Yossi Khayet and Adam Morrison at Tel Aviv University wrote a paper arguing that the Arm weak memory model doesn't offer structural performance benefits compared to TSO. That was through examining the TSO performance behaviour on the Apple M1 and M4 generations and characterising where the observable performance impact came from. They then argued that the performance pitfalls weren't attributable to TSO structural issues and could be addressed on a TSO-centric design. https://dl.acm.org/doi/pdf/10.1145/3779212.3790129?download=true