Apple M2 MacBook vs. AMD/Intel Linux Benchmarks

Initial Apple M2 MacBook Air benchmarks on Asahi Linux for a future article by Michael Larabel.

HTML result view exported from: https://openbenchmarking.org/result/2208099-PTS-APPLEM2896&grs.

Cryptsetup

Serpent-XTS 512b Encryption

Cryptsetup

Serpent-XTS 256b Encryption

Cryptsetup

Serpent-XTS 256b Decryption

x265

Video Input: Bosphorus 1080p

Renaissance

Test: Genetic Algorithm Using Jenetics + Futures

x265

Video Input: Bosphorus 4K

Zstd Compression

Compression Level: 3, Long Mode - Compression Speed

LAME MP3 Encoding

WAV To MP3

Liquid-DSP

Threads: 16 - Buffer Length: 256 - Filter Length: 57

Liquid-DSP

Threads: 8 - Buffer Length: 256 - Filter Length: 57

Renaissance

Test: Apache Spark Bayes

Liquid-DSP

Threads: 1 - Buffer Length: 256 - Filter Length: 57

Renaissance

Test: Finagle HTTP Requests

Etcpak

Benchmark: Multi-Threaded - Configuration: ETC2

Liquid-DSP

Threads: 2 - Buffer Length: 256 - Filter Length: 57

ASTC Encoder

Preset: Exhaustive

Liquid-DSP

Threads: 4 - Buffer Length: 256 - Filter Length: 57

TensorFlow Lite

Model: Mobilenet Quant

Zstd Compression

Compression Level: 8 - Compression Speed

NCNN

Target: CPU - Model: vgg16

ASKAP

Test: tConvolve MT - Gridding

LAMMPS Molecular Dynamics Simulator

Model: 20k Atoms

Etcpak

Benchmark: Single-Threaded - Configuration: ETC2

Zstd Compression

Compression Level: 8, Long Mode - Compression Speed

Apache Spark

Row Count: 1000000 - Partitions: 2000 - Calculate Pi Benchmark

ASKAP

Test: tConvolve OpenMP - Gridding

PyPerformance

Benchmark: python_startup

FLAC Audio Encoding

WAV To FLAC

Cryptsetup

Twofish-XTS 512b Decryption

TensorFlow Lite

Model: Inception ResNet V2

Cryptsetup

Twofish-XTS 256b Decryption

Apache Spark

Row Count: 1000000 - Partitions: 100 - Calculate Pi Benchmark

Cryptsetup

Twofish-XTS 512b Encryption

Primesieve

Length: 1e13

Cryptsetup

Twofish-XTS 256b Encryption

TensorFlow Lite

Model: Mobilenet Float

Apache Spark

Row Count: 1000000 - Partitions: 2000 - Calculate Pi Benchmark Using Dataframe

Apache Spark

Row Count: 1000000 - Partitions: 100 - Calculate Pi Benchmark Using Dataframe

VP9 libvpx Encoding

Speed: Speed 5 - Input: Bosphorus 4K

Primesieve

Length: 1e12

Coremark

CoreMark Size 666 - Iterations Per Second

TNN

Target: CPU - Model: DenseNet

WebP2 Image Encode

Encode Settings: Quality 95, Compression Effort 7

Parallel BZIP2 Compression

FreeBSD-13.0-RELEASE-amd64-memstick.img Compression

POV-Ray

Trace Time

Darktable

Test: Server Room - Acceleration: CPU-only

PyHPC Benchmarks

Device: CPU - Backend: Numpy - Project Size: 1048576 - Benchmark: Equation of State

NCNN

Target: CPU - Model: resnet50

LAMMPS Molecular Dynamics Simulator

Model: Rhodopsin Protein

WebP2 Image Encode

Encode Settings: Quality 75, Compression Effort 7

TNN

Target: CPU - Model: SqueezeNet v1.1

libavif avifenc

Encoder Speed: 2

PyHPC Benchmarks

Device: CPU - Backend: Numpy - Project Size: 4194304 - Benchmark: Isoneutral Mixing

WebP2 Image Encode

Encode Settings: Quality 100, Lossless Compression

PyHPC Benchmarks

Device: CPU - Backend: Numpy - Project Size: 4194304 - Benchmark: Equation of State

NCNN

Target: CPU-v2-v2 - Model: mobilenet-v2

NCNN

Target: CPU - Model: yolov4-tiny

PyHPC Benchmarks

Device: CPU - Backend: Numpy - Project Size: 1048576 - Benchmark: Isoneutral Mixing

GNU Radio

Test: FIR Filter

Darktable

Test: Server Rack - Acceleration: CPU-only

LuaRadio

Test: FM Deemphasis Filter

Zstd Compression

Compression Level: 3 - Compression Speed

GNU Radio

Test: Hilbert Transform

OpenSSL

Algorithm: RSA4096

OpenSSL

Algorithm: RSA4096

GNU Octave Benchmark

TNN

Target: CPU - Model: SqueezeNet v2

Renaissance

Test: In-Memory Database Shootout

OpenSSL

Algorithm: SHA256

Renaissance

Test: Akka Unbalanced Cobwebbed Tree

Google Draco

Model: Lion

GIMP

Test: resize

DaCapo Benchmark

Java Test: Tradesoap

Renaissance

Test: Random Forest

Basis Universal

Settings: UASTC Level 3

libavif avifenc

Encoder Speed: 6

ASKAP

Test: tConvolve MT - Degridding

Apache Spark

Row Count: 1000000 - Partitions: 100 - SHA-512 Benchmark Time

SecureMark

Benchmark: SecureMark-TLS

Apache Spark

Row Count: 1000000 - Partitions: 100 - Group By Test Time

simdjson

Throughput Test: LargeRandom

NCNN

Target: CPU - Model: squeezenet_ssd

Build2

Time To Compile

Google Draco

Model: Church Facade

Basis Universal

Settings: UASTC Level 2

NCNN

Target: CPU - Model: mobilenet

VP9 libvpx Encoding

Speed: Speed 5 - Input: Bosphorus 1080p

ASTC Encoder

Preset: Medium

Ngspice

Circuit: C7552

libavif avifenc

Encoder Speed: 6, Lossless

Zstd Compression

Compression Level: 19 - Decompression Speed

RNNoise

Xmrig

Variant: Monero - Hash Count: 1M

Apache Spark

Row Count: 1000000 - Partitions: 2000 - SHA-512 Benchmark Time

simdjson

Throughput Test: DistinctUserID

Google SynthMark

Test: VoiceMark_100

simdjson

Throughput Test: Kostya

Cryptsetup

PBKDF2-whirlpool

simdjson

Throughput Test: TopTweet

Cryptsetup

PBKDF2-sha512

Renaissance

Test: ALS Movie Lens

Renaissance

Test: Savina Reactors.IO

Zstd Compression

Compression Level: 19, Long Mode - Compression Speed

Renaissance

Test: Apache Spark PageRank

Zstd Compression

Compression Level: 19 - Compression Speed

Cryptsetup

AES-XTS 256b Encryption

Cryptsetup

AES-XTS 256b Decryption