Guys. Can’t anything be done for Krita to still perform well, without
consuming too much energy? The program keeps putting fire on my
CPU like crazy. All the time, and it does so more than Clip Studio and
Photoshop.
Krita is only a few hundreds of megs in size.
CSP is about a Gig.
Photoshop is about a couple Gigs.
Why is Krita heating up the CPU so much for about the same operations?
Mainly drawing and painting with a brush?
What makes you think there is a connection between a programs file size and the CPU power it needs? Im pretty sure I can make a program that will melt your CPU and is just a few kilobytes large.
Can you give some actual numbers or is it just a feeling? My computer’s CPU doesn’t even seem to notice Krita is running, when I paint. Especially compared to all the other stuff running on it at the same time.
Any setting can cause such problems, drivers, other programs, etc. Krita runs very stable and reliable and very resource-saving
It should be related to your canvas size and brush type. When the canvas is larger, there will indeed be more lag and delay, especially when using mixed brushes.
No. Has nothing to do with the canvas.
It comes from my comparison of Krita to CSP, Photoshop or Leonardo.
All the later programs have complex brush engines but they consume less
energy to achieve the task.
Krita on the other hand will put your CPU on fire.
I am on a Core i7 3rd Gen. With an Nvidia GT 640M as my GPU.
12 GB of RAM as well.
Note : Test done with a simple pixel brush engine.
I don’t experience the heating issues with Photoshop, CSP, or Leonardo.
Which is why I wrote the complaint.
It has been years. The program, although faster compared to years ago, still has
plenty of room for optimization.
Mathematicians with Computer Science knowledge are probably needed.
Prove it. Do the benchmarks. Perform the measurements. Give us the hard numbers. Not saying you are wrong but it’s hard to improve on feelings.
My hypothesis is that, krita do most of the calculation part with CPU (except Canvas rendering) as much thread as possible (when MT calculation is available, pixel brush engine for example), keep all CPU cores from idle core parking, and your old laptop cpu will hover in 60-70℃ even in light workload.
(From my observation back in 2020, i7-2670QM laptop with Windows 7. Take it with a big grain of salt)
Can use “Process Lasso” or other cpu tool which shows Core Parking status, and compare the behavior between Krita and other software.
Bruh. I’m not like these ones who make the claim
without reason. It didn’t come out of the blue.
I have a core i7 3rd gen.
Clearly, Krita has greater energy consumption for performance ratio.
( Arguably Photoshop as well, but a bit less )
To measure energy consumption you can use a smart plug at the wall outlet and get various power readings. You can measure that using an app A and then an app B. If you are on a laptop, you could probably remove the battery to only draw power from the power outlet.
You should also consider that the app consuming more power may just be working faster.
If you are bothered by the heat, then you could try setting a power profile so that it doesn’t go all out on the available power budget, eg don’t max out the CPU clocks or all cores.
My CPU is actually underclocked.
I set my profile to use no more than 85% of the max processor power whether on battery or while the
power adapter is plugged in.
And the fact that the software is working faster because it heats up more, is not the case in this context.
Thanks for your suggestion, but it’s not the only way to measure power consumption.
Checking the CPU temp as you work in a specific program is already sufficient proof.
Photoshop and Krita have always been firing up my fan and CPU heat for years.
With Krita being the most greedy in power consumption. Almost like a video game with heavy graphics.
In all the softwares, the test was done with a 10px basic round brush on a 300ppi A4 canvas.
Traditionally we measure power in Watt or energy consumption with something like Watt Hours not by measuring the temperature (admittedly you can calculate watts from temperature but you also have to know other variables like resistance and thermal conductivity (and more)). Otherwise Krita would consume less energy on my CPU because it doesn’t get as warm, not to mention the active and passive cooling that is happening (on your laptop too).
I don’t say Krita doesn’t draw more energy than others only that your way of measuring it is wrong. An honestly I don’t care much because Firefox takes a bigger toll on my CPU than Krita does most of the time. And in the end I wouldn’t see much difference on my electricity bill even if Krita would be twice as efficient, I’m sure.
What would be interesting is to hear from people painting on mobile tablets if their battery drains significantly quicker when using Krita compared to normal usage or other drawing apps with similar capability. That is something I actually find interesting to know. But for my Desktop or Laptop usage I don’t care much. I didn’t notice a significant difference in battery drain when using Krita on Laptop compared to working other tasks (like coding) but I have to admit I don’t draw much on my laptop for other reasons.
You can not say that my way of measuring it is wrong, because I am not using
watts consumption per second for the comparison, but CPU temperature while
doing a specific task : painting in the program.
Just because it’s CPU temperature, doesn’t make the measurement method invalid.
The one who is wrong here is you.
One way to measure it is CPU temperature.(1)
Another is watt consumption per second. (2)
I didn’t say that I would use (2) to measure it only to switch later to (1).
Your initial claim is that “Krita consumes tons of energy”, to prove that ton you better have a number that describes energy consumption in absolute values. For now you have only proven that one thing makes your CPU hotter than another thing but how much power that actually is and if it is a ton of it has yet to be shown.
That’s an hyperbola.
It is literally how you named your own topic.
So I happen to have found a software which helps with the measurement.
And yes the result is in Watts, CPU Power in time.
The system itself consumes about 8-12 watts of CPU power.
It seems that Krita siphons the CPU power quite a lot but not in the most efficient manner
possible yet. At least from a power consumption perspective.
Test done on a core i7 3rd gen.
Maybe the reason for the high temp is because I live in a hot country
or it could be that my PC’s cooling is not as efficient anymore( might have to buy a cooler
pad by a certain time ).
Out of curiosity, can you test with “Canvas Graphics Acceleration” disabled in “Configure Krita” - “Display” ?
With “Canvas Graphics Acceleration” disabled, it seems to consume
at most 20 watts. But would go down up to 16 watts consistently after a while.
Also the power used decreased. My CPU heat went around the same as
Photoshop, so 75-82°C.
It’s weird that it consumes less power with canvas acceleration turned off since this would mean more work needs to be done on the CPU, because canvas acceleration should offload some work to the GPU, not painting though.
Does the tool take GPU usage into account too? Or perhaps Krita uses the CPUs integrated graphics.
Btw, what tool do you use? Would be nice to know so others can do their own benchmarks so we can compare the numbers.