Skip to content

Scripts: Updates to scripts using sof-testbench4 simulation #10545

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
singalsu wants to merge 6 commits into
base: main
Choose a base branch
from
Open

Scripts: Updates to scripts using sof-testbench4 simulation #10545

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
d7879ae
Scripts: sof-testbench-helper: Use memory model for xt-run
singalsu Feb 12, 2026
b430646
Scripts: sof-testbench-helper: Separate input and output rate switches
singalsu Feb 12, 2026
2e90a89
Scripts: sof-testbench-build-profile: Test SRC and ASRC with 44.1 kHz
singalsu Feb 12, 2026
074e730
Scripts: sof-testbench-build-profile: Add more components to profile
singalsu Feb 12, 2026
70b46a7
Tools: Topology: Bench: Change DRC blobs for test to the used ones
singalsu Jan 30, 2026
09315da
Scripts: host-testbench: Add more modules to quick test procedure
singalsu Feb 12, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions scripts/host-testbench.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -122,4 +122,16 @@ test_component asrc 32 32 48000 "$FullTest"
# test with template component
test_component template_comp 32 32 48000 "$FullTest"

# test with Dolby DAX with stub
test_component dolby-dax 32 32 48000 "$FullTest"

# test with level_multiplier
test_component level_multiplier 32 32 48000 "$FullTest"

# test with micsel
test_component micsel 32 32 48000 "$FullTest"

# test with sound_dose
test_component sound_dose 32 32 48000 "$FullTest"

echo "All tests are done!"
11 changes: 10 additions & 1 deletion scripts/sof-testbench-build-profile.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,9 @@ usage() {
echo
}

MODULES_S32="asrc dcblock drc drc_multiband eqfir eqiir gain src tdfb"
MODULES_S32_44K_48K="asrc src"
Copy link

Copilot AI Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The variable name MODULES_S32_44K_48K implies these modules should be profiled at both 44.1k and 48k, but the loop only runs at 44100. Either rename the variable to match the implemented behavior (e.g., MODULES_S32_44K) or add a second run at 48000 to align with the name.

Copilot uses AI. Check for mistakes.
MODULES_S32="dcblock drc drc_multiband dolby-dax eqfir eqiir gain level_multiplier micsel \
sound_dose stft_process_1536_240_ template_comp tdfb"
Copy link

Copilot AI Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The module name stft_process_1536_240_ has a trailing underscore, which is likely unintended and can cause the profiling invocation (-m) to fail if it doesn’t exactly match the component/module name expected by the testbench/topology. If the real module name does not include the trailing _, remove it here.

Suggested change
sound_dose stft_process_1536_240_ template_comp tdfb"
sound_dose stft_process_1536_240 template_comp tdfb"

Copilot uses AI. Check for mistakes.
Copy link
Collaborator Author

@singalsu singalsu Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The topology name it uses is sof-hda-benchmark-stft_process_1536_240_32.tplg. And the simple append of test word length to component name, that now includes FFT size 1536 and hop 240, has the underscore. It's ugly but this is how it is now with test scripts.

MODULES_S24="aria"

if [ -z "${SOF_WORKSPACE}" ]; then
Expand Down Expand Up @@ -61,6 +63,13 @@ $HELPER -x -t development/sof-hda-benchmark-generic.tplg -n 1,2,3 \
-p "$PDIR/profile-$PLATFORM-benchmark.txt" > "$PDIR/log-$PLATFORM-benchmark.txt"

# Profile modules
for mod in $MODULES_S32_44K_48K
do
echo "Profiling $mod ..."
$HELPER -r 44100 -x -m "$mod" \
-p "$PDIR/profile-$PLATFORM-$mod.txt" > "$PDIR/log-$PLATFORM-$mod.txt"
done
Comment on lines +66 to +71
Copy link

Copilot AI Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The variable name MODULES_S32_44K_48K implies these modules should be profiled at both 44.1k and 48k, but the loop only runs at 44100. Either rename the variable to match the implemented behavior (e.g., MODULES_S32_44K) or add a second run at 48000 to align with the name.

Copilot uses AI. Check for mistakes.
Copy link
Collaborator Author

@singalsu singalsu Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The input is 44.1 kHz and output is 48 kHz, so the test really runs at both rates.

Copy link
Collaborator

@lyakh lyakh Feb 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

good idea to add 44100 testing! We should maybe add such a test to PR CI too?..

Copy link
Collaborator Author

@singalsu singalsu Feb 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It would be quite simple with a scripts/hosts-testbench.sh change.


for mod in $MODULES_S32
do
echo "Profiling $mod ..."
Expand Down
11 changes: 7 additions & 4 deletions scripts/sof-testbench-helper.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,8 @@ usage() {
echo " -n <pipelines>, default 1,2"
echo " -o <output wav>, default none"
echo " -p <profiling result text>, use with -x, default none"
echo " -r <rate>, default 48000"
echo " -r <rate>, input rate, default 48000"
echo " -R <rate>, output rate, default 48000"
echo " -t <force topology>, default none, e.g. production/sof-hda-generic.tplg"
echo " -v runs with valgrind, not available with -x"
echo " -x runs testbench with xt-run simulator"
Expand Down Expand Up @@ -55,7 +56,7 @@ PROFILE=false
TPLG0=
VALGRIND=

while getopts "b:c:hi:km:n:o:p:r:t:vx" opt; do
while getopts "b:c:hi:km:n:o:p:r:R:t:vx" opt; do
Copy link

Copilot AI Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A new -R option is introduced for output rate. Make sure the script’s help/usage text documents -R (and clarifies the distinction between -r input rate and -R output rate), otherwise users will not discover the new option from -h.

Copilot uses AI. Check for mistakes.
Copy link
Collaborator Author

@singalsu singalsu Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yep, makes sense

case "${opt}" in
b)
BITS=${OPTARG}
Expand Down Expand Up @@ -89,6 +90,8 @@ while getopts "b:c:hi:km:n:o:p:r:t:vx" opt; do
;;
r)
RATE_IN=${OPTARG}
;;
R)
RATE_OUT=${OPTARG}
;;
Comment on lines 91 to 96
Copy link

Copilot AI Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A new -R option is introduced for output rate. Make sure the script’s help/usage text documents -R (and clarifies the distinction between -r input rate and -R output rate), otherwise users will not discover the new option from -h.

Copilot uses AI. Check for mistakes.
t)
Expand Down Expand Up @@ -147,10 +150,10 @@ if [[ "$XTRUN" == true ]]; then
echo " input: $INFILE1, output: $OUTFILE1, trace: $TRACEFILE, profile: $PROFILETXT"
source "$XTB4_SETUP"
if [[ $PROFILE == true ]]; then
"$XTENSA_PATH"/xt-run --profile="$PROFILEOUT" "$XTB4" $OPTS 2> "$TRACEFILE"
"$XTENSA_PATH"/xt-run --mem_model --profile="$PROFILEOUT" "$XTB4" $OPTS 2> "$TRACEFILE"
"$XTENSA_PATH"/xt-gprof "$XTB4" "$PROFILEOUT" > "$PROFILETXT"
else
"$XTENSA_PATH"/xt-run "$XTB4" $OPTS 2> "$TRACEFILE"
"$XTENSA_PATH"/xt-run --mem_model "$XTB4" $OPTS 2> "$TRACEFILE"
Comment on lines +153 to +156
Copy link

Copilot AI Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

--mem_model is now always passed to xt-run. If this script is expected to work across multiple Xtensa toolchain versions/environments, consider making this flag configurable (env var or script option) or add a small compatibility fallback when xt-run doesn’t recognize --mem_model (e.g., retry without the flag).

Copilot uses AI. Check for mistakes.
Copy link
Collaborator Author

@singalsu singalsu Feb 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not aware of a toolchain version without support for --mem_model.

fi
else
if [ ! -x "$TB4" ]; then
Expand Down
2 changes: 1 addition & 1 deletion tools/topology/topology2/development/tplg-targets-bench.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ set(component_parameters
"BENCH_ASRC_PARAMS=default"
"BENCH_DCBLOCK_PARAMS=default"
"BENCH_DOLBY-DAX_PARAMS=default"
"BENCH_DRC_PARAMS=enabled"
"BENCH_DRC_PARAMS=default_speaker_mic"
Copy link

Copilot AI Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The selector name default_speaker_mic is ambiguous: in capture it maps to a DMIC default, while in playback it maps to a speaker default. Consider renaming to something that reflects this indirection (e.g., default_io), or using explicit names (default_speaker / default_dmic) to make the behavior clearer for anyone reading the bench parameter defaults.

Suggested change
"BENCH_DRC_PARAMS=default_speaker_mic"
"BENCH_DRC_PARAMS=default_io"

Copilot uses AI. Check for mistakes.
Copy link
Collaborator Author

@singalsu singalsu Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The cmake scripts and topologies system has limitations and currently this is the way to apply with a keyword the blob for playback and capture that is in this case different. Default_io in my opinion is less descriptive than my proposal.

"BENCH_DRC_MULTIBAND_PARAMS=default"
"BENCH_EQIIR_PARAMS=loudness"
"BENCH_EQFIR_PARAMS=loudness"
Expand Down
1 change: 1 addition & 0 deletions tools/topology/topology2/include/bench/drc_controls_capture.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@
"default" "include/components/drc/default.conf"
"enabled" "include/components/drc/enabled.conf"
"passthrough" "include/components/drc/passthrough.conf"
"default_speaker_mic" "include/components/drc/dmic_default.conf"
}
}
mixer."1" {
Expand Down
1 change: 1 addition & 0 deletions tools/topology/topology2/include/bench/drc_controls_playback.conf
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@
"default" "include/components/drc/default.conf"
"enabled" "include/components/drc/enabled.conf"
"passthrough" "include/components/drc/passthrough.conf"
"default_speaker_mic" "include/components/drc/speaker_default.conf"
}
}
mixer."1" {
Expand Down