The problem was simple in its horror. On a standard Intel box, NT 4.0 could turn off the display. But it couldn’t sleep. It couldn’t throttle the CPU. And when you hit the power button, you had to pray.
: In original NT 4.0 environments, forcing ACPI often led to performance issues, such as CPU utilization spikes or erratic IRQ assignments (e.g., IRQs appearing as 39 or 125), because the kernel wasn't fully optimized for ACPI tables. Troubleshooting and Installation acpi driver for nt
The next morning, Mark asked, “Will it ship?” The problem was simple in its horror
Lina built a harness she called the “AML asylum.” It sandboxed the interpreter, imposed a 10ms timeout on any method, and mapped fake hardware for the firmware to yell at. Then she wrote the core of the driver—the Acpi.sys dispatcher. It couldn’t throttle the CPU
“The firmware is lying,” she whispered.
That was the dirty secret. Every BIOS vendor implemented ACPI differently. Some forgot to mark the video card as a wake-capable device. Others embedded AML loops that would run for minutes if interpreted naively. One particularly cursed laptop from Compaq had a _BIF (Battery Information) method that called itself recursively.