But 75 is the target temperature, not 85.
Again, if you want a higher target, then set a higher target.
Also rather than reading it to mean what you want, read a little further on:
If fan speed control is not available or already optimal, then GPU clock speed
is only decreased if it goes over the target temperature by the hysteresis
amount, which is set to 3 by default
--temp-hysteresis
If the temperature drops below the target temperature, and engine clock speed
is not at the highest level set at startup, cgminer will raise the clock speed.
Topic: OFFICIAL CGMINER mining software thread for linux/win/osx/mips/arm/r-pi 4.11.0 - page 771. (Read 5805728 times)
September 22, 2011, 12:55:56 AM
September 22, 2011, 12:48:12 AM
Hmm - you want a higher target then set a higher target ... ?
I don't want a higher target.... I want it to do what the readme says it should do.....
It should maintain the highest clock speeds possible (in this case 900mhz for GPU3) until that GPU hits 85 degrees, then it'll start to throttle it back. This isn't happening.
I'm not sure you're grasping this so I'll just wait for con to chime in. Perhaps the readme is just worded incorrectly, or perhaps there's a bug. I'm trying to clarify that.
To your EDIT, I'm not concerned about that. If the temps get that high then I'm happy for it to be throttled back until I can get to it. My issue is that it's throttling it too early.
September 22, 2011, 12:45:22 AM
Hmm - you want a higher target then set a higher target ... ?
Edit: also again that makes no sense.
If it was to put it back up as soon as it gets under 85, then it would go straight back over 85 ... you'd end up jumping up and down OVER the overheat temperature, not the requested temperature.
Edit: also again that makes no sense.
If it was to put it back up as soon as it gets under 85, then it would go straight back over 85 ... you'd end up jumping up and down OVER the overheat temperature, not the requested temperature.
September 22, 2011, 12:39:18 AM
That request makes no sense ...
It would seem illogical to increase the GPU Mhz (or decrease the fan RPM) while the temp is above the requested temperature ... no?
The default requested temp is 75 - while it stays above that, it obviously wouldn't increase the GPU Mhz ...
It would seem illogical to increase the GPU Mhz (or decrease the fan RPM) while the temp is above the requested temperature ... no?
The default requested temp is 75 - while it stays above that, it obviously wouldn't increase the GPU Mhz ...
Welllllll, the readme says:
"GPU control in auto gpu tries to maintain as high a clock speed as possible
while not reaching overheat temperatures."
It's not doing this, strictly speaking
It's not reaching "overheat" temps (classed as 85 degrees by default right?), so it should be keeping the GPU clock as high as possible until overheat is reached, THEN it should throttle it.Going by that, it shouldn't have throttled the GPU at all, as the temps never reached overheat.
September 22, 2011, 12:28:46 AM
That request makes no sense ...
It would seem illogical to increase the GPU Mhz (or decrease the fan RPM) while the temp is above the requested temperature ... no?
The default requested temp is 75 - while it stays above that, it obviously wouldn't increase the GPU Mhz ...
It would seem illogical to increase the GPU Mhz (or decrease the fan RPM) while the temp is above the requested temperature ... no?
The default requested temp is 75 - while it stays above that, it obviously wouldn't increase the GPU Mhz ...
September 22, 2011, 12:17:16 AM
Hmm. So I have 4 GPUs at the mo. Here's my bat file:
start /B /affinity 8 cgminer -o http://127.0.0.1:8337 -u simonk83 -p x -I 9 --auto-fan --auto-gpu --gpu-engine 930,930,900,930 --gpu-memclock 305 2>"%date:~4,2%-%date:~7,2%-%date:~12,2%%time:~0,2%-%time:~3,2%-%time:~6,2%.log"
That 3rd GPU I set to 900 as it runs really hot (stock is 850). Everything has been fine up until now as the temps have been cool in my house, but now that things are starting to heat up I've noticed an issue. This card is hitting over 80 occasionally, and I've now just noticed that cgminer has throttled it back to 850. I was under the impression though that it'd scale it back up to 900 once the temps dropped under the overheat temp (which is 85 isn't it), but it's not doing that. It just stays at 850mhz.
Is it waiting for the temps to drop to (or under) 75 before it scales the clocks back up? If that's the case, it's probably not going to happen Is it possible to change it so that it scales the clocks back up when it drops under the overheat threshold (85)? I would have thought that's what it's supposed to be doing, but it's not (in my case at least).
Ta
EDIT: It's done it to GPU 2 as well. Currently at 885mhz at 78 degrees. Should be at 930.
EDIT2: Just restarted, and it seems to start throttling the GPU at 78 degrees rather than 85.... I don't think this happened with 2.0.2 or earlier....
start /B /affinity 8 cgminer -o http://127.0.0.1:8337 -u simonk83 -p x -I 9 --auto-fan --auto-gpu --gpu-engine 930,930,900,930 --gpu-memclock 305 2>"%date:~4,2%-%date:~7,2%-%date:~12,2%%time:~0,2%-%time:~3,2%-%time:~6,2%.log"
That 3rd GPU I set to 900 as it runs really hot (stock is 850). Everything has been fine up until now as the temps have been cool in my house, but now that things are starting to heat up I've noticed an issue. This card is hitting over 80 occasionally, and I've now just noticed that cgminer has throttled it back to 850. I was under the impression though that it'd scale it back up to 900 once the temps dropped under the overheat temp (which is 85 isn't it), but it's not doing that. It just stays at 850mhz.
Is it waiting for the temps to drop to (or under) 75 before it scales the clocks back up? If that's the case, it's probably not going to happen
Is it possible to change it so that it scales the clocks back up when it drops under the overheat threshold (85)? I would have thought that's what it's supposed to be doing, but it's not (in my case at least).Ta
EDIT: It's done it to GPU 2 as well. Currently at 885mhz at 78 degrees. Should be at 930.
EDIT2: Just restarted, and it seems to start throttling the GPU at 78 degrees rather than 85.... I don't think this happened with 2.0.2 or earlier....
September 21, 2011, 10:15:46 PM
Just a little aside information that some may not realise ...
The value shown after Accept/Reject in the cgminer log is the hex nonce.
If you get a block (pool or solo), then the nonce in the block will match that value exactly.
(this is firstly based on reading the code and working out where the value comes from, and secondly looking at my one and only pool block I have and seeing that it does match as expected)
This means that anyone wondering if they can tell if they got a particular block, they can with reasonable certainty by comparing the block nonce with the log nonce (and of course there should be a LONGPOLL for the block almost straight after it in the log)
Of course it isn't 100% guaranteed since the nonce value is not unique per share, but if it is roughly at the same time as the block and a LONGPOLL shows just after it - it's extremely likely to be the block.
... and another side
My latest pull request makes it show 3/8 of the actual hash rather than the nonce and also says " BLOCK!" after it if it is a valid block difficulty share
(which also may cause some distress if anyone ever actually gets a Rejected block in the log ...)
If you want that branch change before it gets into the main code it's available in my kano branch here:
https://github.com/kanoi/cgminer/tree/kano (main.c, util.c and miner.h)
N.B. that difficulty test code is my original in case anyone wonders where it came from
The value shown after Accept/Reject in the cgminer log is the hex nonce.
If you get a block (pool or solo), then the nonce in the block will match that value exactly.
(this is firstly based on reading the code and working out where the value comes from, and secondly looking at my one and only pool block I have and seeing that it does match as expected)
This means that anyone wondering if they can tell if they got a particular block, they can with reasonable certainty by comparing the block nonce with the log nonce (and of course there should be a LONGPOLL for the block almost straight after it in the log)
Of course it isn't 100% guaranteed since the nonce value is not unique per share, but if it is roughly at the same time as the block and a LONGPOLL shows just after it - it's extremely likely to be the block.
... and another side
My latest pull request makes it show 3/8 of the actual hash rather than the nonce and also says " BLOCK!" after it if it is a valid block difficulty share
(which also may cause some distress if anyone ever actually gets a Rejected block in the log ...)
If you want that branch change before it gets into the main code it's available in my kano branch here:
https://github.com/kanoi/cgminer/tree/kano (main.c, util.c and miner.h)
N.B. that difficulty test code is my original in case anyone wonders where it came from
September 21, 2011, 04:47:05 AM
DISPLAY=:0 cgminer --ndevs
Thanks! That did the trick. September 20, 2011, 10:00:50 PM
You cannot query it inside cgminer AFAIK, ls *.bin and the bin file well indicate what it's running.
September 20, 2011, 09:56:22 PM
I am trying to use the config file, but ran into 3 config switches which refuse to work.
All of the following auto-fan/gpu permutations fail:
Trying to select for a single device is troublesome:
Results in GPU 0 and 2 being disabled.
Results in GPU 0 and 1 being disabled.
Results in GPU 0 and 1 being disabled.
Can you clarify the syntax for these commands?
Also, can you query for what the automatically detected values for vectors and worksize?
All of the following auto-fan/gpu permutations fail:
Code:
"auto-fan" : "true",
Code:
"auto-fan" : "1",
Code:
"auto-fan" : "",
Trying to select for a single device is troublesome:
Code:
"device" : "1",
Code:
"device" : "2",
Code:
"device" : "2",
"device" : "1",
Can you clarify the syntax for these commands?
Also, can you query for what the automatically detected values for vectors and worksize?
September 20, 2011, 09:49:04 PM
DISPLAY=:0 cgminer --ndevs
September 20, 2011, 08:27:03 PM
I can't get cgminer to work from within Linuxcoin. I get the following message when I try to run it:
Phoenix miner detects my GPU and can mine without problems:
This is what aticonfig has to say:
Any ideas on what could be wrong?
Code:
user@linuxcoin:~/Desktop/cgminer-2.0.3$ cgminer --ndevs
[2011-09-21 01:20:58] Error: Getting Device IDs (num)
[2011-09-21 01:20:58] clDevicesNum returned error, none usable
0 GPU devices detected
Phoenix miner detects my GPU and can mine without problems:
Code:
[21/09/2011 01:15:20] Phoenix r116 starting...
[21/09/2011 01:15:20] Connected to server
[21/09/2011 01:15:44] Result: 12c54cbf accepted
[21/09/2011 01:15:49] Result: bb7dccd0 accepted
[21/09/2011 01:16:52] Result: 1f03481a accepted
[21/09/2011 01:16:53] Result: a388f657 accepted
[21/09/2011 01:16:53] Result: 1c510d1c accepted
[217.05 Mhash/sec] [20 Accepted] [0 Rejected] [RPC (+LP)]
This is what aticonfig has to say:
Code:
user@linuxcoin:~/Desktop/cgminer-2.0.3$ DISPLAY=:0 sudo aticonfig --list-adapters
* 0. 01:00.0 ATI Radeon HD 5700 Series
* - Default adapter
Any ideas on what could be wrong?
September 20, 2011, 03:26:10 PM
ahh I did not know that. Thank you! Now I feel like an idiot!
nhehehe thanks again.
nhehehe thanks again.
September 20, 2011, 02:48:13 PM
When solo mining, all shares are rejected, except when you solve a block. This is normal.
September 20, 2011, 02:33:15 PM
I thought for the heck of it I would have my bitcoin wallet accept connections and then set it up as solo mining and only use it as my failover. the idea is cgminer would slip it a few shares every now and then and maybe some day I would get lucky and hit the jackpot (im a gambling man)
I managed to get cgminer to connect to my bitcoin wallet but it says REJECTED for every single share in cgminer
any ideas folks?
thanks
I managed to get cgminer to connect to my bitcoin wallet but it says REJECTED for every single share in cgminer
any ideas folks?
thanks
September 20, 2011, 09:36:44 AM
Stupid Question:
I have an HD6850 running by itself. If I add a second card, say an HD6750, and fire CGMiner back up, will it detect both cards, or do I need to add something to the .bat file? I am using Catalyst 11.8 drivers.
I have an HD6850 running by itself. If I add a second card, say an HD6750, and fire CGMiner back up, will it detect both cards, or do I need to add something to the .bat file? I am using Catalyst 11.8 drivers.
When I added my second card, a 5770, the Catalyst drivers picked it up automatically.
I had to use a dummy plug to simulate a monitor and extend the screen to use it since I was using Win7 and older Catalyst drivers.
But yes the miner saw it and started using it automatically.
Sam
September 20, 2011, 09:20:40 AM
Stupid Question:
I have an HD6850 running by itself. If I add a second card, say an HD6750, and fire CGMiner back up, will it detect both cards, or do I need to add something to the .bat file? I am using Catalyst 11.8 drivers.
I have an HD6850 running by itself. If I add a second card, say an HD6750, and fire CGMiner back up, will it detect both cards, or do I need to add something to the .bat file? I am using Catalyst 11.8 drivers.
September 20, 2011, 07:52:11 AM
I realised you dont need anything fancy just to test 5 different memory speeds, cut and paste is sufficient.
Short/long neat/messy efficient or not doesnt matter for this test, it's a once off thing.
Short/long neat/messy efficient or not doesnt matter for this test, it's a once off thing.
September 20, 2011, 07:45:12 AM
Don't know the *nix equivalent for a bash loop would be, but seems really inefficient to copy-paste the setup over and over, and even then after what 6-8hrs it'd never restart again? I'd think something like
would be wayyyy better and shorter code :]
Code:
:myloop
DISPLAY=:0 ./cgminer -c p1.json -c p2.json & 2>> log
sleep 1h
killall cgminer
goto myloop
would be wayyyy better and shorter code :]
September 20, 2011, 04:54:28 AM
Something like this then?
Though not sure what 2> will actually log. I would prefer to only capture the final summarized output.
Code:
#!/bin/bash
cd ~/cgminer
DISPLAY=:0 ./cgminer -c p1.json -c p2.json & 2>> log
sleep 1h
killall cgminer
DISPLAY=:0 ./cgminer -c p1.json -c p2.json --gpu-memclock 300 & 2>> log
sleep 1h
killall cgminer
DISPLAY=:0 ./cgminer -c p1.json -c p2.json --gpu-memclock 250 & 2>> log
sleep 1h
killall cgminer
DISPLAY=:0 ./cgminer -c p1.json -c p2.json --gpu-memclock 200 & 2>> log
sleep 1h
killall cgminer
DISPLAY=:0 ./cgminer -c p1.json -c p2.json --gpu-memclock 150 & 2>> log
sleep 1h
killall cgminer
DISPLAY=:0 ./cgminer -c p1.json -c p2.json --gpu-memclock 100 & 2>> log
sleep 1h
killall cgminer
DISPLAY=:0 ./cgminer -c p1.json -c p2.json --gpu-memclock 50 & 2>> log
sleep 1h
killall cgminer
Though not sure what 2> will actually log. I would prefer to only capture the final summarized output.
Jump to: