As the size and breadth of demanding applications in high-performance fields continue to grow, so do workstation expectations. The HP ZBook 8 G1a Mobile Workstation delivers pro-level performance with next-gen AMD Ryzen^™ AI 300 Series processors. These processors contain CPU, GPU, and neural processing unit (NPU) architecture, which accelerates AI and machine learning (ML) workloads. A characteristic of the system that can play an important role with AI-enhanced workloads is how many trillions of operations per second (TOPS) an NPU is rated for.

The highest NPU TOPS-rated AMD Ryzen^™ AI 300 Series processor available with the HP ZBook 8 G1a Mobile Workstation is the AMD Ryzen^™ AI 9 HX PRO 375 processor, at 55 NPU TOPS. That’s 5 NPU TOPS more than you’d get with Lenovo^® ThinkPad^® P14S Gen 6 and Dell^™ Pro Max 14 mobile workstations, which contain, at the highest level, AMD Ryzen^™ AI 9 HX PRO 370 processors (50 NPU TOPS). That difference showed up when we compared the performance of these three systems on several different AI-related benchmarks, where the HP system came out on top. Read on to learn more.

What and how we tested

We equipped these 14-inch Windows 11 Pro mobile workstations with the best AMD Ryzen^™ AI 300 Series processor available for each model, integrated AMD Radeon^™ 890M graphics, 64 GB of memory, and 1 TB of storage:

The results we report reflect the specific configurations we tested. Any difference in the configurations—as well as screen brightness, network traffic, and software additions—can affect these results. For a deeper dive into our testing parameters and procedures, see the science behind the report.

We compared general productivity, computer vision, and content creation performance using these benchmarks:

• Cadalyst System Benchmark 2015
• Cinebench 2024
• Geekbench 6
• Maxon RedShift Benchmark
• Procyon^® AI Computer Vision Benchmark
• Revit 2024 RFO Benchmark
• SPECworkstation 4.0

Note: The graphs in this report use different scales to keep a consistent size. Please be mindful of each graph’s data range as you compare.

General productivity performance

When you and your team are spending your days tackling time-sensitive tasks, even small performance bumps based on real-world usage can make a difference.

To reflect the experience of business professionals, we ran two CPU-intensive benchmarks. The Cinebench 2024 benchmark renders a 3D scene with Redshift for Cinema 4D software,⁹ and the Geekbench 6 benchmark executes multiple CPU-intensive tasks simultaneously.¹⁰ Strong CPU multi-core performance enables users to run multiple applications at once and allows background processes to run without slowing down everything else. Strong single-core performance may translate to, among other things, a smoother single application user interface.

AI inference performance

Being able to run computer vision applications on device instead of in the cloud is a game-changer for many organizations. Keeping sensitive data local reduces the risk of data breaches and can even help with compliance efforts.

To measure performance with INT8 (8-bit integer) precision—a level of precision that optimizes for inference times—we ran the UL Procyon AI Computer Vision Benchmark.¹¹ The benchmark measures NPU performance using the following AI inference engines:

• DeepLabv3 for semantic image segmentation tasks.¹²
• Inception-v4 for object detection tasks.¹³
• Real-ESGRAN for image upscaling tasks.¹⁴
• ResNet-50 for image classification tasks.¹⁵
• YOLOv3 is used for real-time object detection and quality inspection within visual data.¹⁶

Content creation performance

Managing increasingly complex projects with multiple contributors, resource-intensive applications, and tight deadlines are just a few pain points that the right technology can help alleviate. By reducing scene rendering, model creation, and collateral export times, your creative team has more time to iterate on new ideas, and you can get projects out the door faster. Plus, smoother navigation enables computer-aided design (CAD) and Revit professionals to immerse themselves within complex 2D and 3D models.

Export files in less time

Just as design professionals need to export work for many different use cases and review stages, the Revit 2024 RFO Benchmark exports different file formats. Its CAD use cases include:

• Design Web Format (DWF) for sharing CAD drawings and design data with stakeholders who may not have CAD software.
• Drawing (DWG) for sharing and collaborating on CAD files with other designers.
• Portable Document Format (PDF) for sharing, printing, and archiving documents that are openable and viewable on almost any device.
• Portable Network Graphics (PNG) for web graphics, digital art, and screenshots where image clarity is crucial.

In these tasks, the HP ZBook 8 G1a provided a clear time advantage. Significantly faster exports mean CAD users can get collateral into review faster, complete projects in less time, and handle more work.

Accelerate industry-specific workflows

Finally, we ran the SPECworkstation 4.0 benchmark. This benchmark measures AI accelerator, CPU, graphics, and storage performance from many perspectives.²⁰ The higher the industry vertical score, the faster you would expect that workstation to perform during real-world usage in that area.

Conclusion

Our hands-on testing shows that investing in the HP ZBook 8 G1a Mobile Workstation powered by an AMD Ryzen^™ AI 9 HX PRO 375 processor brings real-world gains. Across a variety of benchmarks, the 14-inch HP ZBook 8 G1a consistently finished ahead of comparable Lenovo ThinkPad P14s Gen 6 and Dell Pro Max 14 mobile workstations powered by AMD Ryzen^™ AI 9 HX PRO 370 processors. And, while some of the wins were small, others were significant. For example, based on the Revit 2024 RFO Benchmark results, you could save between 3 minutes and over half an hour, per export, depending on the file format. When you spend your days tackling time-sensitive tasks, any advantage, especially those based on real-world usage, can make a big difference.