Intel XeSS 1.0.1 Released With Bug Fixes

[phoronix]

Phoronix: Intel XeSS 1.0.1 Released With Bug Fixes

Last month Intel published the XeSS 1.0 SDK for their Xe Super Sampling technology showcased with Arc Graphics discrete graphics cards. Sadly their initial SDK drop included Windows binaries and wasn't fully open-source. On Friday XeSS 1.0.1 was published with some bug fixes but still not being fully open-source...

https://www.phoronix.com/news/Intel-XeSS-1.0.1-Released

[brucethemoose]

A comment I found on the Intel Simplified Software License, which seemingly sums it up:

...it is not an open source license. It is only a license to use a specific binary, with limitations, and with additional restrictions on reverse engineering. It is not and has never intended to be open source...

What is Intel worried about when restrictively licensing stuff like this and MKL? They aren't selling them. Don't they want the libraries to proliferate so they will sell more hardware?

[CommunityMember]

What is Intel worried about when restrictively licensing stuff like this

It is entirely possible that some of the IP inside the binary is not Intel's to publicly share. Only Intel's lawyers know for sure.

[rmfx]

Where are the sources of this open source software tough ?

[stormcrow]

A comment I found on the Intel Simplified Software License, which seemingly sums it up:




What is Intel worried about when restrictively licensing stuff like this and MKL? They aren't selling them. Don't they want the libraries to proliferate so they will sell more hardware?

The license may not even be fully enforceable even in the US even then. Far as I know no one has tried to enforce EULA prohibitions against software reverse engineering for fear of losing big in the US courts. Losing the ability to appropriately reverse engineer products for interoperability is a cornerstone of the software and electronics industry despite nearly every EULA having such a phrase banning it. Losing on the other side (reverse engineering for interop being formally declared fair use) would be equally problematic, and the hardware DRM interop lawsuits losing to fair use for interoperability purposes defenses hint at what may happen. No one wants to be a test case unless they're utterly insane and arrogant beyond measure (granted some corporations may be stupid enough to test it, but so far even Oracle who originated that clause in EULAs hasn't tried to enforce it in a lawsuit).

But as CommunityMember points out, it may not even contain code they can legally share regardless of the legal minefield of reverse engineering.

[brucethemoose]

It is entirely possible that some of the IP inside the binary is not Intel's to publicly share. Only Intel's lawyers know for sure.

That would be rather strange, since the plan is to eventually release the source IIRC. Unless its a timing issue?