The new 101.4672 Beta drivers do not appear to bring any further improvements or fixes and appear mostly focused on supporting the Starfield game. There are still plenty of issues with Starfield but Intel notes that further fixes are coming in future driver updates.
I get Intel having issues with DX9 ~ DX11 games. These are quite literally the composite of a thousand hacky patches, bug workarounds, engine quirks, and a mix of drivers being developed around game issues and games being developed around driver issues.
But Starfield is DX12… DX 12 was quite literally the standard before Arc was made, during it’s development, and the main graphical API for at least a few more years. It’s also nowhere near as filled of issues as previous ones.
Pretty inexcusable that you buy a full price graphics card, and then need to wait until Intel blesses you with a functional driver to be able to play a game.
I don't really fault Intel for this more than say Bethesda. Both AMD and NVidia still supply a number of game-specific fixes for crashes, glitches, performance improvements in their drivers for new and old games. They've been doing this for decades now and still has to release driver updates to fix game specific issues.
The fact that the API is a "standard" doesn't mean everything is clear-cut. Developers often use these APIs in unexpected ways which were never accounted for or even mentioned in the documentation.
That basically answered my question, is every version of the driver basically an ever increasing if/then of fixes specific to each game that has to be uniquely identified based on something like the executable name or is it more so that they find one oddity that would be fixed across multiple games? It feels like the former as they’re having to do it for most every game that comes out.
That only works when hardware companies have access to the game game before the release to test, and it doesn't sound like that was the case with Starfield. Intel, to their credit, got a patch out before most people even had access.
Many developers probably aren't even testing on Arc because it's new and that only exaggerates the problem. If they don't test, they can't alert Intel to the issue.
My point aren't standards like most people think of them. They're often poorly documented, have lots of ambiguity, and many times they contain bugs/mistakes themselves. Intel potentially fully supported DX12 as far as we know, and these are all just weird exceptions.
Dodaj komentarz