lll

AMD EPYC 8324P 32-Core testing with a AMD Cinnabar (RCB1009C BIOS) and ASPEED on Ubuntu 23.10 via the Phoronix Test Suite.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2401278-NE-LLL53839360
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results

Limit displaying results to tests within:

HPC - High Performance Computing 2 Tests
Large Language Models 2 Tests
Machine Learning 2 Tests

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Geometric Means Per-Suite/Category
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Prefer Vertical Bar Graphs

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Hide
Result
Result
Identifier
View Logs
Performance Per
Dollar
Date
Run
  Test
  Duration
a
January 27
  15 Minutes
b
January 27
  43 Minutes
c
January 27
  41 Minutes
Invert Hiding All Results Option
  33 Minutes

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):


lllOpenBenchmarking.orgPhoronix Test SuiteAMD EPYC 8324P 32-Core @ 2.65GHz (32 Cores / 64 Threads)AMD Cinnabar (RCB1009C BIOS)AMD Device 14a46 x 32 GB DRAM-4800MT/s Samsung M321R4GA0BB0-CQKMG3201GB Micron_7450_MTFDKCB3T2TFSASPEED2 x Broadcom NetXtreme BCM5720 PCIeUbuntu 23.106.5.0-5-generic (x86_64)GNOME ShellX Server 1.21.1.7GCC 13.2.0ext4640x480ProcessorMotherboardChipsetMemoryDiskGraphicsNetworkOSKernelDesktopDisplay ServerCompilerFile-SystemScreen ResolutionLll BenchmarksSystem Logs- Transparent Huge Pages: madvise- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-serialization=2 --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-FTCNCZ/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-FTCNCZ/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v - Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xaa00212 - gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Mitigation of safe RET + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

abcResult OverviewPhoronix Test Suite100%100%101%101%101%Llama.cppLlamafileLZ4 CompressionCacheBench

lllcachebench: Readcachebench: Writecachebench: Read / Modify / Writecompress-lz4: 1 - Compression Speedcompress-lz4: 1 - Decompression Speedcompress-lz4: 3 - Compression Speedcompress-lz4: 3 - Decompression Speedcompress-lz4: 9 - Compression Speedcompress-lz4: 9 - Decompression Speedllama-cpp: llama-2-7b.Q4_0.ggufllama-cpp: llama-2-13b.Q4_0.ggufllama-cpp: llama-2-70b-chat.Q5_0.ggufllamafile: llava-v1.5-7b-q4 - CPUllamafile: mistral-7b-instruct-v0.2.Q8_0 - CPUllamafile: wizardcoder-python-34b-v1.0.Q6_K - CPUabc7612.83500445638.46841487264.314442565.033650.188.23306.928.353471.629.5817.973.4223.5414.755.57615.12937245639.68120287154.946426564.213649.888.233307.628.343467.430.0317.953.4323.5714.765.537615.21480845639.00933387247.993005565.423647.988.093308.628.293468.129.8518.313.4323.7614.755.51OpenBenchmarking.org

CacheBench

This is a performance test of CacheBench, which is part of LLCbench. CacheBench is designed to test the memory and cache bandwidth performance Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Readabc16003200480064008000SE +/- 0.12, N = 3SE +/- 0.07, N = 37612.847615.137615.21MIN: 7612.38 / MAX: 7613.67MIN: 7614.3 / MAX: 7615.59MIN: 7612.23 / MAX: 7615.741. (CC) gcc options: -O3 -lrt
OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Readabc13002600390052006500Min: 7614.9 / Avg: 7615.13 / Max: 7615.29Min: 7615.08 / Avg: 7615.21 / Max: 7615.321. (CC) gcc options: -O3 -lrt

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Writeabc10K20K30K40K50KSE +/- 0.89, N = 3SE +/- 0.76, N = 345638.4745639.6845639.01MIN: 45475.7 / MAX: 45690.04MIN: 45474.92 / MAX: 45693.28MIN: 45474.02 / MAX: 45692.391. (CC) gcc options: -O3 -lrt
OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Writeabc8K16K24K32K40KMin: 45638.71 / Avg: 45639.68 / Max: 45641.46Min: 45637.78 / Avg: 45639.01 / Max: 45640.421. (CC) gcc options: -O3 -lrt

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Read / Modify / Writeabc20K40K60K80K100KSE +/- 87.15, N = 3SE +/- 10.58, N = 387264.3187154.9587247.99MIN: 65718.12 / MAX: 90689.42MIN: 65719.2 / MAX: 90700.03MIN: 65710.28 / MAX: 90697.681. (CC) gcc options: -O3 -lrt
OpenBenchmarking.orgMB/s, More Is BetterCacheBenchTest: Read / Modify / Writeabc15K30K45K60K75KMin: 86987.15 / Avg: 87154.95 / Max: 87279.68Min: 87227.07 / Avg: 87247.99 / Max: 87261.141. (CC) gcc options: -O3 -lrt

LZ4 Compression

This test measures the time needed to compress/decompress a sample file (silesia archive) using LZ4 compression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Compression Speedabc120240360480600SE +/- 0.34, N = 3SE +/- 0.78, N = 3565.03564.21565.421. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Compression Speedabc100200300400500Min: 563.79 / Avg: 564.21 / Max: 564.89Min: 564.49 / Avg: 565.42 / Max: 566.971. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Decompression Speedabc8001600240032004000SE +/- 0.58, N = 3SE +/- 0.41, N = 33650.13649.83647.91. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 1 - Decompression Speedabc6001200180024003000Min: 3649 / Avg: 3649.77 / Max: 3650.9Min: 3647.1 / Avg: 3647.87 / Max: 3648.51. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Compression Speedabc20406080100SE +/- 0.01, N = 3SE +/- 0.10, N = 388.2088.2388.091. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Compression Speedabc20406080100Min: 88.22 / Avg: 88.23 / Max: 88.24Min: 87.93 / Avg: 88.09 / Max: 88.261. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Decompression Speedabc7001400210028003500SE +/- 0.71, N = 3SE +/- 0.62, N = 33306.93307.63308.61. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 3 - Decompression Speedabc6001200180024003000Min: 3306.5 / Avg: 3307.57 / Max: 3308.9Min: 3307.7 / Avg: 3308.6 / Max: 3309.81. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Compression Speedabc714212835SE +/- 0.01, N = 3SE +/- 0.03, N = 328.3528.3428.291. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Compression Speedabc612182430Min: 28.33 / Avg: 28.34 / Max: 28.35Min: 28.23 / Avg: 28.29 / Max: 28.331. (CC) gcc options: -O3

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Decompression Speedabc7001400210028003500SE +/- 0.52, N = 3SE +/- 0.17, N = 33471.63467.43468.11. (CC) gcc options: -O3
OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.4Compression Level: 9 - Decompression Speedabc6001200180024003000Min: 3466.4 / Avg: 3467.43 / Max: 3468Min: 3467.9 / Avg: 3468.07 / Max: 3468.41. (CC) gcc options: -O3

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. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-7b.Q4_0.ggufabc714212835SE +/- 0.22, N = 12SE +/- 0.12, N = 329.5830.0329.851. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-7b.Q4_0.ggufabc714212835Min: 29.49 / Avg: 30.03 / Max: 31.84Min: 29.68 / Avg: 29.85 / Max: 30.081. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-13b.Q4_0.ggufabc510152025SE +/- 0.04, N = 3SE +/- 0.17, N = 317.9717.9518.311. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-13b.Q4_0.ggufabc510152025Min: 17.87 / Avg: 17.95 / Max: 18.01Min: 17.97 / Avg: 18.31 / Max: 18.491. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas

OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-70b-chat.Q5_0.ggufabc0.77181.54362.31543.08723.859SE +/- 0.00, N = 3SE +/- 0.00, N = 33.423.433.431. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
OpenBenchmarking.orgTokens Per Second, More Is BetterLlama.cpp b1808Model: llama-2-70b-chat.Q5_0.ggufabc246810Min: 3.43 / Avg: 3.43 / Max: 3.44Min: 3.42 / Avg: 3.43 / Max: 3.431. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas

Llamafile

Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: llava-v1.5-7b-q4 - Acceleration: CPUabc612182430SE +/- 0.02, N = 3SE +/- 0.06, N = 323.5423.5723.76
OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: llava-v1.5-7b-q4 - Acceleration: CPUabc612182430Min: 23.55 / Avg: 23.57 / Max: 23.61Min: 23.66 / Avg: 23.76 / Max: 23.88

OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: mistral-7b-instruct-v0.2.Q8_0 - Acceleration: CPUabc48121620SE +/- 0.08, N = 3SE +/- 0.09, N = 314.7514.7614.75
OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: mistral-7b-instruct-v0.2.Q8_0 - Acceleration: CPUabc48121620Min: 14.59 / Avg: 14.76 / Max: 14.86Min: 14.58 / Avg: 14.75 / Max: 14.86

OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: wizardcoder-python-34b-v1.0.Q6_K - Acceleration: CPUabc1.24432.48863.73294.97726.2215SE +/- 0.01, N = 3SE +/- 0.01, N = 35.505.535.51
OpenBenchmarking.orgTokens Per Second, More Is BetterLlamafile 0.6Test: wizardcoder-python-34b-v1.0.Q6_K - Acceleration: CPUabc246810Min: 5.51 / Avg: 5.53 / Max: 5.55Min: 5.5 / Avg: 5.51 / Max: 5.52