{"id":912030,"date":"2026-07-21T22:55:04","date_gmt":"2026-07-21T22:55:04","guid":{"rendered":"https:\/\/cri-one.com\/blog\/2026\/07\/21\/wonderphi-compute-put-the-compute-back-in-the-game\/"},"modified":"2026-07-21T23:15:22","modified_gmt":"2026-07-21T23:15:22","slug":"wonderphi-compute-put-the-compute-back-in-the-game","status":"publish","type":"post","link":"https:\/\/cri-one.com\/blog\/2026\/07\/21\/wonderphi-compute-put-the-compute-back-in-the-game\/","title":{"rendered":"WonderPhi Compute \u2014 Put the Compute Back in the Game"},"content":{"rendered":"

Software acceleration for existing hardware.<\/p>\n

Every machine you own has more power than you use. Cores sit idle. GPU memory stays dark. Background processes eat cycles you never asked them to. The operating system has its own agenda, and your work waits in line behind it.<\/p>\n

WonderPhi Compute<\/strong> is the antidote: a constitutional parallel runner that reads a plain-text plan of shell commands and executes them across every CPU core, honestly showing you what it’s doing.<\/p>\n

What Makes It Different<\/h2>\n

Most parallel runners \u2014 GNU parallel, make -j, xargs -P \u2014 grew organically over decades. A hundred flags. No philosophy. WonderPhi Compute is designed the other way: 10 immutable constitutional rules<\/strong>, hardcoded into every binary, verified at startup, and carved in stone.<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
#<\/th>\nCommandment<\/th>\nWhat It Guarantees<\/th>\n<\/tr>\n<\/thead>\n
I<\/td>\nSole Master<\/td>\nNo competing orchestrator overrides WPC mid-run<\/td>\n<\/tr>\n
II<\/td>\nNo Other Gods<\/td>\nExactly one instance per working directory<\/td>\n<\/tr>\n
III<\/td>\nNot in Vain<\/td>\nMalformed plans rejected, never silently mangled<\/td>\n<\/tr>\n
IV<\/td>\nSabbath<\/td>\nIdle workers sleep \u2014 no busy-waiting, no spinning<\/td>\n<\/tr>\n
V<\/td>\nHonour Thy Parents<\/td>\nThe plan format is frozen forever<\/td>\n<\/tr>\n
VI<\/td>\nNot Kill<\/td>\nLog appends only \u2014 a crash never destroys history<\/td>\n<\/tr>\n
VII<\/td>\nNot Adultery<\/td>\nCommands dispatched verbatim \u2014 no silent retry<\/td>\n<\/tr>\n
VIII<\/td>\nNot Steal<\/td>\nMax 64 concurrent, 4 GB\/task, 24-hour runtime cap<\/td>\n<\/tr>\n
IX<\/td>\nNot False Witness<\/td>\nExit codes and elapsed times reported truthfully<\/td>\n<\/tr>\n
X<\/td>\nNot Covet<\/td>\nZero dependencies, no network, no telemetry<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Every version of WonderPhi Compute that will ever exist is v1.0.0<\/strong>. Born once. Never updated. No subscription. No telemetry. No dependencies. That’s the constitutional guarantee.<\/p>\n

How It Works<\/h2>\n

You write a plan file<\/strong> \u2014 a plain-text list of shell commands, one per line. Blank lines are barriers: everything above must finish before anything below starts. Comments start with #<\/code>. That’s the entire format.<\/p>\n

\n# WPC v1\n# Stage 1 \u2014 extract text from every PDF in parallel\npdftotext report001.pdf out001.txt\npdftotext report002.pdf out002.txt\npdftotext report003.pdf out003.txt\n\n# Stage 2 \u2014 aggregate (waits for stage 1)\ncat out*.txt > combined.txt\n<\/pre>\n
Feed that to wpc.exe<\/code>, and it runs every stage across every core simultaneously. A serial script that would run for two hours finishes in ten minutes.<\/strong><\/p>\n
GPU Awareness<\/h2>\n
WonderPhi Compute detects your NVIDIA GPU at startup and exposes $WPC_GPU=1<\/code> to every task’s environment. It doesn’t offload work to the GPU itself \u2014 that’s the job of your commands. But your scripts can branch on GPU availability, choosing the right silicon for the right job.<\/p>\n
Always-On Mode<\/h2>\n
Three cooperating watchers turn WonderPhi into a drop-a-file-and-it-runs service:<\/p>\n