• 15.10.2018, 21:10
  • Registrieren
  • Anmelden
  • Sie sind nicht angemeldet.

 

Lieber Besucher, herzlich willkommen bei: Aqua Computer Forum. Falls dies Ihr erster Besuch auf dieser Seite ist, lesen Sie sich bitte die Hilfe durch. Dort wird Ihnen die Bedienung dieser Seite näher erläutert. Darüber hinaus sollten Sie sich registrieren, um alle Funktionen dieser Seite nutzen zu können. Benutzen Sie das Registrierungsformular, um sich zu registrieren oder informieren Sie sich ausführlich über den Registrierungsvorgang. Falls Sie sich bereits zu einem früheren Zeitpunkt registriert haben, können Sie sich hier anmelden.

Deema

Junior Member

Aquasuite virtual temperature engine re-design

Sonntag, 13. Dezember 2015, 16:37

Hi to Aqua Support Crew!

Does Aquacomputer planning to re-design the virtual temperature processing logics in the way, described below?
- Expand or unlimit the maximum number of virtual temperature sensors to maximum number of controllers in Aquasuite.
- Allow to use virtual temperature sensors as the input for other virtual temperature sensors (recurrence without cycling).
- Allow much operands for functions.
- Allow to write any expression with operations, functions and brackets for virtual temperature calculation in floating point.

For example, let the T1..T8 variables are the physical sensor temperature readings, the ST1..ST8 - the software sensor readings, VT1..VT8 - virtual sensor 'temperatures' (or data of any kind).

Max - Maximum function:
Avg - Average function;
Abs - Absolute function;

Possible expressions might be as follows:
VT1 := Avg( Max( T1, T2 ), ST1 )
VT2 := ( ( VT1 * 25 ) + ( T3 * 75 ) ) / 100
VT3 := Avg( ST1, ST2, ST3, ST4, ST5, ST6, ST7, ST8 )
VT4 := Max( ST1, ST2, ST3, ST4, ST5, ST6, ST7, ST8 ) - Min( ST1, ST2, ST3, ST4, ST5, ST6, ST7, ST8 )
VT5 := 100 * VT4 / VT3
...and so on.

cheers,
Deema

Montag, 14. Dezember 2015, 07:07

Expand or unlimit the maximum number of virtual temperature sensors to maximum number of controllers in Aquasuite
Not possible.
Allow to write any expression with operations, functions and brackets for virtual temperature calculation in floating point
The device has limited ressources. To expand this function we have to limit other functions.
For the most of our coustomers the current funtionality is more than enough.

All funtions are directly computed in the device. The aquasiute is only for visualisatzion and setup.
We have done this and other features (free routing of any input/output, signal) for industrial devices - but the problem is, no one of our (consumer) customers can setup this this.
This would end in a support nightmare.

Deema

Junior Member

Mittwoch, 16. Dezember 2015, 20:06

All funtions are directly computed in the device. The aquasiute is only for visualisatzion and setup.
I thought the 'Aqua Service' in aquasuite package might be a good bridge to deliver software sensors data to aquaero engine. And it might be an 'extender' for the processing engine.
If hardware working in standalone mode - OK, all software sensors are unavailable and extended logics is unavailable also. This is reasonable.

As the minimal improvement, sebastian, will the virtual sensors calculations be possible in the future?
Like VS1 is an average of T1 and T2, VS2 is an average of T3 and T4, and VS3 is a maximum of VS1 and VS2...
That might be a great feature if a components require thermal control in more than single point and cooling of a component group should be based on a worst temperature.

Thank You for the answer!

Donnerstag, 17. Dezember 2015, 03:03

All funtions are directly computed in the device. The aquasiute is only for visualisatzion and setup.
I thought the 'Aqua Service' in aquasuite package might be a good bridge to deliver software sensors data to aquaero engine. And it might be an 'extender' for the processing engine.
If hardware working in standalone mode - OK, all software sensors are unavailable and extended logics is unavailable also. This is reasonable.

As the minimal improvement, sebastian, will the virtual sensors calculations be possible in the future?
Like VS1 is an average of T1 and T2, VS2 is an average of T3 and T4, and VS3 is a maximum of VS1 and VS2...
That might be a great feature if a components require thermal control in more than single point and cooling of a component group should be based on a worst temperature.

Thank You for the answer!

Are you talking about cooling individual components, or separate groups, on an air cooled system?

Donnerstag, 17. Dezember 2015, 07:00

For the future we plan a solver to calculate and mix different sources to a new sensor in the aquasuite service.

Donnerstag, 17. Dezember 2015, 10:12

All funtions are directly computed in the device. The aquasiute is only for visualisatzion and setup.
I thought the 'Aqua Service' in aquasuite package might be a good bridge to deliver software sensors data to aquaero engine. And it might be an 'extender' for the processing engine.
If hardware working in standalone mode - OK, all software sensors are unavailable and extended logics is unavailable also. This is reasonable.

As the minimal improvement, sebastian, will the virtual sensors calculations be possible in the future?
Like VS1 is an average of T1 and T2, VS2 is an average of T3 and T4, and VS3 is a maximum of VS1 and VS2...
That might be a great feature if a components require thermal control in more than single point and cooling of a component group should be based on a worst temperature.

Thank You for the answer!


You should be able to do that now, however, in your example, it seems as though all you need is VS3. What would you control that would require the separation as you describe.

I run a VT sensor in my setup, taking the max of either CPU or GPU to control the fan speed. What you would do in your example is take T1-T4 and create a new VT instead of taking VT sensors to create another VT. I can only presume your example is VS1 is an average of a coupe of CPU cores and VS2 is an average of a couple of GPU cores (or the like) which should generally run at similar loads in most circumstances. But in either case, the max of any is what you want...realistically.

What's your setup out of curiosity that would require these calculations? Also, why would you ever need an average out of any sensor. Surely you want to control your temperature based on a maximum of something. Else, you potentially end up with something really hot and not being adequately cooled while a majority of other things are cool.

The only think I can see is taking individual core readings, at which point I'd take a max of them, certainly not an average.

Dieser Beitrag wurde bereits 2 mal editiert, zuletzt von »Railgun« (17. Dezember 2015, 10:18)

Deema

Junior Member

Freitag, 18. Dezember 2015, 23:00

Are you talking about cooling individual components, or separate groups, on an air cooled system?

Air cooled workstation.

For the future we plan a solver to calculate and mix different sources to a new sensor in the aquasuite service.

Aqua, you are AWESOME!

You should be able to do that now

YES! I simple forgot to re-check that after Aquasuite upgrade. Sorry, it works fine now!

Also, why would you ever need an average out of any sensor.
...
The only think I can see is taking individual core readings, at which point I'd take a max of them, certainly not an average.

You are right, the MAX function much usable than AVERAGE. I found the averaging in Aquasuite and used it in my examples, because it already exists.

What's your setup out of curiosity that would require these calculations?

Nothing extra original, ordinary maniac's workstation.

Case: Fractal Design Define R4, 2x front fans + 1 rear fan (upper and side holes are closed). Possibly will add 1 upper fan, close rear hole and move rear fan to top.
MB: Gigabyte GA-X79-UP4 (LGA2011)
CPU: Core I7-3820 (TDP 125W)
CPU Heatsink: Noctua NH-D15, one fan removed
RAM: 8x Kingston Blu 8 GB modules (64 GB)
Storage: Adaptec 6805 controller, 40mm fan placed on ASIC radiator
HDD: 3x HGST NAS-edition 7K2rpm 4 TB, will add ~3 HDDs in a year or two
Video: Gigabyte GV-N610SL-1GI passive radiator
Audio: old Creative Audigy 2 in old PCI slot
PS: Bequiet! Dark Power Pro 750W
Platform monitoring: HWINFO x64
Tasks (mostly): spherical panorama composition, video editing, photo processing.
Now it prepares a couple of huge topografic maps, 1 week already done, 2 weeks lasting.
Aqua console screenshots in one image are below.

Let's think:
- Number of HDD will rise to 6 and 8 is a case maximum. Thus, we should plan 3 thermal sensors: low / middle / upper. And program one 'controller' for 2 front fans.
- The MB southbridge might be hot as well and controlled with software sensor. It chilled by front fans because of placement. So, it should affect the front fans as well.
- But SB and HDD temperature safe ranges are not equal. Thus, the SB or HDD group temperature must be 'scaled' before front fan 'controller'. Something like T := MAX( Thdds, 0.7 * Tsb ).
- The same situation is at the case upper and rear fan, it affects the PWMs and 2 RAM banks at least. I put 1 sensor at the case top and 2 sensors in each RAM banks.
- In addition, the incoming fans should keep positive internal pressure in the case, to decrease the dust penetration. Thus, RPMs of rear fan should be less than sum of RPMs of front fans (approximately).

As for me, this is a wide area for mathematics and experimenting. :)