I wouldn’t be surprised if it was soldered on because it shares the RAM with the GPU. That’s pretty common these days, and GPUs want both high performance RAM and low latency. There are solutions to significantly lower the trace lengths between expandable RAM and the CPU/GPU, but the standard never took off.
I wouldn’t be surprised if it was soldered on because it shares the RAM with the GPU. That’s pretty common these days, and GPUs want both high performance RAM and low latency. There are solutions to significantly lower the trace lengths between expandable RAM and the CPU/GPU, but the standard never took off.