And Vulkan is now here :)

I have seen this as well , prolly something they wanted to test and have a very large audience able to use it.
 
just wait till next year with their latest phantom video card release (TALK)
By then everything will be amd.

yea yea yea nothing but smoke and mirrors

this will make a big deal until microshaft says it won't then it will just go up in smoke
N won't support it because its not theirs.
I wish that openGL was the only graphics we had.
 
I, for some reason had this memory instantly pop into my head. Had to look it up to share. :D


I always wondered how I could make a tastier 'belima', whatever the hell that is. :p
 
Impressive.
Am I misreading those Phoronix benchmarks? There appears to be significant performance loss when using Vulkan over OGL, in some cases as high as 40%. And while I know Phoronix is a Linux-centric site, a thorough comparison should include Windows w/Vulkan vs. Linux w/Vulkan - not to mention the standard DX11 build of DOTA2.

What did you find impressive about those benchmarks?
 
Am I misreading those Phoronix benchmarks? There appears to be significant performance loss when using Vulkan over OGL, in some cases as high as 40%. And while I know Phoronix is a Linux-centric site, a thorough comparison should include Windows w/Vulkan vs. Linux w/Vulkan - not to mention the standard DX11 build of DOTA2.

What did you find impressive about those benchmarks?

Exactly. It's called sarcasm. All the hoopla about Vulkan and that doesn't look to crash hot.
 
Last edited:
Exactly. It's called sarcasm. All the hoopla about Vulkan and that doesn't look to crash hot.

I said when Mantle was announced that these "low-level APIs" are a pile of crap and nothing more than marketing snow jobs.

There are some nice things like better multi-threading, and DX12 does bring DX up to snuff with AZDO-type principles, etc. But the "low-level" hype was mostly garbage.
 
Am I misreading those Phoronix benchmarks? There appears to be significant performance loss when using Vulkan over OGL, in some cases as high as 40%. And while I know Phoronix is a Linux-centric site, a thorough comparison should include Windows w/Vulkan vs. Linux w/Vulkan - not to mention the standard DX11 build of DOTA2.

What did you find impressive about those benchmarks?

Well Rome wasn't built in a day. All of these companies are learning how to program these APIs on the fly, and literally writing the code that is destined to be added to it for future titles. And to top it off they are porting them into DX9, DX10, and DX11 programs that are were completed and shipped long before a low level API was ever thought of. Think of these games as guinea pigs or proof of concept hacks to show off what it can potentially do. Hate to do the obligatory car analogy, but this is like trying to invent an electric motor to go into a Model A Ford; not shove an electric engine into a Model A Ford. You might have the basic understanding of what you want out of it in the end, but nothing fits properly. So you have to collaborate with others in the industry and get custom code written.

It takes time, understanding, open source resources to share ideas such as seminars and GPUOPEN / whatever the Nvidia open source GameWorks thing is called. Basically these pioneers are teaching the entire industry how to handle problems of a low level API. These guys are true pioneers as this work will go into all Valve games from now on or games using that engine. Same for Unity, CryEngine, Stardock, etc.

Personally I find what they are doing to be remarkable. If they hacked some code into Dota 2 and got this close; imagine if they started Half Life 3 from scratch with the Vulkan API in mind from the onset.
 
This was only tested on a linux os anyway right.....its not like he took the time to test on windows either did he? How many of us are using linux to play this game?

Yes, Phoronix is a Linux only testing website. Everything I've read there was in the context of Linux development.
 
Well Rome wasn't built in a day. All of these companies are learning how to program these APIs on the fly, and literally writing the code that is destined to be added to it for future titles. And to top it off they are porting them into DX9, DX10, and DX11 programs that are were completed and shipped long before a low level API was ever thought of. Think of these games as guinea pigs or proof of concept hacks to show off what it can potentially do. Hate to do the obligatory car analogy, but this is like trying to invent an electric motor to go into a Model A Ford; not shove an electric engine into a Model A Ford. You might have the basic understanding of what you want out of it in the end, but nothing fits properly. So you have to collaborate with others in the industry and get custom code written.

It takes time, understanding, open source resources to share ideas such as seminars and GPUOPEN / whatever the Nvidia open source GameWorks thing is called. Basically these pioneers are teaching the entire industry how to handle problems of a low level API. These guys are true pioneers as this work will go into all Valve games from now on or games using that engine. Same for Unity, CryEngine, Stardock, etc.

Personally I find what they are doing to be remarkable.
You make good points, and I do mostly agree with you. However, I myself struggle to find interest in proof of concept hacks for a Source Engine game. I think people (not necessarily you in particular) need to temper their excitement and expectations here.

There's a lot of hype surrounding Vulkan, especially when it comes to Valve corp. and their use of it. Less than a year ago we had folks proclaiming a new VR gaming future, powered by Source 2 and Vulkan. No doubt, HL3 was around the corner too, and when it's released we'll be blown away by what Valve has been working on. What we got was a Portal-themed tech demo for Vive that runs on Unity.

Instead of expecting revolutionary games from Valve, we need to look at what Valve has been doing for the last ten years and try to predict what they're going to do with Vulkan in the future. I don't expect Valve to use Vulkan to produce next-gen titles, but to help their F2P games run better on the low-end hardware one finds in Russia, China, Oceania and South America - regions where DOTA2 and CSGO are very popular and profitable. I predict seeing Vulkan builds of popular F2P Valve games in the next year or two that will significantly reduce CPU load, hence allowing those titles to better penetrate low income markets.
 
You make good points, and I do mostly agree with you. However, I myself struggle to find interest in proof of concept hacks for a Source Engine game. I think people (not necessarily you in particular) need to temper their excitement and expectations here.

There's a lot of hype surrounding Vulkan, especially when it comes to Valve corp. and their use of it. Less than a year ago we had folks proclaiming a new VR gaming future, powered by Source 2 and Vulkan. No doubt, HL3 was around the corner too, and when it's released we'll be blown away by what Valve has been working on. What we got was a Portal-themed tech demo for Vive that runs on Unity.

Instead of expecting revolutionary games from Valve, we need to look at what Valve has been doing for the last ten years and try to predict what they're going to do with Vulkan in the future. I don't expect Valve to use Vulkan to produce next-gen titles, but to help their F2P games run better on the low-end hardware one finds in Russia, China, Oceania and South America - regions where DOTA2 and CSGO are very popular and profitable. I predict seeing Vulkan builds of popular F2P Valve games in the next year or two that will significantly reduce CPU load, hence allowing those titles to better penetrate low income markets.

in total agreement with you on all points, it definitely makes sense for them to use vulkan in the F2P market for the reasons you stated. it wouldn't surprise me if you see riot follow suit soon afterwards as well. eventually you'll see it move onto other things from valve as well but i think that'll be their primary focus for the next year or so.
 
you guys are 100% correct......its not a bad thing tbh, its just not going to give EVERYONE a performance boost.....those that wont dont need it anyway
 
you guys are 100% correct......its not a bad thing tbh, its just not going to give EVERYONE a performance boost.....those that wont dont need it anyway

Well we have yet to run across a game engine developed from the ground up with DX12 or Vulkan in mind. Also we have yet to experience a game developed from the beginning with one of the new APIs in mind. Even the new Deus Ex title coming this Fall isn't likely designed from the ground up for DX12. It may possess a higher level of DX12 integration; but I'm doubtful that DX12 was even a full spec when the game was started as game development typically takes years.

Next you have to look at performance boost with an open mind. Remember Mantle? It made games developed with it as an option ultra liquid smooth. That came down to improved frame times; NOT boosted frame rates. Again it was a new API slapped onto existing DX9 - DX11 code. Too bad Battlefield stayed broken at that time, before EA publicly announced the revelation, that they discovered that paying customers wanted working games. When it was working, the game was as smooth as glass under Mantle. I'm hoping that the new APIs can bring back those perfectly laid frames.

I can tell you from owning a 120Hz monitor that 120 fps does not always equal perfect smooth game play. Sometimes in poorly designed games like Batman it is nothing more than 120 frames of chopping and stuttering as the game just wildly tosses them in a reckless manner. Great frame times does equal a smooth game even at lower frame rates. Great frame times + 120Hz = gaming nirvana. 120 Hz of stuttering is better than 60Hz of stuttering; but never as good as great frame times. That's why tests like FCAT were so good in discovering frame variance.

It's also why reading [H]ardocp reviews is so awesome. Brent_Justice always lets you know when the frame rate lies in [H]ardocp reviews. There is a lot of substance to those reviews when he says that even though video card X is above 60Hz, the game isn't nearly as smooth as on video card B. Or the SLi and Crossfire implementation in a game makes for a stuttering mess even though the frame rate is off the charts compared to a single card. I'm not saying that this is the case in Dota 2 as this is just proof of concept work at best. But it is another way that you need to be able to analyze the implementation of these low level APIs.

Taking the load off the CPU in these new APIs makes for a much better experience for lower end CPU owners. But if you look at the AOTS testing that [H]ardocp did it even helped the latest Intel chips. It wasn't as great as say the 5 year old AMD chips; but it was a boost. Free boosts are always welcomed. Right now it's hard to sit back and wait for these developers to reinvent the wheel in some cases. Well it feels like that's what they are doing with how long it takes sometimes. ;) But in the long run it's well worth it as I am sure that some brilliant soul will find a way to exploit all those free CPU idle cycles in a new and exciting way.

Check out the thread in the AMD section on the new console porting technique. The hardware functionality is there for AMD hardware as well as Nvidia hardware. Even though it's open source with all the source code freely exposed, I bet someone is going to criticize it because AMD invented it. On Twitter a bunch of developers were retweeting it like the head guy from DICE. I get excited when the actual software developers using the techniques think something is great, even though I like to read all of the PR from both camps also.

Hope I didn't make too many mistakes. Feel free to pick it apart. ;)
 
I wish [H] would do a Vulcan test on dota2 using windows 10 of course....Im not convinced Linux will perform the same as Windows. I guess the reason they dont is fraps being outdated
 
I said when Mantle was announced that these "low-level APIs" are a pile of crap and nothing more than marketing snow jobs.

There are some nice things like better multi-threading, and DX12 does bring DX up to snuff with AZDO-type principles, etc. But the "low-level" hype was mostly garbage.

Agreed. Yes, you get significantly lower CPU usage, which will help weaker CPUs [Core i3s, Pentiums, and FX-4xxx series CPUs in particular]. But then the trade-off: Moving a lot of the memory management out of the driver, where it has historically been VERY optimized by NVIDIA/AMD, and leaving it up to the developers to handle on a per-title basis. Sorry, you're going to have a less then optimal solution, and it will eat away some performance. As I've said from the beginning: DX12/Vulkan are not magic bullets, and the only reason AMD has been pushing for something like this is due to it's relatively poor DX11 driver compared to NVIDIA.
 
Agreed. Yes, you get significantly lower CPU usage, which will help weaker CPUs [Core i3s, Pentiums, and FX-4xxx series CPUs in particular]. But then the trade-off: Moving a lot of the memory management out of the driver, where it has historically been VERY optimized by NVIDIA/AMD, and leaving it up to the developers to handle on a per-title basis. Sorry, you're going to have a less then optimal solution, and it will eat away some performance. As I've said from the beginning: DX12/Vulkan are not magic bullets, and the only reason AMD has been pushing for something like this is due to it's relatively poor DX11 driver compared to NVIDIA.

do you really think AMD alone is enough to somehow set in motion many tens of millions of dollars (and more) worth of engineering effort for no reason other than their DX11 driver sucks chode?
 
do you really think AMD alone is enough to somehow set in motion many tens of millions of dollars (and more) worth of engineering effort for no reason other than their DX11 driver sucks chode?

People try to conspiracy theory everything to death sometimes. It's fun to put on a tinfoil hat though and wonder, "What's out there?"
:playful:
 
No one sees the benefits any more after 1 or 2 benchmarks ? Develop once on Vulkan have minimal work done to make it fit every platform and your product can span across 4 or 5+ different platforms without any serious development time needed for each version ...

All Vulkan needs is some time the magic won't happen overnight (see DX12 titles more "miss" then "hit" atm) ....
 
just wait till next year with their latest phantom video card release (TALK)
By then everything will be amd.

yea yea yea nothing but smoke and mirrors

this will make a big deal until microshaft says it won't then it will just go up in smoke
N won't support it because its not theirs.
I wish that openGL was the only graphics we had.

Don't do drugs kids
 
Also, any news on when DEWM is getting Vulcan? Really looking forward to making my system sing.




I, for some reason had this memory instantly pop into my head. Had to look it up to share. :D


I always wondered how I could make a tastier 'belima', whatever the hell that is. :p



THANK YOU SO MUCH for linking that video. I couldn't for the life of me find it. LOVE that ad.
 
He works for ID Software. :) Saw it on Reddit. I guess after you port one game to Vulkan, it gets easier to do others.

 
Now everyone can check out the Vulkan path for free!
Steam Community :: Group Announcements :: DOOM

DOOM Demo - Now with Vulkan!
JULY 27 - JLEAVEYBETHESDA
We've updated the DOOM Demo on Steam to support Vulkan!

When launching DOOM from Steam, you’ll be prompted with a new menu to choose either OpenGL or Vulkan. To play with Vulkan, select the second menu option. You can also switch between OpenGL and Vulkan in game via the Advanced Settings options menu. The game will restart whenever you switch between OpenGL and Vulkan.

Please be sure to update your drivers to these new version:

NVIDIA: 368.69 - Download the latest NVDIA drivers: Drivers - Download NVIDIA Drivers
.
AMD: 16.7.2 - Download the latest AMD drivers:
Download Drivers

Be sure to review our full FAQ for additional information and troubleshooting:
Steam Community :: Group Announcements :: DOOM
 
Back
Top