Carzy Zarx’s PC Gaming Emporium - Catch Up on All the Latest PC Gaming Related News

[Pemalite]

JEMC said:

Captain_Yuri said:

Intel's Open-Source Vulkan Driver For Ray-Tracing Gets "Like A 100x Improvement"

https://www.phoronix.com/scan.php?page=news_item&px=Intel-Vulkan-RT-100x-Improve

It will be interested to see what Battlemage will be like after they have gotten a lot of the driver issues worked out. With the recent but slow rise of Crypto prices, I get the feeling we are gonna need that third player. Lets just hope they can execute better next time.

From that article:

So what's going on here? The Intel ANV change is to ensure the scratch memory for Vulkan ray-tracing use gets allocated to local memory, a.k.a. the local device memory for the DG2/Alchemist discrete graphics cards.

To now the device allocation for this RT scratch memory didn't have any allocation flags set and so presumably was getting allocated back to general system memory. Now with the "ANV_BO_ALLOC_LOCAL_MEM" flag set, it will ensure that scratch memory buffers are residing on memory local to the Intel GPU.

Yeah, no wonder the performance has improved dramatically after that, even thought it's hard to judge an improvement when we haven't seen it tested in RT games.

It also makes you wonder what other dumb mistakes could be there in the driers waiting to be fixed.

Drivers are complex pieces of software these days... It takes a Shit-ton (Australian standard unit of measurement) to build a competent driver stack.
AMD and nVidia drivers for example have more lines of code than some older Windows kernels which are also monolithic.
AMD and nVidia have also been doing this for decades, so they have a solid foundation and experienced driver teams to draw from.. But even they can sometimes miss things... I.E. Frame pacing with the Radeon 7000 series.

It will take time for them to build drivers optimized for gaming, Intel has never done it before...
They have always optimized their drivers for a very specific and small selection of games that might top out at a couple dozen, professional use cases (I.E. Video encoding) and that is it... They just move onto the next thing.

Honestly though, until their GPU's are releasing on a reliable cadence and have been thoroughly tested... To ensure long term support (As they have a history of giving up) I would avoid at all costs.