Making one's own processor design must be the new "every BigCo needs its own proprietary typeface" from yesteryear.
Are we to believe that Amazon's workload is so unique that an existing design can't fit the bill?
This is just a vanity exercise.
Are we to believe that Amazon's workload is so unique that an existing design can't fit the bill?
This is just a vanity exercise.
POWER from this perspective, does way too much. It's too much chip for AWS's needs, and a big waste of money.
That's what optimization means! You get the smallest/ simplest/cheapest thing that does exactly what you want, and ignore the big bad "do everything" chips that go for megabucks.
Then you can devote chip transistor count toward the instructions you want, and ditch all the dead weight that comes with a traditional chip like power, that tries to make a balanced general purpose cpu.
This results in a chip that isn't good for anything but the specific need, but it's super cost effective and higher performing.
You can easily benchmark this by comparing the Graviton variants with the Intel specific instances.
Graviton 4 has also not been made available for EC2 and RDS instance types in most regions two years after release. My guess is that they are unable to scale production.
You need to run actual workloads that fully utilize all cores and not benchmarks to understand the value of Graviton. People don’t run benchmarks in production; they run real business applications. Graviton is optimized for horizontal scale, not straight line single core performance.
Lol graviton has been in production with fantastic margins for like 7 years. Me thinks maybe you don't know as much as you think you do.