From Whiskey Lake To Meteor Lake: The Intel CPU Linux Performance Evolution

Written by Michael Larabel in Processors on 29 December 2023 at 07:30 AM EST. Page 6 of 12. 21 Comments.
TensorFlow benchmark with settings of Device: CPU, Batch Size: 16, Model: ResNet-50. Core Ultra 7 155H was the fastest.
TensorFlow benchmark with settings of Device: CPU, Batch Size: 16, Model: ResNet-50. Core Ultra 7 155H was the fastest.
TensorFlow benchmark with settings of Device: CPU, Batch Size: 16, Model: ResNet-50. Core Ultra 7 155H was the fastest.

The Core Ultra 7 155H was delivering the best performance for TensorFlow though surprisingly in this particular test the Core i7 1065G7 Ice Lake laptop delivered the best performance-per-Watt.

TensorFlow Lite benchmark with settings of Model: Mobilenet Float. Core Ultra 7 155H was the fastest.
TensorFlow Lite benchmark with settings of Model: Mobilenet Quant. Core Ultra 7 155H was the fastest.
TensorFlow Lite benchmark with settings of Model: Inception V4. Core Ultra 7 155H was the fastest.

Some nice gains in TensorFlow Lite.

Neural Magic DeepSparse benchmark with settings of Model: CV Classification, ResNet-50 ImageNet, Scenario: Asynchronous Multi-Stream. Core Ultra 7 155H was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: CV Classification, ResNet-50 ImageNet, Scenario: Asynchronous Multi-Stream. Core i7 1185G7 was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: CV Segmentation, 90% Pruned YOLACT Pruned, Scenario: Asynchronous Multi-Stream. Core Ultra 7 155H was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: CV Segmentation, 90% Pruned YOLACT Pruned, Scenario: Asynchronous Multi-Stream. Core i7 1185G7 was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: ResNet-50, Sparse INT8, Scenario: Asynchronous Multi-Stream. Core Ultra 7 155H was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: ResNet-50, Sparse INT8, Scenario: Asynchronous Multi-Stream. Core i7 1185G7 was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: BERT-Large, NLP Question Answering, Scenario: Asynchronous Multi-Stream. Core Ultra 7 155H was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: BERT-Large, NLP Question Answering, Scenario: Asynchronous Multi-Stream. Core i7 1185G7 was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: BERT-Large, NLP Question Answering, Sparse INT8, Scenario: Asynchronous Multi-Stream. Core Ultra 7 155H was the fastest.
Neural Magic DeepSparse benchmark with settings of Model: BERT-Large, NLP Question Answering, Sparse INT8, Scenario: Asynchronous Multi-Stream. Core i7 1185G7 was the fastest.

For the DeepSparse AI software the throughput was looking very good for Meteor Lake but for some of the models the latency with Meteor Lake was much higher than with prior generations... Potentially due to some P vs. E core pains.

PyTorch benchmark with settings of Device: CPU, Batch Size: 1, Model: ResNet-50. Core Ultra 7 155H was the fastest.
PyTorch benchmark with settings of Device: CPU, Batch Size: 1, Model: ResNet-50. Core Ultra 7 155H was the fastest.

PyTorch was another benchmark where the AVX-512-enabled Ice Lake laptop was showing the best performance per Watt.


Related Articles
Ampere Computing 2024 Roadmap Update: 256 Core 3nm CPU In 2025
Benchmarking The First RISC-V Cloud Server: Scaleway EM-RV1 Performance
AMD Announces Ryzen PRO 8840 & PRO 8000G Series CPUs
Intel Announces Gaudi 3 AI Accelerator, Intel Xeon 6 Brand
AMD Zen 4 vs. Zen 4C Performance, Zen 4C Core Scaling With Ryzen 5 8500G
Intel Core i3 14100 / i5 14500 vs. AMD Ryzen 5 8500G / 8600G In 500+ Benchmarks