If you are an internal Intel user of the Intel GPA product, please do not post on this forum.
Instead, please use the Intel GPA Inside Blue Forum for any questions, issues, or discussion about the Intel GPA product.
If you are having trouble finding the Inside Blue Forum, feel free to message me and I can send you the link.
When using Intel GPA, using the latest version of the graphics drivers ensures that you are getting the most accurate metrics when analyzing and optimizing your game or graphic-intensive application.
I'm using an Alienware laptop and using Graphics performance analyzer 2017 R4 makes some software (outlook 2016, windows 10 search) crash.
I don't have any problem when Graphics Monitor is not started.
Is there a workaround for that ?
Here is my configuration:
I need to debug a regression between 2 versions of our software. Is there a way to open two graphics frame analyzer at the same time to compare the rendering ?
How can i enable this new feature?
Because when i`m using a new UI on DX11 frame - i didn`t see this new icon, and it`s look like this.. without this new shiny flame icon :(
Just don`t have hotspot icon :(can someone tell me why? or it`s cause i`m using a gtx 1070 card? Aaaand i don`t have any geometry in inputs in base pass section of frame O_O
It looks like I need the installer "gpa_17.3_release_314655_x64_standalone.msi" to upgrade to 2017 R4. I deleted the installer after installation and now I cannot download the installer because the available versions on the download page are only 17.4 and 17.2. How can I upgrade or uninstall my current version?
I'd like to report a crash and get a possible fix.
I'm developing a DirectX11 renderer, where the pixel shader needs to write into UAV, in particular RWStructuredBuffer<float2x4>.
The Analyze Application tool helped me really a lot and it works very well for the performance debugging of my shaders.
Would be great if Intel would create a shader debugger like Nvidia for both the Compute and Graphics shaders (Opengl, Vulkan, Directx 11,12)
That way I can debug the actual values and get the shader algorithm correct before moving onwards to performance stuff.