Spec Check
Will this machine run Physical Analyzer, or just survive it? Drop in your specs and find out.
Which screenshot
One screenshot
Task Manager › Performance
Open Task Manager (Ctrl+Shift+Esc), click Performance, maximize the window so every disk and GPU shows in the left column, and screenshot it with CPU selected.
That one view shows the CPU, total memory, every disk with its type, and every GPU. A few quick questions fill in the rest.
I don’t keep these images. They’re discarded asap. It’s not even about privacy…I’m just not paying to store your crap.
Your screenshot
Spec reference
What the manual asks for.
| Spec | Minimum | Core | Optimal |
|---|---|---|---|
| Processing | Longer processing times | Standard processing times | Fastest processing times |
| Use | For infrequent extractions where speed isn’t a priority. Opens most extractions in reasonable time; standard enrichments may take longer. | Moderate-volume production. Consistent processing times with standard enrichments. | Frequent large extractions with all enrichment combinations, where speed is critical. |
| Processor | Intel i5, Gen 8 | Intel i7 K variant, Gen 10 | Intel i9 K variant, Gen 12+ |
| Memory | 32 GB (64 GB+ recommended), DDR3, 1600 MHz | 64 GB, DDR4, 2666 MHz | 128 GB+, DDR5, 3200 MHz+ |
| Storage | 1x SATA SSD (500 MB/s). SSD is required; HDD hinders performance significantly. | 2x NVMe (3000 MB/s+): one for OS, PA install and evidence (1 TB+), one for the PGSql database (500 GB+). | 3x PCIe NVMe (5000 MB/s+): OS and PA install (2 TB+), evidence (2 TB+), database (1 TB+). |
| Disk size | All-in-one 1 TB. 50 GB free for installation. Add-ons 512 GB (offline maps). | OS/PA/evidence 1 TB+. Database 500 GB+. | OS/PA 2 TB+. Evidence 2 TB+. Database 1 TB+. |
| OS | Windows 11 | Windows 11 | Windows 11 |
| GPU | Not required but recommended for Media Analytics: 1900+ CUDA cores. Recommended NVIDIA RTX A2000 (8/12 GB). | 3500+ CUDA cores. Recommended NVIDIA RTX A4000 (16 GB). | 3500+ CUDA cores. Recommended NVIDIA RTX A4000 (16 GB). |
- — AMD equivalent processors are supported, but CPU-based media classification may be slightly slower.
- — Put the database on the fastest drive, not the OS drive.
- — Keep enough free space for Temp files, or large extractions may not open.
- — VMs need at least the same resources as physical machines.
- — PA uses the GPU for the Image Analytics service (media classification). Without a supported GPU it uses CPU models via Intel OpenVINO. The installer auto-detects and configures a supported GPU.
- — Speech-to-text across an entire extraction requires a CUDA GPU compatible with the Torch library. Without one, only single-file ad-hoc transcription works. v10.11 release notes: transcription is significantly faster with a GPU.
Source: Inseyets PA User Manual v10.11 (2026), System Requirements
Unofficial tool. Requirements from the Inseyets PA v10.11 user manual. Check current Cellebrite documentation before buying hardware.