All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Zhao (Compiler), Yuan" <yuanzhao@ti.com>
To: "Dmytriyenko, Denys" <denys@ti.com>
Cc: "meta-arago@arago-project.org" <meta-arago@arago-project.org>,
	"Robey, Caleb" <c-robey@ti.com>
Subject: Re: [PATCH] tidl-api: update BBAI mcbench script
Date: Thu, 9 Jan 2020 23:18:47 +0000	[thread overview]
Message-ID: <aaa2988f9bdf4dfd95967cea6d6e6d6a@ti.com> (raw)
In-Reply-To: <20200109231143.GP2305@beryl>

No SRCREV bump yet.  The accepted commit stays on our internal develop branch.  It will be externally visible next time we make a tidl-api release (this one missed the PSDK 6.2 window).  - Yuan

-----Original Message-----
From: Dmytriyenko, Denys 
Sent: Thursday, January 9, 2020 5:12 PM
To: Zhao (Compiler), Yuan
Cc: meta-arago@arago-project.org; Robey, Caleb
Subject: Re: [meta-arago] [PATCH] tidl-api: update BBAI mcbench script

On Wed, Jan 08, 2020 at 12:04:12PM -0600, Yuan Zhao wrote:
> Signed-off-by: Caleb Robey <a0232989@ti.com>
> Signed-off-by: Yuan Zhao <yuanzhao@ti.com>
> ---
>  ...p-2-group-layer-use-cases-with-1-dsp.patch | 34 +++++++++++++++++++
>  .../recipes-ti/tidl-api/tidl-examples_git.bb  |  3 +-
>  2 files changed, 36 insertions(+), 1 deletion(-)
>  create mode 100644 meta-arago-extras/recipes-ti/tidl-api/files/0002-replace-2-dsp-2-group-layer-use-cases-with-1-dsp.patch
> 
> diff --git a/meta-arago-extras/recipes-ti/tidl-api/files/0002-replace-2-dsp-2-group-layer-use-cases-with-1-dsp.patch b/meta-arago-extras/recipes-ti/tidl-api/files/0002-replace-2-dsp-2-group-layer-use-cases-with-1-dsp.patch
> new file mode 100644
> index 00000000..0c78ccf0
> --- /dev/null
> +++ b/meta-arago-extras/recipes-ti/tidl-api/files/0002-replace-2-dsp-2-group-layer-use-cases-with-1-dsp.patch
> @@ -0,0 +1,34 @@
> +tidl-api: replace 2 dsp + 2 group layer use cases with 1 dsp
> +
> +    The BBAI only has enough CMEM for 4 EVEs, 1 DSP, and 2 group
> +    layers. In the case of all of our networks, the difference between
> +    1 and 2 dsps is essentially nonexistent. (PLSDK-3189)
> +
> +Upstream-Status: Accepted

Since the patch is marked as "Accepted", shouldn't a simple SRCREV bump be 
enough?


> +Signed-off-by: Caleb Robey <a0232989@ti.com>
> +Signed-off-by: Yuan Zhao <yuanzhao@ti.com>
> +---
> + examples/mcbench/scripts/all_5729.sh | 9 ++++-----
> + 1 file changed, 4 insertions(+), 5 deletions(-)
> +
> +diff --git a/examples/mcbench/scripts/all_5729.sh b/examples/mcbench/scripts/all_5729.sh
> +index defec72..0fee8a1 100755
> +--- a/examples/mcbench/scripts/all_5729.sh
> ++++ b/examples/mcbench/scripts/all_5729.sh
> +@@ -7,8 +7,7 @@ export TIDL_NETWORK_HEAP_SIZE_DSP=56623104
> + export TIDL_NETWORK_HEAP_SIZE_EVE=67108864
> + export TIDL_NETWORK_HEAP_SIZE_DSP=8388608
> + ./mcbench -g 2 -d 1 -e 4 -c ../test/testvecs/config/infer/tidl_config_mobileNet1_lg2.txt -f 50 -i ../test/testvecs/input/preproc_2_224x224_multi.y
> +-./mcbench -g 2 -d 2 -e 4 -c ../test/testvecs/config/infer/tidl_config_mobileNet1_lg2.txt -f 50 -i ../test/testvecs/input/preproc_2_224x224_multi.y
> +-./mcbench -g 2 -d 2 -e 4 -c ../test/testvecs/config/infer/tidl_config_mobileNet2_lg2.txt -f 50 -i ../test/testvecs/input/preproc_2_224x224_multi.y
> +-./mcbench -g 2 -d 2 -e 4 -c ../test/testvecs/config/infer/tidl_config_inceptionNetv1_lg2.txt -f 50 -i ../test/testvecs/input/preproc_0_224x224_multi.y
> +-./mcbench -g 2 -d 2 -e 4 -c ../test/testvecs/config/infer/tidl_config_j11_v2_lg2.txt -f 50 -i ../test/testvecs/input/preproc_0_224x224_multi.y
> +-./mcbench -g 2 -d 2 -e 4 -c ../test/testvecs/config/infer/tidl_config_j11_v2_dense_lg2.txt -f 50 -i ../test/testvecs/input/preproc_0_224x224_multi.y
> ++./mcbench -g 2 -d 1 -e 4 -c ../test/testvecs/config/infer/tidl_config_mobileNet2_lg2.txt -f 50 -i ../test/testvecs/input/preproc_2_224x224_multi.y
> ++./mcbench -g 2 -d 1 -e 4 -c ../test/testvecs/config/infer/tidl_config_inceptionNetv1_lg2.txt -f 50 -i ../test/testvecs/input/preproc_0_224x224_multi.y
> ++./mcbench -g 2 -d 1 -e 4 -c ../test/testvecs/config/infer/tidl_config_j11_v2_lg2.txt -f 50 -i ../test/testvecs/input/preproc_0_224x224_multi.y
> ++./mcbench -g 2 -d 1 -e 4 -c ../test/testvecs/config/infer/tidl_config_j11_v2_dense_lg2.txt -f 50 -i ../test/testvecs/input/preproc_0_224x224_multi.y
> +-- 
> +2.17.1
> +
> diff --git a/meta-arago-extras/recipes-ti/tidl-api/tidl-examples_git.bb b/meta-arago-extras/recipes-ti/tidl-api/tidl-examples_git.bb
> index c7016ad5..e5afad1c 100644
> --- a/meta-arago-extras/recipes-ti/tidl-api/tidl-examples_git.bb
> +++ b/meta-arago-extras/recipes-ti/tidl-api/tidl-examples_git.bb
> @@ -6,7 +6,8 @@ LICENSE = "BSD"
>  include tidl-api.inc
>  require recipes-ti/includes/ti-paths.inc
>  
> -PR = "${INC_PR}.0"
> +PR = "${INC_PR}.1"
> +SRC_URI += "file://0002-replace-2-dsp-2-group-layer-use-cases-with-1-dsp.patch"
>  
>  COMPATIBLE_MACHINE = "dra7xx"
>  PACKAGE_ARCH = "${MACHINE_ARCH}"
> -- 
> 2.24.1
> 
> _______________________________________________
> meta-arago mailing list
> meta-arago@arago-project.org
> http://arago-project.org/cgi-bin/mailman/listinfo/meta-arago


  reply	other threads:[~2020-01-09 23:20 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-08 18:04 [PATCH] tidl-api: update BBAI mcbench script Yuan Zhao
2020-01-09 23:11 ` Denys Dmytriyenko
2020-01-09 23:18   ` Zhao (Compiler), Yuan [this message]
2020-01-10 17:36     ` Denys Dmytriyenko

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=aaa2988f9bdf4dfd95967cea6d6e6d6a@ti.com \
    --to=yuanzhao@ti.com \
    --cc=c-robey@ti.com \
    --cc=denys@ti.com \
    --cc=meta-arago@arago-project.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.