BkkCoins
|
|
September 01, 2012, 03:36:36 AM |
|
I have a minor problem here that maybe someone knows how to fix up. I'm still using 2.6.1 on Linux with 3x 5830 cards. I have 6 threads, 3 x2 each gpu. For the last few weeks thread 1 always shows as "paused". I notice on reboot it will log "enabled", "disabled", "enabled" a few times but after it stabilizes the result on the GPU info screen is that Thread 1 is Alive, paused. So it works ok with thread 0 doing all the work but why? Rebooting doesn't fix it. No settings in conf that indicate it should be paused. Here is the log of startup where this business goes on: Pool 0 http://pool.50btc.com:8332 alive [2012-09-01 10:28:03] Disabling extra threads due to dynamic mode. [2012-09-01 10:28:03] Tune dynamic intensity with --gpu-dyninterval [2012-09-01 10:28:04] Thread 1 being disabled [2012-09-01 10:28:04] Thread 1 being re-enabled [2012-09-01 10:28:04] Pool 1 http://mine2.btcguild.com:8332 alive [2012-09-01 10:28:05] Thread 1 being disabled [2012-09-01 10:28:05] Thread 1 being re-enabled [2012-09-01 10:28:05] Accepted 5d93fd77.124c0eca GPU 2 pool 0 [2012-09-01 10:28:05] Pool 2 http://de.btcguild.com:8332 alive [2012-09-01 10:28:06] Thread 1 being disabled [2012-09-01 10:28:08] API running in IP access mode [2012-09-01 10:28:12] Accepted 8cc6641d.ad99151c GPU 0 pool 0
After this it doesn't change and thread 1 shows as: Last initialised: [2012-09-01 10:28:01] Intensity: 8 Thread 0: 317.0 Mh/s Enabled ALIVE Thread 1: 0.0 Mh/s Enabled ALIVE paused
Any ideas?
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
September 01, 2012, 03:43:50 AM |
|
I have a minor problem here that maybe someone knows how to fix up. I'm still using 2.6.1 on Linux with 3x 5830 cards. I have 6 threads, 3 x2 each gpu. For the last few weeks thread 1 always shows as "paused". I notice on reboot it will log "enabled", "disabled", "enabled" a few times but after it stabilizes the result on the GPU info screen is that Thread 1 is Alive, paused. So it works ok with thread 0 doing all the work but why? Rebooting doesn't fix it. No settings in conf that indicate it should be paused. Here is the log of startup where this business goes on: [2012-09-01 10:28:03] Disabling extra threads due to dynamic mode.After this it doesn't change and thread 1 shows as: Last initialised: [2012-09-01 10:28:01] Intensity: 8 Thread 0: 317.0 Mh/s Enabled ALIVE Thread 1: 0.0 Mh/s Enabled ALIVE paused
Any ideas? Set an intensity
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
os2sam
Legendary
Offline
Activity: 3586
Merit: 1098
Think for yourself
|
|
September 01, 2012, 04:21:12 AM |
|
Some questions about Dynamic Intensity.
1. Did Dynamic always disable a thread?
2. Why does it disable a thread?
3. Why does setting GPU 1 to Dynamic disable a thread on GPU 0?
Thanks, Sam
|
A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing on usenet and in e-mail?
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
September 01, 2012, 04:28:44 AM |
|
Some questions about Dynamic Intensity.
1. Did Dynamic always disable a thread?
2. Why does it disable a thread?
3. Why does setting GPU 1 to Dynamic disable a thread on GPU 0?
Thanks, Sam
1. no 2. because the dyn interval timing means something when only 1 thread is running otherwise they overlap ... if you want high throughput increase the interval 3. it shouldn't...
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
BkkCoins
|
|
September 01, 2012, 06:05:00 AM |
|
I have a minor problem here that maybe someone knows how to fix up. I'm still using 2.6.1 on Linux with 3x 5830 cards. Any ideas?
Set an intensity That's odd. I have intensity 8 set in the conf file. "intensity" : "8,8", "gpu-engine" : "1000,1000", "gpu-fan" : "40-85,40-85,40-85", "gpu-memclock" : "250,250", "gpu-powertune" : "0,0", "gpu-vddc" : "1.161,1.161", "temp-cutoff" : "95,95,95", "temp-overheat" : "85,85,85", "temp-target" : "72,72,72",
But you're right of course. Manually setting intensity 8 after starting made the paused thread go live again. I do have a script that uses the API to alter settings on GPU #2 after startup because it will crash on startup if I put the same settings in the conf file. See conf file has missing arguments for #2. For some reason maybe that API tweaking is causing GPU #0 to think it's intensity is dynamic...? Or not. IDK.
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
September 01, 2012, 06:22:09 AM |
|
I have a minor problem here that maybe someone knows how to fix up. I'm still using 2.6.1 on Linux with 3x 5830 cards. Any ideas?
Set an intensity That's odd. I have intensity 8 set in the conf file. "intensity" : "8,8", "gpu-engine" : "1000,1000", "gpu-fan" : "40-85,40-85,40-85", "gpu-memclock" : "250,250", "gpu-powertune" : "0,0", "gpu-vddc" : "1.161,1.161", "temp-cutoff" : "95,95,95", "temp-overheat" : "85,85,85", "temp-target" : "72,72,72",
But you're right of course. Manually setting intensity 8 after starting made the paused thread go live again. I do have a script that uses the API to alter settings on GPU #2 after startup because it will crash on startup if I put the same settings in the conf file. See conf file has missing arguments for #2. For some reason maybe that API tweaking is causing GPU #0 to think it's intensity is dynamic...? Or not. IDK. I bet you're loading *2* configuration files...
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
BkkCoins
|
|
September 01, 2012, 06:27:21 AM |
|
I bet you're loading *2* configuration files...
I just start it as: cgminer 2>>/var/log/cgminer.log letting it load the default in .cgminer/cgminer.conf Excerpt from last reboot, [2012-09-01 10:27:57] Started cgminer 2.6.1 [2012-09-01 10:27:57] Loaded configuration file /home/miner/.cgminer/cgminer.conf [2012-09-01 10:27:58] Probing for an alive pool
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
September 01, 2012, 06:57:43 AM |
|
Your log shows 3 GPU's but you are only setting some of the settings for 2 of them ... "intensity" : "8,8",
|
|
|
|
BkkCoins
|
|
September 01, 2012, 07:22:30 AM Last edit: September 01, 2012, 12:00:34 PM by BkkCoins |
|
Your log shows 3 GPU's but you are only setting some of the settings for 2 of them ... "intensity" : "8,8",
Yes. As mentioned above I have a problem with GPU#2 not accepting any start up settings (even very conservative ones). It's fussy and will crash if I add the third intensity,engine,memory values. I have a small script that uses the API to set that GPU after startup. eg. #!/bin/bash
export AMDAPPSDKROOT=/home/miner/AMD-APP-SDK-v2.4-lnx32/ export LD_LIBRARY_PATH=${AMDAPPSDKROOT}lib/x86:${LD_LIBRARY_PATH} export DISPLAY=:0
fixcg& cgminer 2>>/var/log/cgminer.log
and fixcg #!/bin/bash sleep 30 echo -n "gpuintensity|2,8" | nc 127.0.0.1 4028 >/dev/null echo -n "gpumem|2,250" | nc 127.0.0.1 4028 >/dev/null echo -n "gpuengine|2,900" | nc 127.0.0.1 4028 >/dev/null
That works and 30 seconds after starting GPU#2 gets fixed and runs fine. But apparently a side effect may be my GPU#0 intensity going dynamic - though I haven't rebooted since manually setting it back to 8, so maybe that's gone away.
|
|
|
|
Askit2
|
|
September 01, 2012, 11:26:24 AM |
|
When It acts up it does this
Proof of work result false booo block info BFL waited (time proper varies a little) Line seems to appear when no nonce is found so not if it says rejected. Pushing work to staged thread sucessfully rolled work Repeat
How many repeats depends. If the work was from the previous block would it come in about 5 second submissions? I think no but only because I assume it would try to submit more then one every 5 seconds to the pool to get credit for stales. What I see will repeat until it stops rolling work (at 120 seconds and 5 seconds per work unit that should leave me with 24 possible work units to get rejected before it runs out of work). The pool is disabled at 13 work units or so being rejected. If the units are truely stales that I am submitting would it count expected stales into the pools rejected work in a row? My thought is not as likely. Maybe it is but somewhere in the 70 or so seconds between LP notice and pool change because of excessive rejects I should submit some work from the current pools LP that my single has been working on. If what I am thinking is accurate I am asking for help from the wrong place and really should ask at EclipseMC thread about it. If I am wrong in my thinking I am sure someone will point out what I am missing.
PS: Did anyone else asking for help on this thread get a PM from lukeJR?
|
|
|
|
sharky112065
|
|
September 01, 2012, 12:48:53 PM |
|
PS: Did anyone else asking for help on this thread get a PM from lukeJR?
As a religious zealot, it is in his nature to try to convert people.
|
Donations welcome: 12KaKtrK52iQjPdtsJq7fJ7smC32tXWbWr
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
September 01, 2012, 12:50:46 PM |
|
PS: Did anyone else asking for help on this thread get a PM from lukeJR?
As a religious zealot, it is in his nature to try to convert people. Come to the dark side ... we have cookies
|
|
|
|
TheHarbinger
Sr. Member
Offline
Activity: 378
Merit: 250
Why is it so damn hot in here?
|
|
September 01, 2012, 01:04:54 PM |
|
2.7.5 test update:
Has been running for about 6 hours now and no sign of the -10 intensity bug so far. This is a good sign, it usually showed up within 30 minutes of CGminer starting previously.
Running one card at intensity 8, and one on dynamic, dynamic card is hooked to a monitor. Test rig is a WIN7 x64 with 2 5870s not in crossfire, it is a daily use desktop, not a dedicated mining rig. Connecting to a LAN P2Pool server.
I will let it run for a full 24 hours and do a follow up post. So far it looks good Con, here's hoping that bug is squashed for good this time.
And now for something completely different..... Some good news, 24+ hours running in the background, with the system going under all sorts of other CPU and GPU stresses, and still no sign of the -10 intensity issue. I'm getting a little bit excited.
|
12Um6jfDE7q6crm1s6tSksMvda8s1hZ3Vj
|
|
|
os2sam
Legendary
Offline
Activity: 3586
Merit: 1098
Think for yourself
|
|
September 01, 2012, 02:02:14 PM |
|
PS: Did anyone else asking for help on this thread get a PM from lukeJR?
As a religious zealot, it is in his nature to try to convert people. Come to the dark side ... we have cookies We are all on the dark side ... cookies optional.
|
A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing on usenet and in e-mail?
|
|
|
jjshabadoo
|
|
September 01, 2012, 03:00:11 PM |
|
Thanks for this work Con and Kano. I just finally went from 2.3.2 windows to 2.7.4 as I was wondering about waiting to work out any bugs in 2.7.5.
I have almost all 5870's and 5970's with one 7970 rig.
Has anyone figured the optimal settings for 5xxx series under Cons new kernel ? Should I change worksize, etc. ? How about SDK, I am still using driver 11.12 and SDK 2.1 for all of my 5xxx series cards and using 12.4 for my 7970's.
Any help would be aprreciated.
|
|
|
|
sharky112065
|
|
September 01, 2012, 06:21:02 PM |
|
Thanks for this work Con and Kano. I just finally went from 2.3.2 windows to 2.7.4 as I was wondering about waiting to work out any bugs in 2.7.5.
I have almost all 5870's and 5970's with one 7970 rig.
Has anyone figured the optimal settings for 5xxx series under Cons new kernel ? Should I change worksize, etc. ? How about SDK, I am still using driver 11.12 and SDK 2.1 for all of my 5xxx series cards and using 12.4 for my 7970's.
Any help would be aprreciated.
I am running 11.9 and SDK 2.5 (Windows 7) for my 5870's. Going from 2.7.0 to 2.7.5, I had to change mem from 170 to 300 and set work size to 256 to get the same hash rate. At 970/300 -w 256 I get 445.6 MH per 5870. Fan is at 75% and temp is 69 to 74 depending on how hot it is outside. I use SDK 2.5 because I also have one 6970 in that rig. I did not have to raise memory or change work size for the 6970 in that rig to maintain the same hash rate as before (960/170 - 430.3 MH) Yes, I know it has the CPU bug, but I use affinity to only use one core. I have tried other driver versions but loose hash rate and have worse stability.
|
Donations welcome: 12KaKtrK52iQjPdtsJq7fJ7smC32tXWbWr
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
September 01, 2012, 06:35:36 PM Last edit: September 01, 2012, 07:45:50 PM by crazyates |
|
Thanks for this work Con and Kano. I just finally went from 2.3.2 windows to 2.7.4 as I was wondering about waiting to work out any bugs in 2.7.5.
I have almost all 5870's and 5970's with one 7970 rig.
Has anyone figured the optimal settings for 5xxx series under Cons new kernel ? Should I change worksize, etc. ? How about SDK, I am still using driver 11.12 and SDK 2.1 for all of my 5xxx series cards and using 12.4 for my 7970's.
Any help would be aprreciated.
I am running 11.9 and SDK 2.5 (Windows 7) for my 5870's. Going from 2.7.0 to 2.7.5, I had to change mem from 170 to 300 and set work size to 256 to get the same hash rate. At 970/300 -w 256 I get 445.6 MH per 5870. Fan is at 75% and temp is 69 to 74 depending on how hot it is outside. I use SDK 2.5 because I also have one 6970 in that rig. I did not have to raise memory or change work size for the 6970 in that rig to maintain the same hash rate as before (960/170 - 430.3 MH) Yes, I know it has the CPU bug, but I use affinity to only use one core. I have tried other driver versions but loose hash rate and have worse stability. I thought 11.11 got rid of the cpu bug, and still had excellent performance? Edit: fixed.
|
|
|
|
LegitBit
|
|
September 01, 2012, 07:14:54 PM |
|
Any news on a Mac Binary?
|
|
|
|
sharky112065
|
|
September 01, 2012, 08:46:21 PM |
|
Thanks for this work Con and Kano. I just finally went from 2.3.2 windows to 2.7.4 as I was wondering about waiting to work out any bugs in 2.7.5.
I have almost all 5870's and 5970's with one 7970 rig.
Has anyone figured the optimal settings for 5xxx series under Cons new kernel ? Should I change worksize, etc. ? How about SDK, I am still using driver 11.12 and SDK 2.1 for all of my 5xxx series cards and using 12.4 for my 7970's.
Any help would be aprreciated.
I am running 11.9 and SDK 2.5 (Windows 7) for my 5870's. Going from 2.7.0 to 2.7.5, I had to change mem from 170 to 300 and set work size to 256 to get the same hash rate. At 970/300 -w 256 I get 445.6 MH per 5870. Fan is at 75% and temp is 69 to 74 depending on how hot it is outside. I use SDK 2.5 because I also have one 6970 in that rig. I did not have to raise memory or change work size for the 6970 in that rig to maintain the same hash rate as before (960/170 - 430.3 MH) Yes, I know it has the CPU bug, but I use affinity to only use one core. I have tried other driver versions but loose hash rate and have worse stability. I thought 11.11 got rid of the cpu bug, and still had excellent performance? Edit: fixed. I was getting a lot of driver not responding messages and pool x not providing work fast enough with other versions. Do what works for you, but the hours I wasted on testing other SDK's and other drivers was nothing but trouble. If it ain't broke, don't fix it.
|
Donations welcome: 12KaKtrK52iQjPdtsJq7fJ7smC32tXWbWr
|
|
|
crazyates
Legendary
Offline
Activity: 952
Merit: 1000
|
|
September 01, 2012, 08:51:12 PM |
|
Thanks for this work Con and Kano. I just finally went from 2.3.2 windows to 2.7.4 as I was wondering about waiting to work out any bugs in 2.7.5.
I have almost all 5870's and 5970's with one 7970 rig.
Has anyone figured the optimal settings for 5xxx series under Cons new kernel ? Should I change worksize, etc. ? How about SDK, I am still using driver 11.12 and SDK 2.1 for all of my 5xxx series cards and using 12.4 for my 7970's.
Any help would be aprreciated.
I am running 11.9 and SDK 2.5 (Windows 7) for my 5870's. Going from 2.7.0 to 2.7.5, I had to change mem from 170 to 300 and set work size to 256 to get the same hash rate. At 970/300 -w 256 I get 445.6 MH per 5870. Fan is at 75% and temp is 69 to 74 depending on how hot it is outside. I use SDK 2.5 because I also have one 6970 in that rig. I did not have to raise memory or change work size for the 6970 in that rig to maintain the same hash rate as before (960/170 - 430.3 MH) Yes, I know it has the CPU bug, but I use affinity to only use one core. I have tried other driver versions but loose hash rate and have worse stability. I thought 11.11 got rid of the cpu bug, and still had excellent performance? Edit: fixed. I was getting a lot of driver not responding messages and pool x not providing work fast enough with other versions. Do what works for you, but the hours I wasted on testing other SDK's and other drivers was nothing but trouble. If it ain't broke, don't fix it. Haha sounds fair. I can't stand not having the latest version of any piece of software, so Im always updating everything.
|
|
|
|
|