RCKTBX Forensic Workstation
Benchmark Engineering Study

BitMindz was founded in 2018 with a singular mission: to provide law enforcement agencies, military units, government bodies, and corporate enterprises with the most advanced, reliable, and performant forensic computing platforms available. Every RCKTBX system is designed, assembled, and validated in the United States.

The story, however, begins earlier. BitMindz Founder and CEO Manny Kressel was employed at a law enforcement agency conducting computer forensic examinations when he recognized a fundamental problem: the systems were slow, missing newer technologies like solid-state disks, NVMe drives, faster memory, and modern PCIe architectures. The goal became clear — make the systems faster. By 2013, Kressel had introduced what was then the fastest forensic workstation on the market to the forensic community, pioneering RAID NVMe configurations and Rapid Mode SSD acceleration before either approach was mainstream.

The forensic computing environment places uniquely demanding requirements on hardware. Examiners routinely process terabytes of evidentiary data, render thousands of images and video files, execute cryptographic hash operations, and conduct password recovery campaigns — often simultaneously. Any bottleneck in the processing pipeline directly impacts case timelines and investigative outcomes. The RCKTBX series is engineered to eliminate those bottlenecks through a systems-level approach.

All RCKTBX systems are benchmarked using PassMark® PerformanceTest (Version 10.0), developed by PassMark Software Pty Ltd. (Sydney, Australia). PassMark PerformanceTest has served as the de facto standard for PC benchmarking since 1998, with over 600,000 systems benchmarked across more than 1,200 CPU architectures. The software produces composite scores across five categories — CPU, 2D Graphics, 3D Graphics, Memory, and Disk — with global percentile rankings against all submitted results.

Storage throughput is measured with ATTO Disk Benchmark v4.01.0f1, providing granular read/write performance across I/O block sizes from 512 bytes to 64 MB. Password recovery throughput is measured using a commercial OpenCL-accelerated GPU recovery platform.

Testing of PCIe 5.0 NVMe drives across multiple current-generation RCKTBX configurations — including different motherboards and processors — consistently produced sequential read/write speeds between 8,000 MB/s and 11,000 MB/s. PCIe 5.0 doubles the maximum data transfer rate of PCIe 4.0, with a theoretical maximum unidirectional bandwidth of 64 GB/s.

In a recent RCKTBX configuration pairing an Intel 13900KF processor with multiple Samsung 990 Pro NVMe drives and a Samsung SSD for the OS volume, the system delivered benchmark scores that outperformed all prior processor-based configurations we had tested. Speeds across the NVMe drives remained consistent even under concurrent benchmark load, with no bandwidth degradation — a key validation of our multi-volume architecture design.

The data across all tested configurations reveals a consistent pattern: RCKTBX systems achieve PassMark ratings in the 90th to 99th global percentile. This is not coincidental — it reflects a systems-engineering philosophy that addresses performance holistically rather than optimizing components in isolation.

A key insight from over a decade of hardware testing is that the weakest link in any forensic processing pipeline is almost always storage. Rotational HDDs and even conventional 2.5" SATA SSDs represent meaningful throughput constraints given modern CPU and GPU capabilities. The RCKTBX three-tier storage architecture mitigates this through RAM-assisted OS drive acceleration, high-throughput NVMe RAID arrays, and dedicated NVMe scratch volumes — each optimized for its specific role.

The GPU password recovery scaling is particularly noteworthy. A single RTX 3090 delivers ~7.8 million candidates per second. Eight GPUs across two nodes exceed 58 million per second — demonstrating near-linear scaling without inter-GPU communication bottlenecks.

The Rapid Mode results merit specific attention. The underlying Samsung SSD has a SATA 6 Gb/s interface ceiling of approximately 600 MB/s — yet Samsung Magician's RAM cache technique produces effective throughput of 29.35 GB/s. That is a 48× improvement over the raw drive specification, with measurable impact on OS responsiveness during active processing sessions.