framework

XLA

The compiler underneath the port; supplies ordinary attention without a shared-memory ceiling, routes bfloat16 through fp32 on pre-Ampere GPUs, and provides the persistent compilation cache.

Current clusters

build1 distinct publisher

Fitting 6.155 GB of Gemma 4 weights lands you on AWS's only Arm GPU family

The cheap fractional-GPU tiers top out at 3 GB of device memory, so the weights pick the instance and hand you a Graviton2 host with a 2018 Turing card. The interesting cost is the toolchain, not the hourly rate.

Publishers:dev.to

Reality

Evidence48
Adoption14
Hype gap−14
Incentives42

build1 distinct publisher

A Pallas kernel tiled for TPU VMEM splits one Gemma 4 port into two checkpoints

One hand-written JAX port runs on TPU v5e, v6e and a Turing T4G, but the quantized fast path stays on TPU and the compute dtype has to be read off the live device, because neither leak ever shows up as red in a log.

Publishers:dev.to

Reality

Evidence54
Adoption14