I just put a window fan in and put two of my computers next to that The door to that room is closed so it can't go up into the house like in the winter when I leave the door open.
I just put a window fan in and put two of my computers next to that The door to that room is closed so it can't go up into the house like in the winter when I leave the door open.
My old house was kinda like that..lots of fans and door mostly closed except in the winter time, it still got into the mid 90's F in the summer time!! It got so bad my wife wouldn't even go down into the basement after a while.
Noticed it was getting a bit warm today - and sure enough as soon as it touched 35C the GPU crunching stopped for 20 minutes, temp then dropped to 34C and so GPU crunching resumed, then temperature rises etc.
11-Jun-2017 13:20:02 [---] Suspending GPU computation - user request
Noticed it was getting a bit warm today - and sure enough as soon as it touched 35C the GPU crunching stopped for 20 minutes, temp then dropped to 34C and so GPU crunching resumed, then temperature rises etc.
11-Jun-2017 13:20:02 [---] Suspending GPU computation - user request
Noticed it was getting a bit warm today - and sure enough as soon as it touched 35C ...
I presume the 'it' is the SSD temperature? You did mention the possibility of getting temperature over the internet from a weather service. Wouldn't the relationship between your particular SSD and a weather service value be a rather tenuous one at best?
You said you use SSD temp as a room temperature indicator so when the SSD temp gets to 35C is that really what your room temperature is?
I know nothing about SSDs - virtually all my drives are >10 year old spinning rust - but are SSDs really that sensitive to temperature that you have to worry about 35C?
Since your SSD is in the "coolest part of the case", wouldn't you be better off making decisions based on something that's in the hottest (eg. CPU core) part of the case? Wouldn't there be a bit of a delay between an increase in core temp and the subsequent reaction of the SSD temp to the extra source of heat?
My experience has been that things tend not to fail if being operated at 'steady state' conditions, even if that means somewhat higher than desired temperatures. I think that cycling things on and off rather frequently might be more damaging in the long run. This is just a gut feeling based on experience.
Noticed it was getting a bit warm today - and sure enough as soon as it touched 35C ...
I presume the 'it' is the SSD temperature?
Yes.
Quote:
You did mention the possibility of getting temperature over the internet from a weather service. Wouldn't the relationship between your particular SSD and a weather service value be a rather tenuous one at best?
At best tenuous? hmmm - no. My observations are
When the window is open the temperature inside is about the same as outside.
Temperature outdoors increases - temperature indoors increases.
Use room air to cool components.
Room temperature increases, the temperature of components increases.
Case temperatures are always higher than room temperatures when running.
Your observations maybe different. You may have air con.
If you have no thermometer and your plan for summer is "not crunch on hot days" it is a lot better that throwing dice, or not crunching at all.
Quote:
You said you use SSD temp as a room temperature indicator so when the SSD temp gets to 35C is that really what your room temperature is?
.. as measured by the thermometer in the SSD, with case fan pulling air - yes. It feels about that, i don't have a room thermometer - but it is hotter when the number goes up and cooler when the number goes down :). I have seen it up at 40C and the room was sauna-esque.
Quote:
I know nothing about SSDs - virtually all my drives are >10 year old spinning rust - but are SSDs really that sensitive to temperature that you have to worry about 35C?
I didn't say i was worried about SSD temps specifically, but its location (in the case) makes it the best room thermometer i have, and based on the heat map i wanted to reduce the peak temps (in the room), i have to sit next to them.
Quote:
Since your SSD is in the "coolest part of the case", wouldn't you be better off making decisions based on something that's in the hottest (eg. CPU core) part of the case? Wouldn't there be a bit of a delay between an increase in core temp and the subsequent reaction of the SSD temp to the extra source of heat?
hmmm - no. The GPU and CPU will throttle their fan speed, to keep at peak output and temps stays roughly the same - i am trying to match the power consumption to the room's ability to shed heat and my tolerance to high temperatures.
Quote:
My experience has been that things tend not to fail if being operated at 'steady state' conditions, even if that means somewhat higher than desired temperatures. I think that cycling things on and off rather frequently might be more damaging in the long run. This is just a gut feeling based on experience.
This is also a concern, and the GPU will take the hit, as other devices will not change temperature. I did consider down clocking the GPU, but that may generate task errors.
To counter the cycling argument the case temperature will be in a smaller temperature range so components will last longer, fans work less etc.
... and today the GPU had a rest for about 3 hours, the temperatures stop at 35C.
I just put a window fan in
)
I just put a window fan in and put two of my computers next to that The door to that room is closed so it can't go up into the house like in the winter when I leave the door open.
mmonnin wrote:I just put a
)
My old house was kinda like that..lots of fans and door mostly closed except in the winter time, it still got into the mid 90's F in the summer time!! It got so bad my wife wouldn't even go down into the basement after a while.
AgentB: Thank you for
)
AgentB:
Thank you for posting that! Hopefully there will be at least some portion of time for testing that on my machines, this century :-/
Noticed it was getting a bit
)
Noticed it was getting a bit warm today - and sure enough as soon as it touched 35C the GPU crunching stopped for 20 minutes, temp then dropped to 34C and so GPU crunching resumed, then temperature rises etc.
I expect it will continue like so for today.
AgentB wrote:Noticed it was
)
:-(
You guy´s need to pass some
)
You guy´s need to pass some time down here, hot and high humidity allmost all the year. LOL.
Hotter than Brazil was?
)
Hotter than Brazil was?
Betreger wrote:Hotter than
)
Where i lived in Brazil at least in the night the temp downs to less than 20C, and in some months the temp goes from 10-26C.
Here it´s hot all the time, even on the night, today´s minimum 24C, today´s max, who cares? it´s hot as a hell. LOL
The humidity, OMG it´s simply amazing, incredible high. So you feel like you are in a steam bath.
AgentB wrote:Noticed it was
)
I presume the 'it' is the SSD temperature? You did mention the possibility of getting temperature over the internet from a weather service. Wouldn't the relationship between your particular SSD and a weather service value be a rather tenuous one at best?
You said you use SSD temp as a room temperature indicator so when the SSD temp gets to 35C is that really what your room temperature is?
I know nothing about SSDs - virtually all my drives are >10 year old spinning rust - but are SSDs really that sensitive to temperature that you have to worry about 35C?
Since your SSD is in the "coolest part of the case", wouldn't you be better off making decisions based on something that's in the hottest (eg. CPU core) part of the case? Wouldn't there be a bit of a delay between an increase in core temp and the subsequent reaction of the SSD temp to the extra source of heat?
My experience has been that things tend not to fail if being operated at 'steady state' conditions, even if that means somewhat higher than desired temperatures. I think that cycling things on and off rather frequently might be more damaging in the long run. This is just a gut feeling based on experience.
Cheers,
Gary.
Gary Roberts wrote:AgentB
)
Yes.
At best tenuous? hmmm - no. My observations are
When the window is open the temperature inside is about the same as outside.
Temperature outdoors increases - temperature indoors increases.
Use room air to cool components.
Room temperature increases, the temperature of components increases.
Case temperatures are always higher than room temperatures when running.
Your observations maybe different. You may have air con.
If you have no thermometer and your plan for summer is "not crunch on hot days" it is a lot better that throwing dice, or not crunching at all.
.. as measured by the thermometer in the SSD, with case fan pulling air - yes. It feels about that, i don't have a room thermometer - but it is hotter when the number goes up and cooler when the number goes down :). I have seen it up at 40C and the room was sauna-esque.
I didn't say i was worried about SSD temps specifically, but its location (in the case) makes it the best room thermometer i have, and based on the heat map i wanted to reduce the peak temps (in the room), i have to sit next to them.
hmmm - no. The GPU and CPU will throttle their fan speed, to keep at peak output and temps stays roughly the same - i am trying to match the power consumption to the room's ability to shed heat and my tolerance to high temperatures.
This is also a concern, and the GPU will take the hit, as other devices will not change temperature. I did consider down clocking the GPU, but that may generate task errors.
To counter the cycling argument the case temperature will be in a smaller temperature range so components will last longer, fans work less etc.
... and today the GPU had a rest for about 3 hours, the temperatures stop at 35C.