Installing AMD Drivers on BAMT?

[philips]

I guess I did install 21 updates.
If I go to /opt/bamt/fix.history it has a list like this.
bamt5_fix_3.tar,1330825106,install
And it goes from 0 to 21 minus #20.  I have read about #20 being messed up so I guess thats why it is not included.
So I guess my problem is not the updates?  

The fixes are fine then.
Have you configured BAMT to use cgminer in bamt.conf with all the necessary settings as described in BAMT thread?

[MrBilling]

The BAMT thread you are talking about is https://bitcointalk.org/index.php?topic=65915.0?

For some reason the GPU fault and red screen were gone when I rebooted.  Everything looked normal but it was not mining.  Then after a few mins it goes back to the red GPU fault screen.  And when I reboot it it says no GPUs are configured.


I set my pool in the pool config.


I have 2 GPUs in right now.  I set the BAMT config the way it should be pretty sure.

When I click restart mining sessions it says.

cgminer api failed



So All I had to do was copy the cgminer folder over and delete the older one to update cgminer?  I dont have to do anything else?


I have not found a guide here to setting the config files.  But they are pretty self explanatory when you read the directions in the configs.


I must be doing something wrong.


Besides setting the config files is there anything else I have to do?


Thanks

[MrBilling]

I ran cgminer just like I would from the command line in Windows and it works.  It will mine BTC and LTC.  Only getting 300-350 KH/s but I havent messed with tuning it yet.


I still have the red background saying the GPU has a fault.

[philips]

The BAMT thread you are talking about is https://bitcointalk.org/index.php?topic=65915.0?

I set my pool in the pool config

So All I had to do was copy the cgminer folder over and delete the older one to update cgminer?  I dont have to do anything else?

Thanks

Yes, that is the thread.

Set the pool in the cgminer line, not in the pool config.

Yes, all you need to do is copy the new cgminer folder in the same place where the old one was.
At the first run make sure you give cgminer plenty of time to compile its kernel for your cards. It may take a while.

Do you have the red screen when mining LTC, or BTC?
While the red screen is on open a terminal and enter the command:

Code:

screen -r cgminer

Is cgminer working?


If, as you say, cgminer is working from the command line but not when managed by BAMT then something is not configured right in your bamt.conf.

[MrBilling]

When I run that screen -r cgminer it says "There is no screen to be resumed matching cgminer."

I ran it from root terminal while at root@bamt-miner:/home/user#


Thanks for the help.  I will def throw you some LTC for all the help.

[philips]

Then my guess is that something is not configured right in bamt.conf.
Probably BAMT tries to start cgminer and something in config prevents it...missing pools?

No need to send anything to me, if you want to donate, donate to loddcrapo (BAMT creator).

[MrBilling]

Here is my bamt.config file



# Big A Miner Thing configuration

# full example of options available in /etc/bamt/bamt.conf



---

settings:

 
 
# miner info used in alerts, etc

 

# id is used to set hostname and identify this rig in farm wide displays
  miner_id: bamt-miner
 
# loc is any addition string you wish to use to identify this rig (location?)
 
miner_loc: unspecified location

 

# sending status to mgpumon..

 

# send directly to IP address (only needed if mgpumon is on different network)
 
# do_direct_status: 192.168.1.1

 

# and/or broadcast on local IP network  1 = yes, 0 = no
 
# recommend to leave enabled, harmless and very handy
 
do_bcast_status: 1

 

# central config managment? 1 = yes, 0 = no  (sensing a pattern here?)
 
do_manage_config: 0

 

# command to run for config updates, this is just one example..
 
# note this is old school, pre autoconf but still a fine way to do things if you prefer it
 
# config_update_cmd: /usr/bin/rsync -aL rsync://192.168.1.1:873/config/miner1/* /etc/bamt/


 


# how long to wait for X to start (seconds) - increase if mining fails to start at boot
 
# slower machines or booting from slow USB key may need as long as 60 seconds here
 
start_mining_init_delay: 20

 

# delay in between starting individual GPUs (seconds) - increase if gpus lock during mining init
 
start_mining_miner_delay: 3

 

# api keys for gpumon, all optional

 

# apikey_btcguild:
 
# apikey_deepbit:
 
# apikey_bitclockers:
 
# apikey_slush:
 
# apikey_bithasher:
 
# apikey_mtred:
 
# apikey_ozcoin:
 
# apikey_nofee:
 
# apikey_rfc:
 
# apikey_polmine:
 
# apikey_btcpool:
 
# apikey_triple:
 
# apikey_abc:


 

# monitoring config..
 
# should we monitor the cards? (send email alerts, etc)  1 = yes, 0 = no
 
do_monitor: 0

 

# smtp settings for sending alerts
 
 

# required..
 
# smtp_host: 127.0.0.1
 
# smtp_to: you@youremail.com
 
 

# optional..
 
# smtp_port: 587
 
# smtp_from: miner@mydomain.com
 
 

# set these to 1 to enable
 
# smtp_tls: 1
 
# smtp_ssl: 1
 
 

# authentication for sasl or tls
 
# smtp_auth_user: youruser
 
# smtp_auth_pass: yourpass
 
 

# don't send emails more often than smtp_min_wait seconds, default 300
 
# smtp_min_wait: 300
 
# include current system status details in alerts? (defaults on, disable for smaller emails)
 
# smtp_include_details: 1
 
# if you don't like the automatic subjects..
 
# smtp_subject: I wanted a custom subject in my alerts...
 
 
 


# start autoconf client - no harm in running client even if not using autoconf server on network
 
do_autoconf_client: 1
 
 

# start mgpumon in background using /etc/bamt/mgpumon.conf?
 
do_mgpumon: 0

 
 


# detect hung GPUs and auto reboot (phoenix only)
 
detect_defunct: 1


 


# use cgsnoop?
 
do_cgsnoop: 0

 
 
 

# cgminer options
 
 

# start cgminer?  set this to 1, and also set cgminer: 1 on each GPU you will manage with cgminer
 
cgminer: 1
 
 

# options to cgminer (must have at least --api-listen for any bamt tools to work)
 
# see cgminer documentation for options.  you may want to put cgminer config in
 
# another file and simply put --config /that/file here.
 
cgminer_opts: -o http://mine.pool-x.eu:8080 -u Helper556.1 -p x --intensity 16 --api-listen --gpu-powertune 5 --gpu-engine 1150 --gpu-

memclock 1000 --auto-fan --temp-target 70
 
 

# tcp port for cgminer.. defaults to 4028
 
# cgminer_port: 4028
   
 
 
 
 



# GPU Configuration

# You'll need one gpuX: section per GPU installed.  Not sure what's installed,

# or the order?  Open a root shell and type:  aticonfig --list-adapters

#

# You can delete gpu sections if you don't have that many cards..







gpu0:
 
# remove this or set to 0 to actually use this card..
 

 


# set to 1 if cgminer will manage this gpu.
 
cgminer: 1
 
 

# overclocking.. all optional 

 

# enable to debug overclocking.. provides info on screen during mine start/restart
 
# debug_oc: 1
 
 
 
# optional command to execute prior to overclocking
 
# pre_oc_cmd:

 

# val_x = the value to set for this profile (0-2)
 
# Profile 2 is used while mining, but setting values
 
# in profile 2 that are lower than the values in the other
 
# profiles is not allowed on some GPUs, so you have to adjust
 
# the other profiles as well.

 
 
# core_speed_0: 1000
 
# core_speed_1: 1050
 
# core_speed_2: 1100
 
# mem_speed_0: 1470
 
# mem_speed_1: 1544
 
# mem_speed_2: 1617
 
# core_voltage_0: 1.125
 
# core_voltage_1: 1.125
 
# core_voltage_2: 1.250

 

# fan_speed: 70

 

# optional command to execute after overclocking
 
# post_oc_cmd:

 

# mining setup

 

# kernel to use.. currently one of:  phatk phatk2 phatk-bamt04 phatk-svn phatk-1.50 poclbm
 
# kernel: phatk2

 

# phoenix style kernel params.  do not include DEVICE=x
 
# kernel_params: BFI_INT VECTORS FASTLOOP=false AGGRESSION=9

 

# optional phoenix queuesize parameter: how many work units to keep queued at all times
 
# queue_size: 2

 

# optional phoenix avgsamples parameter: how many samples to use for hashrate average
 
# avg_samples: 10

 

# IMPORTANT: pool_file must contain a list of pool URLs.. one at least, more if
 
# you like.  You can share the same file for all GPUs, or have one for each.
 
# Miner will start with first URL in file, however if no shares are accepted
 
# for any reason in (pool_timeout) seconds, the miner will move to next URL
 
# in the file.  Wraps around to first URL when last one fails.
 
# You can also set pool specific values, see the docs.

 

pool_file: /etc/bamt/pools

 

# default time to allow phoenix to run without finding a share..
 
# exceed this and we move to next pool in poolfile
 
pool_timeout: 180

 

# monitoring values, used by email alerts and gpumon
 
monitor_temp_lo: 45
 
monitor_temp_hi: 80
 
monitor_load_lo: 80
 
monitor_hash_lo: 125
 
monitor_fan_lo: 1000
 
monitor_reject_hi: 3





gpu1:
 
disabled: 0
 
cgminer: 1
 
# core_speed_0: 1000
 
# core_speed_1: 1050
 
# core_speed_2: 1100
 
# mem_speed_0: 1470
 
# mem_speed_1: 1544
 
# mem_speed_2: 1617
 
# core_voltage_0: 1.125
 
# core_voltage_1: 1.125
 
# core_voltage_2: 1.250

 

fan_speed: 100

 

kernel: phatk2
 
kernel_params: BFI_INT VECTORS FASTLOOP=false AGGRESSION=11

 

pool_file: /etc/bamt/pools
 
pool_timeout: 180

 

monitor_temp_lo: 45
 
monitor_temp_hi: 80
 
monitor_load_lo: 80
 
monitor_hash_lo: 125
 
monitor_fan_lo: 1000
 
monitor_reject_hi: 3

[MrBilling]

Then in my pool config file I just have 1 pool line where at the bottom it says:


#At the very least, change the next line to a valid worker url:
http://XXXXXXX:XXXX@mine.pool-x.eu:8080


But you say not to use this?



So how am I supposed to use BAMT?  If everything were working correctly am I just supposed to start cgminer from a terminal?


Or am I supposed to use the "Start mining sessions"?

[philips]

Before anything else you forgot to add the --scrypt option in cgminer config.
That is a pool for mining LTC.

[MrBilling]

You mean I add --scrypt in the bamt.config file I just posted?


So it would go here?

cgminer_opts: -o http://mine.pool-x.eu:8080 -u Helper556.1 -p x --intensity 16 --api-listen --gpu-powertune 5 --gpu-engine 1150 --gpu-

memclock 1000 --auto-fan --temp-target 70



So like this?


cgminer_opts: -o http://mine.pool-x.eu:8080 -u Helper556.1 -p x --intensity 16 --scrypt --api-listen --gpu-powertune 5 --gpu-engine 1150 --gpu-

memclock 1000 --auto-fan --temp-target 70

[philips]

Yes.

[MrBilling]

OK so I ran

export DISPLAY=:0
export GPU_USE_SYNC_OBJECTS=1
export GPU_MAX_ALLOC_PERCENT=100


Then I manually started cgminer using just --scrypt argument.  Then when cgminer started I manually put the URL and U:P in and now it is mining and accepting shares.

YAY!


But the KH/s is still low.


345 KH/s on I 20

[philips]

Then (for now) use the same arguments in bamt.conf and Restart mining session.
Stop the manually started cgminer first.

[MrBilling]

Sorry for all the newb stuff.


But do you mean put these in the conf?

export DISPLAY=:0
export GPU_USE_SYNC_OBJECTS=1
export GPU_MAX_ALLOC_PERCENT=100

And ifso where would I put them?

[philips]

Sorry for all the newb stuff.


But do you mean put these in the conf?

export DISPLAY=:0
export GPU_USE_SYNC_OBJECTS=1
export GPU_MAX_ALLOC_PERCENT=100

And ifso where would I put them?

No need. These are automatically executed by BAMT before running cgminer.

The line should look something like this:

Code:

cgminer_opts: --scrypt -o http://xxx -u xxx -p xxx --api-listen --auto-fan --temp-target 70

[MrBilling]

OK so I had BAMT running all night.


It runs good but low KH/s.  Like between 350-400.



This is with me just starting it and then increasing the intensity.


If I try to start it and specify the thread concurrency it always gives me this message and wont start.

Loaded config file from xxx/xxx/xxx/.cgminer
Probing for alive pool
Maximum buffer memory device 0 supports says 805306368
Your scrypt settings come to 1245184000
Error -61: clCreateBuffer (padbuffer8), decrease TC or increase LG


I tried doing it by starting cgminer with the --thread concurrency xxxxx -I 20.
And I tried putting the settings in the cgminer config file and then just starting cgminer.

It always gives me the same message.




Just for the info I have tried -I 16-20 and --thread concurrency -18000-21712.
Always same message.


So today I tried -I 20 and --thread concurrency 9000 and it is working and they start running aT 600-620 KH/s but the steadily and slowly go down down down to the 400ish range.



And idea on this?


Much thanks for all the help.

[philips]

Cant say much or anything about LTC mining, I'm mining only BTC or similar.
As long as I can still mine BTC with GPUs I'm not gonna kill my cards even for that extra profit.

Finding the right settings for LTC is a kind of a black art. All I can do is:

1 Point you to this thread with lots of good info: https://bitcointalk.org/index.php?topic=117221.0

2 Provide you with latest cgminer (3.0.0) that contains scrypt (LTC) improvements.

3 Advise you to not ever ever mine LTC with more than 3 cards without using powered risers.

Also, don't forget to read the READMEs included in cgminer´s folder.
Most of the time the answers are there, but no one gives a crap about the Readme, go figure.

As soon as I compile the cgminer (maybe tomorrow) I will notify you here.

[MrBilling]

Cool.


Yea I have the opposite problem.

I read so many readme's, threads, and other info that I get information overload and forget stuff or just can't put it together.


I kind of understand why mining LTC would be more stressful on a GPU but not fully.


As long as I am using powered risers and keep the temps below 80 C I should be ok?

I dunno if it's real or not but on my Windows 8 rig I have 2 cards running right now (7950's) using GUIMiner Scrypt and it says they are getting 810 and 860 KH/s.  Wierd, same settings.   Probably fake or something.

[philips]

Stick with what is written in cgminer´s readme and that guide I linked.

Basically LTC mining brings higher consumption and temperatures.
Cant really say what temperature is safe. I keep my cards around 60, at 80 I would shut them down immediately.
OTOH I remember one guy around here bragging about keeping his cards at 90 for months.

About hashrates, you better compare speeds as they are reported by the pool.
I remember guiminer constantly blowing up the speeds. And if you are using Reaper as miner, that one is even a bigger liar.

[MrBilling]

OK.


I had BAMT running cgminer mining LTC at 600-650 KH/s for a few hours.  Of course I turned it off so I could move it to another room.



So now when I turn it back on and run cgminer it connects and stuff but all my shares are going to HW.  Barely any are getting accepted.




So I thought that running

export DISPLAY=:0
export GPU_USE_SYNC_OBJECTS=1
export GPU_MAX_ALLOC_PRECENT=100

fixes all of that.  but I ran those and it did not.  Been running just a couple mins and I got HW: 64 and A:4.



What is all this about now?


Im really close to having this all figured out.  I was perfectly happy when it was running 600-650 KH/s.  now I just gotta figure this out.