Re:I love ARM
by mumblemumble on Wednesday July 29, 2020 @09:13AM
(#23986247)
In this case, I think that the "sunk cost fallacy" counter is as specious as it is obvious. It completely misplaces the burden of evidence.
Apple has a platform with hundreds of millions of active users, perhaps over a billion actual devices in use. In a situation like that, conservatism is warranted. The guiding principle is simple: If it ain't broke, don't fix it. Meaning that the argument doesn't start with any sort of discussion about technical advantages that RISC-V might have over ARM, or challenging reasons to stay on ARM as if, having dismissed them, the obvious default plan of migrating to RISC-V would kick into action. It starts with trying to demonstrate that something about ARM is broken. Because, if it ain't broke, don't fix it. So. . . is it hindering Apple's strategic or operational priorities in any way?
Also Apple already hold a perpetual license for the ARM architecture, the major costs of their investment in it were spent years or even decades ago.
The obvious counter is the sunk cost fallacy, but what does RISC-V actually have to offer that Apple doesn't already have with ARM? Those sunk costs have bought Apple actual tangible benefits on ARM that RISC-V doesn't have, such as architectural features like the secure enclave, better device integration, better GPU options, more mature device driver support, a stronger development and testing tools ecosystem.
The opportunity for RISC-V is new use cases not already strongly addressed by ARM, or where ARMs more mature ecosystem offers weaker benefits, for companies not already strongly invested in ARM.