DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
February 02, 2012, 10:23:25 PM |
|
[2012-02-02 15:02:40] clGetPlatformsIDs failed (no GPU?) [2012-02-02 15:02:40] 0 GPU devices detected This is a broken SDK installation. Come on conman how could they have broken the installation? It is so simple (one step) we just copied over opencl.dll version 1.1.0.0 from my windows7 machine (to windows xp machine).
|
|
|
|
The00Dustin
|
|
February 02, 2012, 10:24:25 PM |
|
Now having said that, since temp-cutoff doesn't actually adjust engine or memory clock speed, it may well be valid to have it enabled regardless. The default for temp-cutoff, as described in the readme , is 95 degrees. I vote for this. I actually asked about it many pages ago (possibly many moons by now) and didn't see a response. In the meantime, though, without using a config file, can I keep the card at an overclocked speed and cutoff (my card quits working with auto-gpu or auto-fan, which is why I would like temp-cutoff to work without them, but if I could lock the GPU and not turn on auto-fan, that might be a workaround). IOW, since this will presumably downclock before cutoff: --auto-gpu --temp-hysteresis 1 --temp-cutoff 72 --gpu-engine 975 Is this a valid command?: --auto-gpu --temp-hysteresis 1 --temp-cutoff 72 --gpu-engine 975-975 I would test it myself, but I don't really want to have to reboot two or three times because my card messes up, and I would rather fully understand the theory first. Even after a lot of help in the last couple pages, I find it murky, especially since I thought I saw a post that implied the opposite of my understanding of --temp-overheat and --temp-cutoff...
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
February 02, 2012, 10:24:55 PM |
|
[2012-02-02 15:02:40] clGetPlatformsIDs failed (no GPU?) [2012-02-02 15:02:40] 0 GPU devices detected This is a broken SDK installation. Come on conman how could they have broken the installation? It is so simple (one step) we just copied over opencl.dll version 1.1.0.0 from my windows7 machine (to windows xp machine). LOL
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
DeathAndTaxes
Donator
Legendary
Offline
Activity: 1218
Merit: 1079
Gerald Davis
|
|
February 02, 2012, 10:26:10 PM |
|
In the meantime, though, without using a config file, can I keep the card at an overclocked speed and cutoff (my card quits working with auto-gpu or auto-fan, which is why I would like temp-cutoff to work without them, but if I could lock the GPU and not turn on auto-fan, that might be a workaround). IOW, since this will presumably downclock before cutoff: --auto-gpu --temp-hysteresis 1 --temp-cutoff 72 --gpu-engine 975 You need to set --temp-target and --temp-overheat also otherwise it will use the defaults. --temp-target (default 95) <- the temp (adjusted by hysteria) where cgminer starts to DOWNCLOCK the card to prevent overheating. --temp-overheat (default 85) <- the temp (adjusted by hysteria) where cgminer boosts fan to 100% to prevent overheating. --temp-cutoff (default 75) <- the temp (adjusted by hysteria) where cgminer turns off card to prevent overheating. If you don't set any of those 3 the defaults are used. I provided some examples here: https://bitcointalk.org/index.php?topic=28402.msg728093#msg728093
|
|
|
|
kentrolla
|
|
February 02, 2012, 10:26:16 PM |
|
[2012-02-02 15:02:40] clGetPlatformsIDs failed (no GPU?) [2012-02-02 15:02:40] 0 GPU devices detected
to be precise
I am guess this has something to do with it: we just copied over opencl.dll version 1.1.0.0 from my windows7 machine (to windows xp machine). but that worked in 2.1.2. why doesn't that work anymore?
|
█████████████████████████ ████████▀▀████▀▀█▀▀██████ █████▀████▄▄▄▄██████▀████ ███▀███▄████████▄████▀███ ██▀███████████████████▀██ █████████████████████████ █████████████████████████ █████████████████████████ ██▄███████████████▀▀▄▄███ ███▄███▀████████▀███▄████ █████▄████▀▀▀▀████▄██████ ████████▄▄████▄▄█████████ █████████████████████████ | BitList | | █▀▀▀▀ █ █ █ █ █ █ █ █ █ █ █ █▄▄▄▄ | ▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀ . REAL-TIME DATA TRACKING CURATED BY THE COMMUNITY . ▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄ | ▀▀▀▀█ █ █ █ █ █ █ █ █ █ █ █ ▄▄▄▄█ | | List #kycfree Websites |
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
February 02, 2012, 10:27:53 PM |
|
Now having said that, since temp-cutoff doesn't actually adjust engine or memory clock speed, it may well be valid to have it enabled regardless. The default for temp-cutoff, as described in the readme , is 95 degrees. I vote for this. I actually asked about it many pages ago (possibly many moons by now) and didn't see a response. In the meantime, though, without using a config file, can I keep the card at an overclocked speed and cutoff (my card quits working with auto-gpu or auto-fan, which is why I would like temp-cutoff to work without them, but if I could lock the GPU and not turn on auto-fan, that might be a workaround). IOW, since this will presumably downclock before cutoff: --auto-gpu --temp-hysteresis 1 --temp-cutoff 72 --gpu-engine 975 Is this a valid command?: --auto-gpu --temp-hysteresis 1 --temp-cutoff 72 --gpu-engine 975-975 I would test it myself, but I don't really want to have to reboot two or three times because my card messes up, and I would rather fully understand the theory first. Even after a lot of help in the last couple pages, I find it murky, especially since I thought I saw a post that implied the opposite of my understanding of --temp-overheat and --temp-cutoff... That is an awfully low cutoff, and cutoff doesn't throttle unless you set the overheat temp LOWER than it. However, if you specify --auto-gpu without ANY other parameters, like engine speed or memory speed, it will STILL turn off the GPU without adjusting GPU engine clock when it gets to temp cutoff. So, summary: --auto-gpu by itself ends up enabling the cut off feature and nothing else, default is 95 degrees.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
rjk
Sr. Member
Offline
Activity: 448
Merit: 250
1ngldh
|
|
February 02, 2012, 10:28:54 PM |
|
[2012-02-02 15:02:40] clGetPlatformsIDs failed (no GPU?) [2012-02-02 15:02:40] 0 GPU devices detected
to be precise
I am guess this has something to do with it: we just copied over opencl.dll version 1.1.0.0 from my windows7 machine (to windows xp machine). but that worked in 2.1.2. why doesn't that work anymore? OK I'll spell it out. The video driver architecture between Winblows XP and 7 are completely different. XP = Kernel-mode, 7 = User-mode. User-mode drivers trying to run as kernel? Not gonna load/work.
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
February 02, 2012, 10:29:34 PM |
|
does "SICK" for middle card mean a dysfunctional PCI-e slot? the card starts up fine but after about 2 min does this.
Heat? What temps is it getting up to? mid 80's only What happens if you ONLY use the 2nd slots (no other cards)?
haven't tried that yet... It means you're driving your card too hard, likely overclocking it too much for the available voltage and running temperature. cgminer tends to run GPUs extraordinarily hard at similar clocks compared to other mining software.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
The00Dustin
|
|
February 02, 2012, 10:36:37 PM |
|
That is an awfully low cutoff, and cutoff doesn't throttle unless you set the overheat temp LOWER than it. However, if you specify --auto-gpu without ANY other parameters, like engine speed or memory speed, it will STILL turn off the GPU without adjusting GPU engine clock when it gets to temp cutoff.
So, summary: --auto-gpu by itself ends up enabling the cut off feature and nothing else, default is 95 degrees. The reason for the low cutoff is it locks up when it gets above that temp (maybe the sensor is bad, maybe the temp skyrockets when it gets above there and I just never see it higher). I mention --gpu-engine because 975 is a stable over-clocked speed on this card. My concern is that it will be downclocked to factory at --temp-overheat and I don't want that to happen (because it will make the card sick/dead), so --gpu-engine 975-975 would prevent that if it was an option, and if not, I would have to remember how to, and go back to doing, overclock outside of cgminer. That would be fine if I knew it was what I needed to do, but I must ask: See why I would like --temp-cutoff to work without all the other stuff? Also, is DAT correct RE need lower target and overheat than cutoff for cutoff to work? If so, will those do anything if cgminer only has --auto-gpu like you suggest, or would it be acceptable to set them lower by 3 each to get around hysterisis and make cutoff work (on the off chance it ever needs to)?
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
February 02, 2012, 10:39:43 PM |
|
So, summary: --auto-gpu by itself ends up enabling the cut off feature and nothing else, default is 95 degrees. See why I would like --temp-cutoff to work without all the other stuff? ?
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
February 02, 2012, 10:51:47 PM |
|
That is an awfully low cutoff, and cutoff doesn't throttle unless you set the overheat temp LOWER than it. However, if you specify --auto-gpu without ANY other parameters, like engine speed or memory speed, it will STILL turn off the GPU without adjusting GPU engine clock when it gets to temp cutoff.
So, summary: --auto-gpu by itself ends up enabling the cut off feature and nothing else, default is 95 degrees. The reason for the low cutoff is it locks up when it gets above that temp (maybe the sensor is bad, maybe the temp skyrockets when it gets above there and I just never see it higher). I mention --gpu-engine because 975 is a stable over-clocked speed on this card. My concern is that it will be downclocked to factory at --temp-overheat and I don't want that to happen (because it will make the card sick/dead), so --gpu-engine 975-975 would prevent that if it was an option, and if not, I would have to remember how to, and go back to doing, overclock outside of cgminer. That would be fine if I knew it was what I needed to do, but I must ask: See why I would like --temp-cutoff to work without all the other stuff? Also, is DAT correct RE need lower target and overheat than cutoff for cutoff to work? If so, will those do anything if cgminer only has --auto-gpu like you suggest, or would it be acceptable to set them lower by 3 each to get around hysterisis and make cutoff work (on the off chance it ever needs to)? Yeah I'm probably completely wrong with this ... but it sounds like you are saying you have a faulty card and you'd like cgminer to do something other than what works with a correctly operating card ...
|
|
|
|
-ck (OP)
Legendary
Offline
Activity: 4256
Merit: 1645
Ruu \o/
|
|
February 02, 2012, 10:55:20 PM |
|
So, summary: --auto-gpu by itself ends up enabling the cut off feature and nothing else, default is 95 degrees. See why I would like --temp-cutoff to work without all the other stuff? ? Likely, based on what you've described, I'm guessing what you want is: --auto-gpu --temp-cutoff 72 --gpu-engine 975 This will start it at gpu engine speed 975 and will not throttle it, and will enable the thermal cutoff feature if it his 72 degrees.
|
Developer/maintainer for cgminer, ckpool/ckproxy, and the -ck kernel 2% Fee Solo mining at solo.ckpool.org -ck
|
|
|
The00Dustin
|
|
February 03, 2012, 12:15:08 AM |
|
Yeah I'm probably completely wrong with this ... but it sounds like you are saying you have a faulty card and you'd like cgminer to do something other than what works with a correctly operating card ... I can't call you completely wrong on that one. The card has been like this on temperature from day one, so it could be faulty, but even if it's not (5830s have a very wide range of normal from what I have seen in a lot of threads), when 75° is overheating, and 68° is optimal, auto-anything that moves the way cgminer did originally is bound to cause trouble. I simply want cgminer to stop mining on a card without destroying it or forcing a reboot if something extreme happens (IE A/C dies when I’m not around). That having been said... Likely, based on what you've described, I'm guessing what you want is: --auto-gpu --temp-cutoff 72 --gpu-engine 975
This will start it at gpu engine speed 975 and will not throttle it, and will enable the thermal cutoff feature if it his 72 degrees. This is exactly what I want. I was just afraid that the default hysteresis of 3 would apply (hence the setting of 1 in my original question). Since that will do what I want, I obviously misread one of the previous posts. Note that prior to today, I hadn't looked into this (or at the readme) since shortly after the ADL code was added, so my memory may also be a bit foggy on that default hysteresis thing. Regardless, if --auto-gpu won't lower the clock to the original setting when I don't provide a range, those three settings are exactly what I want, and I don't see any reason the code should be changed if it's that easy to set a cut-off only and overclock. The documentation may need changed, but I obviously don't know, I just admitted I hadn't read it since several versions ago.
|
|
|
|
kano
Legendary
Offline
Activity: 4592
Merit: 1851
Linux since 1997 RedHat 4
|
|
February 03, 2012, 12:32:53 AM |
|
Yeah I'm probably completely wrong with this ... but it sounds like you are saying you have a faulty card and you'd like cgminer to do something other than what works with a correctly operating card ... I can't call you completely wrong on that one. The card has been like this on temperature from day one, so it could be faulty, but even if it's not (5830s have a very wide range of normal from what I have seen in a lot of threads), when 75° is overheating, and 68° is optimal, auto-anything that moves the way cgminer did originally is bound to cause trouble. I simply want cgminer to stop mining on a card without destroying it or forcing a reboot if something extreme happens (IE A/C dies when I’m not around). ... I was referring to the comment that downclocking the card makes it sick/die - that's what makes me think it's faulty. Again, yeah I have no idea - but that comment was the strange one. But anyway, ckolivas' comment (or reading the README ) seems to have solved it anyway
|
|
|
|
The00Dustin
|
|
February 03, 2012, 12:53:42 AM |
|
I was referring to the comment that downclocking the card makes it sick/die - that's what makes me think it's faulty. Yeah, I know what you mean, but I'm honestly not sure I ever tried auto-gpu without auto-fan, so I could be wrong on that (I got tired of restarting). My point about the tight range of temperatures was that auto-fan would move around a lot and allow it to get too hot at its' especially low overheat temperature. If the auto-gpu was also causing it, though, it wasn't the downclocking specifically, it was the changing it over and over (and maybe not reacting fast enough, again, this was when ADL support was first added). Regardless, I found a stable fan speed and clock without the auto stuff, and I figure the auto stuff could cause problems (as my gpu could be faulty), and probably wouldn't gain me much, so I don't really want to mess with it again when I can just continue to have a stable miner instead (but I would still like the cutoff safety, which I will now have).
|
|
|
|
BkkCoins
|
|
February 03, 2012, 01:14:16 AM |
|
I've always just used --auto-fan by itself set at either 70C or 75C on my 5830s. cgminer has been quite good at keeping the temperature stable while the fan speed varies according to ambient variations. My understanding is that keeping the temperature constant is the best option for long term longevity.
|
|
|
|
cypherdoc
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
February 03, 2012, 03:13:24 AM |
|
is it possible to underclock the memory down to 300 using cgminer?
|
|
|
|
outsidefactor
Newbie
Offline
Activity: 22
Merit: 0
|
|
February 03, 2012, 03:27:00 AM |
|
is it possible to underclock the memory down to 300 using cgminer?
I clock my 7970s down to 150 with it quite successfully.
|
|
|
|
cypherdoc
Legendary
Offline
Activity: 1764
Merit: 1002
|
|
February 03, 2012, 03:30:08 AM |
|
is it possible to underclock the memory down to 300 using cgminer?
I clock my 7970s down to 150 with it quite successfully. how?
|
|
|
|
outsidefactor
Newbie
Offline
Activity: 22
Merit: 0
|
|
February 03, 2012, 03:36:27 AM |
|
Hello to the cgminer community. I have been trying to implement a background mining setup here at my work (with the consent and support of my boss). I have 20 CAD stations I want to run miners on, but only when the machines are idle (the engineers never shut them off and we source all our power from solar panels on the roof). I have tried various options, from the old RCP screen savers (sooooooo out of date) to commandsaver. Commandsaver was close, but cgminer doesn't seem to like console redirection, plus when the screensaver ends and goes to shut down cgminer it rarely seems to exit cleanly. People directed me to this thread: https://bitcointalk.org/?topic=8403.0, but I can't see how it's any better than command saver and still probably has the clean exiting issue. Any ideas? cgminer is my preference (it's awesomely efficient and only needs one worker account per machine). Maybe if enough of us annoy con he'll fix the console redirection and clean shutdown issues?
|
|
|
|
|