benchmarks for a future article.
Compare your own system(s) to this result file with the
Phoronix Test Suite by running the command:
phoronix-test-suite benchmark 2309060-NE-EXTRATEST87
extra tests
benchmarks for a future article.
d:
Processor: 2 x AMD EPYC 9124 16-Core @ 3.00GHz (32 Cores / 64 Threads), Motherboard: Supermicro H13DSH (1.5 BIOS), Memory: 24 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET, Disk: 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07, Graphics: astdrmfb
OS: AlmaLinux 9.2, Kernel: 5.14.0-284.25.1.el9_2.x86_64 (x86_64), Compiler: GCC 11.3.1 20221121, File-System: ext4, Screen Resolution: 1024x768
g:
Processor: 2 x AMD EPYC 9754 128-Core @ 2.25GHz (256 Cores / 512 Threads), Motherboard: Supermicro H13DSH (1.5 BIOS), Memory: 23 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET, Disk: 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07, Graphics: astdrmfb
OS: AlmaLinux 9.2, Kernel: 5.14.0-284.25.1.el9_2.x86_64 (x86_64), Compiler: GCC 11.3.1 20221121, File-System: ext4, Screen Resolution: 1024x768
h:
Processor: 2 x AMD EPYC 9754 128-Core @ 2.25GHz (256 Cores / 512 Threads), Motherboard: Supermicro H13DSH (1.5 BIOS), Memory: 23 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET, Disk: 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07, Graphics: astdrmfb
OS: AlmaLinux 9.2, Kernel: 5.14.0-284.25.1.el9_2.x86_64 (x86_64), Compiler: GCC 11.3.1 20221121, File-System: ext4, Screen Resolution: 1024x768
2 x AMD EPYC 9334 32-Core:
Processor: 2 x AMD EPYC 9334 32-Core @ 2.70GHz (64 Cores / 128 Threads), Motherboard: Supermicro H13DSH (1.5 BIOS), Memory: 21 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET, Disk: 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07, Graphics: astdrmfb
OS: AlmaLinux 9.2, Kernel: 5.14.0-284.25.1.el9_2.x86_64 (x86_64), Compiler: GCC 11.3.1 20221121, File-System: ext4, Screen Resolution: 1024x768
9334 2p:
Processor: 2 x AMD EPYC 9334 32-Core @ 2.70GHz (64 Cores / 128 Threads), Motherboard: Supermicro H13DSH (1.5 BIOS), Memory: 21 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET, Disk: 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07, Graphics: astdrmfb
OS: AlmaLinux 9.2, Kernel: 5.14.0-284.25.1.el9_2.x86_64 (x86_64), Compiler: GCC 11.3.1 20221121, File-System: ext4, Screen Resolution: 1024x768
93334 rep:
Processor: 2 x AMD EPYC 9334 32-Core @ 2.70GHz (64 Cores / 128 Threads), Motherboard: Supermicro H13DSH (1.5 BIOS), Memory: 21 x 32 GB DDR5-4800MT/s Samsung M321R4GA3BB6-CQKET, Disk: 2 x 1920GB SAMSUNG MZQL21T9HCJR-00A07, Graphics: astdrmfb
OS: AlmaLinux 9.2, Kernel: 5.14.0-284.25.1.el9_2.x86_64 (x86_64), Compiler: GCC 11.3.1 20221121, File-System: ext4, Screen Resolution: 1024x768
Stress-NG 0.16.04
Test: Semaphores
Bogo Ops/s > Higher Is Better
d ......................... 92019395.13 |====
g ......................... 800762640.11 |=====================================
h ......................... 786069338.66 |====================================
2 x AMD EPYC 9334 32-Core . 226498215.44 |==========
9334 2p ................... 212194044.21 |==========
93334 rep ................. 214488001.21 |==========
Stress-NG 0.16.04
Test: Pipe
Bogo Ops/s > Higher Is Better
d ......................... 20249488.61 |====
g ......................... 173499514.94 |=====================================
h ......................... 159491222.34 |==================================
2 x AMD EPYC 9334 32-Core . 39461287.60 |========
9334 2p ................... 42013739.43 |=========
93334 rep ................. 42047609.47 |=========
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 373.03 |======
g ......................... 2613.09 |==========================================
h ......................... 2633.59 |==========================================
2 x AMD EPYC 9334 32-Core . 773.51 |============
9334 2p ................... 779.23 |============
93334 rep ................. 777.85 |============
Neural Magic DeepSparse 1.5
Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 257.42 |======
g ......................... 1780.24 |==========================================
h ......................... 1784.36 |==========================================
2 x AMD EPYC 9334 32-Core . 519.82 |============
9334 2p ................... 521.81 |============
93334 rep ................. 520.32 |============
Neural Magic DeepSparse 1.5
Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 258.56 |======
g ......................... 1778.36 |==========================================
h ......................... 1786.18 |==========================================
2 x AMD EPYC 9334 32-Core . 519.99 |============
9334 2p ................... 522.11 |============
93334 rep ................. 519.73 |============
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 757.48 |======
g ......................... 5150.43 |==========================================
h ......................... 5175.49 |==========================================
2 x AMD EPYC 9334 32-Core . 1522.05 |============
9334 2p ................... 1519.30 |============
93334 rep ................. 1521.75 |============
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 24.72 |======
g ......................... 166.73 |==========================================
h ......................... 168.75 |===========================================
2 x AMD EPYC 9334 32-Core . 49.76 |=============
9334 2p ................... 49.80 |=============
93334 rep ................. 49.87 |=============
Neural Magic DeepSparse 1.5
Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 355.89 |======
g ......................... 2405.88 |==========================================
h ......................... 2429.74 |==========================================
2 x AMD EPYC 9334 32-Core . 718.35 |============
9334 2p ................... 719.77 |============
93334 rep ................. 715.59 |============
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 114.28 |======
g ......................... 770.51 |===========================================
h ......................... 774.41 |===========================================
2 x AMD EPYC 9334 32-Core . 233.91 |=============
9334 2p ................... 234.13 |=============
93334 rep ................. 233.34 |=============
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 113.66 |======
g ......................... 761.87 |===========================================
h ......................... 764.74 |===========================================
2 x AMD EPYC 9334 32-Core . 231.46 |=============
9334 2p ................... 232.23 |=============
93334 rep ................. 231.42 |=============
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 168.72 |======
g ......................... 1128.88 |==========================================
h ......................... 1127.79 |==========================================
2 x AMD EPYC 9334 32-Core . 333.85 |============
9334 2p ................... 334.53 |============
93334 rep ................. 334.27 |============
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2 - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 85.49 |======
g ......................... 566.73 |===========================================
h ......................... 569.32 |===========================================
2 x AMD EPYC 9334 32-Core . 169.27 |=============
9334 2p ................... 168.34 |=============
93334 rep ................. 168.63 |=============
Neural Magic DeepSparse 1.5
Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 19.88 |======
g ......................... 129.44 |==========================================
h ......................... 131.54 |===========================================
2 x AMD EPYC 9334 32-Core . 40.57 |=============
9334 2p ................... 40.55 |=============
93334 rep ................. 40.38 |=============
Neural Magic DeepSparse 1.5
Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 19.81 |======
g ......................... 129.38 |==========================================
h ......................... 131.10 |===========================================
2 x AMD EPYC 9334 32-Core . 40.25 |=============
9334 2p ................... 40.45 |=============
93334 rep ................. 40.29 |=============
Stress-NG 0.16.04
Test: x86_64 RdRand
Bogo Ops/s > Higher Is Better
d ......................... 12013602.06 |======
g ......................... 78673620.14 |======================================
h ......................... 78690548.49 |======================================
2 x AMD EPYC 9334 32-Core . 25697538.80 |============
9334 2p ................... 25688067.52 |============
93334 rep ................. 25663966.02 |============
Stress-NG 0.16.04
Test: Hash
Bogo Ops/s > Higher Is Better
d ......................... 7218607.12 |======
g ......................... 45298625.91 |======================================
h ......................... 45265797.34 |======================================
2 x AMD EPYC 9334 32-Core . 15425278.83 |=============
9334 2p ................... 15412840.21 |=============
93334 rep ................. 15326552.14 |=============
Stress-NG 0.16.04
Test: Context Switching
Bogo Ops/s > Higher Is Better
d ......................... 17931974.92 |======
g ......................... 110432440.32 |=====================================
h ......................... 110991931.60 |=====================================
2 x AMD EPYC 9334 32-Core . 29340312.96 |==========
9334 2p ................... 35971998.98 |============
93334 rep ................. 35280554.42 |============
Stress-NG 0.16.04
Test: Vector Shuffle
Bogo Ops/s > Higher Is Better
d ......................... 25006.73 |======
g ......................... 154267.34 |========================================
h ......................... 154293.33 |========================================
2 x AMD EPYC 9334 32-Core . 52591.77 |==============
9334 2p ................... 52548.25 |==============
93334 rep ................. 52511.07 |==============
Neural Magic DeepSparse 1.5
Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 36.38 |=======
g ......................... 220.91 |===========================================
h ......................... 222.86 |===========================================
2 x AMD EPYC 9334 32-Core . 70.90 |==============
9334 2p ................... 70.21 |==============
93334 rep ................. 70.82 |==============
NCNN 20230517
Target: CPU - Model: regnety_400m
ms < Lower Is Better
d ......................... 33.50 |=======
d ......................... 34.57 |=======
g ......................... 202.34 |===========================================
h ......................... 201.59 |===========================================
2 x AMD EPYC 9334 32-Core . 54.59 |============
9334 2p ................... 53.78 |===========
93334 rep ................. 68.68 |===============
Stress-NG 0.16.04
Test: Floating Point
Bogo Ops/s > Higher Is Better
d ......................... 11939.41 |=======
g ......................... 71782.95 |=========================================
h ......................... 71790.06 |=========================================
2 x AMD EPYC 9334 32-Core . 24103.76 |==============
9334 2p ................... 24101.40 |==============
93334 rep ................. 24129.62 |==============
Stress-NG 0.16.04
Test: CPU Stress
Bogo Ops/s > Higher Is Better
d ......................... 88466.54 |=======
g ......................... 523898.45 |========================================
h ......................... 518600.85 |========================================
2 x AMD EPYC 9334 32-Core . 176344.38 |=============
9334 2p ................... 176621.27 |=============
93334 rep ................. 176710.85 |=============
Stress-NG 0.16.04
Test: Function Call
Bogo Ops/s > Higher Is Better
d ......................... 27731.16 |=======
g ......................... 156381.63 |=======================================
h ......................... 158994.25 |========================================
2 x AMD EPYC 9334 32-Core . 54824.58 |==============
9334 2p ................... 54852.06 |==============
93334 rep ................. 54675.85 |==============
Stress-NG 0.16.04
Test: Matrix Math
Bogo Ops/s > Higher Is Better
d ......................... 173878.70 |=======
g ......................... 989553.80 |========================================
h ......................... 990104.85 |========================================
2 x AMD EPYC 9334 32-Core . 314789.68 |=============
9334 2p ................... 315324.18 |=============
93334 rep ................. 314753.85 |=============
Stress-NG 0.16.04
Test: Vector Floating Point
Bogo Ops/s > Higher Is Better
d ......................... 106715.85 |=======
g ......................... 604724.84 |========================================
h ......................... 605340.53 |========================================
2 x AMD EPYC 9334 32-Core . 198651.88 |=============
9334 2p ................... 198855.75 |=============
93334 rep ................. 198610.93 |=============
Stress-NG 0.16.04
Test: Fused Multiply-Add
Bogo Ops/s > Higher Is Better
d ......................... 32151217.03 |=======
g ......................... 180789739.08 |=====================================
h ......................... 181626937.85 |=====================================
2 x AMD EPYC 9334 32-Core . 58550931.39 |============
9334 2p ................... 58491170.75 |============
93334 rep ................. 58570424.41 |============
Stress-NG 0.16.04
Test: Zlib
Bogo Ops/s > Higher Is Better
d ......................... 2944.54 |=======
g ......................... 16624.07 |=========================================
h ......................... 16625.57 |=========================================
2 x AMD EPYC 9334 32-Core . 5876.15 |==============
9334 2p ................... 5885.26 |===============
93334 rep ................. 5875.45 |==============
Neural Magic DeepSparse 1.5
Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 2661.23 |=======
g ......................... 14854.86 |=========================================
h ......................... 13153.50 |====================================
2 x AMD EPYC 9334 32-Core . 5651.57 |================
9334 2p ................... 5626.77 |================
93334 rep ................. 5618.59 |================
Blender 3.6
Blend File: Classroom - Compute: CPU-Only
Seconds < Lower Is Better
d ......................... 94.88 |============================================
g ......................... 17.31 |========
h ......................... 17.12 |========
2 x AMD EPYC 9334 32-Core . 51.86 |========================
9334 2p ................... 52.77 |========================
93334 rep ................. 51.99 |========================
Stress-NG 0.16.04
Test: Vector Math
Bogo Ops/s > Higher Is Better
d ......................... 234239.60 |=======
g ......................... 1288008.26 |=======================================
h ......................... 1285121.40 |=======================================
2 x AMD EPYC 9334 32-Core . 430109.26 |=============
9334 2p ................... 430141.54 |=============
93334 rep ................. 430096.55 |=============
Blender 3.6
Blend File: Pabellon Barcelona - Compute: CPU-Only
Seconds < Lower Is Better
d ......................... 122.70 |===========================================
g ......................... 22.40 |========
h ......................... 22.35 |========
2 x AMD EPYC 9334 32-Core . 63.37 |======================
9334 2p ................... 63.45 |======================
93334 rep ................. 63.71 |======================
Stress-NG 0.16.04
Test: AVX-512 VNNI
Bogo Ops/s > Higher Is Better
d ......................... 3667964.15 |=======
g ......................... 20133541.11 |======================================
h ......................... 20123210.14 |======================================
2 x AMD EPYC 9334 32-Core . 6626091.93 |=============
9334 2p ................... 6624476.36 |=============
93334 rep ................. 6621974.98 |============
Stress-NG 0.16.04
Test: Memory Copying
Bogo Ops/s > Higher Is Better
d ......................... 13428.64 |========
g ......................... 72825.56 |=========================================
h ......................... 72800.74 |=========================================
2 x AMD EPYC 9334 32-Core . 26146.71 |===============
9334 2p ................... 26144.80 |===============
93334 rep ................. 26129.50 |===============
Stress-NG 0.16.04
Test: Wide Vector Math
Bogo Ops/s > Higher Is Better
d ......................... 1543291.91 |=======
g ......................... 8237581.37 |=======================================
h ......................... 8234456.23 |=======================================
2 x AMD EPYC 9334 32-Core . 2804498.36 |=============
9334 2p ................... 2802185.77 |=============
93334 rep ................. 2802578.87 |=============
Stress-NG 0.16.04
Test: Glibc Qsort Data Sorting
Bogo Ops/s > Higher Is Better
d ......................... 916.74 |========
g ......................... 4846.22 |==========================================
h ......................... 4850.99 |==========================================
2 x AMD EPYC 9334 32-Core . 1675.93 |===============
9334 2p ................... 1674.65 |==============
93334 rep ................. 1676.27 |===============
Stress-NG 0.16.04
Test: Malloc
Bogo Ops/s > Higher Is Better
d ......................... 137476652.93 |=======
g ......................... 719479964.33 |=====================================
h ......................... 725705488.15 |=====================================
2 x AMD EPYC 9334 32-Core . 262321226.78 |=============
9334 2p ................... 263438850.64 |=============
93334 rep ................. 263330738.63 |=============
Stress-NG 0.16.04
Test: Glibc C String Functions
Bogo Ops/s > Higher Is Better
d ......................... 40991307.95 |=======
g ......................... 215429087.61 |=====================================
h ......................... 215308424.81 |=====================================
2 x AMD EPYC 9334 32-Core . 72167140.03 |============
9334 2p ................... 77814243.76 |=============
93334 rep ................. 72810418.71 |=============
Blender 3.6
Blend File: Barbershop - Compute: CPU-Only
Seconds < Lower Is Better
d ......................... 372.72 |===========================================
g ......................... 71.97 |========
h ......................... 71.39 |========
2 x AMD EPYC 9334 32-Core . 201.53 |=======================
9334 2p ................... 201.51 |=======================
93334 rep ................. 201.93 |=======================
Blender 3.6
Blend File: BMW27 - Compute: CPU-Only
Seconds < Lower Is Better
d ......................... 37.93 |============================================
g ......................... 7.51 |=========
h ......................... 7.40 |=========
2 x AMD EPYC 9334 32-Core . 20.65 |========================
9334 2p ................... 20.51 |========================
93334 rep ................. 20.43 |========================
OSPRay 2.12
Benchmark: gravity_spheres_volume/dim_512/scivis/real_time
Items Per Second > Higher Is Better
d ......................... 9.81318 |========
g ......................... 49.51230 |=========================================
h ......................... 50.08790 |=========================================
2 x AMD EPYC 9334 32-Core . 17.43570 |==============
9334 2p ................... 17.33070 |==============
93334 rep ................. 17.27890 |==============
Embree 4.1
Binary: Pathtracer - Model: Asian Dragon
Frames Per Second > Higher Is Better
d ......................... 42.76 |========
g ......................... 217.59 |===========================================
h ......................... 207.25 |=========================================
2 x AMD EPYC 9334 32-Core . 75.28 |===============
9334 2p ................... 75.46 |===============
93334 rep ................. 75.12 |===============
Embree 4.1
Binary: Pathtracer - Model: Asian Dragon Obj
Frames Per Second > Higher Is Better
d ......................... 37.99 |========
g ......................... 193.05 |===========================================
h ......................... 172.21 |======================================
2 x AMD EPYC 9334 32-Core . 67.48 |===============
9334 2p ................... 67.32 |===============
93334 rep ................. 66.85 |===============
OSPRay 2.12
Benchmark: gravity_spheres_volume/dim_512/ao/real_time
Items Per Second > Higher Is Better
d ......................... 10.08 |=========
g ......................... 50.40 |============================================
h ......................... 50.96 |============================================
2 x AMD EPYC 9334 32-Core . 17.81 |===============
9334 2p ................... 17.76 |===============
93334 rep ................. 17.83 |===============
Neural Magic DeepSparse 1.5
Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Asynchronous Multi-Stream
items/sec > Higher Is Better
d ......................... 106.15 |=========
g ......................... 497.90 |========================================
h ......................... 532.46 |===========================================
2 x AMD EPYC 9334 32-Core . 204.41 |=================
9334 2p ................... 204.35 |=================
93334 rep ................. 204.22 |================
Embree 4.1
Binary: Pathtracer - Model: Crown
Frames Per Second > Higher Is Better
d ......................... 37.68 |=========
g ......................... 183.40 |==========================================
h ......................... 187.42 |===========================================
2 x AMD EPYC 9334 32-Core . 69.49 |================
9334 2p ................... 69.67 |================
93334 rep ................. 69.49 |================
Embree 4.1
Binary: Pathtracer ISPC - Model: Asian Dragon
Frames Per Second > Higher Is Better
d ......................... 48.47 |=========
g ......................... 239.42 |===========================================
h ......................... 225.83 |=========================================
2 x AMD EPYC 9334 32-Core . 86.79 |================
9334 2p ................... 86.53 |================
93334 rep ................. 86.56 |================
Embree 4.1
Binary: Pathtracer ISPC - Model: Crown
Frames Per Second > Higher Is Better
d ......................... 39.38 |=========
g ......................... 193.60 |===========================================
h ......................... 194.27 |===========================================
2 x AMD EPYC 9334 32-Core . 72.65 |================
9334 2p ................... 72.64 |================
93334 rep ................. 72.17 |================
Stress-NG 0.16.04
Test: Poll
Bogo Ops/s > Higher Is Better
d ......................... 4337787.72 |========
g ......................... 21346079.30 |======================================
h ......................... 20251576.33 |====================================
2 x AMD EPYC 9334 32-Core . 9104744.15 |================
9334 2p ................... 9067852.19 |================
93334 rep ................. 9078474.01 |================
Blender 3.6
Blend File: Fishy Cat - Compute: CPU-Only
Seconds < Lower Is Better
d ......................... 48.95 |============================================
g ......................... 10.02 |=========
h ......................... 10.02 |=========
2 x AMD EPYC 9334 32-Core . 26.43 |========================
9334 2p ................... 26.15 |========================
93334 rep ................. 26.37 |========================
NCNN 20230517
Target: CPU - Model: blazeface
ms < Lower Is Better
d ......................... 5.63 |=========
d ......................... 5.84 |==========
g ......................... 26.90 |============================================
h ......................... 23.63 |=======================================
2 x AMD EPYC 9334 32-Core . 7.20 |============
9334 2p ................... 8.15 |=============
93334 rep ................. 8.39 |==============
Embree 4.1
Binary: Pathtracer ISPC - Model: Asian Dragon Obj
Frames Per Second > Higher Is Better
d ......................... 40.57 |=========
g ......................... 193.43 |===========================================
h ......................... 184.45 |=========================================
2 x AMD EPYC 9334 32-Core . 72.26 |================
9334 2p ................... 72.77 |================
93334 rep ................. 72.85 |================
Stress-NG 0.16.04
Test: SENDFILE
Bogo Ops/s > Higher Is Better
d ......................... 857324.89 |=========
g ......................... 3761786.29 |=======================================
h ......................... 3745995.16 |=======================================
2 x AMD EPYC 9334 32-Core . 1312330.44 |==============
9334 2p ................... 1465208.81 |===============
93334 rep ................. 1536517.65 |================
OSPRay 2.12
Benchmark: particle_volume/ao/real_time
Items Per Second > Higher Is Better
d ......................... 10.86 |==========
g ......................... 46.66 |============================================
h ......................... 46.76 |============================================
2 x AMD EPYC 9334 32-Core . 20.40 |===================
9334 2p ................... 20.45 |===================
93334 rep ................. 20.39 |===================
OSPRay 2.12
Benchmark: particle_volume/scivis/real_time
Items Per Second > Higher Is Better
d ......................... 10.85 |==========
g ......................... 46.61 |============================================
h ......................... 46.55 |============================================
2 x AMD EPYC 9334 32-Core . 20.33 |===================
9334 2p ................... 20.40 |===================
93334 rep ................. 20.31 |===================
Stress-NG 0.16.04
Test: Crypto
Bogo Ops/s > Higher Is Better
d ......................... 107643.49 |=========
g ......................... 462528.73 |========================================
h ......................... 461785.36 |========================================
2 x AMD EPYC 9334 32-Core . 156553.51 |==============
9334 2p ................... 156859.22 |==============
93334 rep ................. 156863.78 |==============
NCNN 20230517
Target: CPU - Model: shufflenet-v2
ms < Lower Is Better
d ......................... 13.96 |===========
d ......................... 15.46 |============
g ......................... 57.67 |============================================
h ......................... 45.22 |===================================
2 x AMD EPYC 9334 32-Core . 18.63 |==============
9334 2p ................... 18.35 |==============
93334 rep ................. 18.47 |==============
Stress-NG 0.16.04
Test: Forking
Bogo Ops/s > Higher Is Better
d ......................... 1007.46 |=
g ......................... 31064.95 |=========================================
h ......................... 30591.67 |========================================
2 x AMD EPYC 9334 32-Core . 1641.77 |==
9334 2p ................... 1615.00 |==
93334 rep ................. 1616.05 |==
Stress-NG 0.16.04
Test: AVL Tree
Bogo Ops/s > Higher Is Better
d ......................... 410.89 |===========
g ......................... 1564.42 |========================================
h ......................... 1628.91 |==========================================
2 x AMD EPYC 9334 32-Core . 788.21 |====================
9334 2p ................... 793.08 |====================
93334 rep ................. 797.64 |=====================
Stress-NG 0.16.04
Test: CPU Cache
Bogo Ops/s > Higher Is Better
d ......................... 776708.72 |===========================
g ......................... 334039.11 |===========
h ......................... 361173.71 |============
2 x AMD EPYC 9334 32-Core . 1095211.77 |======================================
9334 2p ................... 1135334.90 |=======================================
93334 rep ................. 1096840.32 |======================================
Stress-NG 0.16.04
Test: MMAP
Bogo Ops/s > Higher Is Better
d ......................... 1131.55 |=============
g ......................... 3664.34 |=========================================
h ......................... 3749.47 |==========================================
2 x AMD EPYC 9334 32-Core . 2289.80 |==========================
9334 2p ................... 2299.28 |==========================
93334 rep ................. 2287.69 |==========================
NCNN 20230517
Target: CPU - Model: efficientnet-b0
ms < Lower Is Better
d ......................... 14.76 |=============
d ......................... 15.43 |==============
g ......................... 47.67 |===========================================
h ......................... 48.76 |============================================
2 x AMD EPYC 9334 32-Core . 19.79 |==================
9334 2p ................... 19.92 |==================
93334 rep ................. 19.96 |==================
NCNN 20230517
Target: CPU-v3-v3 - Model: mobilenet-v3
ms < Lower Is Better
d ......................... 11.64 |==============
d ......................... 12.01 |==============
g ......................... 35.96 |==========================================
h ......................... 37.56 |============================================
2 x AMD EPYC 9334 32-Core . 14.87 |=================
9334 2p ................... 15.18 |==================
93334 rep ................. 15.87 |===================
Stress-NG 0.16.04
Test: NUMA
Bogo Ops/s > Higher Is Better
d ......................... 18.35 |===================
g ......................... 14.38 |===============
h ......................... 14.46 |===============
2 x AMD EPYC 9334 32-Core . 42.33 |============================================
9334 2p ................... 41.49 |===========================================
93334 rep ................. 42.10 |============================================
NCNN 20230517
Target: CPU - Model: mnasnet
ms < Lower Is Better
d ......................... 10.70 |================
d ......................... 10.83 |================
g ......................... 29.92 |============================================
h ......................... 29.55 |===========================================
2 x AMD EPYC 9334 32-Core . 13.74 |====================
9334 2p ................... 13.69 |====================
93334 rep ................. 14.24 |=====================
NCNN 20230517
Target: CPU-v2-v2 - Model: mobilenet-v2
ms < Lower Is Better
d ......................... 11.29 |================
d ......................... 11.83 |=================
g ......................... 31.24 |============================================
h ......................... 30.89 |============================================
2 x AMD EPYC 9334 32-Core . 13.94 |====================
9334 2p ................... 14.02 |====================
93334 rep ................. 15.02 |=====================
NCNN 20230517
Target: CPU - Model: FastestDet
ms < Lower Is Better
d ......................... 16.00 |================
g ......................... 44.27 |============================================
h ......................... 42.30 |==========================================
2 x AMD EPYC 9334 32-Core . 16.44 |================
9334 2p ................... 22.29 |======================
93334 rep ................. 21.50 |=====================
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 14.7454 |==========================================
g ......................... 5.4661 |================
h ......................... 5.4481 |================
2 x AMD EPYC 9334 32-Core . 8.9324 |=========================
9334 2p ................... 8.9380 |=========================
93334 rep ................. 8.9437 |=========================
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 67.77 |================
g ......................... 182.64 |===========================================
h ......................... 183.24 |===========================================
2 x AMD EPYC 9334 32-Core . 111.85 |==========================
9334 2p ................... 111.78 |==========================
93334 rep ................. 111.70 |==========================
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 14.6632 |==========================================
g ......................... 5.4478 |================
h ......................... 5.4268 |================
2 x AMD EPYC 9334 32-Core . 8.8864 |=========================
9334 2p ................... 8.8725 |=========================
93334 rep ................. 8.8896 |=========================
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 68.16 |================
g ......................... 183.37 |===========================================
h ......................... 184.09 |===========================================
2 x AMD EPYC 9334 32-Core . 112.47 |==========================
9334 2p ................... 112.65 |==========================
93334 rep ................. 112.43 |==========================
Intel Open Image Denoise 2.0
Run: RT.hdr_alb_nrm.3840x2160 - Device: CPU-Only
Images / Sec > Higher Is Better
d ......................... 1.36 |=================
g ......................... 3.40 |===========================================
h ......................... 3.53 |=============================================
2 x AMD EPYC 9334 32-Core . 2.21 |============================
9334 2p ................... 2.20 |============================
93334 rep ................. 2.19 |============================
Intel Open Image Denoise 2.0
Run: RT.ldr_alb_nrm.3840x2160 - Device: CPU-Only
Images / Sec > Higher Is Better
d ......................... 1.37 |==================
g ......................... 3.28 |===========================================
h ......................... 3.46 |=============================================
2 x AMD EPYC 9334 32-Core . 2.21 |=============================
9334 2p ................... 2.20 |=============================
93334 rep ................. 2.21 |=============================
NCNN 20230517
Target: CPU - Model: squeezenet_ssd
ms < Lower Is Better
d ......................... 24.22 |==================
d ......................... 25.61 |===================
g ......................... 58.66 |============================================
h ......................... 58.41 |============================================
2 x AMD EPYC 9334 32-Core . 25.51 |===================
9334 2p ................... 28.47 |=====================
93334 rep ................. 34.99 |==========================
OSPRay 2.12
Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_time
Items Per Second > Higher Is Better
d ......................... 11.90 |==================
g ......................... 28.35 |============================================
h ......................... 28.25 |============================================
2 x AMD EPYC 9334 32-Core . 20.34 |================================
9334 2p ................... 20.39 |================================
93334 rep ................. 20.33 |================================
Intel Open Image Denoise 2.0
Run: RTLightmap.hdr.4096x4096 - Device: CPU-Only
Images / Sec > Higher Is Better
d ......................... 0.65 |===================
g ......................... 1.50 |============================================
h ......................... 1.54 |=============================================
2 x AMD EPYC 9334 32-Core . 1.04 |==============================
9334 2p ................... 1.03 |==============================
93334 rep ................. 1.04 |==============================
Neural Magic DeepSparse 1.5
Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 71.35 |============================================
g ......................... 30.22 |===================
h ......................... 30.37 |===================
2 x AMD EPYC 9334 32-Core . 44.30 |===========================
9334 2p ................... 44.75 |============================
93334 rep ................. 44.36 |===========================
Neural Magic DeepSparse 1.5
Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 14.01 |===================
g ......................... 33.08 |============================================
h ......................... 32.92 |============================================
2 x AMD EPYC 9334 32-Core . 22.57 |==============================
9334 2p ................... 22.34 |==============================
93334 rep ................. 22.54 |==============================
Neural Magic DeepSparse 1.5
Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 71.84 |============================================
g ......................... 31.45 |===================
h ......................... 30.72 |===================
2 x AMD EPYC 9334 32-Core . 44.74 |===========================
9334 2p ................... 44.63 |===========================
93334 rep ................. 44.66 |===========================
Neural Magic DeepSparse 1.5
Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 13.92 |===================
g ......................... 31.79 |===========================================
h ......................... 32.54 |============================================
2 x AMD EPYC 9334 32-Core . 22.35 |==============================
9334 2p ................... 22.40 |==============================
93334 rep ................. 22.39 |==============================
Dragonflydb 1.6.2
Clients Per Thread: 20 - Set To Get Ratio: 1:100
Ops/sec > Higher Is Better
d ......................... 14250093.71 |=================
2 x AMD EPYC 9334 32-Core . 31139658.49 |======================================
9334 2p ................... 29330317.34 |====================================
93334 rep ................. 29280070.97 |====================================
Dragonflydb 1.6.2
Clients Per Thread: 10 - Set To Get Ratio: 1:100
Ops/sec > Higher Is Better
d ......................... 11860591.65 |=================
2 x AMD EPYC 9334 32-Core . 25802999.03 |======================================
9334 2p ................... 22585212.56 |=================================
93334 rep ................. 24588214.81 |====================================
NCNN 20230517
Target: CPU - Model: resnet18
ms < Lower Is Better
d ......................... 13.65 |====================
d ......................... 13.75 |====================
g ......................... 29.26 |============================================
h ......................... 29.56 |============================================
2 x AMD EPYC 9334 32-Core . 15.16 |=======================
9334 2p ................... 17.49 |==========================
93334 rep ................. 16.22 |========================
NCNN 20230517
Target: CPU - Model: alexnet
ms < Lower Is Better
d ......................... 8.13 |=====================
d ......................... 8.14 |=====================
g ......................... 15.10 |======================================
h ......................... 17.39 |============================================
2 x AMD EPYC 9334 32-Core . 9.39 |========================
9334 2p ................... 10.70 |===========================
93334 rep ................. 10.59 |===========================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 4.9317 |============================
g ......................... 7.3831 |===========================================
h ......................... 7.4589 |===========================================
2 x AMD EPYC 9334 32-Core . 3.5591 |=====================
9334 2p ................... 3.5922 |=====================
93334 rep ................. 3.5347 |====================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 202.50 |===============================
g ......................... 135.40 |=====================
h ......................... 134.02 |====================
2 x AMD EPYC 9334 32-Core . 280.78 |===========================================
9334 2p ................... 278.18 |==========================================
93334 rep ................. 282.72 |===========================================
Dragonflydb 1.6.2
Clients Per Thread: 10 - Set To Get Ratio: 1:10
Ops/sec > Higher Is Better
d ......................... 11686474.26 |==================
2 x AMD EPYC 9334 32-Core . 22176931.98 |==================================
9334 2p ................... 24460221.48 |======================================
93334 rep ................. 23645213.13 |=====================================
Neural Magic DeepSparse 1.5
Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 40.73 |============================================
g ......................... 20.60 |======================
h ......................... 19.74 |=====================
2 x AMD EPYC 9334 32-Core . 26.51 |=============================
9334 2p ................... 26.48 |=============================
93334 rep ................. 26.48 |=============================
Neural Magic DeepSparse 1.5
Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 24.54 |=====================
g ......................... 48.50 |==========================================
h ......................... 50.60 |============================================
2 x AMD EPYC 9334 32-Core . 37.70 |=================================
9334 2p ................... 37.74 |=================================
93334 rep ................. 37.74 |=================================
NCNN 20230517
Target: CPU - Model: vgg16
ms < Lower Is Better
d ......................... 28.69 |=====================
d ......................... 28.57 |=====================
g ......................... 56.34 |==========================================
h ......................... 58.85 |============================================
2 x AMD EPYC 9334 32-Core . 36.65 |===========================
9334 2p ................... 36.90 |============================
93334 rep ................. 42.08 |===============================
NCNN 20230517
Target: CPU - Model: resnet50
ms < Lower Is Better
d ......................... 26.53 |=====================
d ......................... 26.68 |======================
g ......................... 53.62 |===========================================
h ......................... 54.47 |============================================
2 x AMD EPYC 9334 32-Core . 29.62 |========================
9334 2p ................... 30.60 |=========================
93334 rep ................. 41.02 |=================================
Dragonflydb 1.6.2
Clients Per Thread: 20 - Set To Get Ratio: 1:10
Ops/sec > Higher Is Better
d ......................... 15179164.86 |===================
2 x AMD EPYC 9334 32-Core . 31053489.24 |======================================
9334 2p ................... 28043162.30 |==================================
93334 rep ................. 29580028.06 |====================================
NCNN 20230517
Target: CPU - Model: mobilenet
ms < Lower Is Better
d ......................... 23.55 |=======================
d ......................... 23.97 |========================
g ......................... 41.21 |=========================================
h ......................... 44.14 |============================================
2 x AMD EPYC 9334 32-Core . 21.80 |======================
9334 2p ................... 27.41 |===========================
93334 rep ................. 27.95 |============================
NCNN 20230517
Target: CPU - Model: vision_transformer
ms < Lower Is Better
d ......................... 55.05 |======================
d ......................... 55.81 |======================
g ......................... 108.73 |===========================================
h ......................... 93.15 |=====================================
2 x AMD EPYC 9334 32-Core . 53.79 |=====================
9334 2p ................... 66.65 |==========================
93334 rep ................. 61.24 |========================
Stress-NG 0.16.04
Test: Socket Activity
Bogo Ops/s > Higher Is Better
d ......................... 9504.38 |===
g ......................... 110254.19 |=======================================
h ......................... 113347.77 |========================================
2 x AMD EPYC 9334 32-Core . 38147.97 |=============
9334 2p ................... 37674.67 |=============
93334 rep ................. 38727.86 |==============
BRL-CAD 7.36
VGR Performance Metric
VGR Performance Metric > Higher Is Better
d ......................... 544648 |====
g ......................... 6398404 |==========================================
h ......................... 6270577 |=========================================
2 x AMD EPYC 9334 32-Core . 1012537 |=======
9334 2p ................... 994371 |=======
93334 rep ................. 999893 |=======
Stress-NG 0.16.04
Test: MEMFD
Bogo Ops/s > Higher Is Better
d ......................... 912.62 |======================================
g ......................... 97.94 |====
h ......................... 115.83 |=====
2 x AMD EPYC 9334 32-Core . 1004.79 |==========================================
9334 2p ................... 1014.24 |==========================================
93334 rep ................. 1009.47 |==========================================
NCNN 20230517
Target: CPU - Model: googlenet
ms < Lower Is Better
d ......................... 29.03 |======================
d ......................... 30.11 |=======================
g ......................... 57.72 |============================================
h ......................... 55.36 |==========================================
2 x AMD EPYC 9334 32-Core . 31.03 |========================
9334 2p ................... 33.40 |=========================
93334 rep ................. 33.68 |==========================
Neural Magic DeepSparse 1.5
Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 1.1189 |=============================
g ......................... 1.6492 |===========================================
h ......................... 1.6419 |===========================================
2 x AMD EPYC 9334 32-Core . 0.8711 |=======================
9334 2p ................... 0.8677 |=======================
93334 rep ................. 0.8704 |=======================
Neural Magic DeepSparse 1.5
Model: ResNet-50, Sparse INT8 - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 890.94 |=================================
g ......................... 605.47 |======================
h ......................... 608.16 |======================
2 x AMD EPYC 9334 32-Core . 1144.31 |==========================================
9334 2p ................... 1149.13 |==========================================
93334 rep ................. 1145.23 |==========================================
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 11.0868 |==================================
g ......................... 13.8473 |==========================================
h ......................... 12.8286 |=======================================
2 x AMD EPYC 9334 32-Core . 7.5379 |=======================
9334 2p ................... 7.7484 |========================
93334 rep ................. 8.1347 |=========================
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 90.08 |=============================
g ......................... 72.18 |=======================
h ......................... 77.92 |=========================
2 x AMD EPYC 9334 32-Core . 132.50 |===========================================
9334 2p ................... 128.91 |==========================================
93334 rep ................. 122.79 |========================================
SPECFEM3D 4.0
Model: Layered Halfspace
Seconds < Lower Is Better
d ......................... 38.49 |============================================
2 x AMD EPYC 9334 32-Core . 21.67 |=========================
9334 2p ................... 21.54 |=========================
93334 rep ................. 21.69 |=========================
SPECFEM3D 4.0
Model: Mount St. Helens
Seconds < Lower Is Better
d ......................... 14.505645253 |=====================================
2 x AMD EPYC 9334 32-Core . 8.455473534 |======================
9334 2p ................... 8.320091856 |=====================
93334 rep ................. 8.486910520 |======================
Neural Magic DeepSparse 1.5
Model: ResNet-50, Sparse INT8 - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 5.9947 |===========================
g ......................... 8.5843 |======================================
h ......................... 9.6974 |===========================================
2 x AMD EPYC 9334 32-Core . 5.6462 |=========================
9334 2p ................... 5.6693 |=========================
93334 rep ................. 5.6773 |=========================
Neural Magic DeepSparse 1.5
Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 150.28 |=========================
g ......................... 256.16 |===========================================
h ......................... 239.19 |========================================
2 x AMD EPYC 9334 32-Core . 156.06 |==========================
9334 2p ................... 156.28 |==========================
93334 rep ................. 156.19 |==========================
SPECFEM3D 4.0
Model: Water-layered Halfspace
Seconds < Lower Is Better
d ......................... 34.82 |============================================
2 x AMD EPYC 9334 32-Core . 21.39 |===========================
9334 2p ................... 20.86 |==========================
93334 rep ................. 21.15 |===========================
Neural Magic DeepSparse 1.5
Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 7.1456 |===================================
g ......................... 8.7387 |===========================================
h ......................... 8.6016 |==========================================
2 x AMD EPYC 9334 32-Core . 5.3972 |===========================
9334 2p ................... 5.3028 |==========================
93334 rep ................. 5.2862 |==========================
Neural Magic DeepSparse 1.5
Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 139.79 |================================
g ......................... 114.38 |==========================
h ......................... 116.19 |==========================
2 x AMD EPYC 9334 32-Core . 185.02 |==========================================
9334 2p ................... 188.32 |===========================================
93334 rep ................. 188.92 |===========================================
Stress-NG 0.16.04
Test: Futex
Bogo Ops/s > Higher Is Better
d ......................... 4022496.07 |=================================
g ......................... 2951631.96 |========================
h ......................... 2880825.11 |========================
2 x AMD EPYC 9334 32-Core . 4717570.36 |=======================================
9334 2p ................... 4644551.29 |======================================
93334 rep ................. 4752131.37 |=======================================
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 58.83 |============================================
g ......................... 42.04 |===============================
h ......................... 35.85 |===========================
2 x AMD EPYC 9334 32-Core . 37.14 |============================
9334 2p ................... 37.30 |============================
93334 rep ................. 37.57 |============================
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 17.00 |===========================
g ......................... 23.78 |======================================
h ......................... 27.88 |============================================
2 x AMD EPYC 9334 32-Core . 26.92 |==========================================
9334 2p ................... 26.81 |==========================================
93334 rep ................. 26.61 |==========================================
SPECFEM3D 4.0
Model: Homogeneous Halfspace
Seconds < Lower Is Better
d ......................... 18.85 |============================================
2 x AMD EPYC 9334 32-Core . 11.62 |===========================
9334 2p ................... 11.64 |===========================
93334 rep ................. 11.53 |===========================
SVT-AV1 1.7
Encoder Mode: Preset 13 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
g ......................... 203.01 |===========================================
h ......................... 197.03 |==========================================
2 x AMD EPYC 9334 32-Core . 127.14 |===========================
9334 2p ................... 155.11 |=================================
93334 rep ................. 152.58 |================================
SPECFEM3D 4.0
Model: Tomographic Model
Seconds < Lower Is Better
d ......................... 14.877710346 |=====================================
2 x AMD EPYC 9334 32-Core . 9.487145781 |========================
9334 2p ................... 9.419086635 |=======================
93334 rep ................. 9.374563082 |=======================
Neural Magic DeepSparse 1.5
Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 20.71 |============================================
g ......................... 15.02 |================================
h ......................... 14.56 |===============================
2 x AMD EPYC 9334 32-Core . 13.31 |============================
9334 2p ................... 13.30 |============================
93334 rep ................. 13.38 |============================
Neural Magic DeepSparse 1.5
Model: NLP Question Answering, BERT base uncased SQuaD 12layer Pruned90 - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 48.25 |============================
g ......................... 66.52 |=======================================
h ......................... 68.61 |========================================
2 x AMD EPYC 9334 32-Core . 75.10 |============================================
9334 2p ................... 75.15 |============================================
93334 rep ................. 74.68 |============================================
Remhos 1.0
Test: Sample Remap Example
Seconds < Lower Is Better
d ......................... 20.90 |============================================
2 x AMD EPYC 9334 32-Core . 13.51 |============================
9334 2p ................... 13.96 |=============================
93334 rep ................. 13.83 |=============================
Timed Linux Kernel Compilation 6.1
Build: defconfig
Seconds < Lower Is Better
d ......................... 35.28 |============================================
g ......................... 23.06 |=============================
h ......................... 23.05 |=============================
2 x AMD EPYC 9334 32-Core . 24.74 |===============================
9334 2p ................... 24.67 |===============================
93334 rep ................. 24.78 |===============================
Neural Magic DeepSparse 1.5
Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 7.1102 |==========================================
g ......................... 7.2026 |===========================================
h ......................... 7.1792 |===========================================
2 x AMD EPYC 9334 32-Core . 4.8090 |=============================
9334 2p ................... 4.7159 |============================
93334 rep ................. 4.7957 |=============================
Neural Magic DeepSparse 1.5
Model: ResNet-50, Baseline - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 140.46 |=============================
g ......................... 138.76 |============================
h ......................... 139.21 |============================
2 x AMD EPYC 9334 32-Core . 207.72 |==========================================
9334 2p ................... 211.81 |===========================================
93334 rep ................. 208.31 |==========================================
Neural Magic DeepSparse 1.5
Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 7.1228 |==========================================
g ......................... 7.2111 |===========================================
h ......................... 7.1163 |==========================================
2 x AMD EPYC 9334 32-Core . 4.8028 |=============================
9334 2p ................... 4.7942 |=============================
93334 rep ................. 4.8205 |=============================
Neural Magic DeepSparse 1.5
Model: CV Classification, ResNet-50 ImageNet - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 140.20 |=============================
g ......................... 138.59 |=============================
h ......................... 140.44 |=============================
2 x AMD EPYC 9334 32-Core . 207.99 |===========================================
9334 2p ................... 208.34 |===========================================
93334 rep ................. 207.23 |===========================================
Liquid-DSP 1.6
Threads: 64 - Buffer Length: 256 - Filter Length: 512
samples/s > Higher Is Better
d ......................... 504660000 |===========================
g ......................... 659140000 |===================================
h ......................... 652820000 |===================================
2 x AMD EPYC 9334 32-Core . 754210000 |========================================
9334 2p ................... 755970000 |========================================
93334 rep ................. 753120000 |========================================
Laghos 3.1
Test: Sedov Blast Wave, ube_922_hex.mesh
Major Kernels Total Rate > Higher Is Better
d ......................... 263.85 |=============================
2 x AMD EPYC 9334 32-Core . 389.86 |===========================================
9334 2p ................... 388.87 |===========================================
93334 rep ................. 387.15 |===========================================
Stress-NG 0.16.04
Test: Matrix 3D Math
Bogo Ops/s > Higher Is Better
d ......................... 10869.49 |============================
g ......................... 14658.75 |======================================
h ......................... 15802.86 |=========================================
2 x AMD EPYC 9334 32-Core . 14005.29 |====================================
9334 2p ................... 14022.50 |====================================
93334 rep ................. 14009.23 |====================================
Apache Cassandra 4.1.3
Test: Writes
Op/s > Higher Is Better
d ......................... 230402 |=====================================
g ......................... 185134 |==============================
h ......................... 190955 |===============================
2 x AMD EPYC 9334 32-Core . 255016 |=========================================
9334 2p ................... 258638 |=========================================
93334 rep ................. 268984 |===========================================
SVT-AV1 1.7
Encoder Mode: Preset 13 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
g ......................... 599.57 |===========================================
h ......................... 600.76 |===========================================
2 x AMD EPYC 9334 32-Core . 416.17 |==============================
9334 2p ................... 560.79 |========================================
93334 rep ................. 517.44 |=====================================
nekRS 23.0
Input: Kershaw
flops/rank > Higher Is Better
d ......................... 11520900000 |======================================
2 x AMD EPYC 9334 32-Core . 8095630000 |===========================
9334 2p ................... 8147120000 |===========================
93334 rep ................. 8130950000 |===========================
NCNN 20230517
Target: CPU - Model: yolov4-tiny
ms < Lower Is Better
d ......................... 35.75 |================================
d ......................... 40.34 |====================================
g ......................... 48.08 |===========================================
h ......................... 48.82 |============================================
2 x AMD EPYC 9334 32-Core . 34.63 |===============================
9334 2p ................... 40.19 |====================================
93334 rep ................. 43.59 |=======================================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2 - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 19.64 |============================================
g ......................... 14.42 |================================
h ......................... 13.97 |===============================
2 x AMD EPYC 9334 32-Core . 15.02 |==================================
9334 2p ................... 15.09 |==================================
93334 rep ................. 15.08 |==================================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2 - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 50.91 |===============================
g ......................... 69.33 |===========================================
h ......................... 71.57 |============================================
2 x AMD EPYC 9334 32-Core . 66.54 |=========================================
9334 2p ................... 66.25 |=========================================
93334 rep ................. 66.27 |=========================================
nekRS 23.0
Input: TurboPipe Periodic
flops/rank > Higher Is Better
d ......................... 7470490000 |=======================================
2 x AMD EPYC 9334 32-Core . 5441100000 |============================
9334 2p ................... 5426030000 |============================
93334 rep ................. 5437240000 |============================
Stress-NG 0.16.04
Test: System V Message Passing
Bogo Ops/s > Higher Is Better
d ......................... 6851573.20 |=============================
g ......................... 7729966.17 |=================================
h ......................... 7761412.61 |=================================
2 x AMD EPYC 9334 32-Core . 9268236.40 |=======================================
9334 2p ................... 9258853.69 |=======================================
93334 rep ................. 9251814.09 |=======================================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream
ms/batch < Lower Is Better
d ......................... 10.0813 |==========================================
g ......................... 8.0302 |=================================
h ......................... 7.6734 |================================
2 x AMD EPYC 9334 32-Core . 8.0098 |=================================
9334 2p ................... 8.0136 |=================================
93334 rep ................. 7.9992 |=================================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, DistilBERT mnli - Scenario: Synchronous Single-Stream
items/sec > Higher Is Better
d ......................... 99.12 |=================================
g ......................... 124.41 |=========================================
h ......................... 130.19 |===========================================
2 x AMD EPYC 9334 32-Core . 124.72 |=========================================
9334 2p ................... 124.67 |=========================================
93334 rep ................. 124.90 |=========================================
Neural Magic DeepSparse 1.5
Model: CV Segmentation, 90% Pruned YOLACT Pruned - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 436.97 |=================================
g ......................... 573.36 |===========================================
h ......................... 567.86 |===========================================
2 x AMD EPYC 9334 32-Core . 448.73 |==================================
9334 2p ................... 451.05 |==================================
93334 rep ................. 448.89 |==================================
Liquid-DSP 1.6
Threads: 2 - Buffer Length: 256 - Filter Length: 32
samples/s > Higher Is Better
d ......................... 68834000 |=======================================
g ......................... 55708000 |===============================
h ......................... 55923000 |================================
2 x AMD EPYC 9334 32-Core . 72597000 |=========================================
9334 2p ................... 72480000 |=========================================
93334 rep ................. 72532000 |=========================================
OSPRay 2.12
Benchmark: particle_volume/pathtracer/real_time
Items Per Second > Higher Is Better
d ......................... 176.61 |====================================
g ......................... 161.17 |=================================
h ......................... 160.87 |=================================
2 x AMD EPYC 9334 32-Core . 204.93 |==========================================
9334 2p ................... 209.61 |===========================================
93334 rep ................. 208.86 |===========================================
SVT-AV1 1.6
Encoder Mode: Preset 4 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
d ......................... 14.97 |=======================================
g ......................... 13.58 |====================================
h ......................... 12.91 |==================================
2 x AMD EPYC 9334 32-Core . 16.20 |===========================================
9334 2p ................... 16.30 |===========================================
93334 rep ................. 16.68 |============================================
Liquid-DSP 1.6
Threads: 2 - Buffer Length: 256 - Filter Length: 57
samples/s > Higher Is Better
d ......................... 105630000 |======================================
g ......................... 88109000 |================================
h ......................... 88042000 |================================
2 x AMD EPYC 9334 32-Core . 111320000 |========================================
9334 2p ................... 111330000 |========================================
93334 rep ................. 86248000 |===============================
Stress-NG 0.16.04
Test: Mixed Scheduler
Bogo Ops/s > Higher Is Better
d ......................... 35640.36 |=======================================
g ......................... 29932.20 |================================
h ......................... 29481.45 |================================
2 x AMD EPYC 9334 32-Core . 37887.14 |=========================================
9334 2p ................... 37702.22 |=========================================
93334 rep ................. 37534.00 |=========================================
Stress-NG 0.16.04
Test: Pthread
Bogo Ops/s > Higher Is Better
d ......................... 70043.78 |=========================================
g ......................... 54906.43 |================================
h ......................... 55328.71 |================================
2 x AMD EPYC 9334 32-Core . 70405.47 |=========================================
9334 2p ................... 70476.97 |=========================================
Liquid-DSP 1.6
Threads: 4 - Buffer Length: 256 - Filter Length: 512
samples/s > Higher Is Better
d ......................... 49884000 |=======================================
g ......................... 41641000 |=================================
h ......................... 41052000 |================================
2 x AMD EPYC 9334 32-Core . 52315000 |=========================================
9334 2p ................... 52469000 |=========================================
93334 rep ................. 51821000 |========================================
Liquid-DSP 1.6
Threads: 1 - Buffer Length: 256 - Filter Length: 32
samples/s > Higher Is Better
d ......................... 35195000 |=======================================
g ......................... 29219000 |================================
h ......................... 29200000 |================================
2 x AMD EPYC 9334 32-Core . 37170000 |=========================================
9334 2p ................... 37126000 |=========================================
93334 rep ................. 37183000 |=========================================
Liquid-DSP 1.6
Threads: 8 - Buffer Length: 256 - Filter Length: 32
samples/s > Higher Is Better
d ......................... 271730000 |=====================================
g ......................... 228520000 |===============================
h ......................... 228090000 |===============================
2 x AMD EPYC 9334 32-Core . 288730000 |========================================
9334 2p ................... 288430000 |========================================
93334 rep ................. 290320000 |========================================
Liquid-DSP 1.6
Threads: 32 - Buffer Length: 256 - Filter Length: 32
samples/s > Higher Is Better
d ......................... 1073800000 |====================================
g ......................... 914570000 |===============================
h ......................... 911870000 |===============================
2 x AMD EPYC 9334 32-Core . 1158100000 |=======================================
9334 2p ................... 1158700000 |=======================================
93334 rep ................. 1159500000 |=======================================
Liquid-DSP 1.6
Threads: 8 - Buffer Length: 256 - Filter Length: 512
samples/s > Higher Is Better
d ......................... 96449000 |=====================================
g ......................... 83090000 |================================
h ......................... 82749000 |===============================
2 x AMD EPYC 9334 32-Core . 104310000 |========================================
9334 2p ................... 105170000 |========================================
93334 rep ................. 102270000 |=======================================
Liquid-DSP 1.6
Threads: 16 - Buffer Length: 256 - Filter Length: 32
samples/s > Higher Is Better
d ......................... 549390000 |======================================
g ......................... 457820000 |================================
h ......................... 457420000 |===============================
2 x AMD EPYC 9334 32-Core . 575510000 |========================================
9334 2p ................... 579510000 |========================================
93334 rep ................. 580920000 |========================================
Liquid-DSP 1.6
Threads: 16 - Buffer Length: 256 - Filter Length: 512
samples/s > Higher Is Better
d ......................... 195510000 |=====================================
g ......................... 165180000 |================================
h ......................... 165390000 |================================
2 x AMD EPYC 9334 32-Core . 208630000 |========================================
9334 2p ................... 207810000 |========================================
93334 rep ................. 209370000 |========================================
Liquid-DSP 1.6
Threads: 4 - Buffer Length: 256 - Filter Length: 32
samples/s > Higher Is Better
d ......................... 136500000 |======================================
g ......................... 114550000 |================================
h ......................... 114390000 |================================
2 x AMD EPYC 9334 32-Core . 144210000 |========================================
9334 2p ................... 144800000 |========================================
93334 rep ................. 143830000 |========================================
Liquid-DSP 1.6
Threads: 1 - Buffer Length: 256 - Filter Length: 57
samples/s > Higher Is Better
d ......................... 52765000 |=======================================
g ......................... 44187000 |================================
h ......................... 44183000 |================================
2 x AMD EPYC 9334 32-Core . 54155000 |========================================
9334 2p ................... 55624000 |=========================================
93334 rep ................. 55882000 |=========================================
Liquid-DSP 1.6
Threads: 1 - Buffer Length: 256 - Filter Length: 512
samples/s > Higher Is Better
d ......................... 12652000 |=======================================
g ......................... 10576000 |================================
h ......................... 10567000 |================================
2 x AMD EPYC 9334 32-Core . 13207000 |=========================================
9334 2p ................... 13357000 |=========================================
93334 rep ................. 13320000 |=========================================
Liquid-DSP 1.6
Threads: 64 - Buffer Length: 256 - Filter Length: 57
samples/s > Higher Is Better
d ......................... 1778600000 |===============================
g ......................... 2076600000 |====================================
h ......................... 2072000000 |====================================
2 x AMD EPYC 9334 32-Core . 2247200000 |=======================================
9334 2p ................... 2231600000 |=======================================
93334 rep ................. 2194300000 |======================================
Liquid-DSP 1.6
Threads: 2 - Buffer Length: 256 - Filter Length: 512
samples/s > Higher Is Better
d ......................... 24900000 |=======================================
g ......................... 20660000 |================================
h ......................... 20837000 |=================================
2 x AMD EPYC 9334 32-Core . 25706000 |========================================
9334 2p ................... 25885000 |=========================================
93334 rep ................. 26081000 |=========================================
Liquid-DSP 1.6
Threads: 32 - Buffer Length: 256 - Filter Length: 512
samples/s > Higher Is Better
d ......................... 387460000 |=====================================
g ......................... 331240000 |================================
h ......................... 329210000 |================================
2 x AMD EPYC 9334 32-Core . 413970000 |========================================
9334 2p ................... 414060000 |========================================
93334 rep ................. 415200000 |========================================
Stress-NG 0.16.04
Test: Mutex
Bogo Ops/s > Higher Is Better
d ......................... 438280.16 |======================================
g ......................... 364035.57 |================================
h ......................... 363892.10 |================================
2 x AMD EPYC 9334 32-Core . 457807.79 |========================================
9334 2p ................... 456942.16 |========================================
93334 rep ................. 457061.39 |========================================
Liquid-DSP 1.6
Threads: 64 - Buffer Length: 256 - Filter Length: 32
samples/s > Higher Is Better
d ......................... 2045600000 |===================================
g ......................... 1827600000 |===============================
h ......................... 1822100000 |===============================
2 x AMD EPYC 9334 32-Core . 2284700000 |=======================================
9334 2p ................... 2291300000 |=======================================
93334 rep ................. 2291700000 |=======================================
SVT-AV1 1.7
Encoder Mode: Preset 4 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
g ......................... 9.831 |==================================
h ......................... 9.901 |==================================
2 x AMD EPYC 9334 32-Core . 12.060 |==========================================
9334 2p ................... 12.358 |===========================================
93334 rep ................. 12.250 |===========================================
Stress-NG 0.16.04
Test: Atomic
Bogo Ops/s > Higher Is Better
d ......................... 236.99 |========================================
g ......................... 204.44 |===================================
h ......................... 203.49 |==================================
2 x AMD EPYC 9334 32-Core . 253.60 |===========================================
9334 2p ................... 254.16 |===========================================
93334 rep ................. 254.23 |===========================================
Liquid-DSP 1.6
Threads: 8 - Buffer Length: 256 - Filter Length: 57
samples/s > Higher Is Better
d ......................... 331200000 |======================================
g ......................... 279840000 |================================
h ......................... 277250000 |================================
2 x AMD EPYC 9334 32-Core . 346260000 |========================================
9334 2p ................... 336320000 |=======================================
93334 rep ................. 329180000 |======================================
Neural Magic DeepSparse 1.5
Model: NLP Token Classification, BERT base uncased conll2003 - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 794.04 |===================================
g ......................... 969.24 |===========================================
h ......................... 952.87 |==========================================
2 x AMD EPYC 9334 32-Core . 776.91 |==================================
9334 2p ................... 779.18 |===================================
93334 rep ................. 779.83 |===================================
Liquid-DSP 1.6
Threads: 16 - Buffer Length: 256 - Filter Length: 57
samples/s > Higher Is Better
d ......................... 650410000 |=======================================
g ......................... 542300000 |================================
h ......................... 546690000 |================================
2 x AMD EPYC 9334 32-Core . 658580000 |=======================================
9334 2p ................... 674330000 |========================================
93334 rep ................. 656510000 |=======================================
Neural Magic DeepSparse 1.5
Model: NLP Document Classification, oBERT base uncased on IMDB - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 795.85 |===================================
g ......................... 969.88 |===========================================
h ......................... 954.94 |==========================================
2 x AMD EPYC 9334 32-Core . 780.60 |===================================
9334 2p ................... 781.54 |===================================
93334 rep ................. 782.00 |===================================
SVT-AV1 1.6
Encoder Mode: Preset 12 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
d ......................... 199.52 |===========================================
g ......................... 189.16 |=========================================
h ......................... 194.86 |==========================================
2 x AMD EPYC 9334 32-Core . 163.44 |===================================
9334 2p ................... 160.63 |===================================
93334 rep ................. 161.05 |===================================
Liquid-DSP 1.6
Threads: 4 - Buffer Length: 256 - Filter Length: 57
samples/s > Higher Is Better
d ......................... 174280000 |=====================================
g ......................... 151010000 |================================
h ......................... 155360000 |=================================
2 x AMD EPYC 9334 32-Core . 185390000 |========================================
9334 2p ................... 185360000 |========================================
93334 rep ................. 186670000 |========================================
Liquid-DSP 1.6
Threads: 32 - Buffer Length: 256 - Filter Length: 57
samples/s > Higher Is Better
d ......................... 1234500000 |=====================================
g ......................... 1057300000 |================================
h ......................... 1065000000 |================================
2 x AMD EPYC 9334 32-Core . 1293000000 |=======================================
9334 2p ................... 1275000000 |======================================
93334 rep ................. 1302400000 |=======================================
SVT-AV1 1.6
Encoder Mode: Preset 4 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
d ......................... 5.053 |=======================================
g ......................... 4.615 |====================================
h ......................... 4.593 |====================================
2 x AMD EPYC 9334 32-Core . 5.599 |============================================
9334 2p ................... 5.643 |============================================
93334 rep ................. 5.650 |============================================
SVT-AV1 1.6
Encoder Mode: Preset 13 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
d ......................... 195.83 |===========================================
g ......................... 191.67 |==========================================
h ......................... 177.36 |=======================================
2 x AMD EPYC 9334 32-Core . 162.86 |====================================
9334 2p ................... 163.43 |====================================
93334 rep ................. 160.08 |===================================
SVT-AV1 1.7
Encoder Mode: Preset 4 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
g ......................... 4.124 |====================================
h ......................... 4.204 |=====================================
2 x AMD EPYC 9334 32-Core . 4.914 |===========================================
9334 2p ................... 5.012 |============================================
93334 rep ................. 5.030 |============================================
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 140.58 |====================================
g ......................... 167.42 |===========================================
h ......................... 166.76 |===========================================
2 x AMD EPYC 9334 32-Core . 137.91 |===================================
9334 2p ................... 137.48 |===================================
93334 rep ................. 137.97 |===================================
Neural Magic DeepSparse 1.5
Model: CV Detection, YOLOv5s COCO, Sparse INT8 - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 139.61 |====================================
g ......................... 165.51 |===========================================
h ......................... 164.66 |===========================================
2 x AMD EPYC 9334 32-Core . 136.60 |===================================
9334 2p ................... 136.46 |===================================
93334 rep ................. 136.81 |====================================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2 - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 186.65 |====================================
g ......................... 224.65 |===========================================
h ......................... 223.85 |===========================================
2 x AMD EPYC 9334 32-Core . 188.74 |====================================
9334 2p ................... 189.52 |====================================
93334 rep ................. 189.33 |====================================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, DistilBERT mnli - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 94.62 |====================================
g ......................... 113.02 |===========================================
h ......................... 113.20 |===========================================
2 x AMD EPYC 9334 32-Core . 95.74 |====================================
9334 2p ................... 95.54 |====================================
93334 rep ................. 95.60 |====================================
Neural Magic DeepSparse 1.5
Model: NLP Sentiment Analysis, 80% Pruned Quantized BERT Base Uncased - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 44.91 |=====================================
g ......................... 53.12 |============================================
h ......................... 52.58 |============================================
2 x AMD EPYC 9334 32-Core . 44.49 |=====================================
9334 2p ................... 44.41 |=====================================
93334 rep ................. 44.68 |=====================================
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering, Sparse INT8 - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 42.83 |=======================================
g ......................... 48.88 |============================================
h ......................... 48.51 |============================================
2 x AMD EPYC 9334 32-Core . 41.31 |=====================================
9334 2p ................... 41.02 |=====================================
93334 rep ................. 41.10 |=====================================
Neural Magic DeepSparse 1.5
Model: BERT-Large, NLP Question Answering - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 641.61 |====================================
g ......................... 756.62 |===========================================
h ......................... 751.72 |===========================================
2 x AMD EPYC 9334 32-Core . 635.64 |====================================
9334 2p ................... 635.65 |====================================
93334 rep ................. 635.20 |====================================
SVT-AV1 1.7
Encoder Mode: Preset 12 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
g ......................... 179.41 |=========================================
h ......................... 187.68 |===========================================
2 x AMD EPYC 9334 32-Core . 158.71 |====================================
9334 2p ................... 159.26 |====================================
93334 rep ................. 161.67 |=====================================
Neural Magic DeepSparse 1.5
Model: NLP Text Classification, BERT base uncased SST2, Sparse INT8 - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 21.09 |=====================================
g ......................... 24.79 |============================================
h ......................... 24.68 |============================================
2 x AMD EPYC 9334 32-Core . 21.00 |=====================================
9334 2p ................... 21.03 |=====================================
93334 rep ................. 21.00 |=====================================
SVT-AV1 1.6
Encoder Mode: Preset 12 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
d ......................... 419.85 |=====================================
g ......................... 490.15 |===========================================
h ......................... 485.59 |===========================================
2 x AMD EPYC 9334 32-Core . 442.55 |=======================================
9334 2p ................... 442.44 |=======================================
93334 rep ................. 416.15 |=====================================
Neural Magic DeepSparse 1.5
Model: CV Classification, ResNet-50 ImageNet - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 61.81 |======================================
g ......................... 71.82 |============================================
h ......................... 71.54 |============================================
2 x AMD EPYC 9334 32-Core . 61.46 |======================================
9334 2p ................... 61.22 |======================================
93334 rep ................. 61.49 |======================================
Neural Magic DeepSparse 1.5
Model: ResNet-50, Baseline - Scenario: Asynchronous Multi-Stream
ms/batch < Lower Is Better
d ......................... 62.10 |======================================
g ......................... 71.76 |============================================
h ......................... 71.58 |============================================
2 x AMD EPYC 9334 32-Core . 61.49 |======================================
9334 2p ................... 61.25 |======================================
93334 rep ................. 61.45 |======================================
SVT-AV1 1.7
Encoder Mode: Preset 12 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
g ......................... 511.80 |===========================================
h ......................... 507.02 |===========================================
2 x AMD EPYC 9334 32-Core . 462.33 |=======================================
9334 2p ................... 458.52 |=======================================
93334 rep ................. 443.32 |=====================================
SVT-AV1 1.6
Encoder Mode: Preset 13 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
d ......................... 545.20 |=======================================
g ......................... 600.23 |===========================================
h ......................... 581.07 |==========================================
2 x AMD EPYC 9334 32-Core . 562.31 |========================================
9334 2p ................... 520.17 |=====================================
93334 rep ................. 593.63 |===========================================
SVT-AV1 1.7
Encoder Mode: Preset 8 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
g ......................... 125.22 |======================================
h ......................... 124.70 |======================================
2 x AMD EPYC 9334 32-Core . 142.19 |===========================================
9334 2p ................... 140.17 |==========================================
93334 rep ................. 137.61 |==========================================
SVT-AV1 1.6
Encoder Mode: Preset 8 - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
d ......................... 136.07 |=======================================
g ......................... 134.88 |=======================================
h ......................... 134.38 |======================================
2 x AMD EPYC 9334 32-Core . 149.11 |===========================================
9334 2p ................... 150.35 |===========================================
93334 rep ................. 148.85 |===========================================
VVenC 1.9
Video Input: Bosphorus 4K - Video Preset: Fast
Frames Per Second > Higher Is Better
d ......................... 6.768 |========================================
2 x AMD EPYC 9334 32-Core . 7.397 |============================================
9334 2p ................... 7.447 |============================================
93334 rep ................. 7.358 |===========================================
VVenC 1.9
Video Input: Bosphorus 1080p - Video Preset: Fast
Frames Per Second > Higher Is Better
d ......................... 19.17 |========================================
2 x AMD EPYC 9334 32-Core . 20.73 |===========================================
9334 2p ................... 20.88 |============================================
93334 rep ................. 21.01 |============================================
Stress-NG 0.16.04
Test: Cloning
Bogo Ops/s > Higher Is Better
d ......................... 1170.31 |=======================================
g ......................... 1139.87 |======================================
h ......................... 1209.53 |=========================================
2 x AMD EPYC 9334 32-Core . 1245.21 |==========================================
9334 2p ................... 1190.00 |========================================
93334 rep ................. 1188.64 |========================================
VVenC 1.9
Video Input: Bosphorus 4K - Video Preset: Faster
Frames Per Second > Higher Is Better
d ......................... 12.68 |========================================
2 x AMD EPYC 9334 32-Core . 13.73 |============================================
9334 2p ................... 13.68 |============================================
93334 rep ................. 13.80 |============================================
VVenC 1.9
Video Input: Bosphorus 1080p - Video Preset: Faster
Frames Per Second > Higher Is Better
d ......................... 33.57 |=========================================
2 x AMD EPYC 9334 32-Core . 36.03 |============================================
9334 2p ................... 35.58 |===========================================
93334 rep ................. 34.76 |==========================================
SVT-AV1 1.7
Encoder Mode: Preset 8 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
g ......................... 92.64 |==========================================
h ......................... 94.83 |===========================================
2 x AMD EPYC 9334 32-Core . 89.91 |=========================================
9334 2p ................... 96.11 |============================================
93334 rep ................. 94.73 |===========================================
SVT-AV1 1.6
Encoder Mode: Preset 8 - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
d ......................... 72.35 |=========================================
g ......................... 75.15 |===========================================
h ......................... 73.25 |==========================================
2 x AMD EPYC 9334 32-Core . 77.22 |============================================
9334 2p ................... 76.50 |============================================
93334 rep ................. 77.19 |============================================
Kripke 1.2.6
Throughput FoM > Higher Is Better
d ......................... 372072000 |======================================
2 x AMD EPYC 9334 32-Core . 391567400 |========================================
9334 2p ................... 393276500 |========================================
93334 rep ................. 375715500 |======================================
Laghos 3.1
Test: Triple Point Problem
Major Kernels Total Rate > Higher Is Better
d ......................... 195.40 |==========================================
2 x AMD EPYC 9334 32-Core . 198.94 |===========================================
9334 2p ................... 199.20 |===========================================
93334 rep ................. 197.19 |===========================================
Dragonflydb 1.6.2
Clients Per Thread: 50 - Set To Get Ratio: 1:100
Ops/sec > Higher Is Better
d ......................... 17583489.38 |======================================
Dragonflydb 1.6.2
Clients Per Thread: 50 - Set To Get Ratio: 1:10
Ops/sec > Higher Is Better
d ......................... 15253621.53 |======================================
AOM AV1 3.7
Encoder Mode: Speed 0 Two-Pass - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 0 Two-Pass - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 4 Two-Pass - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 4 Two-Pass - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 6 Two-Pass - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 6 Two-Pass - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 6 Realtime - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 6 Realtime - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 8 Realtime - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 8 Realtime - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 9 Realtime - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 9 Realtime - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 10 Realtime - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 10 Realtime - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 11 Realtime - Input: Bosphorus 4K
Frames Per Second > Higher Is Better
AOM AV1 3.7
Encoder Mode: Speed 11 Realtime - Input: Bosphorus 1080p
Frames Per Second > Higher Is Better
Stress-NG 0.16.04
Test: IO_uring
Dragonflydb 1.6.2
Clients Per Thread: 60 - Set To Get Ratio: 1:100
Dragonflydb 1.6.2
Clients Per Thread: 60 - Set To Get Ratio: 1:10
Timed Linux Kernel Compilation 6.1
Build: allmodconfig
Seconds < Lower Is Better