From 897cc14968e017b1f48f376f7f7cefc515c5fe88 Mon Sep 17 00:00:00 2001
From: Andreas Nevalainen <andreas.nevalainen@arm.com>
Date: Wed, 28 Oct 2020 15:42:08 +0100
Subject: MLBEDSW-3222: Bias tensors in fast storage

For IFM streamed cascades bias tensors are read several times.
Moves these tensors to fast storage and add DMA commands.

Change-Id: I630f6275986c1b5e3f126c925b11e22500fb1128
Signed-off-by: Andreas Nevalainen <andreas.nevalainen@arm.com>
---
 OPTIONS.md | 9 +++++++++
 1 file changed, 9 insertions(+)

(limited to 'OPTIONS.md')

diff --git a/OPTIONS.md b/OPTIONS.md
index a7b513c8..92201518 100644
--- a/OPTIONS.md
+++ b/OPTIONS.md
@@ -72,6 +72,15 @@ system's SRAM this optimisation is required.
 vela network.tflite --cascading False
 ```
 
+### Keep scale placement
+
+Prevents scheduler from placing scale tensors for IFM streamed passes in SRAM
+and keeps these in flash.  
+
+```bash
+vela network.tflite --keep-scale-placement
+```
+
 ### Force Block Config
 
 Force a specific block configuration in the format HxWxC, where H, W, and C are
-- 
cgit v1.2.1