ThreadX status and future

[TDJ]

Hi All,

Not so long ago ST has selected Microsoft ThreadX as one of its its strategic technologies.
In my opinion it is a very good RTOS. The problem is that it was supported by Microsoft only, no other community project integrates with it, e.g. mcuboot etc. It is almost like ThreadX does not exist.
Now, Microsoft ThreadX is Eclipse-ThreadX. It is open-source, MIT licensed but last Github commit was made 6 months ago. The latest version is 6.4.1 while ST is still stacked with version 6.2.0.

No significant ST 's commitment is visible. It does not look good - regretfully, I need to say.

Am I missing something?

I need to decide whether I should use it in my project or abandon while I still have a choice.

[TDJ]

One more thought. The problem with ST, as I see it, is that as much as possible it tries to remain hardware company and maintain very minimum software it exclusively owns like CubeXX, TouchGFX etc. In some ways, maturity-wise ST is where Microsoft used to be during Steve Ballmer era.
The same time competition truly engages in community projects like mcuboot, trustedfirmware or lvgl where ST's presence is at best only pretended. ThreadX seems to fall into this pattern. Probably it was selected in the first place mainly because competitors could not offer it free. Since solutions we build consist of both hardware and software and time-to-market matters a lot, ST, although makes excellent hardware, to me by no means is a clear winner because competitors provide much more help with software.

[Pavel A.]

> because competitors provide much more help with software.

Really?

[TDJ]

@Pavel A. E.g. look at lvgl 9.x and vector graphics support.
ST's DMA2D? It is supported only because I coded it and someone just ported it from rev 8.3.
The same time other vector renderers are supported.

[Andrew Neil]

It's all just 3rd-party software - you don't have to rely on ST for any of it ...

[Pavel A.]

@Andrew Neil Developing for graphic hardware depends on support and documentation of the vendor. For example new graphic controllers of H5 and U5 (?).  Even situation with USB and Ethernet is far from great (well, 3rd party IPs...). But are competitors much better?

[TDJ]

@Andrew Neil 
True, but that is not my point.
It is not about what you have to or not.
It is about how long it takes and how how expensive it is to build a quality product using ST MCU and then maintain it.

[Andrew Neil]

---

@Pavel A. wrote:

 are competitors much better?

---

I think not.

[TDJ]

@Andrew Neil, @Pavel A. That depends on a particular area, project and expectations but if competitors were much butter, I would not be here and ST board would hold emergency meetings because of stocks falling down sharply. But I am neither ST VP nor stake holder and as an engineer I often have reasons to be unhappy. ThreadX uncertain future and U5 GPU2D docs are just two of them.

[TDJ]

@Pavel A. There is Cortex-M33 MPU available from another, well-known Dutch vendor with much more SRAM than ST's STM32U5G9, faster clock, GUP2D well documented and LVGL supported. I do not know yet if it is overall better (software support in particular) but it looks promising, so I will find out - I just ordered evaluation board.

GPU2D in some ways seems to be a story similar to ThreadX. It has been confirmed (AN5632) that GPU2D is highly proprietary NemaGFX. No full documentation will be ever available.