Introduce
Sakura Pi RK3308B is our first officially released SBC.
It is a low-power, cost-effective development board powered by the Rockchip RK3308B SoC.
Processor
Thanks to the ARM® Cortex-A35 processor, the RK3308B boasts numerous advantages such as low power consumption, low cost, and minimal heat generation.
In modern mobile device architectures, the Cortex-A35 often serves as a 'little core' (efficiency core/E-core), striking a graceful balance between handling complex tasks and maintaining long standby times with its efficient performance.
Features
The basic specifications of the Sakura Pi RK3308B are as follows
- Processor: RK3308B
- Memory: 512MB DDR3
- Wireless WLAN: Supports IEEE 802.11a/b/g/n/ac, dual-band 2.4GHz/5GHz
- Bluetooth: Supports BLE 4.2
- eMMC: Optional 0GB / 8GB / 16GB
- Indicator LEDs: WS2812 (SPI) × 4
- LCD Interface: RGB666
Pinout
The Sakura Pi RK3308B features a 40-pin header compatible with the Raspberry Pi Zero W. By default, U-Boot and the kernel use UART2 as the output interface.
Benchmark
UnixBench
| CPU Freq | Single Core | Multi Core |
|---|---|---|
| 1.0GHz | 172.2 | 563.4 |
| 1.3GHz | 210.3 | 671.9 |
1.0GHz UnixBench
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: sakurapi-rk3308b: GNU/Linux
OS: GNU/Linux -- 6.12.23-current-rockchip64 -- #2 SMP PREEMPT Thu Apr 10 20:39:41 HKT 2025
Machine: aarch64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
CPU 1: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
CPU 2: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
CPU 3: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
19:08:56 up 56 min, 1 user, load average: 0.73, 0.73, 0.42; runlevel 2025-04-11
------------------------------------------------------------------------
Benchmark Run: Fri Apr 11 2025 19:08:56 - 19:37:02
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 4390531.8 lps (10.0 s, 7 samples)
Double-Precision Whetstone 1204.4 MWIPS (10.0 s, 7 samples)
Execl Throughput 288.4 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 93939.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 27259.1 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 244850.1 KBps (30.0 s, 2 samples)
Pipe Throughput 221670.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 17347.5 lps (10.0 s, 7 samples)
Process Creation 543.1 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 952.2 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 290.6 lpm (60.1 s, 2 samples)
System Call Overhead 307830.4 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 4390531.8 376.2
Double-Precision Whetstone 55.0 1204.4 219.0
Execl Throughput 43.0 288.4 67.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 93939.0 237.2
File Copy 256 bufsize 500 maxblocks 1655.0 27259.1 164.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 244850.1 422.2
Pipe Throughput 12440.0 221670.8 178.2
Pipe-based Context Switching 4000.0 17347.5 43.4
Process Creation 126.0 543.1 43.1
Shell Scripts (1 concurrent) 42.4 952.2 224.6
Shell Scripts (8 concurrent) 6.0 290.6 484.4
System Call Overhead 15000.0 307830.4 205.2
========
System Benchmarks Index Score 172.2
------------------------------------------------------------------------
Benchmark Run: Fri Apr 11 2025 19:37:02 - 20:05:16
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 17375717.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4755.7 MWIPS (10.0 s, 7 samples)
Execl Throughput 893.7 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 350019.2 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 106703.9 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 610322.2 KBps (30.0 s, 2 samples)
Pipe Throughput 877759.9 lps (10.0 s, 7 samples)
Pipe-based Context Switching 113403.0 lps (10.0 s, 7 samples)
Process Creation 1898.4 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 2244.3 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 296.1 lpm (60.6 s, 2 samples)
System Call Overhead 1215141.6 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 17375717.0 1488.9
Double-Precision Whetstone 55.0 4755.7 864.7
Execl Throughput 43.0 893.7 207.8
File Copy 1024 bufsize 2000 maxblocks 3960.0 350019.2 883.9
File Copy 256 bufsize 500 maxblocks 1655.0 106703.9 644.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 610322.2 1052.3
Pipe Throughput 12440.0 877759.9 705.6
Pipe-based Context Switching 4000.0 113403.0 283.5
Process Creation 126.0 1898.4 150.7
Shell Scripts (1 concurrent) 42.4 2244.3 529.3
Shell Scripts (8 concurrent) 6.0 296.1 493.4
System Call Overhead 15000.0 1215141.6 810.1
========
System Benchmarks Index Score 563.4
1.3GHz UnixBench
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: sakurapi-rk3308b: GNU/Linux
OS: GNU/Linux -- 6.12.23-current-rockchip64 -- #2 SMP PREEMPT Thu Apr 10 20:39:41 HKT 2025
Machine: aarch64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
CPU 1: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
CPU 2: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
CPU 3: ARM Cortex-A35 (48.0 bogomips)
CPU Features: fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
05:51:01 up 3 min, 1 user, load average: 0.36, 0.21, 0.10; runlevel 2025-04-12
------------------------------------------------------------------------
Benchmark Run: Sat Apr 12 2025 05:51:01 - 06:19:05
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 5640815.9 lps (10.0 s, 7 samples)
Double-Precision Whetstone 1548.9 MWIPS (10.1 s, 7 samples)
Execl Throughput 333.2 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 116631.1 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 34717.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 283747.8 KBps (30.0 s, 2 samples)
Pipe Throughput 286177.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 22383.8 lps (10.0 s, 7 samples)
Process Creation 612.4 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1111.3 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 325.8 lpm (60.1 s, 2 samples)
System Call Overhead 395002.8 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 5640815.9 483.4
Double-Precision Whetstone 55.0 1548.9 281.6
Execl Throughput 43.0 333.2 77.5
File Copy 1024 bufsize 2000 maxblocks 3960.0 116631.1 294.5
File Copy 256 bufsize 500 maxblocks 1655.0 34717.5 209.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 283747.8 489.2
Pipe Throughput 12440.0 286177.6 230.0
Pipe-based Context Switching 4000.0 22383.8 56.0
Process Creation 126.0 612.4 48.6
Shell Scripts (1 concurrent) 42.4 1111.3 262.1
Shell Scripts (8 concurrent) 6.0 325.8 543.1
System Call Overhead 15000.0 395002.8 263.3
========
System Benchmarks Index Score 210.3
------------------------------------------------------------------------
Benchmark Run: Sat Apr 12 2025 06:19:05 - 06:47:14
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 22403396.8 lps (10.0 s, 7 samples)
Double-Precision Whetstone 6128.9 MWIPS (10.1 s, 7 samples)
Execl Throughput 977.2 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 419419.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 136432.9 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 638890.6 KBps (30.0 s, 2 samples)
Pipe Throughput 1134168.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 145101.5 lps (10.0 s, 7 samples)
Process Creation 2031.2 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 2503.7 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 330.8 lpm (60.5 s, 2 samples)
System Call Overhead 1565701.4 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 22403396.8 1919.7
Double-Precision Whetstone 55.0 6128.9 1114.3
Execl Throughput 43.0 977.2 227.3
File Copy 1024 bufsize 2000 maxblocks 3960.0 419419.9 1059.1
File Copy 256 bufsize 500 maxblocks 1655.0 136432.9 824.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 638890.6 1101.5
Pipe Throughput 12440.0 1134168.3 911.7
Pipe-based Context Switching 4000.0 145101.5 362.8
Process Creation 126.0 2031.2 161.2
Shell Scripts (1 concurrent) 42.4 2503.7 590.5
Shell Scripts (8 concurrent) 6.0 330.8 551.3
System Call Overhead 15000.0 1565701.4 1043.8
========
System Benchmarks Index Score 671.9
LLVMPipe GLSL
info
GitHub Repository astefanutti/kms-glsl
kms-glsl can rendering glsl shaders via GPU，but RK3308B doesn't has a GPU hardware thus causes Linux forces use LLVMPipe to render it.
We use LLVMPipe to render glsl shaders, conducting stress tests at a CPU frequency of 1.3GHz. Without any additional cooling, the junction temperature reaches around 140 degrees Celsius.
