Intel Arc Graphics Running On Fully Open-Source Linux Driver

[fagzal]

I was actually looking forward to this chip, because... what are the Linux GPU options anyway?

- Integrated Intel : works nice, open source, but slow
- Integrated AMD : works nice, open source, but no motherboards have 3+ HDMI/DVI (also not the fastest)
- AMD Radeon : too much power draw
- NVidia GeForce : proprietary drivers a PITA

ARC would totally fill this niche where you have multiple monitors and want an open source stack, but no gaming, e.g. video editing / transcoding /etc. I'm rooting for Intel in this case

[mitcoes]

It seems Intel CPUs will perform better with Intel GPUs than AMD ones because of resizable bar.
I would like to read tests about it.

[kylew77]

Ugh. P5800X is still waymore than I want to spend on a single component.

I couldn't afford that so I got the boot drive option, the P1600X 118GB. It is still over 200 dollars though! Going to use it for a system to learn OpenBSD and NetBSD on which both use the FFS/UFS filesystem and Michael has demonstrated that it is generally slower than ext4 so I figured the extra umph would be worth the cost.

[Fanboy80]

I believed "Fully Open-Source" for 20 seconds

[pinguinpc]

Sounds like someone's never heard of XeSS.

https://www.anandtech.com/show/16896...ing-technology

Another thing dont care, only stay interested on av1 hardware encoding and older games using dxvk - d3d8to9 - zink

[Rabiator]

I was actually looking forward to this chip, because... what are the Linux GPU options anyway?

- Integrated Intel : works nice, open source, but slow
- Integrated AMD : works nice, open source, but no motherboards have 3+ HDMI/DVI (also not the fastest)
- AMD Radeon : too much power draw
- NVidia GeForce : proprietary drivers a PITA

ARC would totally fill this niche where you have multiple monitors and want an open source stack, but no gaming, e.g. video editing / transcoding /etc. I'm rooting for Intel in this case

• ARC: still beta quality, open source, performance erratic (assuming similar maturity as on Windows). I'm looking forward to Michael's tests.
• AMD Radeon : Power draw is reasonable for the performance, tends to better than Nvidia in that respect these days.
• NVidia GeForce : Opening up the drivers now, we will see how fast that goes. By comparison, IIRC AMD announced they would go open source in 2007. It took them several years to fully get there.

[Venemo]

" Fully Open-Source Linux Driver" and "the usual firmware caveat" is a nonsense.
Radeon (r100, r200, r300g, r600g, radeonsi) and AMDGPU (Radeonsi) are not fully opensource sice they need a proprietary blob firmware.
The same for intel.

The drivers are fully open source, the firmware are not. This has been the situation for a long time now.

[cqcallaw]

...with Linux 6.0 the DG2 class support isn't exposed by default but requires setting the i915.force_probe=[PCI-ID] module option to force the driver to initialize the graphics card. Mesa 22.2 for ANV Vulkan and Iris OpenGL will in turn work out-of-the-box with current cards when booting with the necessary kernel support in place.

If fussing with PCI IDs becomes too much of a nuisance, i915.force_probe also accepts a "*" wildcard

[qarium]

I can't say you're wrong, but PA RISC never really challenged Intel in the markets where x86 played. It was only perhaps on obstacle in some of the markets where Intel wanted to be, but it would be incorrect to say that HP was their only competition. At the time of Itanium's inception, DEC Alpha, Sun SPARC, IBM POWER, and MIPS were all credible threats, as well.

And by the time of Itanium's demise, HP ate Compaq, which ate DEC. I forget if it was HP or Compaq who was responsible for dismantling DEC and even selling off parts of it to Intel. Oracle ate Sun and bled it dry. SGI ate MIPS and imploded. That left IBM POWER as the only viable competitor, and of course IBM is its own worst enemy. Anyway, the point is that Itanium didn't kill any of these guys. Itanium only killed PA RISC by getting HP to stop investing in it (which isn't to say it wouldn't have eventually met some other demise, but that's what killed it).

well... in this game maybe not the better technology wins instead the surviver writes the history book...
Itanium clearly kicked HP out of the cpu business...
and when we see that all of these competitiors dies and are death even if they had very very very good cpus...

MIPS is death but had good technology...
DEC Alpha where really futuristic and very good cpus but they died.
Sun SPARC had some good cpus but is in fact death...

some people believe that a product need to be a success to be a success for evil companies like intel but in fact in this chess game an evil company like intel can produce a big shit-show and even FAKE a product and make it FAIL if only intel remove 1 competitor they won in the end and this by a failing product.

we live in an evil world... people are naive if they belive the "better" product wins...

to my knowlege DEC Alpha where in fact "better" products and they lost. it was beautifull 64 bit design and the clock speeds compared to intel where super high... intel did 166mhz cpus alpha did 300-500mhz cpus at the same time

today people claim apple m2 is shit for gaming and they proof it with some x86 game run on rosetta2 in x86 emulation mode and also it is dx11 game need to be translated to metal another emulation mode...
and people see this as "proof" that ARM is bad for gaming. https://youtu.be/c3exhJpNQbU

thats the same logik what killed DEC Alpha... they claimed the Alpha is slower at running x86 code in emulation mode.

we life in a full joke world really people can no longer compile an opensource game for ARM to compare it to x86 cpus and people can not even use a vulkan translation to metal to make a fair GPU api comparison ...

without this fake world: "DEC Alpha run x86 code in emulation mode slow" "apple m2 runs x86 code slow in rosetta2 emulation mode" really without this fake argument intel would already be broke 20 years ago...

people really has to stop this bullshit. intel and the x86 ISA war is complete and pure evil...

[qarium]

Itanium launched with all the big server OEMs (SGI, Bull, Fujitsu, IBM, Dell, etc.) but that soon wound down to only one, HP. It was never really competitive, the customers had Itanium pushed on them by HP. And when a single enterprise ISV (Oracle) announced to drop support it was game over.
Intel did make a profit from Itanium, but only because HP paid for all its development. This was revealed from HP-Oracle lawsuit.
Other than that, it nearly turned into Rambus style situation as Intel refused to extend x86 to 64-bit until AMD came and forced them to adopt amd64 (under changing names, Yamhill -> IA32e -> EM64T -> Intel 64).
It was HP who drove and paid for everything. This is also why Intel didn't care while going from broad server OEM support to just HP.
I don't think you got that right.
DEC was bought by Compaq, which was then bought by HP. It was HP which killed Alpha and PA-RISC in favor of Itanium. The team who previously worked on Alpha went on to work for Intel. SGI pivoted from MIPS to Itanium. Dell had no business with any of the other CPU architectures.

right HP paid for everything .... and intel had no reason to make itanium a success it is a fact that a failed itanium would give intel more power ... :.. intel did use itanium to remove HP from the CPU market... in my point of view intel is a evil company.

there is a golden rule: if you go into bed with: intel or microsoft or nvidia you will die a horrible death.