GUIDELLM2PCP(1)

✨ GUIDELLM2PCP(1) ✨

okay so bestie, this is a General Commands Manual page and the star of the show is guidellm2pcp — let's get into it 💅

💖 NAME 💖

guidellm2pcp — GuideLLM benchmark results importer 🦋

🌸 SYNOPSIS 🌸

guidellm2pcp [-hv] [-a archive] [-H hostname] results [...]

✨ DESCRIPTION ✨

guidellm2pcp is a performance metrics importer tool, and what it does is iconic: it takes the JSON results of a GuideLLM benchmark and builds a PCP archive out of them. literally building something out of nothing, main character energy 💫

You get an archive spanning the time for which the benchmark and its run iterations were active. Discrete PCP metrics get created for each recorded aspect of a run, and they're logged once at the start and once at the end of each iteration, followed by a mark record. every single one, no skipping, periodt 👑

The whole point? This lets PCP tools report on these metrics in a meaningful way alongside all the other system metrics — kernel, hardware, vLLM, and so on — that were gathered during the benchmark. so you're not just looking at one lil thing, you're seeing the whole vibe check together 🌈

The only mandatory command line argument is the JSON results file from a GuideLLM invocation. that's it, that's the requirement. Now — this JSON can optionally be wrapped with additional Model Furnace metadata, and if that's detected, extra PCP labels get extracted and stored (stuff like model and inference server names). cute little bonus, lowkey obsessed 🎀

💅 OPTIONS 💅

Here's the available command line options, bestie:

• -a archive, --archive=archive — Specifies the name of the output PCP archive to be produced. By default, an archive name using the timestamp of the start of the benchmark run will be used. so if you don't pick a name, it just names itself after when the run started 🕐
• -H hostname, --hostname=hostname — Sets the hostname that will be stamped into the PCP archive label header. that's your name in the header, own it 💜
• -h, --help — Display usage message and exit. the classic "pls explain yourself" move 🙈
• -v, --verbose — Report progress as each run is identified and its values saved. for when you need the play-by-play commentary fr fr 🔥

🌟 SEE ALSO 🌟

PCPIntro(1), pminfo(1), pmrep(1), LOGARCHIVE(5), and PMNS(5).

go read those too, iykyk 💖

and that's the whole page, bestie!! this one came from Performance Co-Pilot, part of PCP, and it's the GUIDELLM2PCP(1) page — slay, ate, no crumbs left 🫶✨💅