Microsoft Store install
Use the Store listing for managed package delivery and updates without handling the executable manually.
Download VRAM Check. Verify it before you run it.
On Windows, install the verified package from Microsoft Store. Direct binaries remain available with a published checksum, while macOS and Linux keep their own platform-specific release paths.
Local benchmarkInference runs on the machine you are testing.
Explicit publishingYou choose whether a result is submitted publicly.
Trusted distributionWindows installs through Microsoft Store; direct artifacts remain traceable.
Choose the build that matches the machine.
Only active, verified binary paths are labeled as official downloads. Everything else stays visibly marked as a source build.
Prefer the direct Windows file? Verify it in under a minute.
Microsoft Store handles package trust automatically. These steps protect the optional direct-download path.
- 1Download both direct-release files
Keep the Windows executable and its `.sha256` file in the same folder.
- 2Calculate the local hash
Use the PowerShell command below against the downloaded executable.
- 3Compare all 64 characters
Run the file only when your result matches the published SHA256 exactly.
PowerShellVerification command
Get-FileHash .\vramcheck-windows.exe -Algorithm SHA256Expected51b208df0c752ca26b3041b69d46e42576950db8691b0d06b1cd6d6bc018b1d4Choose the trust path that fits your workflow.
Microsoft Store is recommended on Windows. A directly downloaded executable can still trigger SmartScreen, so verify its SHA256 before launch.
Direct executable
A matching SHA256 proves the file has the same cryptographic identity as the artifact VRAM Check published. If it differs, do not run it.
Building from source
macOS remains a source build until notarization is available. Linux switches to an official binary only after its independent hash and provenance gate passes.
Inspect the method before you compare the result.
The benchmark policy explains verification, ranking eligibility, and the difference between measured performance and estimated fit.