NeoChrom/GPU2D registers documentation

[robvos]

Will this peripheral ever be documented in detail? i.e. registers, interrupts, etc.

Or even a lightweight/low level library would be acceptable.

We may choose STM32U5 for a product but not if we are forced to use TouchGFX for good performance.

[JMYGI]

I am sorry to hear TDJ - you are most welcome to contact me directly on jorgen.mygind@st.com for further sharing and possible alignments - thank you 

[TDJ]

@JMYGI The problem, as I see it, is much wider. Recent years ST, instead of supporting community open-source software, selects/builds proprietary closed-source technologies which lock customers into ST's solutions like ThreadX, TouchGFX, and now NemaGFX. It is a bright red flag. Even from mcuboot ST had to create own customized SBSFU solution.

Ofc, ST is for-profit company and it has to select the best strategy primarily to please the stakeholders, rather than engineers, but when such strategies are too aggressive clients simply go somewhere else. Especially because the whole software industry is more mature than it was 20yrs ago while, in my opinion, ST makes similar mistakes Microsoft made during Steve Ballmer era.

[Jack3]

A Dutch company, three letters with an 'X' in the middle has it all!

At an even higher performormance (M7 @ 1GHz) + GPU2D!

The IDE is also eclipse-based, so you instantly will feel home!

We made the switch!

[Jack3]

Three letters, with an 'X' in the middle?;)

[DmitryR]

@Jack3, I can give you more definite clue: just look what acceleration hardware is now supported by LVGL ;)

Regards,

Dmitry

[TDJ]

Does anyone know if full NeoChrom/GPU2D (NemaGFX) documentation is already available or WHEN will it be available?

[DmitryR]

Hi @TDJ ,

LVGL team has already made a port for STM32U5 chips that uses NemaGFX. You may probably use it as is. If it is not enough for you then looking through the code will probably make most NemaGFX interface clear for you. Of course it uses precompiled NemaGFX libraries while Think Silicon does not disclose the source. If your project needs full source code for security or certification purposes than you probably should not rely on STM32U5 chips. Hope that helps.

Regards,

Dmitry

[TDJ]

@DmitryR I know this sad story but still 'NeoChromSDK' was promised by @JMYGI in August 2024 so I was just wondering what the current status was since it was almost first STM32U5G9 release anniversary.

[DmitryR]

@TDJ , we all know how fast STM can be. I have once found a simple bug in CubeMX, it was fixed in a year. Recently I have found severe bug in AzureRTOS and they have just told me to go away despite the fact that the bug is in the STM32 glue code that is not present in main ThreadX repository. So I have just stopped monitoring it, I think that NemaGFX will sooner appear in ThreadX GUIX than in CubeMX ;)

Regards,

Dmitry

[TDJ]

@DmitryR I know, I found this syntax error in HAL and it was just fixed - almost a year later. I think we all noticed that ST microcontroller business is on decline and, I think, actually struggles to survive. During recent Electronica 2024 in Munich ST microcontrollers were largely not present. Im afraid there are some bad 'exciting news' to be shared soon. ST simply failed to realize that good hardware for their customers is not good enough, that software is of at least equal importance since what counts at the end is the final solution and time to market. I do understand that probably not being able to develop graphics engine on their own ST had to purchase this technology trying to stay competitive, but it just was not available for them to purchase with full docs they could share. I do understand that. Luckily, NemaGFX just appeared in LVGL (ofc with precompiled lib) and for my current project it still may be a viable technology. At least for this last project that I already started over a year ago being a victim of the fine art of making customers hear what they want to hear without both lying and revealing the ugly truth for as long as it gets.