It is not just a way of writing ring buffers. It's a way of implementing concurrent non-blocking single-reader single-writer atomic ring buffers with only atomic load and store (and memory barriers).
The author says that non-power-of-two is not possible, but I'm pretty sure it is if you use a conditional instead of integer modulus.
I first learnt of this technique from Phil Burk, we've been using it in PortAudio forever. The technique is also widely known in FPGA/hardware circles, see:
"Simulation and Synthesis Techniques for Asynchronous
FIFO Design", Clifford E. Cummings, Sunburst Design, Inc.
I think unfortunately we sometimes ascribe to powers of two supernatural powers that are really about caches being built in powers of two.
Intel is still 64 byte cache lines as they have been for quite a long time but they also do some shenanigans on the bus where they try to fetch two lines when you ask for one. So there’s ostensibly some benefit of aligning data particularly on linear scans to 128 byte alignment for cold cache access.
But there's a reason that caches are always sized in powers of two as well, and that same reason is applicable to high-performance ring buffers: Division by powers of two is easy and easy is fast. It's reliably a single cycle, compared to division by arbitrary 32bit integers which can be 8-30 cycles depending on CPU.
Also, there's another benefit downstream of that one: Powers of two work as a schelling point for allocations. Picking powers of two for resizable vectors maximizes "good luck" when you malloc/realloc in most allocators, in part because e.g. a buddy allocator is probably also implemented using power-of-two allocations for the above reason, but also for the plain reason that other users of the same allocator are more likely to have requested power of two allocations. Spontaneous coordination is a benefit all its own. Almost supernatural! :)
A couple of the comments to the article suggest using 64-bit numbers, which is exactly the right solution. 2^64 nanoseconds=584.55 years - overflow is implausible for any realistic use case. Even pathological cases will struggle to induce wraparound at a human timescale.
(People will probably moan at the idea of restarting the process periodically rather than fixing the issue properly, but when the period would be something like 50 years I don't think it's actually a problem.)
> using 64-bit numbers, which is exactly the right solution
On a 64-bit platform, sure. When you're working on ring buffers with an 8-bit microcontroller, using 64-bit numbers would be such an overhead that nobody would even think of it.
Non-power-of-two is only really feasible of the total number of inserts will fit in your post/ack counters. Otherwise you have to implement overflow manually which may or may not be possible to do with the available atomic primitives on your architecture.
I first encountered this structure at a summer internship at a company making data switches.
Regardless of correctness, as a DSP dork I really identified with the question: "What kind of a monster would make a non-power of two ring anyway?" I remember thinking similarly when requesting a power of two buffer from a 3rd party audio hardware device and having it correct to a nearby non-power of two. Latency adding ringbuffer to the rescue.
> The author says that non-power-of-two is not possible, but I'm pretty sure it is if you use a conditional instead of integer modulus.
I don't see why it wouldn't be, it's just computationally expensive to take the modulo value of the pointer rather than just masking off the appropriate number of bits.
> It is not just a way of writing ring buffers. It's a way of implementing concurrent non-blocking single-reader single-writer atomic ring buffers with only atomic load and store (and memory barriers).
That may or may not be part of the actual definition of a ring buffer, but every ring buffer I have written had those goals in mind.
And the first method mentioned in the article fully satisfies this, except for the one missing element mentioned by the author. Which in practice, often is not only not a problem, but simplifies the logic so much that you make up for it in code space.
Or, for example, say you have a 256 character buffer. You really, really want to make sure you don't waste that one character. So you increase the size of your indices. Now they are 16 bits each instead of 8 bits, so you've gained the ability to store 256 bytes by having 260 bytes of data, rather than 255 bytes by having 258 bytes of data.
Obviously, if you have a 64 byte buffer, there is no such tradeoff, and the third example wins (but, whether your are doing the first or third example, you still have to mask the index data off at some point, whether it's on an increment or a read).
> The author says that non-power-of-two is not possible, but I'm pretty sure it is if you use a conditional instead of integer modulus.
There's "not possible" and then "not practical."
Sure, you could have a 50 byte buffer, but now, if your indices are ever >= 50, you're subtracting 50 before accessing the array, so this will increase the code space (and execution time).
> The [index size > array size] technique is also widely known in FPGA/hardware circles
Right, but in those hardware circles, power-of-two _definitely_ matters. You allocate exactly one extra bit for your pointers, and you never bother manually masking them or taking a modulo or anything like that -- they simply roll over.
If you really, really need to construct something like a 6 entry FIFO in hardware, then you have techniques available to you that mere mortal programmers could not use efficiently at all. For example, you could construct a drop-through FIFO, where every element traverses every storage slot (with a concomitant increase in minimum latency to 6 clock cycles), or you could construct 4 bit indices that counted 0-1-2-3-4-5-8-9-10-11-12-13-0-1-2 etc.
Most ring buffers, hardware or software, are constructed as powers of two, and most ring buffers either (a) have so much storage that one more element wouldn't make any difference, or (b) have the ability to apply back pressure, so one more element wouldn't make any difference.
It's a silly offhand remark at the end of the article, but anybody who is genuinely interested in whether they've been tying their shoes wrong will enjoy Ian's shoelace site: https://www.fieggen.com/shoelace/
> So there I was, implementing a one element ring buffer. Which, I'm sure you'll agree, is a perfectly reasonable data structure.
It is, but, IMO, shouldn’t use the code for “a n-element ring buffer, with n set to 1”, similarly to how an array of booleans in many languages shouldn’t be implemented as “an arrayof Foos, with Foo set to bool”.
C++ has std::bitset and std::vector and Java similarly has BitSet and Array because using the generic code for arrays of bits is too wasteful.
Similarly, a one-element ring buffer is either full or it is empty. Why use two indexes to encode a single boolean?
> C++ has std::bitset and std::vector and Java similarly has BitSet and Array because using the generic code for arrays of bits is too wasteful.
Rather infamously, C++ tried to be clever here and std::vector<bool> is not just a vector-of-bools but instead a totally different vector-ish type that lacks many of the important properties of every other instantiation of std::vector. Yes, a lot of the time you want the space efficiency of a dynamic bitset, rather than wasting an extra 7 bits per element. But also quite often you do want the behavior of a "real" std::vector for true/false values, and then you have to work around it manually (usually via std::vector<uint8_t> or similar) to get the expected behavior.
It was for a dynamically growing ring buffer that also did short-object optimization. The natural implementation was to have the capacity and the offsets stored in fixed locations and with a fixed width, and have the variable part be a union of pointer or inline byte buffer.
Depending on the element width, you'd have space for different amounts of data in the inline buffer. Sometimes 1, sometimes a few more. Specializing for a one-element inline buffer would be quite complex with limited gains.
In retrospect trying to use that as a running gag for the blog post did not work well without actually giving the full context, but the full context would have been a distraction.
That depends on your architecture and access pattern. In case of sequential access, packed bools may perform better due to arithmetic being usually way cheaper than memory operations.
The 2*bufSize gives you an extra bit (beyond representing bufSize) that lets you disambiguate empty vs full. And if it is a constant power of two (e.g. via C++ template), then you can see how this just compiles into a bitmask instead, like the author's version. You read and write the buffer at (rdPtr%bufSize) and (wrPtr%bufSize) respectively.
Mostly Type 1 and overflow is a diagnostic log at most. Losing all stale unprocessed data and leaving a ready empty buffer behind is often the desired outcome.
Type 3 is probably banned on most codebases because of the integer overflow.
> Losing all stale unprocessed data and leaving a ready empty buffer behind is often the desired outcome.
Yeah, the Type 3 example could conceivably make it so that you intermix old and new data if you overflow, rather than just dumping a whole buffer.
Especially when your full() function checks for exact equality, like the one in the article does.
And if you remove the asserts, and then somehow underflow? God help you. You'll be pulling 4 billion entries you never actually stored out of the buffer, just repeating previously stored garbage over and over.
> Type 3 is probably banned on most codebases because of the integer overflow.
Not only this, but the purported code reduction benefits associated with type 3 are only superficial, and won't actually appear in any assembly listing.
Signed integer overflow is definitely a problem however. Something as simple as incrementing a user-provided int can lead to UB (if the user provides INT_MAX).
And yet here I sit, writing ring buffers, and never thinking about this idea. Probably because of the power of two issue. Which isn't actually a problem because as he points out, who would do that? But it makes me think that it's a restriction that it just isn't.
But in all honesty, look for more embedded jobs, then. We can certainly use the help.
What do you work with (if you don't mind answering)? I'm looking for a change and like low-level stuff about as much as I like any other level. I've done some cycle-accurate NES emulation and VM implementation stuff - I'm not much of a DSA guy but performance and efficiency appeal to me.
Jokes on me, when I need them, I don't feel like writing them so I just pick up an old one and tweak it. Or just tell Claude to build me one and it one shots it.
Not only that, but you can also always form a normal (ptr, size) slice reference to any piece of the ring buffer, even when it wraps. This is really helpful for Eigen arrays that you need to rotate.
Technically each side needs an index plus a single bit. The bit is a counter, you increment it on every wrap. It overflows, but this is correct, we only need the last bit. Initially it is 0. By comparing the indexes and the bit you tell apart all cases and do not lose an entry.
(I think this was published in one of Llang's papers but in a rather obscure language.)
Which does that and that's what Llang suggests as well, I remember 2 in some power in his formula. I myself find the separate one-bit counter easier to understand: each side counts pages, but they don't need the full number, only the difference between their counters and the difference can be at most one, so one bit is sufficient. If the counters are same, the actors are on the same page, if they are different, then the writer is one page ahead.
Readit News
The author says that non-power-of-two is not possible, but I'm pretty sure it is if you use a conditional instead of integer modulus.
I first learnt of this technique from Phil Burk, we've been using it in PortAudio forever. The technique is also widely known in FPGA/hardware circles, see:
"Simulation and Synthesis Techniques for Asynchronous FIFO Design", Clifford E. Cummings, Sunburst Design, Inc.
https://twins.ee.nctu.edu.tw/courses/ip_core_04/resource_pdf...
Intel is still 64 byte cache lines as they have been for quite a long time but they also do some shenanigans on the bus where they try to fetch two lines when you ask for one. So there’s ostensibly some benefit of aligning data particularly on linear scans to 128 byte alignment for cold cache access.
Also, there's another benefit downstream of that one: Powers of two work as a schelling point for allocations. Picking powers of two for resizable vectors maximizes "good luck" when you malloc/realloc in most allocators, in part because e.g. a buddy allocator is probably also implemented using power-of-two allocations for the above reason, but also for the plain reason that other users of the same allocator are more likely to have requested power of two allocations. Spontaneous coordination is a benefit all its own. Almost supernatural! :)
(People will probably moan at the idea of restarting the process periodically rather than fixing the issue properly, but when the period would be something like 50 years I don't think it's actually a problem.)
I think you have that backwards. If something needs to be done every week, it will get done every week. That's not a problem.
If something needs to be done every fifty years, you'll be lucky if it happens once.
On a 64-bit platform, sure. When you're working on ring buffers with an 8-bit microcontroller, using 64-bit numbers would be such an overhead that nobody would even think of it.
I first encountered this structure at a summer internship at a company making data switches.
I don't see why it wouldn't be, it's just computationally expensive to take the modulo value of the pointer rather than just masking off the appropriate number of bits.
The problem is incrementing past the index integer type limit.
Consider a simple example with ring buffer size 9, and 16bit indices:
When you increment the write index from 0xffff to 0, your "masked index" jumps from 6 (0xffff % 9) to 0 (instead of 7).
There is no elegant fix that I'm aware of (using a very wide index type, like possibly a uint64, is extremely non-elegant).
That may or may not be part of the actual definition of a ring buffer, but every ring buffer I have written had those goals in mind.
And the first method mentioned in the article fully satisfies this, except for the one missing element mentioned by the author. Which in practice, often is not only not a problem, but simplifies the logic so much that you make up for it in code space.
Or, for example, say you have a 256 character buffer. You really, really want to make sure you don't waste that one character. So you increase the size of your indices. Now they are 16 bits each instead of 8 bits, so you've gained the ability to store 256 bytes by having 260 bytes of data, rather than 255 bytes by having 258 bytes of data.
Obviously, if you have a 64 byte buffer, there is no such tradeoff, and the third example wins (but, whether your are doing the first or third example, you still have to mask the index data off at some point, whether it's on an increment or a read).
> The author says that non-power-of-two is not possible, but I'm pretty sure it is if you use a conditional instead of integer modulus.
There's "not possible" and then "not practical."
Sure, you could have a 50 byte buffer, but now, if your indices are ever >= 50, you're subtracting 50 before accessing the array, so this will increase the code space (and execution time).
> The [index size > array size] technique is also widely known in FPGA/hardware circles
Right, but in those hardware circles, power-of-two _definitely_ matters. You allocate exactly one extra bit for your pointers, and you never bother manually masking them or taking a modulo or anything like that -- they simply roll over.
If you really, really need to construct something like a 6 entry FIFO in hardware, then you have techniques available to you that mere mortal programmers could not use efficiently at all. For example, you could construct a drop-through FIFO, where every element traverses every storage slot (with a concomitant increase in minimum latency to 6 clock cycles), or you could construct 4 bit indices that counted 0-1-2-3-4-5-8-9-10-11-12-13-0-1-2 etc.
Most ring buffers, hardware or software, are constructed as powers of two, and most ring buffers either (a) have so much storage that one more element wouldn't make any difference, or (b) have the ability to apply back pressure, so one more element wouldn't make any difference.
Deleted Comment
It is, but, IMO, shouldn’t use the code for “a n-element ring buffer, with n set to 1”, similarly to how an array of booleans in many languages shouldn’t be implemented as “an arrayof Foos, with Foo set to bool”.
C++ has std::bitset and std::vector and Java similarly has BitSet and Array because using the generic code for arrays of bits is too wasteful.
Similarly, a one-element ring buffer is either full or it is empty. Why use two indexes to encode a single boolean?
Rather infamously, C++ tried to be clever here and std::vector<bool> is not just a vector-of-bools but instead a totally different vector-ish type that lacks many of the important properties of every other instantiation of std::vector. Yes, a lot of the time you want the space efficiency of a dynamic bitset, rather than wasting an extra 7 bits per element. But also quite often you do want the behavior of a "real" std::vector for true/false values, and then you have to work around it manually (usually via std::vector<uint8_t> or similar) to get the expected behavior.
Depending on the element width, you'd have space for different amounts of data in the inline buffer. Sometimes 1, sometimes a few more. Specializing for a one-element inline buffer would be quite complex with limited gains.
In retrospect trying to use that as a running gag for the blog post did not work well without actually giving the full context, but the full context would have been a distraction.
Notably, this is not the case. C++ std::vector is specialised for bools to pack bits into words, causing an untold array (heh) of headaches.
And "wasteful" is doing a lot of lifting here. In terms of memory usage? Yes. In terms of CPU? The other way around.
That depends on your architecture and access pattern. In case of sequential access, packed bools may perform better due to arithmetic being usually way cheaper than memory operations.
For non-power of two, just checked our own very old circular byte buffer library code and using the notation from this article, it is:
The 2*bufSize gives you an extra bit (beyond representing bufSize) that lets you disambiguate empty vs full. And if it is a constant power of two (e.g. via C++ template), then you can see how this just compiles into a bitmask instead, like the author's version. You read and write the buffer at (rdPtr%bufSize) and (wrPtr%bufSize) respectively.It feels like 90% swe jobs these days are about writing CRUD wrappers.
Mostly Type 1 and overflow is a diagnostic log at most. Losing all stale unprocessed data and leaving a ready empty buffer behind is often the desired outcome.
Type 3 is probably banned on most codebases because of the integer overflow.
Yeah, the Type 3 example could conceivably make it so that you intermix old and new data if you overflow, rather than just dumping a whole buffer.
Especially when your full() function checks for exact equality, like the one in the article does.
And if you remove the asserts, and then somehow underflow? God help you. You'll be pulling 4 billion entries you never actually stored out of the buffer, just repeating previously stored garbage over and over.
> Type 3 is probably banned on most codebases because of the integer overflow.
Not only this, but the purported code reduction benefits associated with type 3 are only superficial, and won't actually appear in any assembly listing.
Signed integer overflow is definitely a problem however. Something as simple as incrementing a user-provided int can lead to UB (if the user provides INT_MAX).
But in all honesty, look for more embedded jobs, then. We can certainly use the help.
Makes the code trivial
I've been writing ring buffers wrong all these years - https://news.ycombinator.com/item?id=13175832 - Dec 2016 (167 comments)
(I think this was published in one of Llang's papers but in a rather obscure language.)