docs(benchmarks): the fresh-process table says it times a build handed to the solver, not a solve

CHANGELOG.md

@@ -7,6 +7,7 @@ it releases that version ([RELEASING.md](https://github.com/fluxopt/specsolve/bl
 
 ## Upcoming version
 
+- docs: the fresh-process table says it times a build handed to the solver, not a solve ([#2041](https://github.com/fluxopt/specsolve/pull/2041))
 - docs: the published benchmarks are re-measured at the new model sizes, five times each tenfold step ([#2039](https://github.com/fluxopt/specsolve/pull/2039))
 - docs: the benchmarks page opens on its charts, counts sizes as the columns and rows that reach the solver, and tables what each model declares ([#2042](https://github.com/fluxopt/specsolve/pull/2042))
 - docs: the benchmark charts share one pair of axes across every model, and a box dragged on any chart zooms all of them ([#2040](https://github.com/fluxopt/specsolve/pull/2040))

README.md

@@ -38,8 +38,8 @@ which states the math and checks it before any data exists.
   reads it where it lies and solves the same question again, and a directory
   of runs reads as one set of tables.
   [Archiving a solve →](https://specsolve.readthedocs.io/en/latest/howto/archiving/)
-- **Fast from the first call.** A script that solves one small model finishes
-  in about half a second, about half of linopy's time
+- **Fast from the first call.** A script that builds one small model and hands
+  it to the solver finishes in about half a second, about half of linopy's time
   ([#2031](https://github.com/fluxopt/specsolve/pull/2031)). The solver stays
   loaded: `update()` puts new numbers on it, and the next solve carries on from
   where the last one ended.

bench/report.py

@@ -400,7 +400,7 @@ def order(key: tuple[str, str]) -> tuple[float, str, str]:
 
 
 def fresh(rows: dict[Key, Row]) -> str:
-    """What a script that solves one model pays, launch to exit, on `FRESH_SINK`, as a heat table.
+    """What a script that builds one model pays, launch to exit, on `FRESH_SINK`, as a heat table.
 
     One row per arm and one column per case, each case at the largest rung the
     fresh phase measured, so a reader compares libraries down a column. Wall

docs/about/benchmarks.md

@@ -105,11 +105,11 @@ There it is 2.7× at 1M columns, 1.75 GB against 0.65 GB, and the memory
 budget stops linopy at that size
 ([#2039](https://github.com/fluxopt/specsolve/pull/2039)).
 
-## A script that solves one model
+## A script that builds one model
 
 The charts leave out what a script pays to start: the interpreter, the import
-and the first build. This table times a new process from launch to exit, on
-HiGHS.
+and the first build. This table times a new process from launch to exit. The
+process builds the model, hands it to HiGHS and exits, without a solve.
 
 <!-- bench:fresh -->
 
@@ -124,8 +124,8 @@ HiGHS.
 
 <!-- bench:/fresh -->
 
-A script that solves one small model with specsolve finishes in about half a
-second, about half of linopy's time
+A script that builds one small model with specsolve and hands it to the solver
+finishes in about half a second, about half of linopy's time
 ([#2031](https://github.com/fluxopt/specsolve/pull/2031)).
 
 ## Not measured yet