Strange power numbers with ASAP7

[oharboe]

OpenROAD v2.0-17847-g43e602d7b

https://drive.google.com/file/d/1bq3ULEK2iD4Q0-ilI1IE4K_QX9qfVse3/view?usp=sharing

to inspect results untar and run:

. vars-fs-mul-mod-q-asap7-base.sh
ODB_FILE=results/asap7/fs-mul-mod-q/base/6_final.odb openroad -gui flow/scripts/open.tcl

>>> report_power
Group                  Internal  Switching    Leakage      Total
                          Power      Power      Power      Power (Watts)
----------------------------------------------------------------
Sequential             1.16e-03   7.67e-06   3.04e-08   1.17e-03   0.3%
Combinational          2.06e-01   1.95e-01   2.05e-06   4.01e-01  99.7%
Clock                  8.88e-05   4.72e-05   9.05e-09   1.36e-04   0.0%
Macro                  0.00e+00   0.00e+00   0.00e+00   0.00e+00   0.0%
Pad                    0.00e+00   0.00e+00   0.00e+00   0.00e+00   0.0%
----------------------------------------------------------------
Total                  2.07e-01   1.95e-01   2.09e-06   4.02e-01 100.0%
                          51.5%      48.5%       0.0%

Originally posted by @oharboe in #6320 (reply in thread)

[oharboe]

@jjcherry56 @maliberty @evmanz Standard format bug report.

[oharboe]

Tried with parallaxsw/OpenSTA@87418f2

Group                  Internal  Switching    Leakage      Total
                          Power      Power      Power      Power (Watts)
----------------------------------------------------------------
Sequential             1.16e-03   7.67e-06   3.04e-08   1.17e-03   0.3%
Combinational          2.06e-01   1.95e-01   2.05e-06   4.01e-01  99.7%
Clock                  8.88e-05   4.72e-05   9.05e-09   1.36e-04   0.0%
Macro                  0.00e+00   0.00e+00   0.00e+00   0.00e+00   0.0%
Pad                    0.00e+00   0.00e+00   0.00e+00   0.00e+00   0.0%
----------------------------------------------------------------
Total                  2.07e-01   1.95e-01   2.09e-06   4.02e-01 100.0%
                          51.5%      48.5%       0.0%

[oharboe]

@maliberty Unchanged after parallaxsw/OpenSTA@87418f2

[tspyrou]

From discussion #6320

[jjcherry56]

I gave up trying to run this because there is so much garbage like this
set env(BC_CCS_DFF_LIB_FILE) "$::env(FLOW_HOME)/../foo/test-case/../../flow/platforms/asap7/lib/NLDM/asap7sc7p5t_SEQ_RVT_FF_ccs_220123.lib"

in
vars-fs-mul-mod-q-asap7-base.tcl

[evmanz]

Here is the archive made with make detail_route_issue

[oharboe]

I gave up trying to run this because there is so much garbage like this set env(BC_CCS_DFF_LIB_FILE) "$::env(FLOW_HOME)/../foo/test-case/../../flow/platforms/asap7/lib/NLDM/asap7sc7p5t_SEQ_RVT_FF_ccs_220123.lib"

in vars-fs-mul-mod-q-asap7-base.tcl

Sorry about that, make issue generated broken vars-fs-mul-mod-q-asap7-base.* files and I only fixed the .sh variant.

I never use the .gdb or .tcl versions, only the .sh version.

I would give the vars-*.tcl/.db files from make issue a wide berth and I've suggested that they should be retired as they obviously haven't been tested in quite some time The-OpenROAD-Project/OpenROAD-flow-scripts#2670

@maliberty FYI

[jjcherry56]

vars_mumble.tcl is the only useful file as far as I am concerned.
I debug inside emacs by sourcing vars.tcl in .openroad so shell scripts are worthless.
Currently I have to set env(FLOW_HOME) and FLOW_HOME for the broken vars.tcl file to work.

[maliberty]

still relevant?

[evmanz]

still relevant?

yep.

I tried the latest last week, and the power situation is the same.

[tspyrou]

@evmanz were you able to run make issue so that it will be straightforward for us to reproduce what you are doing?
We would like to look at this.

[oharboe]

@evmanz Did I transport the problem correctly into this github issue? parallaxsw/OpenSTA#201

[evmanz]

@evmanz Did I transport the problem correctly into this github issue? parallaxsw/OpenSTA#201

yep!

I am a bit confused. I am using OpenROAD-flow-scripts, and there is no make issue in OpenROAD-flow-scripts/flow/Makefile. Is the issue target in a different location?

@tspyrou, @oharboe did it.

[maliberty]

util/utils.mk

[evmanz]

util/utils.mk

Well, it has %_issue, which means it's an issue per stage, right? Like, make cts_issue. Correct?

[maliberty]

Yes, you can package an issue from any step