTUSA

sr. member

Activity: 504

Merit: 250

Quote from: PandaPanda! on February 17, 2017, 10:47:12 AM

My farm is a mess with v12.0 - i'm moving back to v11.1 until claymore sorts out all the issues with crashing GPUs etc.

The strange part is that it seems v12.0 actually messed up my Fury cards somehow so now when I revert to 11.x they still crash. Could be some kind of odd coincidence but it seems unlikely.

IOTUSA

sr. member

Activity: 504

Merit: 250

Question: I have a few Bios modded Rx 470/480 that I'm trying to get running under Windows 10 ver. 1607. On install they show as 'problem detected, code 43' in device manager. Do I have to use modded driver files and disable signature check etc. to solve the issue? I tried 15.12 and 16.3 and I get the issue with both.

Anyone of you scholars that know?

PandaPanda!

newbie

Activity: 35

Merit: 0

My farm is a mess with v12.0 - i'm moving back to v11.1 until claymore sorts out all the issues with crashing GPUs etc.

IOTUSA

sr. member

Activity: 504

Merit: 250

Quote from: Redhotdog on February 17, 2017, 07:53:40 AM

Quote from: d57heinz on February 17, 2017, 07:24:56 AM

Same here as well. Gpu 1 incorrect data. 7970. Finally pulled the card and put in a 390x. All is well. Funny was that after running v12. That 7970 does same now on v11.1. Think it wiped out my card. Def some bugs in the new release. Save yourself some headaches and move to v11.1 if you can't get stability. At least it will run more than a few hours. Also same rig when it would drop to 0 hash on a different random card before replacing 7970 it would start getting socket errors. Then pc was no longer accessible via team viewer. Had to manually reboot Computer. Great effort claymore. I am getting over 400 sols on 390x. Those seem to have no issues running v12.

BR

Also rolled back to 11.1
My rig consists of 3 RX470 and one R9 390 (the above mentioned GPU 1). Tried to increase voltage. Nothing helped.

I have a 6 x Fury rig that also started behaving real odd. I keep getting 'GPU 3' crashing for no reason. Driver reinstall, and stock settings have not helped.

fr4nkthetank

legendary

Activity: 2294

Merit: 1182

Now the money is free, and so the people will be

for the incorrect data thingy, I had the same problem with 280x but at i -2 it seems to be good. One card is good at -3. So lower i it should work.

Redhotdog

newbie

Activity: 4

Merit: 0

Quote from: d57heinz on February 17, 2017, 07:24:56 AM

Same here as well. Gpu 1 incorrect data. 7970. Finally pulled the card and put in a 390x. All is well. Funny was that after running v12. That 7970 does same now on v11.1. Think it wiped out my card. Def some bugs in the new release. Save yourself some headaches and move to v11.1 if you can't get stability. At least it will run more than a few hours. Also same rig when it would drop to 0 hash on a different random card before replacing 7970 it would start getting socket errors. Then pc was no longer accessible via team viewer. Had to manually reboot Computer. Great effort claymore. I am getting over 400 sols on 390x. Those seem to have no issues running v12.

BR

Also rolled back to 11.1
My rig consists of 3 RX470 and one R9 390 (the above mentioned GPU 1). Tried to increase voltage. Nothing helped.

d57heinz

legendary

Activity: 1453

Merit: 1011

Bitcoin Talks Bullshit Walks

Quote from: Redhotdog on February 17, 2017, 05:53:57 AM

Constantly getting "GPU #1 returned incorrect data!" on v12.0
v11.1 works fine.
Virtual memory size = 16Gb

SET GPU_FORCE_64BIT_PTR 1
SET GPU_MAX_HEAP_SIZE 100
SET GPU_USE_SYNC_OBJECTS 1
SET GPU_MAX_ALLOC_PERCENT 100
SET GPU_SINGLE_ALLOC_PERCENT 100

are set

UPD. In my case they are set via SETX but I think it doesn't matter.

Same here as well. Gpu 1 incorrect data. 7970. Finally pulled the card and put in a 390x. All is well. Funny was that after running v12. That 7970 does same now on v11.1. Think it wiped out my card. Def some bugs in the new release. Save yourself some headaches and move to v11.1 if you can't get stability. At least it will run more than a few hours. Also same rig when it would drop to 0 hash on a different random card before replacing 7970 it would start getting socket errors. Then pc was no longer accessible via team viewer. Had to manually reboot Computer. Great effort claymore. I am getting over 400 sols on 390x. Those seem to have no issues running v12.

BR

pwpwpw

hero member

Activity: 908

Merit: 503

Quote from: ownmind on February 17, 2017, 01:38:53 AM

Quote from: golden0707 on February 17, 2017, 12:50:24 AM

Hi ALL!

So, having 4*R9 280X with 8 Gb RAM and 32 Gb Virtual mem. CM 12.0 runs only with -i 2 stable. With -i 3 or -i 4 after ~1 hr. one card hows 0 hash and miner restarts . With default -i ( -i 6) miner writes low virtual mem. and lower intensity to 5 and i get same situation as with -i 3 , -i 4. Only way to run it stable is -i 2 option Huh

? My hash rate lower by ~10-15 hashes per card with this intensity. Anyone in this situation ?

pwpwpw,thesaltonsea and me ownmind. We all in one boat. Claymore or someone, can you explain this situation? Please, take a look at problem.

https://bitcointalksearch.org/topic/m.17878165

4x 280x, tried to run -i 0 all cards still fail, looked at v11.1 had settings at i 6,6,4,5, now I'm trying 5,5,0,1 with asm 1,1,0,1 so the 3rd card is completely turned down, doing 210H/s (way worse than V11), the rest do, 290 290 270, this is barely faster but somehow one card simply dies with this new miner and another can only go with low setting, using 40GB virtual memory, 4GB system ram, tried the cpu with and without OC, no difference, wiped and reinstalled drivers no difference, meanwhile v11.1 flawless so it's not a physical loose plug or something.

Schwacke

member

Activity: 64

Merit: 10

Quote from: AzzAz on February 17, 2017, 05:56:42 AM

Quote from: Schwacke on February 17, 2017, 05:19:48 AM

Quote from: Mitja1998 on February 17, 2017, 01:58:16 AM

Quote from: Schwacke on February 16, 2017, 06:16:41 PM

Quote from: Mitja1998 on February 16, 2017, 04:30:56 PM

Quote from: Schwacke on February 16, 2017, 04:15:04 PM

Quote from: Mitja1998 on February 16, 2017, 04:02:57 PM

modded rx 470 -285 h/s (80W in Afterburner), modded rx480 - 305 h/s (86W in Afterburner), modded HD7950 - 245 h/s (around 110w I think). Very nice results Claymore, you made equihash most profitable algo at nicehash again.

would you pls post your settings?

BR

rx470 4gb: one has 1256 mhz and other has 1260 mhz stock clocks, mem clock is 1750 mhz. Voltage is 985 mV.
rx480 8gb: core clock is 1266 mhz stock and mem clock is 2000 mhz. Voltage is 961 mV.
hd7950 3gb: core clock is 900 mhz, mem clock is 1250 mhz all stock. Voltage is 956 mV.

Thanks alot. but one thing i dont get i was searching and searching and then i gaved up.
In Afterburner the Core Voltage (mV) is set to 0. what do i change there? i can change from -100 to +100.

BR

You change mVolts. If your card is running on 1120 mV and you change it to -100 then you get 1020 mV. But I couldn't lower it more so I changed it in gpu's bios to 985 mV

is it save to change it to -100? sry for the noob question but i have no clue of undervolting and what will happen if i undervolt to much.

BR

depends of the card, some will hang.Some may work @ -100 even.

ok if it only will hang its ok. i thought it will burn something maybe.

Ty

AzzAz

legendary

Activity: 1030

Merit: 1006

This:
GPU_FORCE_64BIT_PTR 1
GPU_MAX_HEAP_SIZE 100
GPU_USE_SYNC_OBJECTS 1
GPU_MAX_ALLOC_PERCENT 100
GPU_SINGLE_ALLOC_PERCENT 100
must be entered this way:
SET GPU_FORCE_64BIT_PTR=1
SET GPU_MAX_HEAP_SIZE=100
SET GPU_USE_SYNC_OBJECTS=1
SET GPU_MAX_ALLOC_PERCENT=100
SET GPU_SINGLE_ALLOC_PERCENT=100

@Claymore: can it be put into readme this way? For newcomers.

AzzAz

legendary

Activity: 1030

Merit: 1006

Quote from: Schwacke on February 17, 2017, 05:19:48 AM

Quote from: Mitja1998 on February 17, 2017, 01:58:16 AM

Quote from: Schwacke on February 16, 2017, 06:16:41 PM

Quote from: Mitja1998 on February 16, 2017, 04:30:56 PM

Quote from: Schwacke on February 16, 2017, 04:15:04 PM

Quote from: Mitja1998 on February 16, 2017, 04:02:57 PM

modded rx 470 -285 h/s (80W in Afterburner), modded rx480 - 305 h/s (86W in Afterburner), modded HD7950 - 245 h/s (around 110w I think). Very nice results Claymore, you made equihash most profitable algo at nicehash again.

would you pls post your settings?

BR

rx470 4gb: one has 1256 mhz and other has 1260 mhz stock clocks, mem clock is 1750 mhz. Voltage is 985 mV.
rx480 8gb: core clock is 1266 mhz stock and mem clock is 2000 mhz. Voltage is 961 mV.
hd7950 3gb: core clock is 900 mhz, mem clock is 1250 mhz all stock. Voltage is 956 mV.

Thanks alot. but one thing i dont get i was searching and searching and then i gaved up.
In Afterburner the Core Voltage (mV) is set to 0. what do i change there? i can change from -100 to +100.

BR

You change mVolts. If your card is running on 1120 mV and you change it to -100 then you get 1020 mV. But I couldn't lower it more so I changed it in gpu's bios to 985 mV

is it save to change it to -100? sry for the noob question but i have no clue of undervolting and what will happen if i undervolt to much.

BR

depends of the card, some will hang.Some may work @ -100 even.

Redhotdog

newbie

Activity: 4

Merit: 0

Constantly getting "GPU #1 returned incorrect data!" on v12.0
v11.1 works fine.
Virtual memory size = 16Gb

SET GPU_FORCE_64BIT_PTR 1
SET GPU_MAX_HEAP_SIZE 100
SET GPU_USE_SYNC_OBJECTS 1
SET GPU_MAX_ALLOC_PERCENT 100
SET GPU_SINGLE_ALLOC_PERCENT 100

are set

UPD. In my case they are set via SETX but I think it doesn't matter.

Schwacke

member

Activity: 64

Merit: 10

Quote from: Mitja1998 on February 17, 2017, 01:58:16 AM

Quote from: Schwacke on February 16, 2017, 06:16:41 PM

Quote from: Mitja1998 on February 16, 2017, 04:30:56 PM

Quote from: Schwacke on February 16, 2017, 04:15:04 PM

Quote from: Mitja1998 on February 16, 2017, 04:02:57 PM

modded rx 470 -285 h/s (80W in Afterburner), modded rx480 - 305 h/s (86W in Afterburner), modded HD7950 - 245 h/s (around 110w I think). Very nice results Claymore, you made equihash most profitable algo at nicehash again.

would you pls post your settings?

BR

rx470 4gb: one has 1256 mhz and other has 1260 mhz stock clocks, mem clock is 1750 mhz. Voltage is 985 mV.
rx480 8gb: core clock is 1266 mhz stock and mem clock is 2000 mhz. Voltage is 961 mV.
hd7950 3gb: core clock is 900 mhz, mem clock is 1250 mhz all stock. Voltage is 956 mV.

Thanks alot. but one thing i dont get i was searching and searching and then i gaved up.
In Afterburner the Core Voltage (mV) is set to 0. what do i change there? i can change from -100 to +100.

BR

You change mVolts. If your card is running on 1120 mV and you change it to -100 then you get 1020 mV. But I couldn't lower it more so I changed it in gpu's bios to 985 mV

is it save to change it to -100? sry for the noob question but i have no clue of undervolting and what will happen if i undervolt to much.

BR

Schwacke

member

Activity: 64

Merit: 10

Quote from: Lister on February 16, 2017, 11:03:46 PM

modded rx 470 -290 h/s

do u have a link for the mod pls.

Thank you.

golden0707

full member

Activity: 189

Merit: 100

Quote from: theoulis on February 17, 2017, 04:56:54 AM

very nice job with V12.
but now 280x with the 3GB can accept only -i5. 290 with 4GB can accept -i6. gpu mem starts to be a bottleneck!

What kind of CPU you have ?

Topic: Claymore's ZCash/BTG AMD GPU Miner v12.6 (Windows/Linux) - page 152. (Read 3839201 times)