better benchmark

JAlcocerT · Mar 23, 2024 · fa7f655 · fa7f655
1 parent 45ffaad
commit fa7f655
Show file tree

Hide file tree

Showing 5 changed files with 30 additions and 3 deletions.
diff --git a/Z_MQTT/MQTT_DASH/readme.md b/Z_MQTT/MQTT_DASH/readme.md
@@ -0,0 +1 @@
+## DASH Webhooks + MQTT
diff --git a/Z_MQTT/Readme.md b/Z_MQTT/Readme.md
@@ -0,0 +1,6 @@
+* Edge Computing - https://github.com/qijianpeng/awesome-edge-computing
+* Domotica para pobres - https://forocoches.com/foro/showthread.php?t=6655749
+
+* MicroPython - https://awesome-micropython.com/
+
+* https://github.com/ahundt/awesome-robotics
diff --git a/_posts/2023-05-02-pi-vs-orange.md b/_posts/2023-05-02-pi-vs-orange.md
@@ -271,6 +271,8 @@ https://cpu.userbenchmark.com/Compare/Intel-Core-i7-1185G7-vs-AMD-Ryzen-5-4600G/
 
 ### CPU Consumption
 
+The orange idles ~2W
+
 
 ### Build Consumption
 

diff --git a/_posts/2024-03-15-minipc-vs-pi.md b/_posts/2024-03-15-minipc-vs-pi.md
@@ -77,15 +77,27 @@ And these are the temperatures [registered by NetData](https://fossengineer.com/
 
 Comparing N95 (x86) with the Rockchip RK3588S (ARM64).
 
+#### Temperatures
+
+```sh
+sudo stress --cpu  8 --timeout 120
+```
+
+* Orange Pi 5 - 80C & 8w peak power (no fan enabled) and it quickly goes back to the ~45C after the test
+* BMax B4 (N95, fan enabled that goes to full speed) - 66C and 15W peak power
+
+
+#### Phoronix
+
 For Synthetic benchmarks I have used [phoronix](#how-to-benchmark-with-phoronix-test-suite):
 
 ![BMAX B4 - Temperature during Docker Build](/img/minipc-vs-pis/n95-cpu-phoronix.png)
 _Intel N95 4 cores with phoronix Open Source Benchmark_
 
 ![BMAX B4 - Temperature during Docker Build](/img/minipc-vs-pis/orangepi5-cpu-phoronix.png)
-_The Orange Pi 8 Cores is a beast_
-
+_The Orange Pi 8 Cores is a beast scoring 38s_
 
+For reference, I [benchmarked *bigger* CPUs here](https://jalcocert.github.io/Linux/docs/linux__cloud/benchmark/). Plot twist, both CPUs (specially the Rockchip, has nothing to envy)
 
 ---
 

diff --git a/_posts/2024-12-01-ollama-raspberry-pi.md b/_posts/2024-12-01-ollama-raspberry-pi.md
@@ -6,6 +6,12 @@ categories: [IoT & Data Analytics, AI]
 tags: [Self-Hosting, Docker]
 ---
 
+## Best Performing LLMs for a RPi
+
 ---
 
-## FAQ
+## FAQ
+
+### How to use Ollama with a Raspberry Pi
+
+You can [SelfHost Ollama with Docker](https://fossengineer.com/selfhosting-llms-ollama/) and choose a small enough model that it will fit in the RAM.