C Is Not a Low-Level Language (2018)

This is fine, it’s a term of art and those don’t need to be immediately obvious.

I don’t like the title of this article for that reason, though. Really a better title would be something like “a modern x86 processor is not a PDP-11.” The subtitle is perfect basically.

Edit: also IMO it is not really fair to beat up on C for this, the problem is not really one of low-level-ness. A language that actually exposed the complexity of speculative execution and all that could be pretty high level. It would just be harder to read in a linear text editor, right? We’d be better off drawing the dependency graph or something.

  int main() {
    __m512i vecA = _mm512_setr_epi32(0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15);
    __m512i vecB = _mm512_setr_epi32(0,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75);
    unsigned short mask = 0;

    __asm__ (
        "vp2intersectd %[B], %[A], %%k2"
        : "=@cck2" (mask)
        : [A] "v" (vecA), [B] "v" (vecB)
        : "k3"
    );

    printf("Intersection Mask: 0x%04X\n", mask);
    return 0;
  }

Obviously if you rely on implicit behavior from the compiler to optimize your code you are no longer "low level". But if you can quickly and easily drop into machine-level instructions to provide explicit implementation semantics, and the language indeed makes that relatively simple and easy to do, that sure seems "low level" to me

I actually believe Mojo is the only modern language not designed to pretend every computer is a PDP-11. C has been so successful that many succeeding languages just did C things as a matter of course.

In Mojo, everything is designed with the complexity of the modern computer in mind, and at every stage the programmer has complete control of outcomes. You decide what gets inlined, what gets passed in registers, what gets unrolled, etc. The language has excellent ... ney, probably the best portable SIMD support there is; all integers are built on top of SIMD, and the scalar integers are just SIMD with length of 1. You get complete control of what gets compiled as well due to powerful compile-time programming that is similar to, but more powerful than Zig's (imo, because you can supply a lot more information). While C and Rust allow inline asm, Mojo goes further by letting you supply inline MLIR and LLVM as well, so in situations that warrant it, you can tell the compiler to compile to a specific LLVM intrinsic. The language also does not assume you're compiling to run on just one machine; every modern computer is heterogeneous by nature and may contain multiple programmable units, so the compilation pipeline is designed to allow compiling certain parts of code for one target and other parts for other targets... as one compilation unit.

    #define array(T, N) struct array##T##N { T value[N]; }

    void copy(array(int, 32)* x, array(int, 32)* y) {
        *x = *y;
    }

    int main() {
        array(int, 32) x;
        array(int, 32) y = { 1, 2, 3, 4 };
        copy(&x, &y);
    }

https://en.wikipedia.org/wiki/C29_(C_standard_revision)

---

C is a general-purpose programming language with features economy of expression, modern flow control and data structures, and a rich set of operators. C is not a "very high level" language, nor a "big" one, and is not specialized to any particular area of application.

But its absence of restrictions and its generality make it more convenient and effective for many tasks than supposedly more powerful languages.

Current ISAs have so much machinery underneath that it’s hard to tell when you’re talking to the iron and when you’re talking to the microcode

You can still argue that Forth on a Forth CPU is a genuinely low-level language. :)

I don’t think byte code qualifies as human readable but it is closer to the metal obviously

So the question is if we really want lower level ISAs. Probably not?

There are many ways in which our current ISAs are actually thoughtfully optimized for superscalar out-of-order processors. Just look at all of the big differences from 32 bit arm to 64 bit arm, which all exist to make execution faster on superscalar processors.

And yet they are still perfectly implementable in cheap microcontrollers. The Cortex-A53, available in boards for a little over $15, is a simple 2-wide perfectly in-order design, without a physical register page beyond the ISA register. Basically, it is a simple Pentium-type chip.

The Apple M chips are some of the most impressive feats of out-of-order superscalar micro-engineering ever. And yet both of these can run the same software with the same ISA. This is enormously valuable.

I fail to see how any sort of much lower level access to the machine would be portable across price ranges and microarchitecture generations. I also fail to see how it would provide a non-trivial speedup over C code pattern recommendations and targeted extensions (eg. vector extensions).

IMHO, C is the lowest level a procedural programming language can get.

https://ulanguage.org

What level would you say this language was? Is it a low-level systems language, or is it also usable for writing web sites?

This is embarrassingly bad.

pre- and post- increment operators cleanly lined up with... the programmer's conceptualization and objectives--the index is/was frequently used in other contexts than loop bounds and indexing. if that's not your conceptualization, don't use that operator. whether you are on a PDP-11 makes no difference.