Forerunner 955 poor performance with Analog sample (drawBackground ~10ms)

I ran the Analog sample by switching the buffers to modern form:

_offscreenBuffer = Graphics.createBufferedBitmap({
    :width => dc.getWidth(),
    :height => dc.getHeight(),
        :palette=> [
            Graphics.COLOR_DK_GRAY,
            Graphics.COLOR_LT_GRAY,
            Graphics.COLOR_BLACK,
            Graphics.COLOR_WHITE
        ] as Array<ColorValue>
}).get();

And date buffer:

_dateBuffer = Graphics.createBufferedBitmap({
        :width=>dc.getWidth(),
        :height=>Graphics.getFontHeight(Graphics.FONT_MEDIUM)
}).get();

After that it compiles and runs. However when I profile it, I don't get good numbers:

drawBackground takes nearly 10 000 us on average with with Forerunner 955 whereas with Fenix 6 it takes just 4000 us.

That's more than twice as bad with Forerunner 955 than Fenix 6. Is something off? I also notice that emulator does not show antialiasing on FR955 but it does on older devices.

Thanks

Top Replies

jim_m_58 over 2 years ago in reply to Ciantic +1 verified

Understand the display is updated by row. The more rows, the bigger impact. Think about a second hand with an analog face. The lowest impact will be at 3 and 9, and the biggest at 6 and 12.

Top right quarter…
Ciantic over 2 years ago in reply to jim_m_58 +1

> Top right quarter and top half half screen would be the same number of rows.

Now I get it! Thanks a lot. You seem to be king of Monkey C around here, reading other posts.

I do get the sense though…
FlowState over 2 years ago in reply to Ciantic +1

Ciantic said:
Garmin should own the Monkey C and make their devs to write the stock Watch Faces and code them in Monkey C. Or indeed let us build the stuff with what ever they use to build their watch faces…

All Replies

0 jim_m_58 over 2 years ago

The 955 is a CIQ 4 device, with a GPU. setAntiAlias has no effect on that device, as it's handled by the GPU. Same for the f7 devices. These devices also have the graphic pool, which is why the change to BufferedBitmap is needed

0 Ciantic over 2 years ago in reply to jim_m_58

Okay, that seems unrelated though.

I have now tested with other devices the Analog sample:

FR255 = 1250 us

FR955 = 11000 us!

Forerunner 955 is 10 times worse than FR255 with same code.

The simulator might be broken for FR955?

0 Ciantic over 2 years ago

Here is the issue in images:

Forerunner 955 its nearly 10ms

Forerunner 255 only 1ms:

Both with 62 second profiling time.

0 jim_m_58 over 2 years ago in reply to Ciantic

Even 10ms isn't that bad. Try running "View Watch face diagnostics" when the WF is in low power mode and see what you find. The avg for onPartialUpdate needs to be under 30ms for example.

The 955 and 255 FW is very much the same (like the 245 and 945), and I suspect what you are seeing is based on low/high power and onPartialUpdate.

0 Ciantic over 2 years ago in reply to jim_m_58

That's why I chose FR255 to test against, it's basically same thing as FR955.

FR955 with low power mode, and 62 profiling second snapshot:

FR955 without low power mode, and 62 profiling second snapshot:

FR255 with low power mode, and 62 profiling second snapshot:

FR255 without low power mode, and 62 second profiling snapshot:

But nope, FR955 in simulator is consistently 10 times worse than FR255.

Have someone at Garmin added FR955 to simulator and typed wrong GPU/CPU settings? It can't be 10 times worse than FR255

0 _psx_ over 2 years ago in reply to Ciantic

I reported this behaviour as bug months ago. Nobody even has answered.

GPU devices should be handled a bit different to run efficient.

0 FlowState over 2 years ago in reply to _psx_

_psx_ said:
I reported this behaviour as bug months ago. Nobody even has answered.

You mean this bug? It's marked as fixed now.

0 Ciantic over 2 years ago in reply to FlowState

Interesting, I don't know even what the bug is? Did they type wrong CPU/GPU combo to FR955 in simulator?

Because to me it looks like FR255 gives more accurate numbers from the profiler. FR255 is ~4 times better than Fenix 6 in Profiler, whereas FR955 is 10 times worse than FR255.

0 _psx_ over 2 years ago in reply to FlowState

maybe it's fixed but as you see problem still exists

I don't make tests on device assuming sim shows good or at least proportional values.

So, unfortunately for devices with graphics pool you have to add some code to run efficient but I'm not sure if it meet older device - I'll make some tests. We have new devices with gpu and they should have beautiful graphics and be faster - in my opinion graphics pool spoils something - the better solution was increasing memory for apps.

0 FlowState over 2 years ago in reply to _psx_

_psx_ said:
maybe it's fixed but as you see problem still exists

Yeah I was just curious if that was the same bug, that's all. If they say it's fixed and it's not really fixed, that's even worse than not getting a response.