You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit bc6e396
Browse filesBrowse the repository at this point in the historyBrowse files
With SKIP_REPORT_METRICS=1, report_metrics reported nothing at all, so a
flow that skips it for speed also lost design area, instance counts and
utilization, though those are a walk over the instances that needs no
timer.
Report them unconditionally, before the skip. SKIP_REPORT_METRICS now
skips only what costs time: the timing update behind setup and hold WS
and TNS, fmax, the max slew/cap/fanout checks and clock skew; the power
analysis; and the reports/<stage>_*.rpt text reports.
The calls that take measurable time even on asap7/gcd, 5 to 200 ms each
on a few hundred cells, run under log_cmd, so on a large design the log
names any that takes five seconds or more. The four register path
searches go through `log_cmd eval { ... }`, so the log prints the
command rather than the expanded register list.
On asap7/gcd, metrics on (DETAILED_METRICS=1) and off give
byte-identical .odb files at every stage; with SKIP_REPORT_METRICS=1
every stage's metrics .json has the design area and no timing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Øyvind Harboe <oyvind.harboe@zylin.com>
Copy file name to clipboardExpand all lines: docs/user/FlowVariables.md
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -291,7 +291,7 @@ configuration file.
291
291
| <aname="SKIP_LAST_GASP"></a>SKIP_LAST_GASP| Do not use last gasp optimization to fix timing violations (default: use gate last gasp).||
292
292
| <aname="SKIP_PIN_SWAP"></a>SKIP_PIN_SWAP| Do not use pin swapping as a transform to fix timing violations (default: use pin swapping).||
293
293
| <aname="SKIP_REPAIR_TIE_FANOUT"></a>SKIP_REPAIR_TIE_FANOUT| Skip repair_tie_fanout at floorplan step.| 0|
294
-
| <aname="SKIP_REPORT_METRICS"></a>SKIP_REPORT_METRICS| If set to 1, then metrics, report_metrics does nothing. Useful to speed up builds.| 0|
294
+
| <aname="SKIP_REPORT_METRICS"></a>SKIP_REPORT_METRICS| If set to 1, report_metrics writes only the metrics that cost no analysis: design area, instance counts and utilization. Timing, power, clock skew, the max slew/cap/fanout checks and the reports/<stage>_*.rpt text reports are skipped; they need a timing update and a power analysis at every stage that calls it. The counters the tools record as they run (placement, legalization, routing, repair, warnings) are written either way.| 0|
295
295
| <aname="SKIP_VT_SWAP"></a>SKIP_VT_SWAP| Do not perform VT swap to improve QoR (default: do VT swap).||
296
296
| <aname="SLEW_MARGIN"></a>SLEW_MARGIN| Specifies a slew margin when fixing max slew violations. This option allows you to overfix.||
297
297
| <aname="SWAP_ARITH_OPERATORS"></a>SWAP_ARITH_OPERATORS| Improve timing QoR by swapping ALU and MULT arithmetic operators.| 0|
0 commit comments