enhancementgood first issuehelp wantedperformance
仓库指标
- 星标
- (264 个星标)
- PR 合并指标
- (平均合并 8天 15小时) (30 天内合并 10 个 PR)
描述
It would be useful for performance profiling to provide FLOP count estimates for CeedOperatorLinearAssemblePointBlockDiagonal, CeedOperatorLinearAssemble, and CeedOperatorLinearAssembleDiagonal. This is quite backend dependent, and also may differ based on whether the QFunction assembly data is stale. We should determine how to be as accurate as possible with that info, but even a rough estimate would be useful.
This is mostly just counting.