AMD EPYC 9684X 3D V-Cache Benchmark

AMD EPYC 9684X 96-Core testing by Michael Larabel for a future article. Various benchmarks conducted with the EPYC 9684X 1P and then repeated after disabling 3D V-Cache from the BIOS to see direct comparison of 3DV impact. Plus monitoring CPU thermal / power / frequency for future follow-up article.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2307201-PTS-GENOAX3D86
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:

BLAS (Basic Linear Algebra Sub-Routine) Tests 3 Tests
C++ Boost Tests 3 Tests
Chess Test Suite 2 Tests
Timed Code Compilation 6 Tests
C/C++ Compiler Tests 4 Tests
CPU Massive 13 Tests
Creator Workloads 9 Tests
Fortran Tests 7 Tests
Game Development 4 Tests
HPC - High Performance Computing 20 Tests
Linear Algebra 2 Tests
Machine Learning 5 Tests
Molecular Dynamics 7 Tests
MPI Benchmarks 7 Tests
Multi-Core 18 Tests
NVIDIA GPU Compute 3 Tests
Intel oneAPI 4 Tests
OpenMPI Tests 15 Tests
Programmer / Developer System Benchmarks 8 Tests
Python 2 Tests
Raytracing 2 Tests
Renderers 3 Tests
Scientific Computing 10 Tests
Software Defined Radio 2 Tests
Server CPU Tests 9 Tests
Texture Compression 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
No Box Plots
On Line Graphs With Missing Data, Connect The Line Gaps

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs
Condense Test Profiles With Multiple Version Results 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
Default
July 17 2023
  1 Day, 46 Minutes
3DV Disabled
July 19 2023
  1 Day, 4 Hours, 53 Minutes
Invert Hiding All Results Option
  1 Day, 2 Hours, 50 Minutes
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):

AMD EPYC 9684X 3D V-Cache Benchmark Suite 1.0.0 System Test suite extracted from AMD EPYC 9684X 3D V-Cache Benchmark. pts/openfoam-1.2.0 incompressible/simpleFoam/drivaerFastback/ -m M Input: drivaerFastback, Medium Mesh Size - Execution Time pts/libxsmm-1.0.1 64 64 64 M N K: 64 pts/npb-1.4.5 sp.C Test / Class: SP.C pts/libxsmm-1.0.1 32 32 32 M N K: 32 pts/stress-ng-1.10.0 --tree -1 --tree-method avl --no-rand-seed Test: AVL Tree pts/openvino-1.2.0 -m models/intel/vehicle-detection-0202/FP16/vehicle-detection-0202.xml -d CPU Model: Vehicle Detection FP16 - Device: CPU pts/askap-2.1.0 tConvolveMPI Test: tConvolve MPI - Degridding pts/heffte-1.0.0 r2c fftw float 512 512 512 Test: r2c - Backend: FFTW - Precision: float - X Y Z: 512 pts/openfoam-1.2.0 incompressible/simpleFoam/drivaerFastback/ -m L Input: drivaerFastback, Large Mesh Size - Execution Time pts/npb-1.4.5 lu.C Test / Class: LU.C pts/npb-1.4.5 bt.C Test / Class: BT.C pts/lczero-1.6.0 -b eigen Backend: Eigen pts/xmrig-1.1.0 --bench=1M Variant: Monero - Hash Count: 1M pts/heffte-1.0.0 c2c fftw float 512 512 512 Test: c2c - Backend: FFTW - Precision: float - X Y Z: 512 pts/npb-1.4.5 is.D Test / Class: IS.D pts/libxsmm-1.0.1 128 128 128 M N K: 128 pts/stress-ng-1.10.0 --cache -1 --no-rand-seed Test: CPU Cache pts/npb-1.4.5 mg.C Test / Class: MG.C pts/npb-1.4.5 cg.C Test / Class: CG.C pts/heffte-1.0.0 c2c fftw double 256 256 256 Test: c2c - Backend: FFTW - Precision: double - X Y Z: 256 pts/heffte-1.0.0 r2c fftw double 512 512 512 Test: r2c - Backend: FFTW - Precision: double - X Y Z: 512 pts/pyhpc-3.0.0 --device cpu -b numpy -s 4194304 benchmarks/equation_of_state/ Device: CPU - Backend: Numpy - Project Size: 4194304 - Benchmark: Equation of State pts/lczero-1.6.0 -b blas Backend: BLAS pts/heffte-1.0.0 r2c fftw double 256 256 256 Test: r2c - Backend: FFTW - Precision: double - X Y Z: 256 pts/incompact3d-2.0.2 input_193_nodes.i3d Input: input.i3d 193 Cells Per Direction pts/draco-1.6.0 -i church.ply Model: Church Facade pts/stress-ng-1.10.0 --malloc -1 --no-rand-seed Test: Malloc pts/openfoam-1.2.0 incompressible/simpleFoam/drivaerFastback/ -m L Input: drivaerFastback, Large Mesh Size - Mesh Time pts/wrf-1.0.1 -i conus 2.5km Input: conus 2.5km pts/stress-ng-1.10.0 --matrix -1 --no-rand-seed Test: Matrix Math pts/gromacs-1.8.0 mpi-build water-cut1.0_GMX50_bare/1536 Implementation: MPI CPU - Input: water_GMX50_bare pts/embree-1.5.0 pathtracer_ispc -c crown/crown.ecs Binary: Pathtracer ISPC - Model: Crown pts/palabos-1.0.0 400 Grid Size: 400 pts/gpaw-1.2.0 carbon-nanotube Input: Carbon Nanotube pts/embree-1.5.0 pathtracer_ispc -c asian_dragon_obj/asian_dragon.ecs Binary: Pathtracer ISPC - Model: Asian Dragon Obj pts/libxsmm-1.0.1 256 256 256 M N K: 256 pts/embree-1.5.0 pathtracer_ispc -c asian_dragon/asian_dragon.ecs Binary: Pathtracer ISPC - Model: Asian Dragon pts/lulesh-1.1.1 pts/askap-2.1.0 tConvolveMT Test: tConvolve MT - Degridding pts/pyhpc-3.0.0 --device cpu -b numpy -s 4194304 benchmarks/isoneutral_mixing/ Device: CPU - Backend: Numpy - Project Size: 4194304 - Benchmark: Isoneutral Mixing pts/stress-ng-1.10.0 --pipe -1 --no-rand-seed Test: Pipe pts/stress-ng-1.10.0 --futex -1 --no-rand-seed Test: Futex pts/askap-2.1.0 tConvolveMT Test: tConvolve MT - Gridding pts/ngspice-1.0.0 ~/iscas85Circuits/85/c7552p/c7552_ann.net Circuit: C7552 pts/astcenc-1.4.0 -exhaustive -repeats 2 Preset: Exhaustive pts/stress-ng-1.10.0 --fma -1 --no-rand-seed Test: Fused Multiply-Add pts/palabos-1.0.0 500 Grid Size: 500 pts/minife-1.0.0 -‐nx 264 --ny 256 -‐nz 256 Problem Size: Small pts/build-linux-kernel-1.15.0 allmodconfig Build: allmodconfig pts/deepsparse-1.5.0 zoo:nlp/question_answering/bert-base/pytorch/huggingface/squad/12layer_pruned90-none --scenario async Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Asynchronous Multi-Stream pts/astcenc-1.4.0 -thorough -repeats 10 Preset: Thorough pts/mt-dgemm-1.2.0 Sustained Floating-Point Rate pts/build-godot-4.0.0 Time To Compile pts/build-llvm-1.5.0 Ninja Build System: Ninja pts/palabos-1.0.0 1000 Grid Size: 1000 pts/ospray-studio-1.1.0 --cameras 2 2 --resolution 3840 2160 --spp 1 --renderer pathtracer Camera: 2 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer pts/stress-ng-1.10.0 --vecfp -1 --no-rand-seed Test: Vector Floating Point pts/ospray-studio-1.1.0 --cameras 2 2 --resolution 3840 2160 --spp 16 --renderer pathtracer Camera: 2 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer pts/ospray-studio-1.1.0 --cameras 1 1 --resolution 3840 2160 --spp 16 --renderer pathtracer Camera: 1 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer pts/ospray-studio-1.1.0 --cameras 3 3 --resolution 3840 2160 --spp 1 --renderer pathtracer Camera: 3 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer pts/draco-1.6.0 -i lion.ply Model: Lion pts/ospray-studio-1.1.0 --cameras 3 3 --resolution 3840 2160 --spp 32 --renderer pathtracer Camera: 3 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer pts/ospray-studio-1.1.0 --cameras 1 1 --resolution 3840 2160 --spp 1 --renderer pathtracer Camera: 1 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer pts/ospray-studio-1.1.0 --cameras 2 2 --resolution 3840 2160 --spp 32 --renderer pathtracer Camera: 2 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer pts/blender-3.6.0 -b ../barbershop_interior_gpu.blend -o output.test -x 1 -F JPEG -f 1 -- --cycles-device CPU Blend File: Barbershop - Compute: CPU-Only pts/ospray-2.12.0 --benchmark_filter=gravity_spheres_volume/dim_512/ao/real_time Benchmark: gravity_spheres_volume/dim_512/ao/real_time pts/ospray-2.12.0 --benchmark_filter=gravity_spheres_volume/dim_512/scivis/real_time Benchmark: gravity_spheres_volume/dim_512/scivis/real_time pts/ospray-studio-1.1.0 --cameras 3 3 --resolution 3840 2160 --spp 16 --renderer pathtracer Camera: 3 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer pts/heffte-1.0.0 c2c fftw double 512 512 512 Test: c2c - Backend: FFTW - Precision: double - X Y Z: 512 pts/ospray-studio-1.1.0 --cameras 1 1 --resolution 3840 2160 --spp 32 --renderer pathtracer Camera: 1 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer pts/namd-1.2.1 ATPase Simulation - 327,506 Atoms pts/build-nodejs-1.3.0 Time To Compile pts/stress-ng-1.10.0 --memcpy -1 --no-rand-seed Test: Memory Copying pts/hpcg-1.3.0 --nx=144 --ny=144 --nz=144 --rt=60 X Y Z: 144 144 144 - RT: 60 pts/ospray-2.12.0 --benchmark_filter=particle_volume/scivis/real_time Benchmark: particle_volume/scivis/real_time pts/npb-1.4.5 ep.D Test / Class: EP.D pts/ospray-2.12.0 --benchmark_filter=particle_volume/ao/real_time Benchmark: particle_volume/ao/real_time pts/build-llvm-1.5.0 Build System: Unix Makefiles pts/openvino-1.2.0 -m models/intel/person-detection-0106/FP32/person-detection-0106.xml -d CPU Model: Person Detection FP32 - Device: CPU pts/deepsparse-1.5.0 zoo:cv/segmentation/yolact-darknet53/pytorch/dbolya/coco/pruned90-none --scenario async Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream pts/blender-3.6.0 -b ../classroom_gpu.blend -o output.test -x 1 -F JPEG -f 1 -- --cycles-device CPU Blend File: Classroom - Compute: CPU-Only pts/srsran-2.1.0 tests/benchmarks/phy/upper/channel_processors/pusch_processor_benchmark -m throughput_total -R 100 -P pusch_scs15_50MHz_256qam_max Test: PUSCH Processor Benchmark, Throughput Total pts/astcenc-1.4.0 -medium -repeats 20 Preset: Medium pts/numpy-1.2.1 pts/openvino-1.2.0 -m models/intel/person-vehicle-bike-detection-2004/FP16/person-vehicle-bike-detection-2004.xml -d CPU Model: Person Vehicle Bike Detection FP16 - Device: CPU pts/ospray-2.12.0 --benchmark_filter=gravity_spheres_volume/dim_512/pathtracer/real_time Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_time pts/amg-1.1.0 pts/stress-ng-1.10.0 --sem -1 --no-rand-seed Test: Semaphores pts/tensorflow-2.1.0 --device cpu --batch_size=512 --model=googlenet Device: CPU - Batch Size: 512 - Model: GoogLeNet pts/blender-3.6.0 -b ../pavillon_barcelone_gpu.blend -o output.test -x 1 -F JPEG -f 1 -- --cycles-device CPU Blend File: Pabellon Barcelona - Compute: CPU-Only pts/blender-3.6.0 -b ../fishy_cat_gpu.blend -o output.test -x 1 -F JPEG -f 1 -- --cycles-device CPU Blend File: Fishy Cat - Compute: CPU-Only pts/openfoam-1.2.0 incompressible/simpleFoam/drivaerFastback/ -m M Input: drivaerFastback, Medium Mesh Size - Mesh Time pts/hpcg-1.3.0 --nx=160 --ny=160 --nz=160 --rt=60 X Y Z: 160 160 160 - RT: 60 pts/stress-ng-1.10.0 --cpu -1 --cpu-method all --no-rand-seed Test: CPU Stress pts/build-php-1.6.0 Time To Compile pts/openvino-1.2.0 -m models/intel/person-detection-0106/FP16/person-detection-0106.xml -d CPU Model: Person Detection FP16 - Device: CPU pts/stress-ng-1.10.0 --mutex -1 --no-rand-seed Test: Mutex pts/liquid-dsp-1.6.0 -n 192 -b 256 -f 512 Threads: 192 - Buffer Length: 256 - Filter Length: 512 pts/askap-2.1.0 tHogbomCleanOMP Test: Hogbom Clean OpenMP pts/stress-ng-1.10.0 --vecmath -1 --no-rand-seed Test: Vector Math pts/stress-ng-1.10.0 --vecwide -1 --no-rand-seed Test: Wide Vector Math pts/cloverleaf-1.1.0 Lagrangian-Eulerian Hydrodynamics pts/blender-3.6.0 -b ../bmw27_gpu.blend -o output.test -x 1 -F JPEG -f 1 -- --cycles-device CPU Blend File: BMW27 - Compute: CPU-Only pts/deepsparse-1.5.0 zoo:cv/detection/yolov5-s/pytorch/ultralytics/coco/base-none --scenario async Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream pts/remhos-1.0.0 -m ./data/inline-quad.mesh -p 14 -rs 2 -rp 1 -dt 0.0005 -tf 0.6 -ho 1 -lo 2 -fct 3 Test: Sample Remap Example pts/deepsparse-1.5.0 zoo:cv/classification/resnet_v1-50/pytorch/sparseml/imagenet/base-none --scenario async Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream pts/build-gem5-1.0.0 Time To Compile pts/openvino-1.2.0 -m models/intel/vehicle-detection-0202/FP16-INT8/vehicle-detection-0202.xml -d CPU Model: Vehicle Detection FP16-INT8 - Device: CPU pts/hpcg-1.3.0 --nx=192 --ny=192 --nz=192 --rt=60 X Y Z: 192 192 192 - RT: 60 pts/ngspice-1.0.0 ~/iscas85Circuits/85/c2670/c2670_ann.net Circuit: C2670 pts/petsc-1.0.0 streams Test: Streams pts/stress-ng-1.10.0 --matrix-3d -1 --no-rand-seed Test: Matrix 3D Math pts/askap-2.1.0 tConvolveOMP Test: tConvolve OpenMP - Degridding pts/askap-2.1.0 tConvolveOMP Test: tConvolve OpenMP - Gridding pts/askap-2.1.0 tConvolveMPI Test: tConvolve MPI - Gridding pts/stockfish-1.4.0 Total Time pts/incompact3d-2.0.2 input_129_nodes.i3d Input: input.i3d 129 Cells Per Direction pts/hpcg-1.3.0 --nx=104 --ny=104 --nz=104 --rt=60 X Y Z: 104 104 104 - RT: 60