Request for Future Performance Upgrades in the ESP32 Series: ESP32-Wroom-32F2X-4Core 800MHz

[Knottis]

Hi Espressif team and community,

I’ve been working with ESP32 for several years and truly appreciate the platform’s flexibility and the great developer ecosystem around it. However, I — and many others in the community — are starting to feel the limitations when it comes to performance and capacity, especially for more demanding applications where real-time performance and parallel processing are critical.

With that in mind, I would like to raise the following suggestions and questions:

1. More cores and higher clock speeds:
A future ESP32 version with 4 cores running at 800 MHz or more would significantly help with multitasking and intensive computation workloads.

2. Dedicated RF modules for Wi-Fi and Bluetooth:
Having separate radio modules for Wi-Fi and Bluetooth would allow both to run simultaneously at maximum throughput without interfering with each other — something that is still a limitation in the current architecture.

3. Increased internal memory (SRAM) for the CPU:
There is a strong need for more internal RAM directly accessible by the CPU, without relying on slower and sometimes unstable external PSRAM under high load.

Background and concerns:

• There has been little to no development in this direction for many years, at least for those of us who require more performance and power.

• Many developers are now switching to Raspberry Pi or similar platforms just to get the performance they need, even when they would prefer to stay in the microcontroller space.

• It seems that Espressif has recently focused mainly on low-power, minimalist IoT devices — which of course have their place — but this leaves an unmet need in the higher-performance segment.

My questions to Espressif and the community:

• Are there any plans for a more powerful ESP32 variant, or is Espressif moving away from this segment altogether?

• Are there any prototypes or development roadmaps addressing these needs?

• What can we as developers and users do to help prioritize this kind of product direction?

I’d greatly appreciate any insights — from Espressif directly or from others in the community who may have information or suggestions on alternative approaches.

Best regards,

[Sprite]

Not sure if you have looked at the ESP32-P4 yet? You could see that as a direction we could go into wrt cpu power and memory. We certainly aren't thinking about abandoning the higher-end segment. (And yes, the CPU there is 'only' 400MHz, but wrt instructions/MHz it's a lot better than the LX6 in your hypothetical block diagram.)

Can I ask what specifically you need all this CPU power and RAM for? It certainly does help development if we know there's a market for certain classes of chips.

[Knottis]

The ESP32 is in serious need of a Major Upgrade.
Its current performance and feature set are no longer on par with competing products, which now set a much higher standard for speed, stability, and modern functionality.

If Espressif wants the ESP32 to remain relevant and attractive, real development efforts are required – not just minor tweaks.

[Sprite]

Its current performance and feature set are no longer on par with competing products, which now set a much higher standard for speed, stability, and modern functionality.

This is not a desktop or workstation market, however: better performance is only one of the many axes amongst which value is measured, and rarely the most important one. Again, can I ask you what applications you would have in mind for such a 4-core 800MHz beast that cannot be fulfilled by the current stock of chips, including the -P4? You need to understand we're running a commercial company here; as much as I'd like us to make whatever chips strike our fancy, in the end we need to have a non-trivial amount of customers (as in: more than just hobbyists) for them in order to justify the costs of producing them.

[username]

Knottis, I think your application is well beyond any embedded micro. You need to more look at some embedded linux platforms to get those features you want.

If Espressif wants the ESP32 to remain relevant and attractive, real development efforts are required – not just minor tweaks.

This is totally false. Espressif has pretty much dominated the embedded market. They out perform most everything out there, and there cost is also the lowest. Bang for buck your just not going to find anything currently out there to be it.

[thefury]

real development efforts are required

Real development efforts, hmmm. Like, say, putting together the industry's best toolchain and development framework that not only brings in hundreds of libraries but also works across several different architectures seamlessly without the user having to write their own system startup code?

Real development efforts, like managing thousands of issues and feature requests on GitHub?

Real development efforts, like creating the industry's best low cost high performance microcontrollers, most of which have wireless built in? (which is the exception rather than the rule with other MCU families)

Real development efforts, like providing all the software support necessary to marry two ESP32 class chips together to combine the best of both worlds?

Yeah, but what have they done for us lately?

If you can't do what you need to do with a P4 (and some companion wireless chip if you need wireless), then you're likely asking too much from a microcontroller and need to move to an application processor and all the complexities that that entails.

Why don't you ask your LLM to do some real development efforts instead of worthless fan fiction posts for a dream product that doesn't make sense to build?

[Sprite]

Yeah, but what have they done for us lately?

Thanks for defending Espressif, but in honesty, I don't mind requests like this and other (better-defended) requests for improvement actually have lead to changes in the silicon that was made after. All I ask for are use cases so I can explain why we need to do the work to change our designs. I am sure that OP has some use cases in mind, I can't imagine they made this post without it; it's unfortunately they seem to have left the forums without disclosing them.

[thefury]

There certainly are some super high end microcontrollers elsewhere in the market (Cortex M7 600mhz to 1ghz) that will run circles around the P4 at raw speed, but it's not a mainstream kind of thing that the ESP32 series is so good at serving.

I don't fathom a lot of people needing quad core 800 MHz on a microcontroller platform. I can just about max out the P4 with graphics and audio processing. The further up you go in performance, the more you need a more a operating system to really take advantage of it. It sounds like they basically want you to make a Raspberry Pi competitor with Bluetooth & WiFi that runs full Linux. Like every other Pi clone out there?