Appendix 1 - GPU Recommendations for Adept Viewer
GPU acceleration plays a key role in the Adept Viewer’s ability to efficiently convert and process complex file types. The following specifications outline the recommended GPUs for the Synergis Adept Viewer when performing file conversion tasks. These guidelines are intended to help IT administrators select appropriate hardware for both on-premises and hosted cloud deployments. The listed GPUs have been validated for stability, performance, and driver support across Windows environments.
Generally our recommendations are based on the graphics rendering APIs available.
Minimum: WebGL 1.0 which is based on OpenGL ES 2.0
Recommended: WebGL 2.0 which is based on OpenGL ES 3.0
This third-party resource may be used to check the capabilities of your workstation or server: https://systemdiagnose.com/tools/webgl-support/.
Synergis Adept Viewer - GPU Specification for file conversion
Use Case: Multi-user server or desktop workload, no real-time rendering, files <300 MB, <1M objects, conversions <3 minutes.
Tier | GPU Model or Class | VRAM (GB) | TDP (W) | Driver Support | Notes (Win11 / Windows Server / Linux) |
|---|---|---|---|---|---|
Minimum | NVIDIA RTX A2000 | 6 | 70 | NVIDIA Enterprise (Win11/Server/Linux) | Entry-level workstation GPU. Stable and efficient for small to mid-size files. |
Minimum | AMD Radeon Pro W6600 | 8 | ~100 | AMD Pro (Win11/Server/Linux) | Good memory headroom for moderate models. Certified for server OS. |
Recommended | NVIDIA RTX A4000 | 16 | 140 | NVIDIA Enterprise (Win11/Server/Linux) | Workstation-grade performance, great for larger files and moderate concurrency. |
Recommended | NVIDIA Tesla T4 | 16 | 70 | NVIDIA Datacenter (Server/Linux only) | Ideal for rack servers. Headless, ECC memory, great for parallel conversion jobs. |
Recommended | AMD Radeon Pro W6800 | 32 | 250 | AMD Pro (Win11/Server/Linux) | High-end GPU for large models or heavy multi-user batch processing. |
Cloud | Azure NC-series (e.g., NC6, NC12) | 16+ | 150–300 | NVIDIA Datacenter | Uses Tesla-class GPUs (e.g., V100, A100). Ideal for headless compute tasks. Avoid NV-series (graphics-focused). |
Cloud | AWS G4dn (NVIDIA T4) | 16 | 70 | NVIDIA Datacenter | G4dn instances with Tesla T4 are well-suited for DWG-to-SCZ conversion. Reliable and cost-effective. |
Platform Notes
Windows 11 Desktop: All listed GPUs are supported with professional drivers.
Windows Server OS: Use only workstation/datacenter GPUs with certified enterprise drivers (Quadro/RTX A, Tesla, Radeon Pro).
Linux Server: Fully supported via NVIDIA or AMD Pro drivers; Tesla T4 is especially suited for headless workloads. Not applicable to Adept WINTER '25.
IT Procurement Notes (for Desktop, Server, or Cloud-Hosted)
GPU Class: Use only workstation- or datacenter-class GPUs (e.g., NVIDIA RTX A-series, Tesla, or AMD Radeon Pro). Avoid consumer GPUs like GeForce GTX.
Graphics Language Support: The deployed system must support OpenGL 4.6 or later, Direct3D 12 or later, and OpenGL ES 3.2 or later via GPU hardware or software drivers. Driver installation and configuration must also enable backward compatibility with older OpenGL 1.0 and WinDirect3D 9. Note: All recommended GPUs inherently support all of these standards.
Driver Support: Ensure drivers are from the NVIDIA Enterprise or AMD Pro channels for compatibility and stability on Windows Server and Linux.
Form Factor: For on-prem servers, verify that the GPU fits a 1U or 2U rack height (Tesla T4 is optimal for 1U).
Cloud Usage: If using AWS or Azure, select VMs with Tesla T4/V100/A100-class GPUs (e.g., AWS G4dn or Azure NC-series).
Power & Cooling: High-performance GPUs may require adequate PSU wattage and airflow in rack deployments (especially A4000 and W6800).