Emulating the RPI PIOs instead of the TI PRUs is really a miss.
The PRUs really get a bunch right. Very specifically, the ability to broadside dump the ENTIRE register file in a single cycle from one PRU to the other is gigantic. It's the single thing that allows you to transition the data from a hard real-time domain to a soft real-time domain and enables things like the industrial Ethernet protocols or the BeagleLogic, for example.
What are your thoughts on efficiency? BIO vs PIO implementing, say, 68k 16-bit-wide bus slave. I know i can support 66MHz 68K bus clock with PIO at 300MHz. How much clock speed would BIO need?
It depends a lot upon where the processing is happening. For example, you could do something where all the data is pre-processed and you're just blasting bits into a GPIO register with a pair of move instructions. In which case you could get north of 60MHz, but I think that's sort of cheating - you'll run out of pre-processed data pretty quickly, and then you have to take a delay to generate more data.
The 25MHz number I cite as the performance expectation is "relaxed": I don't want to set unrealistic expectations on the core's performance, because I want everyone to have fun and be happy coding for it - even relatively new programmers.
However, with a combination of overclocking and optimization, higher speeds are definitely on the horizon. Someone on the Baochip Discord thought up a clever trick I hadn't considered that could potentially get toggle rates into the hundreds of MHz's. So, there's likely a lot to be discovered about the core that I don't even know about, once it gets into the hands of more people.
I specified slave specifically because slave is a LOT harder. Master is always easy. Waiting for someone else’s clock and then capturing and replying asap is the hard part. Especially if as a slave you need to simulate a read.
On rp2350 it is pio (wait for clock) -> pio (read address bus) -> dma (addr into lower bits of dma source for next channel) -> dma (Data from SRAM to PIO) -> pio (write data to data bus) chain and it barely keeps up.
I loved this article and had wanted to play with PIO for a long time (or at least, learn from it through playing!).
One thing jumped out here - I assumed CISC inside PIO had a mental model of "one instruction by cycle" and thus it was pretty easy to reason about the underlying machine (including any delay slots etc...).
For this RISC model using C, we are now reasoning about compiled code which has a somewhat variable instruction timing (1-3 cycles) and that introduces an uncertainty - the compiler and understanding its implementation.
I think this means that the PIO is timing-first, as timing == waveform where BIO is clarity-first with C as the expression and then explicit hardware synchronization.
I like both models! I am wondering about the quantum delays however that are being used to set the deadlines - here, human derived wait delays are utilized knowledge of the compiled instructions to set the timing.
Might there not be a model of 'preparing the next hardware transaction' and then 'waiting for an external synchronization' such as an external signal or internal clock, so we don't need to count the instruction cycles so precisely. On the external signal side, I guess the instruction is 'wait for GPIO change' or something, so the value is immediately ready (int i = GPIO_read_wait_high(23) or something) and the external one is doing the same, but synchronizing (GPIO_write_wait_clock( 24, CLOCK_DEF)) as an alternative to the explicit quantum delays.
This might be a shadow register / latch model in more generic terms - prep the work in shadow, latch/commit on trigger.
The idea of the wait-to-quantum register is that it gets you out of cycle-counting hell at the expense of sacrificing a few cycles as rounding errors. But yes, for maximum performance you would be back to cycle counting.
That being said - one nice thing about the BIO being open source is you can run the verilog design in Verilator. The simulation shows exactly how many cycles are being used, and for what. So for very tight situations, the open source RTL nature of the design opens up a new set of tools that were previously unavailable to coders. You can see an example of what it looks like here: https://baochip.github.io/baochip-1x/ch00-00-rtl-overview.ht...
Of course, there's a learning curve to all new tools, and Verilator has a pretty steep curve in particular. But, I hope people give the Verilator simulations a try. It's kind of neat just to be able to poke around inside a CPU and see what it's thinking!
This is actually super cool, you can use those as both math accelerators and as io, and them being in lockstep you can kind of use them as int only shader units. I don't know how this is useful yet.
Btw I am curious what about edge cases. Maybe I have missed that from the article but what is the size of the FIFO?
Or the more dangerous part that is you have complex to determine timing now for complex cases like each reqd from FIFO is and ISR and you have until the next read from the FIFO amount of instructions otherwise you would stall the system and that looks to me too hard to debug.
I appreciate the intro, motivation and comparison to the PIO of the RP2040/2350. How would this compare to the (considerably older, slower, but more flexible) Parallax P8X32A ("Propeller")?
IIRC the Propeller is an eight thread barrel CPU with the same number of pipeline stages. So it "retires" just one instruction per cycle. All PIO state machines can run every cycle so they should be considered very small CPU cores. You can think of them as channel I/O co-processors for a microcontroller instead of a mainframe.
> Above is the logic path isolated as one of the longest combination paths in the design, and below is a detailed report of what the cells are.
which is an argument that "fpga_pio" is badly implemented or that PIO is unsuitable for FPGA impls. Real silicon does not need to use a shitton of LUT4s to implement this logic and it can be done much more efficiently and closes timing at higher clocks (as we know since PIO will run near a GHz)
As a side note about speed comparisons - please keep in mind the faster speeds cited for the PIO are achieved through overclocking.
The BIO should also be able to overclock. It won't overclock as well as the PIO, for sure - the PIO stores its code in flip-flops, which performance scales very well with elevated voltages. The BIO uses a RAM macro, which is essentially an analog part at its heart, and responds differently to higher voltages.
That being said, I'm pretty confident that the BIO can run at 800MHz for most cases. However, as the manufacturer I have to be careful about frequency claims. Users can claim a warranty return on a BIO that fails to run at 700MHz, but you can't do the same for one that fails to run at 800MHz - thus whenever I cite the performance of the BIO, I always stick it at the number that's explicitly tested and guaranteed by the manufacturing process, that is, 700MHz.
Third-party overclockers can do whatever they want to the chip - of course, at that point, the warranty is voided!
PIO is unsuitable for FPGA impls, that's what the article says.
> If you’re thinking about using it in an FPGA, you’d be better off skipping the PIO and just implementing whatever peripherals you want directly using RTL.
Yea, I think the point is that if you’re implementing in FPGA in any case, a dedicated state machine is going to be a lot smaller than PIO or BIO. But if you’re making a standard part with hardcoded functionality then BIO is going to be smaller than PIO.
Yes, my point is that the article throws a lot of shade at PIO while the real issue is that the author is trying to shove a third-party FPGA reimpl of it into a place it never belonged. PIO itself is a perfectly good design for what it does and where it does it.
Actually, the PIO does what it does very well! There is no "worse" or "better" - just different.
Because it does what it does so well, I use the PIO as the design study comparison point. This requires taking a critical view of its architecture. Such a review doesn't mean its design is bad - but we try to take it apart and see what we can learn from it. In the end, there are many things the PIO can do that the BIO can't do, and vice-versa. For example, the BIO can't do the PIO's trick of bit-banging DVI video signals; but, the PIO isn't going to be able to protocol processing either.
In terms of area, the larger area numbers hold for both an ASIC flow as well as the FPGA flow. I ran the design through both sets of tools with the same settings, and the results are comparable. However, it's easier to share the FPGA results because the FPGA tools are NDA-free and everyone can replicate it.
That being said, I also acknowledge in the article that it's likely there are clever optimizations in the design of the actual PIO that I did not implement. Still, barrel shifters are a fairly expensive piece of hardware whether in FPGA or in ASIC, and the PIO requires several of them, whereas the BIO only has one. The upshot is that the PIO can do multiple bit-shifts in a single clock cycle, whereas the BIO requires several cycles to do the same amount of bit-shifting. Again, neither good or bad - just different trade-offs.
Thanks btw for saying clearly that BIO is not suitable for DVI output. I was curious about this and was planning to ask on social media.
I've done some fun stuff in PIO, in particular the NRZI bit stuffing for USB (12Mbps max). That's stretching it to its limit. Clearly there will be things for which BIO is much better.
I suspect that a variant of BIO could probably do DVI by optimizing for that specific use case (in particular, configuring shifters on the output FIFO), but I'm not sure it's worth the lift.
> The build script compiles C code down to a clang intermediate assembly, which is then handed off to a Python script that translates it into a Rust macro which is checked into Xous as a buildable artifact using its pure-Rust toolchain.
Ah yes, the good ol “we solved the C problem by turning it into four other problems” pipeline
Hello again HN, I'm bunnie! Unfortunately, time zones strike again...I'll check back when I can, and respond to your questions.
Have you looked at TI's PRU at all?
Emulating the RPI PIOs instead of the TI PRUs is really a miss.
The PRUs really get a bunch right. Very specifically, the ability to broadside dump the ENTIRE register file in a single cycle from one PRU to the other is gigantic. It's the single thing that allows you to transition the data from a hard real-time domain to a soft real-time domain and enables things like the industrial Ethernet protocols or the BeagleLogic, for example.
very cool. tiny processors everywhere. but be nice to PIO. PIO is good :)
Agreed! The PIO is great at what it does. I drew a lot of inspiration from it.
What are your thoughts on efficiency? BIO vs PIO implementing, say, 68k 16-bit-wide bus slave. I know i can support 66MHz 68K bus clock with PIO at 300MHz. How much clock speed would BIO need?
It depends a lot upon where the processing is happening. For example, you could do something where all the data is pre-processed and you're just blasting bits into a GPIO register with a pair of move instructions. In which case you could get north of 60MHz, but I think that's sort of cheating - you'll run out of pre-processed data pretty quickly, and then you have to take a delay to generate more data.
The 25MHz number I cite as the performance expectation is "relaxed": I don't want to set unrealistic expectations on the core's performance, because I want everyone to have fun and be happy coding for it - even relatively new programmers.
However, with a combination of overclocking and optimization, higher speeds are definitely on the horizon. Someone on the Baochip Discord thought up a clever trick I hadn't considered that could potentially get toggle rates into the hundreds of MHz's. So, there's likely a lot to be discovered about the core that I don't even know about, once it gets into the hands of more people.
I specified slave specifically because slave is a LOT harder. Master is always easy. Waiting for someone else’s clock and then capturing and replying asap is the hard part. Especially if as a slave you need to simulate a read.
On rp2350 it is pio (wait for clock) -> pio (read address bus) -> dma (addr into lower bits of dma source for next channel) -> dma (Data from SRAM to PIO) -> pio (write data to data bus) chain and it barely keeps up.
I loved this article and had wanted to play with PIO for a long time (or at least, learn from it through playing!).
One thing jumped out here - I assumed CISC inside PIO had a mental model of "one instruction by cycle" and thus it was pretty easy to reason about the underlying machine (including any delay slots etc...).
For this RISC model using C, we are now reasoning about compiled code which has a somewhat variable instruction timing (1-3 cycles) and that introduces an uncertainty - the compiler and understanding its implementation.
I think this means that the PIO is timing-first, as timing == waveform where BIO is clarity-first with C as the expression and then explicit hardware synchronization.
I like both models! I am wondering about the quantum delays however that are being used to set the deadlines - here, human derived wait delays are utilized knowledge of the compiled instructions to set the timing.
Might there not be a model of 'preparing the next hardware transaction' and then 'waiting for an external synchronization' such as an external signal or internal clock, so we don't need to count the instruction cycles so precisely. On the external signal side, I guess the instruction is 'wait for GPIO change' or something, so the value is immediately ready (int i = GPIO_read_wait_high(23) or something) and the external one is doing the same, but synchronizing (GPIO_write_wait_clock( 24, CLOCK_DEF)) as an alternative to the explicit quantum delays.
This might be a shadow register / latch model in more generic terms - prep the work in shadow, latch/commit on trigger.
Anyway, great work Bunnie!
The idea of the wait-to-quantum register is that it gets you out of cycle-counting hell at the expense of sacrificing a few cycles as rounding errors. But yes, for maximum performance you would be back to cycle counting.
That being said - one nice thing about the BIO being open source is you can run the verilog design in Verilator. The simulation shows exactly how many cycles are being used, and for what. So for very tight situations, the open source RTL nature of the design opens up a new set of tools that were previously unavailable to coders. You can see an example of what it looks like here: https://baochip.github.io/baochip-1x/ch00-00-rtl-overview.ht...
Of course, there's a learning curve to all new tools, and Verilator has a pretty steep curve in particular. But, I hope people give the Verilator simulations a try. It's kind of neat just to be able to poke around inside a CPU and see what it's thinking!
You could always get around the compiler uncertainty using a RV assembler, no? These IO programs are not long or terribly sophisticated.
This is actually super cool, you can use those as both math accelerators and as io, and them being in lockstep you can kind of use them as int only shader units. I don't know how this is useful yet.
Btw I am curious what about edge cases. Maybe I have missed that from the article but what is the size of the FIFO?
Or the more dangerous part that is you have complex to determine timing now for complex cases like each reqd from FIFO is and ISR and you have until the next read from the FIFO amount of instructions otherwise you would stall the system and that looks to me too hard to debug.
I appreciate the intro, motivation and comparison to the PIO of the RP2040/2350. How would this compare to the (considerably older, slower, but more flexible) Parallax P8X32A ("Propeller")?
The Propeller 2 would be an interesting comparison as well, with it's own smart pins playing a similar role.
IIRC the Propeller is an eight thread barrel CPU with the same number of pipeline stages. So it "retires" just one instruction per cycle. All PIO state machines can run every cycle so they should be considered very small CPU cores. You can think of them as channel I/O co-processors for a microcontroller instead of a mainframe.
Thanks for making this a blog post!
Have some on the way! Can't wait!
> Above is the logic path isolated as one of the longest combination paths in the design, and below is a detailed report of what the cells are.
which is an argument that "fpga_pio" is badly implemented or that PIO is unsuitable for FPGA impls. Real silicon does not need to use a shitton of LUT4s to implement this logic and it can be done much more efficiently and closes timing at higher clocks (as we know since PIO will run near a GHz)
As a side note about speed comparisons - please keep in mind the faster speeds cited for the PIO are achieved through overclocking.
The BIO should also be able to overclock. It won't overclock as well as the PIO, for sure - the PIO stores its code in flip-flops, which performance scales very well with elevated voltages. The BIO uses a RAM macro, which is essentially an analog part at its heart, and responds differently to higher voltages.
That being said, I'm pretty confident that the BIO can run at 800MHz for most cases. However, as the manufacturer I have to be careful about frequency claims. Users can claim a warranty return on a BIO that fails to run at 700MHz, but you can't do the same for one that fails to run at 800MHz - thus whenever I cite the performance of the BIO, I always stick it at the number that's explicitly tested and guaranteed by the manufacturing process, that is, 700MHz.
Third-party overclockers can do whatever they want to the chip - of course, at that point, the warranty is voided!
PIO is unsuitable for FPGA impls, that's what the article says.
> If you’re thinking about using it in an FPGA, you’d be better off skipping the PIO and just implementing whatever peripherals you want directly using RTL.
Yea, I think the point is that if you’re implementing in FPGA in any case, a dedicated state machine is going to be a lot smaller than PIO or BIO. But if you’re making a standard part with hardcoded functionality then BIO is going to be smaller than PIO.
Yes, my point is that the article throws a lot of shade at PIO while the real issue is that the author is trying to shove a third-party FPGA reimpl of it into a place it never belonged. PIO itself is a perfectly good design for what it does and where it does it.
Actually, the PIO does what it does very well! There is no "worse" or "better" - just different.
Because it does what it does so well, I use the PIO as the design study comparison point. This requires taking a critical view of its architecture. Such a review doesn't mean its design is bad - but we try to take it apart and see what we can learn from it. In the end, there are many things the PIO can do that the BIO can't do, and vice-versa. For example, the BIO can't do the PIO's trick of bit-banging DVI video signals; but, the PIO isn't going to be able to protocol processing either.
In terms of area, the larger area numbers hold for both an ASIC flow as well as the FPGA flow. I ran the design through both sets of tools with the same settings, and the results are comparable. However, it's easier to share the FPGA results because the FPGA tools are NDA-free and everyone can replicate it.
That being said, I also acknowledge in the article that it's likely there are clever optimizations in the design of the actual PIO that I did not implement. Still, barrel shifters are a fairly expensive piece of hardware whether in FPGA or in ASIC, and the PIO requires several of them, whereas the BIO only has one. The upshot is that the PIO can do multiple bit-shifts in a single clock cycle, whereas the BIO requires several cycles to do the same amount of bit-shifting. Again, neither good or bad - just different trade-offs.
Thanks btw for saying clearly that BIO is not suitable for DVI output. I was curious about this and was planning to ask on social media.
I've done some fun stuff in PIO, in particular the NRZI bit stuffing for USB (12Mbps max). That's stretching it to its limit. Clearly there will be things for which BIO is much better.
I suspect that a variant of BIO could probably do DVI by optimizing for that specific use case (in particular, configuring shifters on the output FIFO), but I'm not sure it's worth the lift.
It didn't read that way, to me.
3 comments on this, from 2d ago, https://news.ycombinator.com/item?id=47469744
> The build script compiles C code down to a clang intermediate assembly, which is then handed off to a Python script that translates it into a Rust macro which is checked into Xous as a buildable artifact using its pure-Rust toolchain.
Ah yes, the good ol “we solved the C problem by turning it into four other problems” pipeline