March 7, 2026
Bench wars, meme scars
Does Apple's M5 Max really "Destroy" a 96-Core Threadripper?
Tech forum erupts: “M5 Max DESTROYS Threadripper? Or just a clickbait benchmark”
TLDR: A splashy claim that Apple’s M5 Max beats a 96‑core Threadripper using Geekbench 6 sparked a brawl over what benchmarks should measure. Commenters say Geekbench’s tests don’t use many cores, making big CPUs look weak, while others argue that mirrors real-life tasks—turning this into realism versus raw power.
Tom’s Hardware lit the fuse with a flashy claim: Apple’s 18‑core M5 Max “destroys” a 96‑core AMD Threadripper using Geekbench 6 multi‑core scores. The comment section? Pure fireworks. The OP says Geekbench 6’s multi‑core test stops scaling on big chips and can even get worse as you add more cores—cue accusations of clickbait benchmarking and “Apple vs. AMD” tribal warfare. One user called it simple: don’t trust Geekbench unless you’re comparing the same chip family across generations. Another countered that Geekbench chose to mirror everyday tasks, where most apps don’t use 96 cores anyway.
Then the drama hit peak nerd: a fiery takedown for failing to mention Amdahl’s Law (plain English: the parts of a job that can’t be split across many cores are the bottleneck, so more cores won’t always help). One link making the rounds claims Geekbench 6 multicore is “[broken by design]” with charts to prove it: dev.to/dkechag. Meanwhile, the top meme is MC Hammer’s “When you measure, include the measurer,” roasting the test itself. The core battle: Should benchmarks reflect real‑world tasks (favoring Apple in this case) or raw throughput (where Threadripper usually shines)? Either way, the crowd’s verdict is split—and very, very loud.
Key Points
- •Tom’s Hardware ran a front-page article comparing Apple’s 18‑core M5 Max to a 96‑core Ryzen Threadripper using Geekbench 6 multi-core scores.
- •The Tom’s Hardware piece briefly noted Geekbench’s scaling limitations, but the article argues the severity wasn’t clearly conveyed.
- •The author claims Geekbench 6 multi-core is unsuitable for large CPUs because some tests stop scaling beyond 4–8 cores.
- •They state overall Geekbench 6 multi-core scores can decrease as additional cores are added, misrepresenting high-core CPU performance.
- •The author references a prior detailed breakdown and questions whether such benchmarking supports sensational headlines by major tech sites.