feat: run benchmarks on nightly too
Compare changes
+ 44
− 6
@@ -146,6 +146,15 @@ version:
@@ -439,10 +448,8 @@ jug_xl:singularity:nightly:
@@ -455,10 +462,8 @@ benchmarks:reconstruction:
@@ -472,6 +477,39 @@ benchmarks:physics:
We are only running benchmarks on default builds, not nightly. This means that lots can go wrong when we update the EICRECON_VERSION
since it could cover a large number of commits. Surfacing errors earlier with benchmarks against nightly is preferred (even though we should really trigger the benchmarks from EICrecon).