First contact

Our in-house GPU lab engine connects your terminal to an eight-GPU sandbox: device inventory, process ownership, memory allocation, editable configuration and service state. Each supported command reads or changes the same machine.

Practice NVIDIA command-line workflows entirely in your browser. The command emulator covers the missions in this lab; CUDA kernel execution and performance benchmarking require real hardware. Reset any time for a fresh node.

Completed mission IDs and run summaries are saved on this device. Machine state and terminal history disappear when you leave or reload.