On 14 Jun 2026 17:22:22 GMT, ram@zedat.fu-berlin.de (Stefan Ram)
wrote:
[quoting a chatbot]

| Or, consider a loop:
|
| for (int i = 0; i < n; i++) {
| arr[i] = 0;
| }
|
| If out-of-bounds array access had defined behavior, the compiler would
| have to insert a bounds check ("if (i >= array_length)") on every single
| iteration. Because out-of-bounds access is UB, the compiler can assume
| n is always within bounds. This allows it to unroll the loop,
| vectorize it using SIMD instructions, and process 8 or 16 elements per
| CPU cycle, yielding massive performance gains.

Not so. If n is <= the length of arr (or the usable portion of an
array that arr is a pointer to), and also <= MAX_INT so the loop is well-defined, then no check is needed on each assignment, only before
starting the loop.

Well, there are some tests that can be taken out of loops (as
in Java), but other tests can't.

Not in the language defined by JLS and JVMS; those check every array subscripting operation, period. OTOH an implementation can recognize
this pattern and hoist the check in JITted or otherwise postcompiled
code, and I think HotSpot in particular does.

--- PyGate Linux v1.5.19
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)

dave_thompson_2@comcast.net writes:

On 14 Jun 2026 17:22:22 GMT, ram@zedat.fu-berlin.de (Stefan Ram)
wrote:
[quoting a chatbot]

| Or, consider a loop:
|
| for (int i = 0; i < n; i++) {
| arr[i] = 0;
| }
|
| If out-of-bounds array access had defined behavior, the compiler would
| have to insert a bounds check ("if (i >= array_length)") on every single >> | iteration. Because out-of-bounds access is UB, the compiler can assume
| n is always within bounds. This allows it to unroll the loop,
| vectorize it using SIMD instructions, and process 8 or 16 elements per
| CPU cycle, yielding massive performance gains.

Not so. If n is <= the length of arr (or the usable portion of an
array that arr is a pointer to), and also <= MAX_INT so the loop is well-defined, then no check is needed on each assignment, only before starting the loop.

Maybe. The first part of the quoted paragraph (apparently from
a chatbot) discusses the consequences *if* out-of-bounds array
access had defined behavior, which it clearly doesn't in C. But it
doesn't necessarily follow that there has to be a bounds check on
every iteration. That depends on what the hypothetical defined
behavior actually is.

I can't think of a plausible or useful defined behavior that would
require an explicit check on every iteration, but I haven't given
it much thought.

(In Ada, for example, an out-of-bounds array access is specified
to raise an exception, which terminates execution of the enclosing
construct. Ada compilers can certainly hoist bounds checks out
of loops.)

[...]

--
Keith Thompson (The_Other_Keith) Keith.S.Thompson+u@gmail.com
void Void(void) { Void(); } /* The recursive call of the void */

--- PyGate Linux v1.5.19
* Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)