Your opinion on the GD32 devices

[technix]

GigaDevice have a series of Cortex-M microcontrollers that bears suspiciously similar names to the STM32 line. What is your opinion on those? I have a few samples of GD32F103CB (should be similar to the usual STM32F103CB) and GD32F150C8 (seem like being similar to the STM32F2 series?) so maybeI can experiment with something on those...

[thm_w]

What is your reason for looking at the samples, cost? Unless you are driving down to an extremely low cost and can put in a lot of time to work on the differences, I don't think its worth it.
You can get STM32F103CBT6 for $1.50 (digikey $8). I see GD32F103CBT6 for $1.00, and not available from as many sellers.

It has become clear over the last year, that the GD32 is not a direct replacement for the STM32, Despite GD's claims. so I guess I should just remove support for it from the repo.
I don't have time to resolve amy such issues, and GD32 boards are no longer available cheaply, so it seems pointless to continue with them.

http://www.stm32duino.com/viewtopic.php?t=1643

[hans]

I think they have been discussed before on forum that they are cloned STM32s. More people have burned their fingers..

Typical one clones is that analog performance is horrid.
Don't expect code compatibility despite 'claims' - if ST create silicon errata, so will GigaDevice have some.

[donotdespisethesnake]

My understanding is that GigaDevice Semiconductor make Flash chips, and decided to get into MCUs. They decided the best way to do that is create register-compatible versions of a market leader, who are very popular brand in China, namely STM, (I guess because they are also low cost). GigaDevice are a legitimate ARM licensee, so it is possible they got a few silicon engineers in a clean room and told them "create a compatible chip". It's also possible they have reverse engineered from a sample of STM chip, which may or may not be legal.

Whatever method they did choose, I am sure they found creating a top quality design is not that easy. This seems to be evident in the silicon problems with GD32 devices.

STM chips are already pretty good value in the Cortex M market, so saving a few cents on a rather poor copy just doesn't seem to make any sense. Add to that the problems of getting support, and that GD software seems to be a blatant copy of STM software, there isn't a whole lot going for it. I guess if I was a Chinese manufacturer wanting to save every cent, didn't care much about quality, then I might consider GD32.

The background to this is that the Chinese government are investing $ billions into their semiconductor industry, to try and reduce the amount they pay to foreign companies for semiconductor devices. The Chinese realise that state-run companies are not the answer, so they funnel grants and investment funds into investment trusts, who then fund private startups as well as established companies. Eventually they might have enough domestic skills to create their own IP, but so far they are in the "copy and clone" phase.

[technix]

I just took a sample GD32F103CBT6 and mounted it on one of my SushiBits One PCB (intended for STM32F103CBT6 but those are supposed to be pin-to-pin compatible.) Maybe I can experiment with it somehow against a genuine STM32F103CBT6 and a STM32F303CCT6?