OpenMW 0.39.0

Anything related to PR, release planning and any other non-technical idea how to move the project forward should be discussed here.
dontpokethebear3893
Posts: 15
Joined: 02 Jun 2014, 14:21

Re: OpenMW 0.39.0

Post by dontpokethebear3893 » 18 Feb 2016, 04:40

scrawl wrote:

All rendering relies on a certain amount of CPU work to feed the GPU. The upcoming Vulkan API is looking to remedy that by giving the coder more direct access to the graphics hardware.

I am not familiar with the particular AMD CPU's you have, but if it's a lower end model then that would explain the performance issues.
Hi, sorry it took me so long to get back to you

So about the Vulkan API... now that it's out, any plans??

my main issue is that the game throttles based on CPU performance yet isn't properly utilizing multi-threading
I have an AMD 8350 so it's fairly high end (for an AMD processor).
Image

Image

User avatar
scrawl
Posts: 2152
Joined: 18 Feb 2012, 11:51

Re: OpenMW 0.39.0

Post by scrawl » 18 Feb 2016, 14:43

I think it is the combination of an AMD CPU (low IPC compared to Intel) with AMD GPU (unoptimised CPU-heavy OpenGL drivers) that leads to the low performance in OpenMW. In a nutshell:

Intel + Nvidia = fast
Intel + AMD = meh
AMD + Nvidia = meh
AMD + AMD = bad
my main issue is that the game throttles based on CPU performance yet isn't properly utilizing multi-threading
OpenMW uses multi-threading but when the OpenGL thread is the bottleneck, then there is not that much we can do.

I'm thinking of adding back the "reflect X" settings for the water shader, this way you could turn off e.g. NPCs appearing in reflections, which are not that noticable visually, but have a big impact on performance due to the way they're made up of multiple parts. I believe most games use lower detail in reflections, but people won't notice if it's done right.
So about the Vulkan API... now that it's out, any plans??
It is in my (long term) plans. For now I'll wait and see how the Vulkan ecosystem evolves. Good news is that we're in the same boat with hundreds of other people and companies that rely on OpenSceneGraph. I'm curious to see what kind of migration paths people will come up with.

I'm also curious when we're going to see the first OpenGL implementation on top of Vulkan. Being in control of the OpenGL code would be a huge bonus when it comes to mitigating slow OpenGL drivers. And it would potentially allow people to further optimize e.g. by turning off the validation all together, or removing OpenGL features that they don't need (e.g. multiple contexts, which are sort of useless but could have a major performance impact when every OpenGL call you make has to go through some form of thread-local storage first), adding a "command stream" feature like Nvidia's __GL_THREADED_OPTIMIZATIONS (except that it'd be cross-vendor)...
Edit: some more ramblings on Reddit.

dontpokethebear3893
Posts: 15
Joined: 02 Jun 2014, 14:21

Re: OpenMW 0.39.0

Post by dontpokethebear3893 » 23 Feb 2016, 04:40

scrawl wrote: AMD + AMD = bad
Yeah, I was afraid that was the issue. I built my machine about 4 years ago now and regret the AMD cpu ever since, but I'm too in love with AMD gpu's believe it or not to move on. Thanks for explaining all that to me, I'm a huge fan of the project and have been following for 3 years if not more. Really excited for where things are going.

ezze
Posts: 492
Joined: 21 Nov 2013, 13:20

Re: OpenMW 0.39.0

Post by ezze » 23 Feb 2016, 09:51

I had the same issue before moving to Intel + Nvidia. But I was always confused by this, obviously the AMD has good engineers... how it can work so badly?

SquireNed
Posts: 402
Joined: 21 Dec 2013, 22:18

Re: OpenMW 0.39.0

Post by SquireNed » 23 Feb 2016, 13:36

ezze wrote:I had the same issue before moving to Intel + Nvidia. But I was always confused by this, obviously the AMD has good engineers... how it can work so badly?
It just really comes down to Intel having a slight technological advantage; I use an AMD CPU, and they've got a brilliant power-price ratio, but Intel has hyperthreading and other proprietary technology: my FX-8350 has more raw power than an Intel CPU, but doesn't seem to have some of the management features that Intel's been working on for as long as I can remember (I still remember excitedly having my first hyperthreaded processor back when I was ten? twelve?).

Basically, AMD's raw power works really well for something like Blender (though you really should be using an Nvidia card for that, but that's besides the point) but Intel has slightly better per-core and I think its implementation of the architecture might be slightly superior.

User avatar
Lagahan
Posts: 47
Joined: 16 Aug 2014, 11:24
Location: Ireland

Re: OpenMW 0.39.0

Post by Lagahan » 23 Feb 2016, 22:25

ezze wrote:I had the same issue before moving to Intel + Nvidia. But I was always confused by this, obviously the AMD has good engineers... how it can work so badly?
Here's a couple of videos that explain what happened to AMD's performance and why they aren't on par at the moment with Intel. Dude goes really in depth as to why AMD don't perform as well in some situations.

https://www.youtube.com/watch?v=d9aaGyqm2m8

Second one covers AMD's next architecture coming out this year.
https://www.youtube.com/watch?v=OvF3BJTLgRQ
Specs: Core i7 8700k @5GHz, EVGA GTX2080ti XC SLI,
32GB RAM @3.2GHz, Windows 10 1803

User avatar
psi29a
Posts: 4917
Joined: 29 Sep 2011, 10:13
Location: Belgium
Gitlab profile: https://gitlab.com/psi29a/
Contact:

Re: OpenMW 0.39.0

Post by psi29a » 24 Feb 2016, 10:57

I got an email back from the OSG Debian Maintainers:
alberto wrote: Hello, psi29a:

we tried some time ago to upload 3.3.x, weeks before 3.4 were
available. This new package version was rejected because the FTP team
realized that the copyright file was in a really bad shape. After that,
I wanted to fix that before packaging 3.4 and 3.2.2.

I have almost finished to update that copyright file, so 3.2.2 should be
out soon, in about a week at most. Updating to 3.4 should also be
almost trivial.

For v3.4, we planned creating a new source package (openscenegraph-3.4)
in order for both versions to co-exist. Although OpenMW will
indeed benefit from this new version, there are some other users that
rely on 3.2.x in order to avoid that their packages FTBFS.

Thank you for your help. I will definitely look at your packaging for
bringing in your improvements.

Regards,
Alberto
So Debian maintainers are held up by the arm-chair lawyers on getting newer versions of OSG into main.

This means that we'll have to wait with releasing 0.39 on Debian until OSG 3.4 is uploaded. Looks like Ogre3D all over again. ;)

User avatar
Atahualpa
Posts: 1129
Joined: 09 Feb 2016, 20:03

Re: OpenMW 0.39.0

Post by Atahualpa » 11 Mar 2016, 11:10

How is the current status of v0.39.0? Zini, you stated that we are almost ready to start the release (source: Morrowind Modding Interview). What exactly is missing?

Needless to say I'm curious because of the release commentary videos. I'll start to work on the v0.38.0 videos tomorrow, i.e. with a bit of luck they can be released before Maundy Thursday. If we want to publish v0.39.0 and the corresponding commentary videos at the same time, that would be between 2nd and 8th April. Does this look feasible to you?

User avatar
Zini
Posts: 5538
Joined: 06 Aug 2011, 15:16

Re: OpenMW 0.39.0

Post by Zini » 11 Mar 2016, 11:14

We are having two pretty important OpenMW-CS features coming up, but they will need a bit more time. I am undecided if we should wait a bit longer or go ahead with starting the RC phase next week. Opinions?

User avatar
psi29a
Posts: 4917
Joined: 29 Sep 2011, 10:13
Location: Belgium
Gitlab profile: https://gitlab.com/psi29a/
Contact:

Re: OpenMW 0.39.0

Post by psi29a » 11 Mar 2016, 12:06

What's the ETA? I'm comfortable with waiting, but for personal reasons. :P I'm blocked on the debian side anyway due to the OSG situation.

Post Reply