You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I happened to be doing some testing of different sha256 impls and found that in my test mbedtls sha256 with esp32/arduino v2.0.3 was slower than v2.0.2, taking about 80 microseconds compared to 43 microseconds on v2.0.2, so nearly twice as slow.
I remember seeing this issue that I think is saying from 2.0.3 mbedtls will use the ESP32 hardware acceleration for things like mbedtls sha256 hashes.
Its not a big deal I'm just curious - shouldn't v2.0.3 be faster at this if its using hardware acceleration now?
I do have a testcase to show this if interested but would have to do some work to pullout just this sha256 test into a standalone project.
Sketch
.
Debug Message
.
Other Steps to Reproduce
.
I have checked existing issues, online documentation and the Troubleshooting Guide
I confirm I have checked existing issues, online documentation and Troubleshooting guide.
The text was updated successfully, but these errors were encountered:
@torntrousers actually it's slower in hardware, because the clock speed of the hardware is 80MHz and the CPU at 240MHz can calculate a bit faster. If the CPU was also at 80MHz, then hardware would have been twice faster or so...
Board
M5Stack EduKit
Device Description
.
Hardware Configuration
.
Version
v2.0.3
IDE Name
platformio
Operating System
Windows
Flash frequency
40MHz
PSRAM enabled
yes
Upload speed
115200
Description
I happened to be doing some testing of different sha256 impls and found that in my test mbedtls sha256 with esp32/arduino v2.0.3 was slower than v2.0.2, taking about 80 microseconds compared to 43 microseconds on v2.0.2, so nearly twice as slow.
I remember seeing this issue that I think is saying from 2.0.3 mbedtls will use the ESP32 hardware acceleration for things like mbedtls sha256 hashes.
Its not a big deal I'm just curious - shouldn't v2.0.3 be faster at this if its using hardware acceleration now?
I do have a testcase to show this if interested but would have to do some work to pullout just this sha256 test into a standalone project.
Sketch
Debug Message
Other Steps to Reproduce
.
I have checked existing issues, online documentation and the Troubleshooting Guide
The text was updated successfully, but these errors were encountered: