Raising the Bar: WEKA and HPE Achieve Unmatched SPECstorage Performance
![](https://www.weka.io/wp-content/uploads/files/2025/02/blog-og-unmatched-specstorage-performance-1200x675-1.jpg)
In today’s data-driven world, performance claims are easy to make but difficult to prove. That’s why industry-standard benchmarks are crucial, providing an objective, verifiable way to measure storage solutions under real-world conditions. Benchmarks like SPECstorage® Solution 2020 don’t just test theoretical peak speeds – they evaluate how well a system handles the most demanding workloads, from AI and genomics to software development and video analytics.
Breaking records in these benchmarks isn’t just about bragging rights — it’s a testament to the ability of a storage solution to scale, perform, sustain efficiency in production environments, and handle multiple workloads. Modern data-intensive workloads, including those involving AI and ML, demand more than just powerful GPUs, ultra-fast networks and high-performance storage — they require a seamless end-to-end pipeline that ensures data flows efficiently to and from processing units. Organizations rely on industry standard benchmark results to make informed decisions, ensuring they invest in technology that delivers on its promises and maximizes return on their infrastructure investments.
#1 Across the Board
WEKA on the industry-leading HPE Alletra Storage Server 4110, powered by Intel® Xeon® Processors, set new records on January 28, 2025, securing the No. 1 ranking across all five SPECstorage® Solution 2020 benchmark workloads: AI_IMAGE, EDA_BLENDED, GENOMICS, SWBUILD, and VDA.
Our combined solution not only raised the bar by setting new records for jobs and streams across all workloads – it also delivered significantly lower latency – in some cases up to 6.5x lower than previous records. Latency is key to storage efficiency, directly affecting how quickly applications access and process data. These improvements translate to faster AI training, reduced semiconductor simulation delays, quicker genomic analysis, and more responsive video analytics. By minimizing response times, WEKA on HPE ensures workloads run at peak efficiency, accelerating insights and time to solution, and eliminating bottlenecks that traditionally slow high-performance computing and leave GPUs underutilized.
The results showcase the combined power of WEKA’s data platform and HPE storage servers in handling the most demanding, data-intensive workloads. The records were achieved using a single, consistent configuration across the five SPECstorage® benchmarks without the need for workload-specific tuning. WEKA’s high-performance software running on HPE Alletra Storage Server 4110 ensures peak performance out of the box, enabling organizations to run artificial intelligence (AI), software development, EDA, genomics, and video analytics workloads on the same infrastructure—simplifying operations while maximizing scalability and efficiency.
Why SPECstorage® Matters
The SPECstorage® benchmark suite is the gold standard for evaluating storage performance across different workloads. These benchmarks simulate real-world applications to measure throughput, latency, and scalability, providing organizations with a reliable metric for comparing storage solutions. Breaking SPECstorage® records is more than just a technical feat; it’s a testament to a solution’s ability to meet the most demanding enterprise workloads with unmatched efficiency and speed. By surpassing all previous benchmarks, WEKA and HPE not only showcase advanced storage capabilities, but also the fastest file performance, providing organizations with a proven solution to tackle the most data-intensive workloads with confidence.
Summary of Results
![](https://www.weka.io/wp-content/uploads/files/2025/02/image.png)
Source: SPECstorage® Solution 2020 Solution results (January 28, 2025). spec.org/storage2020/results/.
The Software Behind the Records
WEKA Data Platform
The WEKA® Data Platform is a software solution designed to modernize enterprise data stacks for AI and other performance-intensive workloads with high-performance, massively scalable, distributed data storage. Its AI-native, data pipeline-oriented architecture, eliminates traditional storage bottlenecks and delivers exceptional performance at scale, accelerating GPUs, AI, ML, and HPC workloads while improving efficiency and reducing operational complexity.
With zero-tuning architecture, WEKA delivers high-speed, low-latency performance out of the box, removing the need for manual optimizations or workload-specific tuning. WEKA’s software is compatible with AMD or Intel x86-based servers, NVMe SSDs, and leading hyperscaler clouds, providing seamless data portability and flexibility in deployment. It supports a wide range of use cases, including AI, software development, electronic design automation (EDA), genomics, video analytics, and financial modeling, helping organizations achieve faster insights, better resource efficiency, and reduced operational complexity.
The Industry Leading Storage Server Behind the Records
HPE Alletra Storage Server 4110
The HPE Alletra Storage Server 4110 is purpose-built to run the most performance demanding, data storage-intensive workloads, empowering customers to drive their data-driven initiatives to success. It delivers exceptional performance with an optimal balance of compute, memory, network bandwidth and drive performance– offering greater data throughput, processing, and capacity in a high-density 1U form factor that deploys seamlessly.
Designed with zero-trust security at its foundation – from silicon to software, factory to cloud – the HPE Alletra Storage Server 4110 is uncompromising at its core and provides protection to customers, organizations, and data against increasingly sophisticated and dangerous threats. Whether purchased outright or consumed as-a-service through HPE GreenLake Flex Solutions, IT administrators can easily monitor and operate their data infrastructure with intuitive fleet management SaaS and rich REST APIs that automate lifecycle management and simplify firmware updates.
- Architecture: 12 x HPE Alletra Storage Server 4110 nodes
- Specifications (per node):
- 1U ultra-dense form factor
- 16 x 3.84TB PCIe Gen5 NVMe E3.S SSDs
- 2 x 5th Gen Intel® Xeon® Gold 6548N Processors
- 512GB DDR5 memory at 5,600 MT/s
- 2 x NVIDIA® CX-7 400Gb 1-port NDR adapters
WEKA on HPE: Key Points
WEKA delivers the fastest file performance through our powerful partnership with HPE. We bring together WEKA’s high-performance data platform software with HPE’s high-performance data storage servers in our joint solution – HPE Solutions with WEKA – available from HPE and HPE authorized partners.
Backed by joint engineering, HPE Solutions with WEKA maximizes GPU infrastructure efficiency and returns by eliminating legacy storage bottlenecks. This one-stop shop collaboration has shattered all SPECstorage® 2020 world records across diverse workloads, empowering organizations to push the boundaries of performance.
With transformative scalability, efficiency, and reliability, HPE Solutions with WEKA are redefining what is possible for data-intensive applications, helping businesses unlock new levels of innovation and achieve their most ambitious goals.
SPECstorage® Solution 2020 Results
![](https://www.weka.io/wp-content/uploads/files/2025/02/image-1.png)
Common Use Cases for Each Workload
1. AI_IMAGE
- New Record: 5,000 AI jobs, achieving 2,174,983 operations/sec and 488,945 MB/sec.
- Comparison: This result surpasses the previous best of 2,400 jobs on AWS Cloud (WEKA) from Mar. 13, 2024, with a 2.1x improvement in scalability and over 2x higher throughput, all while maintaining an exceptional ORT of 0.4 ms.
- Use Case: AI workloads such as image recognition, video analysis, and autonomous systems require massive data processing capabilities with ultra-low latency. This record proves the unparalleled suitability of HPE Alletra Storage Servers and WEKA for accelerating AI training and inference pipelines.
2. EDA_BLENDED
- New Record: 17,000 job sets, achieving 7,650,428 operations/sec and 123,442 MB/sec at an ORT of 0.18 ms.
- Comparison: This result is over 2.1x better in scalability and 2.45x faster in throughput compared to NetApp’s record of 8,100 job sets from May 04, 2023, while achieving 6.5x lower latency.
- Use Case: Electronic design automation (EDA) for semiconductor design and verification demands high-performance storage to manage complex simulations and workflows. WEKA and HPE’s ability to deliver this performance with just 15 HPE load generator servers highlights its exceptional cost-efficiency for large-scale EDA environments.
3. GENOMICS
- New Record: 5,600 jobs, achieving 5,403,070 operations/sec and 458,896 MB/sec with an ORT of 0.57 ms.
- Comparison: This result outpaces WEKA’s AWS Cloud performance of 2,200 jobs from Mar. 13, 2024, delivering 2.5x better scalability and throughput while maintaining lower latency.
- Use Case: Genomics workloads, such as genome sequencing and analysis, require massive I/O throughput and low latency to process terabytes of data. The results demonstrate WEKA and HPE’s ability to meet the demands of life sciences and healthcare organizations with cutting-edge solutions.
4. SWBUILD
- New Record: 6,200 builds, achieving 3,097,860 operations/sec and 25,203 MB/sec at an ORT of 0.86 ms.
- Comparison: This result surpasses NetApp’s record of 6,120 builds on Nov. 16, 2023, achieving slightly higher scalability but with 1.8x faster response times and higher throughput.
- Use Case: Software build environments, especially for large-scale CI/CD pipelines, require efficient storage for rapid compilation and artifact management. HPE’s solution ensures developers can iterate faster, improving productivity and reducing time-to-market for software products.
5. VDA (Video Analytics)
- New Record: 14,400 streams, achieving 144,098 operations/sec and 66,321 MB/sec with an ORT of 1.29 ms.
- Comparison: This result beats WEKA’s AWS Cloud performance of 12,000 streams from Mar. 13, 2024, offering 1.2x better scalability, 1.2x higher throughput, and over 2.4x lower latency (ORT of 1.29 ms vs. 3.16 ms).
- Use Case: Video analytics workloads, including real-time surveillance and media processing, demand high throughput and low latency to handle thousands of streams simultaneously. HPE’s solution delivers the performance needed to power next-generation video platforms.
Conclusion
WEKA on HPE’s record-breaking performance across all five SPECstorage® workloads proves the ability to push the limits of high-performance, scalable storage. Delivering unmatched speed, efficiency, and reliability, HPE Solutions with WEKA meets the demands of AI, semiconductor design, genomics, software development, and video analytics—all without the need for complex tuning or specialized hardware.
As a leader in high-performance data platforms, WEKA has partnered with HPE to deliver the fastest file storage solution to accelerate and maximize GPU environments, ensuring customers can scale effortlessly, optimize resource efficiency, and accelerate time to insight. This partnership isn’t just about breaking records — it’s about giving organizations the confidence to tackle the most data-intensive challenges with ease.
SPEC® and the benchmark name SPECstorage® are registered trademarks of the Standard Performance Evaluation Corporation. For more information about the SPECstorage® Solution 2020 benchmark, visit https://www.spec.org/storage2020/