MetaComputing AI PC Framework 13 Mainboard with Cix CP8180

 ## Basic information - Board URL (official): https://metacomputing.io/products/metacomputing-aipc - Board purchased from: (Provided for review by MetaComputing) - Board purchase date: Delivered 2026-04-01 - Board specs (as tested): 16GB RAM - Board price (as tested): $798 ($1,280 with Framework 13 Chassis) ## Linux/system information ``` # output of `fastfetch` .-/+oossssoo+/-. jgeerling@framework-arm `:+ssssssssssssssssss+:` ----------------------- -+ssssssssssssssssssyyssss+- OS: Ubuntu 25.04 aarch64 .ossssssssssssssssssdMMMNysssso. Host: MC FML13V04 Board 1.0 /ssssssssssshdmmNNmmyNMMMMhssssss/ Kernel: 6.6.89-cix-build-generic +ssssssssshmydMMMMMMMNddddyssssssss+ Uptime: 8 mins /sssssssshNMMMyhhyyyyhmNMMMNhssssssss/ Packages: 1908 (dpkg) .ssssssssdMMMNhsssssssssshNMMMdssssssss. Shell: bash 5.2.37 +sssshhhyNMMNyssssssssssssyNMMMysssssss+ Resolution: 2256x1504 ossyNMMMNyMMhsssssssssssssshmmmhssssssso Terminal: /dev/pts/1 ossyNMMMNyMMhsssssssssssssshmmmhssssssso CPU: CIX P1 CP8180 (12) @ 2.600GHz +sssshhhyNMMNyssssssssssssyNMMMysssssss+ Memory: 2363MiB / 15092MiB .ssssssssdMMMNhsssssssssshNMMMdssssssss. /sssssssshNMMMyhhyyyyhdNMMMNhssssssss/ +sssssssssdmydMMMMMMMMddddyssssssss+ /ssssssssssshdmNNNNmyNMMMMhssssss/ .ossssssssssssssssssdMMMNysssso. -+sssssssssssssssssyyyssss+- `:+ssssssssssssssssss+:` .-/+oossssoo+/-. # output of `uname -a` Linux framework-arm 6.6.89-cix-build-generic #2 SMP PREEMPT Fri Jan 30 10:32:39 CST 2026 aarch64 aarch64 aarch64 GNU/Linux ``` ### System topology <img width="1252" height="324" alt="Image" src="https://github.com/user-attachments/assets/ff4e85d5-94b7-4f73-8716-5e58068bb1c5" /> Note: `lstopo` results may be missing some information on new and strange SoCs. ## Benchmark results ### CPU - Geekbench 6: 1306 single / 6477 multi - https://browser.geekbench.com/v6/cpu/17448233 ([run 2](https://browser.geekbench.com/v6/cpu/17452181), [run 3 - latest BIOS](https://browser.geekbench.com/v6/cpu/17534169)) - 56.391 Gflops at 26.4W, for **2.14 Gflops/W** ([geerlingguy/top500-benchmark](https://github.com/geerlingguy/top500-benchmark) [HPL result](https://github.com/geerlingguy/top500-benchmark/issues/94)) ### Power - Idle power draw (at wall): 7.9 W (10.2W with screen at 50% brightness) - Maximum simulated power draw (`stress-ng --matrix 0`): 27.5 W - During Geekbench multicore benchmark: 25.7 W - During `top500` HPL benchmark: 26.4 W Power measurements taken with battery at 100%, with charger plugged in an additional 8 hours so no extra trickle charging was happening, and with display off. ### Disk #### Inland TN320 NVMe SSD (240 GB) [comment]: # (Run `lsblk -o NAME,FSTYPE,LABEL,MOUNTPOINT,SIZE,MODEL` to get model) | Benchmark | Result | | -------------------------- | ------ | | iozone 4K random read | 50.82 MB/s | | iozone 4K random write | 155.39 MB/s | | iozone 1M random read | 1337.00 MB/s | | iozone 1M random write | 1282.01 MB/s | | iozone 1M sequential read | 1748.13 MB/s | | iozone 1M sequential write | 1409.63 MB/s | ``` wget https://raw.githubusercontent.com/geerlingguy/pi-cluster/master/benchmarks/disk-benchmark.sh chmod +x disk-benchmark.sh sudo MOUNT_PATH=/ TEST_SIZE=1g ./disk-benchmark.sh ``` Run benchmark on any attached storage device (e.g. eMMC, microSD, NVMe, SATA) and add results under an additional heading. ### Network `iperf3` results: #### WiFi 6 (Intel AX200) - `iperf3 -c $SERVER_IP`: 156 Mbps - `iperf3 -c $SERVER_IP --reverse`: 181 Mbps - `iperf3 -c $SERVER_IP --bidir`: 12.9 Mbps up, 115 Mbps down (Be sure to test all interfaces, noting any that are non-functional.) ## GPU ### glmark2 `glmark2-es2` / `glmark2-es2-wayland` results: ``` ======================================================= glmark2 2023.01 ======================================================= OpenGL Information GL_VENDOR: ARM GL_RENDERER: Mali-G720-Immortalis GL_VERSION: OpenGL ES 3.2 v1.r53p0-00eac0.51e7b43a8ca32d45d7e15e03e4ee52d8 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: 3450 FrameTime: 0.290 ms [build] use-vbo=true: FPS: 4746 FrameTime: 0.211 ms [texture] texture-filter=nearest: FPS: 4898 FrameTime: 0.204 ms [texture] texture-filter=linear: FPS: 4659 FrameTime: 0.215 ms [texture] texture-filter=mipmap: FPS: 4808 FrameTime: 0.208 ms [shading] shading=gouraud: FPS: 4686 FrameTime: 0.213 ms [shading] shading=blinn-phong-inf: FPS: 4705 FrameTime: 0.213 ms [shading] shading=phong: FPS: 4535 FrameTime: 0.221 ms [shading] shading=cel: FPS: 4547 FrameTime: 0.220 ms [bump] bump-render=high-poly: FPS: 3315 FrameTime: 0.302 ms [bump] bump-render=normals: FPS: 5045 FrameTime: 0.198 ms [bump] bump-render=height: FPS: 4652 FrameTime: 0.215 ms [effect2d] kernel=0,1,0;1,-4,1;0,1,0;: FPS: 4637 FrameTime: 0.216 ms [effect2d] kernel=1,1,1,1,1;1,1,1,1,1;1,1,1,1,1;: FPS: 4209 FrameTime: 0.238 ms [pulsar] light=false:quads=5:texture=false: FPS: 4710 FrameTime: 0.212 ms [desktop] blur-radius=5:effect=blur:passes=1:separable=true:windows=4: FPS: 2882 FrameTime: 0.347 ms [desktop] effect=shadow:windows=4: FPS: 3292 FrameTime: 0.304 ms [buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=map: FPS: 496 FrameTime: 2.020 ms [buffer] columns=200:interleave=false:update-dispersion=0.9:update-fraction=0.5:update-method=subdata: FPS: 510 FrameTime: 1.962 ms [buffer] columns=200:interleave=true:update-dispersion=0.9:update-fraction=0.5:update-method=map: FPS: 747 FrameTime: 1.339 ms [ideas] speed=duration: FPS: 2201 FrameTime: 0.454 ms [jellyfish] <default>: FPS: 3852 FrameTime: 0.260 ms [terrain] <default>: FPS: 615 FrameTime: 1.628 ms [shadow] <default>: FPS: 3948 FrameTime: 0.253 ms [refract] <default>: FPS: 1257 FrameTime: 0.796 ms [conditionals] fragment-steps=0:vertex-steps=0: FPS: 4374 FrameTime: 0.229 ms [conditionals] fragment-steps=5:vertex-steps=0: FPS: 4382 FrameTime: 0.228 ms [conditionals] fragment-steps=0:vertex-steps=5: FPS: 4489 FrameTime: 0.223 ms [function] fragment-complexity=low:fragment-steps=5: FPS: 4434 FrameTime: 0.226 ms [function] fragment-complexity=medium:fragment-steps=5: FPS: 4360 FrameTime: 0.229 ms [loop] fragment-loop=false:fragment-steps=5:vertex-steps=5: FPS: 4066 FrameTime: 0.246 ms [loop] fragment-steps=5:fragment-uniform=false:vertex-steps=5: FPS: 4670 FrameTime: 0.214 ms [loop] fragment-steps=5:fragment-uniform=true:vertex-steps=5: FPS: 4717 FrameTime: 0.212 ms ======================================================= glmark2 Score: 3723 ======================================================= ``` ### vkmark `vkmark` results: ``` 1. Install vkmark with `sudo apt install -y vkmark` 2. Run `vkmark` (with `DISPLAY=:0` prepended if running over SSH) 3. Replace this block of text with the results. ``` > **Note**: `vkmark` needs to be [compiled from source](https://github.com/geerlingguy/sbc-reviews/issues/76) on Debian 12 and earlier. ### GravityMark GravityMark results:  [7,627](https://gravitymark.tellusim.com/report/?id=8ed048387ccca9a89ba294aed959523bd64bd534) Note: These benchmarks require an active display on the device. Not all devices may be able to run `glmark2-es2`, so in that case, make a note and move on! ### AI / LLM Inference Basic AI inference results: See full results: https://github.com/geerlingguy/ai-benchmarks/issues/51 | model | size | params | backend | threads | test | t/s | | ------------------------------ | ---------: | ---------: | ---------- | ------: | --------------: | -------------------: | | llama 3B Q4_K - Medium | 1.87 GiB | 3.21 B | CPU | 12 | pp512 | 81.27 ± 0.06 | | llama 3B Q4_K - Medium | 1.87 GiB | 3.21 B | CPU | 12 | pp4096 | 58.97 ± 0.04 | | llama 3B Q4_K - Medium | 1.87 GiB | 3.21 B | CPU | 12 | tg128 | 13.10 ± 0.06 | | llama 3B Q4_K - Medium | 1.87 GiB | 3.21 B | CPU | 12 | pp4096+tg128 | 50.59 ± 0.00 | ## Memory `tinymembench` results: <details> <summary>Click to expand memory benchmark result</summary> ``` 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 : 13887.3 MB/s (1.1%) C copy backwards (32 byte blocks) : 13686.0 MB/s (0.5%) C copy backwards (64 byte blocks) : 13867.0 MB/s (0.6%) C copy : 15684.9 MB/s (1.5%) C copy prefetched (32 bytes step) : 14599.2 MB/s (0.3%) C copy prefetched (64 bytes step) : 14234.6 MB/s C 2-pass copy : 15382.0 MB/s C 2-pass copy prefetched (32 bytes step) : 13179.9 MB/s C 2-pass copy prefetched (64 bytes step) : 15746.1 MB/s C fill : 39950.4 MB/s C fill (shuffle within 16 byte blocks) : 39952.0 MB/s C fill (shuffle within 32 byte blocks) : 39949.8 MB/s C fill (shuffle within 64 byte blocks) : 39951.3 MB/s NEON 64x2 COPY : 17247.2 MB/s (1.5%) NEON 64x2x4 COPY : 17018.3 MB/s (0.3%) NEON 64x1x4_x2 COPY : 16745.4 MB/s (0.7%) NEON 64x2 COPY prefetch x2 : 17181.4 MB/s (0.5%) NEON 64x2x4 COPY prefetch x1 : 17393.5 MB/s NEON 64x2 COPY prefetch x1 : 17411.1 MB/s NEON 64x2x4 COPY prefetch x1 : 17442.9 MB/s (0.1%) --- standard memcpy : 16635.8 MB/s (2.2%) standard memset : 47816.0 MB/s (9.9%) --- NEON LDP/STP copy : 17382.3 MB/s (1.2%) NEON LDP/STP copy pldl2strm (32 bytes step) : 15138.1 MB/s (0.4%) NEON LDP/STP copy pldl2strm (64 bytes step) : 15755.3 MB/s NEON LDP/STP copy pldl1keep (32 bytes step) : 16795.6 MB/s (0.3%) NEON LDP/STP copy pldl1keep (64 bytes step) : 17102.4 MB/s (0.6%) NEON LD1/ST1 copy : 16931.5 MB/s (0.4%) NEON STP fill : 47705.7 MB/s (5.9%) NEON STNP fill : 45452.0 MB/s ARM LDP/STP copy : 16811.5 MB/s (0.9%) ARM STP fill : 47097.0 MB/s (13.1%) ARM STNP fill : 46251.6 MB/s (0.6%) ========================================================================== == 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, [MADV_NOHUGEPAGE] 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 : 1.0 ns / 1.6 ns 262144 : 1.5 ns / 2.2 ns 524288 : 1.9 ns / 2.5 ns 1048576 : 18.9 ns / 26.9 ns 2097152 : 27.2 ns / 33.9 ns 4194304 : 31.9 ns / 36.2 ns 8388608 : 37.0 ns / 37.2 ns 16777216 : 43.6 ns / 48.5 ns 33554432 : 125.0 ns / 162.3 ns 67108864 : 186.9 ns / 218.0 ns block size : single random read / dual random read, [MADV_HUGEPAGE] 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 : 1.0 ns / 1.5 ns 262144 : 1.5 ns / 2.2 ns 524288 : 1.9 ns / 2.5 ns 1048576 : 17.8 ns / 26.3 ns 2097152 : 25.6 ns / 32.6 ns 4194304 : 30.9 ns / 34.6 ns 8388608 : 31.4 ns / 35.2 ns 16777216 : 41.5 ns / 46.2 ns 33554432 : 119.2 ns / 156.3 ns 67108864 : 167.7 ns / 200.0 ns ``` </details> ### Core to Core Memory Latency <img width="1920" height="1080" alt="Image" src="https://github.com/user-attachments/assets/f9cd9f5e-2a5f-44d4-84ab-b2a6e1016105" /> ## `sbc-bench` results https://github.com/ThomasKaiser/sbc-bench/issues/131 ## Phoronix Test Suite Results from [pi-general-benchmark.sh](https://gist.github.com/geerlingguy/570e13f4f81a40a5395688667b1f79af): - pts/encode-mp3: 8.320 sec - pts/x264 1080p: 52.67 fps - pts/x264 4K: 12.53 fps - pts/phpbench: 626464 - pts/build-linux-kernel (defconfig): 812.792 sec