Repository navigation
Raspberry Pi 500+ #81
Description
Basic information
- Board URL (official): https://www.raspberrypi.com/products/raspberry-pi-500-plus/
- Board purchased from: (Provided by Raspberry Pi)
- Board purchase date: September 25, 2025
- Board specs (as tested): 16 GB RAM / 256 GB NVMe SSD
- Board price (as tested): $200
Linux/system information
# output of `screenfetch`
_,met$$$$$gg. pi@raspberrypi
,g$$$$$$$$$$$$$$$P. OS: Debian 12 bookworm
,g$$P"" """Y$$.". Kernel: aarch64 Linux 6.12.20+rpt-rpi-2712
,$$P' `$$$. Uptime: 2m
',$$P ,ggs. `$$b: Packages: 1951
`d$$' ,$P"' . $$$ Shell: dash
$$P d$' , $$P Disk: 13G / 243G (6%)
$$: $$. - ,d$$' CPU: ARM Cortex-A76 @ 4x 2.4GHz
$$\; Y$b._ _,d$P' GPU:
Y$$. `.`"Y$$$$P"' RAM: 785MiB / 16215MiB
`$$b "-.__
`Y$$
`Y$$.
`$$b.
`Y$$b.
`"Y$b._
`""""
# output of `uname -a`
Linux raspberrypi 6.12.20+rpt-rpi-2712 #1 SMP PREEMPT Debian 1:6.12.20-1+rpt1~bpo12+1 (2025-03-19) aarch64 GNU/Linux
Benchmark results
CPU
- Geekbench 6: (883 single / 2111 multi - https://browser.geekbench.com/v6/cpu/14699737)
- 35.579 Gflops at 12.9 W for 2.76 Gflops/W (geerlingguy/top500-benchmark HPL result)
Power
- Idle power draw (at wall): 2.9W
- Idle power draw (at wall - LEDs on white at full brightness): 7W
- Maximum simulated power draw (
stress-ng --matrix 0): 9.4 W - During Geekbench multicore benchmark: 11.7 W
- During
top500HPL benchmark: 12.9 W
Disk
Raspberry Pi 256 GB SSD (SAMSUNG MZ9LQ256HBJD-00BVL)
| Benchmark | Result |
|---|---|
| iozone 4K random read | 49.38 MB/s |
| iozone 4K random write | 189.98 MB/s |
| iozone 1M random read | 742.21 MB/s |
| iozone 1M random write | 749.75 MB/s |
| iozone 1M sequential read | 797.70 MB/s |
| iozone 1M sequential write | 750.91 MB/s |
Raspberry Pi 32GB microSD
| Benchmark | Result |
|---|---|
| iozone 4K random read | 18.08 MB/s |
| iozone 4K random write | 6.59 MB/s |
| iozone 1M random read | 89.05 MB/s |
| iozone 1M random write | 30.19 MB/s |
| iozone 1M sequential read | 89.64 MB/s |
| iozone 1M sequential write | 30.29 MB/s |
Network
iperf3 results:
Built-in Ethernet
iperf3 -c $SERVER_IP: 938 Mbpsiperf3 -c $SERVER_IP --reverse: 942 Mbpsiperf3 -c $SERVER_IP --bidir: 933 Mbps up, 936 Mbps down
Built-in WiFi
iperf3 -c $SERVER_IP: 258 Mbpsiperf3 -c $SERVER_IP --reverse: 228 Mbpsiperf3 -c $SERVER_IP --bidir: 85.4 Mbps up, 145 Mbps down
(Be sure to test all interfaces, noting any that are non-functional.)
GPU
glmark2
glmark2-es2-wayland results:
=======================================================
glmark2 2023.01
=======================================================
OpenGL Information
GL_VENDOR: Broadcom
GL_RENDERER: V3D 7.1.10.2
GL_VERSION: OpenGL ES 3.1 Mesa 24.2.8-1~bpo12+rpt2
Surface Config: buf=32 r=8 g=8 b=8 a=8 depth=24 stencil=0 samples=0
Surface Size: 800x600 windowed
=======================================================
[build] use-vbo=false: FPS: 2291 FrameTime: 0.437 ms
[build] use-vbo=true: FPS: 2994 FrameTime: 0.334 ms
[texture] texture-filter=nearest: FPS: 2370 FrameTime: 0.422 ms
[texture] texture-filter=linear: FPS: 2353 FrameTime: 0.425 ms
[texture] texture-filter=mipmap: FPS: 2406 FrameTime: 0.416 ms
[shading] shading=gouraud: FPS: 2508 FrameTime: 0.399 ms
[shading] shading=blinn-phong-inf: FPS: 2135 FrameTime: 0.468 ms
[shading] shading=phong: FPS: 1767 FrameTime: 0.566 ms
[shading] shading=cel: FPS: 1719 FrameTime: 0.582 ms
[bump] bump-render=high-poly: FPS: 1294 FrameTime: 0.773 ms
[bump] bump-render=normals: FPS: 2608 FrameTime: 0.383 ms
[bump] bump-render=height: FPS: 2432 FrameTime: 0.411 ms
[effect2d] kernel=0,1,0;1,-4,1;0,1,0;: FPS: 958 FrameTime: 1.044 ms
[effect2d] kernel=1,1,1,1,1;1,1,1,1,1;1,1,1,1,1;: FPS: 388 FrameTime: 2.583 ms
[pulsar] light=false:quads=5:texture=false: FPS: 2487 FrameTime: 0.402 ms
[desktop] blur-radius=5:effect=blur:passes=1:separable=true:windows=4: FPS: 239 FrameTime: 4.192 ms
[desktop] effect=shadow:windows=4: FPS: 883 FrameTime: 1.134 ms
[buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=map: FPS: 462 FrameTime: 2.165 ms
[buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata: FPS: 446 FrameTime: 2.246 ms
[buffer] columns=200:interleave=true:update-dispersion=0.9:update-fraction=0.5:update-method=map: FPS: 481 FrameTime: 2.083 ms
[ideas] speed=duration: FPS: 1848 FrameTime: 0.541 ms
[jellyfish] <default>: FPS: 989 FrameTime: 1.011 ms
[terrain] <default>: FPS: 69 FrameTime: 14.664 ms
[shadow] <default>: FPS: 312 FrameTime: 3.215 ms
[refract] <default>: FPS: 104 FrameTime: 9.659 ms
[conditionals] fragment-steps=0:vertex-steps=0: FPS: 2746 FrameTime: 0.364 ms
[conditionals] fragment-steps=5:vertex-steps=0: FPS: 1899 FrameTime: 0.527 ms
[conditionals] fragment-steps=0:vertex-steps=5: FPS: 2690 FrameTime: 0.372 ms
[function] fragment-complexity=low:fragment-steps=5: FPS: 2269 FrameTime: 0.441 ms
[function] fragment-complexity=medium:fragment-steps=5: FPS: 1529 FrameTime: 0.654 ms
[loop] fragment-loop=false:fragment-steps=5:vertex-steps=5: FPS: 2189 FrameTime: 0.457 ms
[loop] fragment-steps=5:fragment-uniform=false:vertex-steps=5: FPS: 2190 FrameTime: 0.457 ms
[loop] fragment-steps=5:fragment-uniform=true:vertex-steps=5: FPS: 1452 FrameTime: 0.689 ms
=======================================================
glmark2 Score: 1620
=======================================================
vkmark
vkmark results:
=======================================================
vkmark 2025.01
=======================================================
Vendor ID: 0x14E4
Device ID: 0x55701C33
Device Name: V3D 7.1.10.2
Driver Version: 100671496
Device UUID: 62ba0c091d6a41ea52f57d647c1d0ff0
=======================================================
[vertex] device-local=true: FPS: 2088 FrameTime: 0.479 ms
[vertex] device-local=false: FPS: 1777 FrameTime: 0.563 ms
[texture] anisotropy=0: FPS: 1654 FrameTime: 0.605 ms
[texture] anisotropy=16: FPS: 1592 FrameTime: 0.628 ms
[shading] shading=gouraud: FPS: 1485 FrameTime: 0.673 ms
[shading] shading=blinn-phong-inf: FPS: 1369 FrameTime: 0.730 ms
[shading] shading=phong: FPS: 1188 FrameTime: 0.842 ms
[shading] shading=cel: FPS: 1192 FrameTime: 0.839 ms
[effect2d] kernel=edge:gs FPS: 747 FrameTime: 1.339 ms
[effect2d] kernel=blur: FPS: 356 FrameTime: 2.809 ms
[desktop] <default>: FPS: 804 FrameTime: 1.244 ms
[cube] <default>: FPS: 2040 FrameTime: 0.490 ms
[clear] <default>: FPS: 2231 FrameTime: 0.448 ms
=======================================================
vkmark Score: 1424
=======================================================
Note:
vkmarkneeds to be compiled from source on Debian 12 and earlier.
GravityMark and other benchmarks with external AMD 7900 XT
I tested an AMD Radeon 7900 XT on the Pi 500 using an Oculink to M.2 adapter, and here are my test results for that:
- glmark2: 8011
- vkmark: 12510
- GravityMark: 65,842 (pretty decent matched to other runs)
LLMs and AI
Testing on the Pi 500+: geerlingguy/ai-benchmarks#24 (Llama 3.2:3b - 5.55 tokens/s)
Testing with an eGPU: AMD RX 7900 XT 20GB.
Memory
tinymembench results:
Click to expand memory benchmark result
tinymembench v0.4.10 (simple benchmark for memory throughput and latency)
==========================================================================
== Memory bandwidth tests ==
== ==
== Note 1: 1MB = 1000000 bytes ==
== Note 2: Results for 'copy' tests show how many bytes can be ==
== copied per second (adding together read and writen ==
== bytes would have provided twice higher numbers) ==
== Note 3: 2-pass copy means that we are using a small temporary buffer ==
== to first fetch data into it, and only then write it to the ==
== destination (source -> L1 cache, L1 cache -> destination) ==
== Note 4: If sample standard deviation exceeds 0.1%, it is shown in ==
== brackets ==
==========================================================================
C copy backwards : 6091.2 MB/s (0.8%)
C copy backwards (32 byte blocks) : 5973.3 MB/s
C copy backwards (64 byte blocks) : 5972.4 MB/s
C copy : 5894.2 MB/s (0.1%)
C copy prefetched (32 bytes step) : 5891.3 MB/s
C copy prefetched (64 bytes step) : 5897.5 MB/s (0.4%)
C 2-pass copy : 5337.8 MB/s (0.5%)
C 2-pass copy prefetched (32 bytes step) : 5896.0 MB/s (0.3%)
C 2-pass copy prefetched (64 bytes step) : 5898.1 MB/s
C fill : 9751.2 MB/s (0.3%)
C fill (shuffle within 16 byte blocks) : 9763.8 MB/s (0.4%)
C fill (shuffle within 32 byte blocks) : 9734.1 MB/s (0.5%)
C fill (shuffle within 64 byte blocks) : 9690.8 MB/s
NEON 64x2 COPY : 5818.6 MB/s
NEON 64x2x4 COPY : 5837.8 MB/s
NEON 64x1x4_x2 COPY : 5847.8 MB/s
NEON 64x2 COPY prefetch x2 : 5596.1 MB/s (0.2%)
NEON 64x2x4 COPY prefetch x1 : 5707.2 MB/s
NEON 64x2 COPY prefetch x1 : 5573.0 MB/s
NEON 64x2x4 COPY prefetch x1 : 5707.8 MB/s
---
standard memcpy : 5865.0 MB/s (0.2%)
standard memset : 9738.5 MB/s (0.4%)
---
NEON LDP/STP copy : 5851.2 MB/s
NEON LDP/STP copy pldl2strm (32 bytes step) : 5954.0 MB/s
NEON LDP/STP copy pldl2strm (64 bytes step) : 5956.5 MB/s
NEON LDP/STP copy pldl1keep (32 bytes step) : 5959.0 MB/s
NEON LDP/STP copy pldl1keep (64 bytes step) : 5958.6 MB/s
NEON LD1/ST1 copy : 5845.9 MB/s
NEON STP fill : 9753.7 MB/s (0.7%)
NEON STNP fill : 9745.6 MB/s (0.8%)
ARM LDP/STP copy : 5844.7 MB/s
ARM STP fill : 9757.9 MB/s (0.5%)
ARM STNP fill : 9778.5 MB/s (0.1%)
==========================================================================
== Framebuffer read tests. ==
== ==
== Many ARM devices use a part of the system memory as the framebuffer, ==
== typically mapped as uncached but with write-combining enabled. ==
== Writes to such framebuffers are quite fast, but reads are much ==
== slower and very sensitive to the alignment and the selection of ==
== CPU instructions which are used for accessing memory. ==
== ==
== Many x86 systems allocate the framebuffer in the GPU memory, ==
== accessible for the CPU via a relatively slow PCI-E bus. Moreover, ==
== PCI-E is asymmetric and handles reads a lot worse than writes. ==
== ==
== If uncached framebuffer reads are reasonably fast (at least 100 MB/s ==
== or preferably >300 MB/s), then using the shadow framebuffer layer ==
== is not necessary in Xorg DDX drivers, resulting in a nice overall ==
== performance improvement. For example, the xf86-video-fbturbo DDX ==
== uses this trick. ==
==========================================================================
NEON LDP/STP copy (from framebuffer) : 1605.5 MB/s
NEON LDP/STP 2-pass copy (from framebuffer) : 1456.7 MB/s
NEON LD1/ST1 copy (from framebuffer) : 1618.8 MB/s
NEON LD1/ST1 2-pass copy (from framebuffer) : 1465.1 MB/s
ARM LDP/STP copy (from framebuffer) : 1615.2 MB/s
ARM LDP/STP 2-pass copy (from framebuffer) : 1464.6 MB/s
==========================================================================
== Memory latency test ==
== ==
== Average time is measured for random memory accesses in the buffers ==
== of different sizes. The larger is the buffer, the more significant ==
== are relative contributions of TLB, L1/L2 cache misses and SDRAM ==
== accesses. For extremely large buffer sizes we are expecting to see ==
== page table walk with several requests to SDRAM for almost every ==
== memory access (though 64MiB is not nearly large enough to experience ==
== this effect to its fullest). ==
== ==
== Note 1: All the numbers are representing extra time, which needs to ==
== be added to L1 cache latency. The cycle timings for L1 cache ==
== latency can be usually found in the processor documentation. ==
== Note 2: Dual random read means that we are simultaneously performing ==
== two independent memory accesses at a time. In the case if ==
== the memory subsystem can't handle multiple outstanding ==
== requests, dual random read has the same timings as two ==
== single reads performed one after another. ==
==========================================================================
block size : single random read / dual random read
1024 : 0.0 ns / 0.0 ns
2048 : 0.0 ns / 0.0 ns
4096 : 0.0 ns / 0.0 ns
8192 : 0.0 ns / 0.0 ns
16384 : 0.0 ns / 0.0 ns
32768 : 0.0 ns / 0.0 ns
65536 : 0.0 ns / 0.0 ns
131072 : 1.1 ns / 1.5 ns
262144 : 1.6 ns / 2.0 ns
524288 : 1.9 ns / 2.2 ns
1048576 : 8.3 ns / 11.2 ns
2097152 : 12.1 ns / 14.5 ns
4194304 : 50.9 ns / 75.7 ns
8388608 : 79.5 ns / 104.9 ns
16777216 : 93.9 ns / 114.3 ns
33554432 : 103.3 ns / 119.3 ns
67108864 : 108.6 ns / 122.5 ns
sbc-bench results
Run sbc-bench and paste a link to the results here:
wget https://raw.githubusercontent.com/ThomasKaiser/sbc-bench/master/sbc-bench.sh
sudo /bin/bash ./sbc-bench.sh -r
Phoronix Test Suite
Results from pi-general-benchmark.sh:
- pts/encode-mp3: 11.784 sec
- pts/x264 4K: 4.57 fps
- pts/x264 1080p: 19.30 fps
- pts/phpbench: 435919
- pts/build-linux-kernel (defconfig): 2197.595 sec
Activity
Issue #81: Add Raspberry Pi 500 plus.
geerlingguy commentedon Sep 25, 2025
More info: