Intel added PREFETCHW to their Broadwell processors launched in 2014, years after AMD dropped all 3DNow! instructions except the prefetch instructions. That timeline strongly suggests that the instructions aren't no-ops and likely are used by some popular software.

Do you mean storing the next index/offset and having the pointer value calculated as late as possible by adding the starting address (and maybe multiplying the index by sizeof)? That would probably defeat/mislead Intel's prefetcher, as described at https://www.intel.com/content/www/us/en/developer/articles/t...

Please do me the courtesy of not misrepresenting what I have said and clarified. I have not said that AndroidAPS or Loop or any other DIY medical device software is stupid. I have only said that it is something a reputable business would consider risky and dangerous.

What I have said is stupid is: your attempt to use that software as an argument in favor of loosening Apple's restrictions on iOS applications. The apps in question are such a legal nightmare that they're not available in any major app store, for iOS or Android. No business with a competent legal department would want to be directly involved with those software projects, unless they were planning to incorporate them into their own product and get it certified by regulators like the FDA. If Apple or any other company in a similar situation decided to open up their platform and even mentioned that one of the benefits of that change was to enable this DIY medical device use case, that would probably put them at substantial legal risk.

Regardless of how useful you find such software, you're not going to get a corporation on board by drawing their attention to such a big legal risk. You're also unlikely to win over government regulators, since they're likely to be of the opinion that medical devices should be regulated.

You've identified a reason why some customers may want Apple to change their strategy, but you've completely failed to provide a reason why Apple would want to change, or why a government would want to force Apple to change.

> A more interesting example might be if each slot in the target array has the next index to go to in addition to the value, then you will introduce a dependency chain preventing this from happening.

However, on some processors there's a data-dependent prefetcher that will notice the pointer-like value and start prefetching that address before the CPU requests it.

Obviously there's no point in trying to entirely prevent people from DIYing dangerous devices, medical or otherwise. But that's not what's at issue here. The problem here is the absolutely stupid idea of using a desire for dangerous DIY as the argument in favor of loosening restrictions already in place. It's pretty much the least-compelling argument possible.