NVIDIA Continues Ramping Up Their Vulkan Articles

[TheOne]

Since AMD apparently invented all of this stuff in the first place and is light years ahead of Nvidia, have one of their developers post a detailed technical explanation about how there are no disadvantages whatsoever to going to Vulkan.

Actually that's a pretty good idea, I would like to see an AMD developer thoughts of OpenGL vs Vulkan. In any case, Michael published a link to a Vulkan talk made by a Intel Developer and he seemed much more optimistic about Vulkan been a much more better option than OpenGL http://video.fosdem.org/2016/k1105/v...pen-source.mp4. In contrast, I can't say the same about the nvidia vulkan blog post, they where instead telling developers how to make OpenGL applications similar to Vulkan by using Nvidia OpenGL extensions, sparkling it with "this way your application will be more ready to support vulkan, (if that is feasible for you)"...

[johnc]

I don't get it why AMD lets NVIDIA steal their thunder when it comes to Vulkan. Vulkan is basically their brain child, but apart from a slide where they talked about their closed source Vulkan driver that will be open source later we haven't heard anything about Vulkan from them since it was announced. Even ImaginationTechnologies and Intel showed some demos and had a few blog posts and seminars. For me as a consumer that is telling me that AMD is much less invested in Vulkan than NVIDIA and that I better buy a NVIDIA card. In the meantime you can hear them banging the DirectX12 drum all the time.
It would have been so cheap for them to improve their reputation with the Linux community, by just showing a demo where Vulkan on Linux makes an AMD card perform at its hardware limit. The unclear driver situation doesn't make it any better. With NVIDIA it's pretty clear which cards will be supported and how they will be supported.

Well AMD does have a pretty close relationship with MS, so it's understandable.

[Michael]

Good news for AMD: Their next-generation Polaris chips are probably going to have a 4-6 month lead before Nvidia gets Pascal onto the market.
In that time, AMD is free to get their Vulkan Linux drivers up to snuff.

4~6 months lead of Polaris over Pascal? Haha.

[Michael]

Then, when Nvidia launches Pascal we'll see who has better Vulkan support. Once again, AMD -- who supposedly came up with 90% of Vulkan before it even launched and ought to have a huge advantage here -- will have plenty of time to show that they really care about Linux by getting first-rate drivers together long before Nvidia has its next generation product commercially available.

Mantle isn't even close to 90% of Vulkan... It's just a big myth. While it started off as based on Mantle, it's a very different design at this point.

[emblemparade]

As an experienced programmer, I didn't see one iota of negativity in any of those Nvidia articles. Instead, I saw a pretty smart analysis of the tradeoffs between two different APIs that was clearly written by an experienced software engineer and not a marketing droid.

You're basically right, but I think the NVIDIA blogs really do have an unenthusiastic, unimaginative tone. It's true that there's a tradeoff in complexity, but I imagine that not many people will be using "raw" Vulkan. The API lends itself very easily to creating higher-level wrapping layers. And actually, it's so low-level that there can be very interesting competition between various approaches to wrappers. For example, one wrapper might want to duplicate the raw "texture" experience as in OpenGL, another might present a more innovative approach to mapping pixels to images.

The bottom line is that the "complexity" of Vulkan will end up being unimportant.

What will be important for a long time is that many mobile devices, consoles, and older PCs just do not have support for Vulkan, and won't get it. If you're developing a crossplatform application, that will me your major tradeoff.

[blackout23]

...While they are saying that Vulkan will not improve situation much and will just complicate stuff, other companies are showcasing demos of opengl vs Vulkan where Vulkan performs a lot better.

You should read the article again. Nowhere does it say that in general it won't improve the situation much. The article says that in scenarios where you are already basically GPU bound Vulkan won't add magical GPU Cores to your card and overclock your VRAM. It's not inherently faster than any other graphics API just like any API before it.
Of course if you want to make a demo that shows off the strengths of Vulkan vs. OpenGL you create a scenario where OpenGL by its design will perform poorly. Which is what Imagination Technologies with their Vulkan Gnomes vs OpenGL ES Gnomes demo did. Just like most benchmarks that's not a 100% realistic real world gaming workload though.

[bug77]

You should read the article again. Nowhere does it say that in general it won't improve the situation much. The article says that in scenarios where you are already basically GPU bound Vulkan won't add magical GPU Cores to your card and overclock your VRAM. It's not inherently faster than any other graphics API just like any API before it.
Of course if you want to make a demo that shows off the strengths of Vulkan vs. OpenGL you create a scenario where OpenGL by its design will perform poorly. Which is what Imagination Technologies with their Vulkan Gnomes vs OpenGL ES Gnomes demo did. Just like most benchmarks that's not a 100% realistic real world gaming workload though.

On top of that, I would point people the Techpowerup's game reviews. They review the performance separately and that makes it obvious most games today are GPU limited.

However, if it Vulkan is not inherently faster than OpenGL if, between the hardware, drivers and graphics API, developing for Vulkan turns out to be faster/simpler, then we still gain.

[Sethox]

All we do (the grunts that are not under the disclosure agreement of Vulkan API) is just speculating. No Hype, No anger.. just wait and see.
Still though interesting article nonetheless.

[chuckula]

4~6 months lead of Polaris over Pascal? Haha.

I'm talking about predicted launch windows. Additionally, AMD has even said that the first Polaris parts out the door will not be the ultra high-end chips, but more consumer focused chips. Nvidia, because of its supercomputer business, is more likely to roll out a larger chip "earlier" (relatively speaking) but the actual launch date will be after the first Polaris chips have been out for a while.

Both companies are attacking the market segments that they think will make them money with these plans.

[juno]

Keep in mind though, AMD's initial Vulkan support will be a closed blob. We won't see that code until some time after Pascal, so even when Nvidia is in their preferred position, AMD will still be angling into their preferred position.

I'm not so sure about that anymore. Alex said at the xdc that it will be open source "eventually".
With the release of Vulkan being delayed, I see the a chance of free drivers initially at Vulkan launch. Especially when you keep in mind that the khronos group seems to have delayed the release in favor of day one support of all vendors