hexagon: HMX quantized matmul rework (#23368)

* hmx-mm: update debug logging in hmx-mm

* hmx-mm: update dequant logic to use HVX_vector_x2/4

* hmx-mm: remove non-pipelined version of the quantize matmul

It seems that we don't reall need non-pipelined version

* hmx-mm: use activation depth mode and update naming

Co-authored-by: Kim-Chyan Gan <kgan@qti.qualcomm.com>

* hex-mm: minor hmx matmul naming updates

* hmx-mm: remove unused vars

* snapdragon: scripts bump default ubatch-size to 1K

* hexagon: combine HMX and power and clock settings into a single set_power call

* hmx-mm: remove leftover of the scale repl helper

* hexagon: fix editconf error

---------

Co-authored-by: Kim-Chyan Gan <kgan@qti.qualcomm.com>
This commit is contained in:
Max Krasnyansky
2026-05-20 07:39:01 -07:00
committed by GitHub
parent e947228222
commit c9872a2575
12 changed files with 212 additions and 523 deletions

View File

@@ -45,5 +45,5 @@ adb $adbserial $adbhost shell " \
ADSP_LIBRARY_PATH=$basedir/$branch/lib \
$ndev $nhvx $opmask $verbose $profile $hb ./$branch/bin/llama-bench --device $device --mmap 0 -m $basedir/../gguf/$model \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--ubatch-size 256 -fa 1 -ngl 99 $cli_opts $@ \
--ubatch-size 1024 -fa 1 -ngl 99 $cli_opts $@ \
"

View File

@@ -73,6 +73,6 @@ adb $adbserial $adbhost shell " \
$verbose $sched $opmask $profile $nhvx $hmx $ndev $hb $opbatch $opqueue $opflt $vmem $mbuf \
./$branch/bin/llama-cli --no-mmap -m $basedir/../gguf/$model \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--ctx-size 8192 --ubatch-size 256 -fa on \
--ctx-size 8192 --ubatch-size 1024 -fa on \
-ngl 99 --device $device $cli_opts $@ \
"

View File

@@ -69,6 +69,6 @@ adb $adbserial $adbhost shell " \
$verbose $sched $opmask $profile $nhvx $hmx $ndev $hb $opbatch $opqueue $opflt $vmem $mbuf \
./$branch/bin/llama-completion --no-mmap -m $basedir/../gguf/$model \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--ctx-size 8192 --ubatch-size 256 -fa on \
-ngl 99 --device $device $cli_opts $@ \
--ctx-size 8192 --ubatch-size 1024 -fa on \
-ngl 99 --device $device $cli_opts $@ \
"

View File

@@ -66,6 +66,6 @@ adb $adbserial $adbhost shell " \
--mmproj $basedir/../gguf/$mmproj \
--image $basedir/../gguf/$image \
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 \
--ctx-size 8192 --ubatch-size 256 -fa on \
--ctx-size 8192 --ubatch-size 1024 -fa on \
-ngl 99 --device $device -v $cli_opts $@ \
"

View File

@@ -45,4 +45,4 @@ $env:ADSP_LIBRARY_PATH="$basedir\lib"
& "$basedir\bin\llama-bench.exe" `
--mmap 0 -m $basedir\..\..\gguf\$model `
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 `
--batch-size 128 -ngl 99 --device $device $cli_opts
--ubatch-size 1024 -ngl 99 --device $device $cli_opts

View File

@@ -49,5 +49,5 @@ $env:ADSP_LIBRARY_PATH="$basedir\lib"
& "$basedir\bin\llama-cli.exe" `
--no-mmap -m $basedir\..\..\gguf\$model `
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 `
--ctx-size 8192 --ubatch-size 256 -fa on `
--ctx-size 8192 --ubatch-size 1024 -fa on `
-ngl 99 --device $device $cli_opts

View File

@@ -49,5 +49,5 @@ $env:ADSP_LIBRARY_PATH="$basedir\lib"
& "$basedir\bin\llama-completion.exe" `
--no-mmap -m $basedir\..\..\gguf\$model `
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 `
--ctx-size 8192 --batch-size 256 -fa on `
--ctx-size 8192 --ubatch-size 1024 -fa on `
-ngl 99 -no-cnv --device $device $cli_opts

View File

@@ -64,5 +64,5 @@ $env:ADSP_LIBRARY_PATH="$basedir\lib"
--mmproj $basedir\..\..\gguf\$mmproj `
--image $basedir\..\..\gguf\$image `
--poll 1000 -t 6 --cpu-mask 0xfc --cpu-strict 1 `
--ctx-size 8192 --ubatch-size 256 -fa on `
--ctx-size 8192 --ubatch-size 1024 -fa on `
-ngl 99 --device $device -v $cli_opts