Sexy

1630 killer.

Fo sho gonna do great for old games emulation, I like it gets my money vote 😛

1630 was also great i liked it.

Undying:

1630 killer.

except in directx legacy/9/10/11, opengl, and I bet even vulkan :V

It WILL get better at the older stuff, it's easy to point out it's faults but have a little faith, Intel isn't stupid they know what they got to fix, they've got XeSS workin well in SoTR on AMDs card isn't ment to work for all gpus but better on arc Support will grow

fixing those stuff will requires years. intel never been good on fixing stuff on older graphics gens. instead intel has always been prone to cut stuff before quitting driver support.

It's exciting times we live in with it all about to launch, of course it won't compete with the next gen from the other 2 but if its performance is good enough for me to start with I'm in I've been nvidia since my GeForce 2mx pci upgrade from voodoo 3 but do badly want a change

Alessio1989:

except in directx legacy/9/10/11, opengl, and I bet even vulkan :V

DX9 and 10 are non-issues, as these GPUs ought to be plenty powerful enough to handle any of such game. On Windows, there's hardly any games that are only OpenGL so that can be ignored. On Linux, OpenGL support should be just fine; that team is completely different. I think Intel has been prioritizing Vulkan, in addition to DX12. DX11 on Linux won't be any issue since DXVK often yields better-than-native performance. So really that just leaves DX11 on Windows, which unfortunately is enough to doom this product since most people will be playing at least some DX11 games on Windows.

schmidtbag:

DX9 and 10 are non-issues, as these GPUs ought to be plenty powerful enough to handle any of such game. On Windows, there's hardly any games that are only OpenGL so that can be ignored. On Linux, OpenGL support should be just fine; that team is completely different. I think Intel has been prioritizing Vulkan, in addition to DX12. DX11 on Linux won't be any issue since DXVK often yields better-than-native performance. So really that just leaves DX11 on Windows, which unfortunately is enough to doom this product since most people will be playing at least some DX11 games on Windows.

I hope you are right I am kind of expecting a disaster in dx9 and older and opengl even denying to run em ... Will see!

Venix:

I hope you are right I am kind of expecting a disaster in dx9 and older and opengl even denying to run em ... Will see!

Intel is using a translation layer for DX9 and presumably DX10 too for both Windows and Linux. Assuming MS didn't screw up their translation layer, it should work totally fine. I would say the GTX 980 was the last high-end GPU that was probably still being optimized for DX9 games, since games like Borderlands: Pre Sequel were released around the same time, and that was a DX9-only game. The A750 on paper is very roughly equivalent in performance (you can basically ignore memory since not even the 980 was lacking in VRAM for DX9 titles). So, if someone were to benchmark that game with the A750 against the 980, that would be a good way to determine whether Intel's DX9 performance is good enough. Remember: unlike the 980, the A750 will have had no optimizations, so if it even comes within a few percentile of the 980, I'd consider that a success. I consider OpenGL to be a non-issue. I'm confident it will be fine on Linux, and there's hardly anything in Windows that's OpenGL-only, let alone worth anyone's attention today. Worst-case scenario, someone could always port Zink to Windows, which has probably already been done by now. So long as they have OpenGL 2.x support in Windows, I don't think they'll need anything else. In any case, while I don't think Intel should abandon OpenGL or anything older than DX11, resources are finite and they have to start somewhere. Even AMD has practically given up OpenGL on Windows for the past decade, and they actually have professional-grade GPUs (a lot of pro software uses OpenGL). The last noteworthy DX9-only game released was in 2017 (A Hat in Time). These games are old and graphically simple enough that all Intel has to do is just make sure the API calls are working and throw transistors and GHz at the games rather than try to optimize. Given the situation Intel is in, I think it's wise that they prioritize DX12 and Vulkan, though, I do still think they should be dumping a little more money into DX11. DX11 games are still too common for them to brush aside.

schmidtbag:

DX9 and 10 are non-issues, as these GPUs ought to be plenty powerful enough to handle any of such game. On Windows, there's hardly any games that are only OpenGL so that can be ignored. On Linux, OpenGL support should be just fine; that team is completely different. I think Intel has been prioritizing Vulkan, in addition to DX12. DX11 on Linux won't be any issue since DXVK often yields better-than-native performance. So really that just leaves DX11 on Windows, which unfortunately is enough to doom this product since most people will be playing at least some DX11 games on Windows.

With dx9/10/ogl and older dx versions the problem is not the performance that is enough as you say, the problem is not having glitches or games crashing. DX11 on games is still a thing. Majority of most online games played today still run on DX11 backends (https://www.battlemetrics.com/) only few of them have working and solid DX12/vulkan backends. Also for gamers that don't are so casual to use just an igpu, retro-compatibility is still a thing.

Alessio1989:

With dx9/10/ogl and older dx versions the problem is not the performance that is enough as you say, the problem is not having glitches or games crashing. DX11 on games is still a thing. Majority of most online games played today still run on DX11 backends (https://www.battlemetrics.com/) only few of them have working and solid DX12/vulkan backends.

I honestly don't see glitches being an issue. While I have very limited faith in MS, I don't think they're going to screw up their own translation layer that badly. In the event a game is found to not work, it will most likely be MS to figure out the patches since it's their translation layer that isn't working. This can be pretty easily proven by just using a competitor's GPU using the same translation layer to see if they suffer the same issues. In Linux, you can choose between either D9VK or Gallium Nine, both of which yield good stability and performance. Perhaps some of the more poorly developed titles will encounter glitches but I really don't see people playing those that often. While I think it's perfectly acceptable to expect Arc to be able to play DX9, DX10, and OGL games, I don't think it's wise for anyone to buy one with the intention to play them. But yeah like I said, DX11 is a big deal that Intel really does need to fix. I'm sure that alone is the only reason Arc has been delayed for so long.

Also for gamers that don't are so casual to use just an igpu, retro-compatibility is still a thing.

Yeah, I was considering saying in my last post that most people would be better off just getting a 5600G if they intend to play stuff from 8+ years ago, or, casual games.

Alessio1989:

except in directx legacy/9/10/11, opengl, and I bet even vulkan :V

With their Gen 12 IGPs, their Vulkan driver seems to be the only one that doesn't produce crap frame times. E.g. Strange Brigade D3D12 runs like crap, while Vulkan is fine.

schmidtbag:

I honestly don't see glitches being an issue. While I have very limited faith in MS, I don't think they're going to screw up their own translation layer that badly. In the event a game is found to not work, it will most likely be MS to figure out the patches since it's their translation layer that isn't working. This can be pretty easily proven by just using a competitor's GPU using the same translation layer to see if they suffer the same issues. In Linux, you can choose between either D9VK or Gallium Nine, both of which yield good stability and performance. Perhaps some of the more poorly developed titles will encounter glitches but I really don't see people playing those that often. While I think it's perfectly acceptable to expect Arc to be able to play DX9, DX10, and OGL games, I don't think it's wise for anyone to buy one with the intention to play them. But yeah like I said, DX11 is a big deal that Intel really does need to fix. I'm sure that alone is the only reason Arc has been delayed for so long. Yeah, I was considering saying in my last post that most people would be better off just getting a 5600G if they intend to play stuff from 8+ years ago, or, casual games.

sure the translation layers now aren't just surface projection libs anymore, but games on DX9 and older versions (aren't those thunked/translated already by a OS dlls?) required a lot of tricky bug fixes in drivers etc... not sure how a translation layer can fix this. While decades of know-how by AMD/ATI and NVIDIA driver teams have them. I can still run like a charm games of 9x era on a RDNA2 graphics card as well opengl 1.x games. I know this doesn't matter today for young games but still is a thing for a good amount of people buying dedicated graphics cards.

Alessio1989:

except in directx legacy/9/10/11, opengl, and I bet even vulkan :V

Intel has better vulkan support than AMD. Intel also has less issues with d3d10 or 11 than amd.

Alessio1989:

not sure how a translation layer can fix this

frankly speaking, it won't. without the vendor driver workarounds all these games are going to be demonstrating their graphical or crash issues by running directly to ddi spec, there are already workarounds being added to the 9on12 public repo to workaround that manifest because the game developer screwed something up. https://github.com/microsoft/D3D9On12/issues/6 having performance tested some titles with the 9on12 ddi, they are almost always slower than using a native d3d9 driver.

Astyanax:

Intel has better vulkan support than AMD. Intel also has no issues with d3d10 or 11.

You apparently don't own any Intel Gen 12 graphics, hilarious.