summaryrefslogtreecommitdiff
path: root/docs/sections/deployment.md
diff options
context:
space:
mode:
authorKshitij Sisodia <kshitij.sisodia@arm.com>2021-05-07 16:08:14 +0100
committerKshitij Sisodia <kshitij.sisodia@arm.com>2021-05-07 17:28:51 +0100
commitf9c19eaa9ab11e4409679fc6d2862c89410493a7 (patch)
treeb791a4c03f1fe986a2ac32593a3dc817ae3f247a /docs/sections/deployment.md
parent2181d0ac35f30202985a877950c88325ff665f6b (diff)
downloadml-embedded-evaluation-kit-f9c19eaa9ab11e4409679fc6d2862c89410493a7.tar.gz
MLECO-1860: Support for Arm GNU Embedded Toolchain
This patch enables compilation of ML use cases bare-metal applications using Arm GNU Embedded Toolchain. The GNU toolchain can be used instead of the Arm Compiler that was already supported. The GNU toolchain is also set as the default toolchain when building applications for the MPS3 target. Note: The version of GNU toolchain must be 10.2.1 or higher. Change-Id: I5fff242f0f52d2db6c75d292f9fa990df1aec978 Signed-off-by: Kshitij Sisodia <kshitij.sisodia@arm.com>
Diffstat (limited to 'docs/sections/deployment.md')
-rw-r--r--docs/sections/deployment.md18
1 files changed, 9 insertions, 9 deletions
diff --git a/docs/sections/deployment.md b/docs/sections/deployment.md
index a6e9c3c..10acbcf 100644
--- a/docs/sections/deployment.md
+++ b/docs/sections/deployment.md
@@ -150,9 +150,9 @@ For details, on platform set up, please see the relevant documentation. For `Arm
For MPS3 board, instead of loading the axf file directly, the executable blobs
generated under the *sectors/<use_case>* subdirectory need to be
-copied over to the MP3 board's micro SD card. Also, every use case build
-generates a corresponding images.txt file which is used by the MPS3 to
-understand which memory regions the blobs are to be loaded into.
+copied over to the MP3 board's micro SD card. Also, *sectors/images.txt* file is
+used by the MPS3 to understand which memory regions the blobs are to be loaded
+into.
Once the USB A <--> B cable between the MPS3 and the development machine
is connected and the MPS3 board powered on, the board should enumerate
@@ -160,10 +160,10 @@ as a mass storage device over this USB connection.
There might be two devices also, depending on the version of the board
you are using. The device named `V2M-MPS3` or `V2MMPS3` is the `SD card`.
-If the axf/elf file is within 1MiB, it can be flashed into the FPGA
-memory directly without having to break it down into separate load
-region specific blobs. However, with neural network models exceeding
-this size, it becomes necessary to follow this approach.
+If the axf/elf file is within the ITCM load size limit, it can be copied into
+the FPGA memory directly without having to break it down into separate load
+region specific blobs. However, with neural network models exceeding this size,
+it becomes necessary to follow this approach.
1. For example, the image classification use case will produce:
@@ -181,7 +181,7 @@ this size, it becomes necessary to follow this approach.
cp -av ./bin/sectors/img_class/* /media/user/V2M-MPS3/SOFTWARE/
```
-2. The generated `\<use-case\>_images.txt` file needs to be copied
+2. The `./bin/sectors/images.txt` file needs to be copied
over to the MPS3. The exact location for the destination will depend
on the MPS3 board's version and the application note for the bit
file in use.
@@ -190,7 +190,7 @@ application note directory named "ETHOSU", to replace the images.txt
file:
```commandline
- cp ./bin/images-img_class.txt /media/user/V2M-MPS3/MB/HBI0309C/ETHOSU/images.txt
+ cp ./bin/sectors/images.txt /media/user/V2M-MPS3/MB/HBI0309C/ETHOSU/images.txt
```
3. Open the first serial port available from MPS3, for example,