More to come here...
The Freedom CPU project has a log here too now :-)
As the FC1 moved from a single-issue superpipeline to a superscalar architecture, instruction decoding faces new challenge. There is the potential for up to 4 instructions to be issued for every clock cycle and this is a lot !
Depending on the actual implementation and object code, 1, 2, 3 or 4 instructions might be issued to any of the 4 parallel execution units ("globules") and this requires significant efforts. Some processors like the Alpha 21264 managed this pretty well, two decades ago, so there is a hope it is doable. Later, the 21264 added OOX but there is no point to go this far for now.
Anyway there is pressure on both the implementations (1, 2, 3 or 4-issue wide versions) and the code generators to create efficient code that runs optimally everywhere.
One option I had considered for a while was to create a pseudo-instruction (or meta-instruction) that indicates the properties of the following instruction bundle. This is pretty under-efficient though and would waste code size for single-issue implementations.
I think I have found a much better option, explained below.
The instructions operate on one globule at a time, over the local 16-registers set. That's 4×2 source address bits for the source operands. The destination address requires 6 bits, for the 4-bits local address and two more bits to indicate the destination globule.
It is necessary to extend this system because globules would not be able to communicate so one source register gets more bits to access the global register set. Of course, an implementation might not be happy with it because it adds a 3rd read port to the register set for the sake of intercommunication and that would still not allow several instructions to read from the same globule at once, but this can be solved with hazard detection logic.
I consider adding another bit to the destination address field, which is a "duplicate instruction" flag. I borrow the idea from #RISC Relay CPU but instead of a serial operation (used in @roelh's architecture as a "repeat for another cycle" order), FC1 does it in parallel, sending the same opcode and register address to a globule pair. This should help with pointer increments or flushing/shuffling registers with less code. Not sure it will be used in the end but there is an opportunity here.
Now, this is where the fun part starts.
A group of parallelly-executable instruction requires no additional information than what is already encoded in the instruction flow. Since I have reduced the addressing range, it's easier to check for data read and write hazards. Only 4 bits per instruction are required to evaluate the data dependencies, 2 of them indicate which globule is addressed.
Therefore, a group is formed by a sequence of instructions where the destination globule (2MSB of destination address) is stricly in increasing sequence.
This is illustrated by the following random sequence where the numbers are grouped on one line according to the above rule.
$ strings /dev/urandom |grep '[1-4]'|sed 's/[^1234]//g' 4 4 13 13 4 1 14 13 2 2 14 2 14 24 4 13 4 14 4 2 2 1 1 14 2 13 13 4
When two consecutive instructions operate on the same destination globule, there is an obvious hazard and the second instruction must be delayed.
The logic that evaluates this condition requires only a few gates and can run very fast.
This first example with random data shows that not all instructions can be grouped (one half can) and the groups don't reach 3 members (in this example at least) so a first implementation can safely use only 2 instruction decoders, not 4. A wider decode unit will require more instruction set bandwidth and a better compiler, but more importantly : source code with enough available ILP.
It's less obvious if some instructions can be paired differently, for example across increasing sequences. If 4 is followed by 1, thy could be both exectuted at the same time if no write hazard is detected, but...Read more »
Lately, I've discussed with @★ STMAN ★who is very knowldgeable about computer security and I'm starting to see old problems with a fresh, new and promising angle. Cue opening theme.
Code is not data and vice versa. FC1 is a Harvard system that does not allow self-modifying code. You can't execute data, so there is no risk from the stack, for example.
Each globule might contain enough space to store registers for 2 or 4 different threads so context switching is pretty immediate. Upon context switching, you can choose to completely switch the context or to keep the current register set (good if you need to provide informations). An instruction provides a mask to clear/wipe any register containing unneeded information for the callee.
Forget about Linux, forget about POSIX. FC1 is aiming at speed and security, which Linux can't garantee. In turn, this frees us from the tyrany of "having to port GCC".
My favorite approach borrows ideas from microkernels. Thank you Hurd people :-) But STMAN convinced me to go even further. Actually, the most critical code (which handles vital parts) should be as short as possible (for speed AND security audit AND non-modifiability in a small ROM). To this end, hardware must support fast and safe support for all the necessary primitives. This is implemented, for example, with specific instructions with inherent checks against user-unaccessible task properties.
You'll hate me, I know, but I swear it's for your own good ;-)
I'll enforce an idea I have tinkered with while thinking about the YASEP: the executable code size of a thread is limited to 64KB.
This does not apply to other things, like data sizes. But honestly, when your executable code reaches 64KB (or 16K instructions), you're definitely not doing a trivial program. The chances of having a bug is not nil. An exploit can hide.
The F-CPU forces you to compartiment your functions and make your application modular. The modules can communicate faster than with a classic architecture so there is no excuse to not enforce safe coding practices.
For example, I borrow other ideas from the YASEP: the InterProcess Call/Entry/Return trinity of instructions lets you call any publicly accessible code entry point safely, for the caller and the callee. With almost no penalty compared to a classicl call/return, and much less risks. Use it, though you won't be able to abuse it ;-)
An old riddle might get finally solved...
How do you transmit data from one thread to another ? Safely yet fast ? I've been researching about this for 14 years now. STMAN offered a clue : avoid aliasing by design. It sounds weird at first but this actually solves many things, since aliasing creates more problems than it solves. Usually, aliasing lets us upload executable code or share a data segment between two or more threads. But if these two cases are solved, there is no need to maintain a risky "feature".
Loading executable code in the instruction space can be achieved with special instructions that can only run in a restricted more. Data sharing is a more complex case but might be solved at last.
More than 15 years have passed since FC0 was drafted. It's a very nice and venerable architecture but time has shown some of its practical limitations. The proposed FC1 addresses most of them, thanks to the experience gained since. The #AMBAP: A Modest Bitslice Architecture Proposal has been the most influential inspiration lately but it's just the melon on the already huge cake of my design explorations. Some things remain the same as in FC0, some things have changed and some have been radically altered.
I keep everything that makes sense and is characteristic of the project.
I simply dropped the load/store instructions altogether. Since I have solved some compiler issues with the YASEP, I can now confidently use the same techniques with F-CPU.
Each scalar globule have half of their registers dedicated to memory access, with the same register-mapped memory principles. With 4 A/D register pairs per globule, there can be two dual-ported SRAM blocks (or cache) per globule.
This changes everything. With 8 pairs of data/address registers, the instantaneous bandwidth is not comparable with standard CPU cores of this class (scalar in-order). The split "globule" architecture means that 4 data can be read and 2 data written. Simultaneously. With only two instructions in one clock cycle. This greatly compensates for the lower number of registers.
This new structure solves many issues I had with the FC0.
First, there was this huge, slow crossbar... Now it's gone and the most usual instructions save one cycle (ADD/SUB/ROP2 have no latency, though the SHL and MUL units are shared and might add some latency).
Then there was the memory system. FC0's was complex and very architecture-dependant.
Also that large register set with 3R2W and out-of-order-completion scheduling : gone.
The FC1 can scale up (as FC0) but also down (32 bits), not just in data width but also in IPC : it's easy to design a decoder for 1 or 2 instructions per cycle, as well as more when the SIMD globules are implemented.
A smaller register set (per computation unit) means that it's possible to implement a larger physical file, either for improved cross-globule latency, or for implementing "multithreading" (barrel CPU).
I believe the FC1 will be faster and more efficient, as well as easier to design.
There is a major difference with YASEP though. It is not practical to perform address register post-updates in the same instruction because
The YASEP has special hardware to update the pointers but the F-CPU is too general for this. The solution is in the organisation and allocation of the registers, with cross-linked addresses and data:
This allows pairs of instructions to execute in a single cycle,...Read more »