Llama.cpp

Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage.

To run this test with the Phoronix Test Suite, the basic command is: phoronix-test-suite benchmark llama-cpp.

Project Site

github.com

Source Repository

github.com

Test Created

10 January 2024

Last Updated

23 November 2024

Test Maintainer

Michael Larabel

Test Type

System

Average Install Time

51 Seconds

Average Run Time

17 Minutes, 28 Seconds

Test Dependencies

C/C++ Compiler Toolchain + BLAS (Basic Linear Algebra Sub-Routine)

Accolades

10k+ Downloads

Supported Platforms

* Uploading of benchmark result data to OpenBenchmarking.org is always optional (opt-in) via the Phoronix Test Suite for users wishing to share their results publicly.
** Data based on those opting to upload their test results to OpenBenchmarking.org and users enabling the opt-in anonymous statistics reporting while running benchmarks from an Internet-connected platform.
Data updated weekly as of 17 December 2024.

Revision History

pts/llama-cpp-2.0.0 [View Source] Sat, 23 Nov 2024 17:33:55 GMT
Update against Llama.cpp upstream, switch to llama-bench executor.

pts/llama-cpp-1.1.0 [View Source] Sun, 02 Jun 2024 10:36:25 GMT
Update against Llama.cpp upstream, switch to Llama 3 model.

pts/llama-cpp-1.0.0 [View Source] Wed, 10 Jan 2024 18:02:20 GMT
Initial commit of llama.cpp CPU benchmark.

Suites Using This Test

Machine Learning

HPC - High Performance Computing

Large Language Models

Performance Metrics

Analyze Test Configuration:

Llama.cpp b4154

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128

OpenBenchmarking.org metrics for this test profile configuration based on 67 public results since 23 November 2024 with the latest data as of 15 December 2024.

Below is an overview of the generalized performance for components where there is sufficient statistically significant data based upon user-uploaded results. It is important to keep in mind particularly in the Linux/open-source space there can be vastly different OS configurations, with this overview intended to offer just general guidance as to the performance expectations.

Component

Percentile Rank

# Compatible Public Results

Tokens Per Second (Average)

AMD EPYC 9575F 64-Core

99th

51.8 ^{+/- 0.1}

AMD EPYC 9655P 96-Core

93rd

46.0 ^{+/- 0.3}

ARMv8 Neoverse-V2 72-Core

81st

20.3 ^{+/- 0.9}

Mid-Tier

75th

< 19.0

ARMv8 Neoverse-N1 128-Core

73rd

19.0

AMD Ryzen Threadripper 7980X 64-Cores

67th

15.5

AMD Ryzen AI 9 365

60th

10.3

AMD Ryzen AI 9 HX 370

54th

10.0 ^{+/- 0.1}

AMD Ryzen 9 9950X 16-Core

51st

9.3

Median

50th

9.3

Intel Core Ultra 7 256V

42nd

8.7

AMD Ryzen 7 9800X3D 8-Core

37th

7.3

AMD Ryzen 7 7840HS

31st

7.3

Low-Tier

25th

< 7.3

Intel Core Ultra 7 155H

25th

7.2 ^{+/- 0.1}

AMD EPYC 4484PX 12-Core

25th

7.1

AMD EPYC 4564P 16-Core

12th

6.9

Detailed Performance Overview

Based on OpenBenchmarking.org data, the selected test / test configuration (Llama.cpp b4154 - Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128) has an average run-time of 5 minutes. By default this test profile is set to run at least 3 times but may increase if the standard deviation exceeds pre-defined defaults or other calculations deem additional runs necessary for greater statistical accuracy of the result.

Notable Instruction Set Usage

Notable instruction set extensions supported by this test, based on an automatic analysis by the Phoronix Test Suite / OpenBenchmarking.org analytics engine.

Instruction Set

Support

Instructions Detected

Advanced Vector Extensions (AVX)

Used by default on supported hardware.
Found on Intel processors since Sandy Bridge (2011).
Found on AMD processors since Bulldozer (2011).

VZEROUPPER VEXTRACTF128 VPERMILPS VBROADCASTSS VPERM2F128 VINSERTF128 VBROADCASTSD VMASKMOVPS

Advanced Vector Extensions 2 (AVX2)

Used by default on supported hardware.
Found on Intel processors since Haswell (2013).
Found on AMD processors since Excavator (2016).

VPBROADCASTQ VEXTRACTI128 VPBROADCASTD VPERMQ VGATHERQPS VPBROADCASTW VINSERTI128 VPBROADCASTB VPSLLVD VPERMD VPSRLVD VPERM2I128

Advanced Vector Extensions 512 (AVX512)

Used by default on supported hardware.

(ZMM REGISTER USE)

AVX Vector Neural Network Instructions (AVX-VNNI)

Used by default on supported hardware.

VPDPBUSD

FMA (FMA)

Used by default on supported hardware.
Found on Intel processors since Haswell (2013).
Found on AMD processors since Bulldozer (2011).

VFMADD231PS VFMADD132PS VFNMADD231PS VFMADD213PS VFNMADD231SS VFMADD132SS VFNMADD213SS VFMADD132SD VFMADD231SS VFMADD213SS

The test / benchmark does honor compiler flag changes.

Last automated analysis: 6 June 2024

This test profile binary relies on the shared libraries libopenblas.so.0, libm.so.6, libc.so.6, libgfortran.so.5, libquadmath.so.0.

Tested CPU Architectures

This benchmark has been successfully tested on the below mentioned architectures. The CPU architectures listed is where successful OpenBenchmarking.org result uploads occurred, namely for helping to determine if a given test is compatible with various alternative CPU architectures.

CPU Architecture

Kernel Identifier

Verified On

Intel / AMD x86 64-bit

x86_64

(Many Processors)

ARMv8 64-bit

aarch64

ARMv8 Neoverse-N1 128-Core, ARMv8 Neoverse-V2 72-Core

nnn 4 Systems - 22 Benchmark Results	ARMv8 Neoverse-V2 - Pegatron JIMBO P4352 - 1 x 480GB LPDDR5-6400MT Ubuntu 24.04 - 6.8.0-50-generic-64k - NVIDIA
9950X EOY 2024 3 Systems - 127 Benchmark Results	AMD Ryzen 9 9950X 16-Core - ASRock X870E Taichi - AMD Device 14d8 Ubuntu 24.04 - 6.8.0-50-generic - GNOME Shell 46.0
dddda 3 Systems - 79 Benchmark Results	AMD Ryzen AI 9 365 - ASUS Zenbook S 16 UM5606WA_UM5606WA UM5606WA v1.0 - AMD Device 1507 Ubuntu 24.10 - 6.12.0-rc7-phx-eraps - GNOME Shell 47.0
9575F smoke 2 Systems - 116 Benchmark Results	AMD EPYC 9575F 64-Core - Supermicro Super Server H13SSL-N v1.01 - AMD 1Ah Ubuntu 24.10 - 6.12.0-rc7-linux-pm-next-phx - GNOME Shell 47.0
granite-test-output.txt 1 System - 2 Benchmark Results	AMD EPYC 9K84 96-Core - Tencent Cloud CVM v3.0 - Intel 440FX 82441FX PMC TencentOS Server 3.1 - 5.4.119-19.0009.44 - GCC 8.5.0 20210514 + Clang 17.0.6
Llama3bench 1 System - 4 Benchmark Results	AMD Ryzen Threadripper 2990WX 32-Core - ASRock X399 Phantom Gaming 6 - AMD 17h AlmaLinux 9.5 - 5.14.0-503.14.1.el9_5.x86_64 - GNOME Shell 40.10
nov 4 Systems - 110 Benchmark Results	Intel Core i7-1280P - MSI Prestige 14Evo A12M MS-14C6 - Intel Alder Lake PCH Ubuntu 24.10 - 6.11.0-rc6-phx - GNOME Shell 47.0
x3d sun 2 Systems - 90 Benchmark Results	AMD Ryzen 7 9800X3D 8-Core - ASRock X870E Taichi - AMD Device 14d8 Ubuntu 24.04 - 6.8.0-49-generic - GNOME Shell 46.0
ai ryzen 5 Systems - 64 Benchmark Results	AMD Ryzen AI 9 HX 370 - ASUS Zenbook S 16 UM5606WA_UM5606WA UM5606WA v1.0 - AMD Device 1507 Ubuntu 24.10 - 6.11.0-rc6-phx - GNOME Shell 47.0
ddddai 4 Systems - 64 Benchmark Results	AMD Ryzen 7 7840HS - Framework Laptop 16 - AMD Device 14e8 Ubuntu 24.04 - 6.8.0-48-generic - GNOME Shell 46.0
ampere arm tests 2 Systems - 77 Benchmark Results	ARMv8 Neoverse-N1 - System76 Thelio Astra - Ampere Computing LLC Altra PCI Root Complex A Ubuntu 24.04 - 6.8.0-48-generic-64k - GNOME Shell 46.0
sunny 2 Systems - 82 Benchmark Results	Intel Core Ultra 7 256V - ASUS Zenbook S 14 UX5406SA_UX5406SA UX5406SA v1.0 - Intel Device a87f Ubuntu 24.10 - 6.12.0-rc3-phx-aipt - GNOME Shell 47.0
ai-vino 3 Systems - 16 Benchmark Results	Intel Core Ultra 7 155H - MTL Swift SFG14-72T Coral_MTH - Intel Device 7e7f Ubuntu 24.10 - 6.11.0-rc6-phx - GNOME Shell 47.0
new sun 3 Systems - 32 Benchmark Results	AMD EPYC 9655P 96-Core - Supermicro Super Server H13SSL-N v1.01 - AMD 1Ah Ubuntu 24.10 - 6.12.0-rc7-linux-pm-next-phx - GNOME Shell 47.0
llama fun 1 System - 48 Benchmark Results	AMD Ryzen Threadripper 7980X 64-Cores - System76 Thelio Major - AMD Device 14a4 Ubuntu 24.04 - 6.8.0-48-generic - GNOME Shell 46.0

9575F smoke 2 Systems - 116 Benchmark Results	AMD EPYC 9575F 64-Core - Supermicro Super Server H13SSL-N v1.01 - AMD 1Ah Ubuntu 24.10 - 6.12.0-rc7-linux-pm-next-phx - GNOME Shell 47.0
new sun 3 Systems - 32 Benchmark Results	AMD EPYC 9655P 96-Core - Supermicro Super Server H13SSL-N v1.01 - AMD 1Ah Ubuntu 24.10 - 6.12.0-rc7-linux-pm-next-phx - GNOME Shell 47.0
ampere arm tests 2 Systems - 77 Benchmark Results	ARMv8 Neoverse-N1 - System76 Thelio Astra - Ampere Computing LLC Altra PCI Root Complex A Ubuntu 24.04 - 6.8.0-48-generic-64k - GNOME Shell 46.0
ai ryzen 5 Systems - 64 Benchmark Results	AMD Ryzen AI 9 HX 370 - ASUS Zenbook S 16 UM5606WA_UM5606WA UM5606WA v1.0 - AMD Device 1507 Ubuntu 24.10 - 6.11.0-rc6-phx - GNOME Shell 47.0
sunny 2 Systems - 82 Benchmark Results	Intel Core Ultra 7 256V - ASUS Zenbook S 14 UX5406SA_UX5406SA UX5406SA v1.0 - Intel Device a87f Ubuntu 24.10 - 6.12.0-rc3-phx-aipt - GNOME Shell 47.0
sunny 2 Systems - 82 Benchmark Results	Intel Core Ultra 7 256V - ASUS Zenbook S 14 UX5406SA_UX5406SA UX5406SA v1.0 - Intel Device a87f Ubuntu 24.10 - 6.12.0-rc3-phx-aipt - GNOME Shell 47.0
x3d sun 2 Systems - 90 Benchmark Results	AMD Ryzen 7 9800X3D 8-Core - ASRock X870E Taichi - AMD Device 14d8 Ubuntu 24.04 - 6.8.0-49-generic - GNOME Shell 46.0
ddddai 4 Systems - 64 Benchmark Results	AMD Ryzen 7 7840HS - Framework Laptop 16 - AMD Device 14e8 Ubuntu 24.04 - 6.8.0-48-generic - GNOME Shell 46.0
llama cpp update 3 Systems - 12 Benchmark Results	AMD Ryzen Threadripper 7980X 64-Cores - System76 Thelio Major - AMD Device 14a4 Ubuntu 24.04 - 6.8.0-48-generic - GNOME Shell 46.0
ai-vino 3 Systems - 16 Benchmark Results	Intel Core Ultra 7 155H - MTL Swift SFG14-72T Coral_MTH - Intel Device 7e7f Ubuntu 24.10 - 6.11.0-rc6-phx - GNOME Shell 47.0
ampere arm tests 2 Systems - 77 Benchmark Results	ARMv8 Neoverse-N1 - System76 Thelio Astra - Ampere Computing LLC Altra PCI Root Complex A Ubuntu 24.04 - 6.8.0-48-generic-64k - GNOME Shell 46.0
x3d sun 2 Systems - 90 Benchmark Results	AMD Ryzen 7 9800X3D 8-Core - ASRock X870E Taichi - AMD Device 14d8 Ubuntu 24.04 - 6.8.0-49-generic - GNOME Shell 46.0
llama cpp grace 4 Systems - 12 Benchmark Results	ARMv8 Neoverse-V2 - Pegatron JIMBO P4352 - 1 x 480GB LPDDR5-6400MT Ubuntu 24.04 - 6.8.0-49-generic-64k - GCC 13.2.0 + Clang 18.1.3 + CUDA 11.8
nov 4 Systems - 110 Benchmark Results	Intel Core i7-1280P - MSI Prestige 14Evo A12M MS-14C6 - Intel Alder Lake PCH Ubuntu 24.10 - 6.11.0-rc6-phx - GNOME Shell 47.0
9575F smoke 2 Systems - 116 Benchmark Results	AMD EPYC 9575F 64-Core - Supermicro Super Server H13SSL-N v1.01 - AMD 1Ah Ubuntu 24.10 - 6.12.0-rc7-linux-pm-next-phx - GNOME Shell 47.0

Llama.cpp

Project Site

Source Repository

Test Created

Last Updated

Test Maintainer

Test Type

Average Install Time

Average Run Time

Test Dependencies

Accolades

Supported Platforms

Revision History

Suites Using This Test

Performance Metrics

Llama.cpp b4154

Backend: CPU BLAS - Model: Llama-3.1-Tulu-3-8B-Q8_0 - Test: Text Generation 128

Notable Instruction Set Usage

Tested CPU Architectures

Recent Test Results

Most Popular Test Results