ISI Planning to convert RF2 from DX9 to DX10 or DX11 like iRacing did?

[stonec]

iRacing has limitless money for game development while ISI has one guy working on graphics engine, that's the difference. Also I think ISI is working on other game at moment, so even less resources to do some whole game engine rewrite that DX11 requires. I remember in 2013 when first Silverstone version came out, it ran quite poorly and it was justified to criticize rF2 for poor performance back then. After that ISI has only optimized the engine without bringing any new FPS consuming effects, new track versions have improved FPS stability, GPU drivers have improved, and finally average GPU performance according to Moore's law has tripled since then. I think the performance complaints 3 years ago were very justified, but not so much anymore since rF2 hasn't got any heavier to run.

 

[WhiteShadow]

iRacing has limitless money for game development while ISI has one guy working on graphics engine, that's the difference. Also I think ISI is working on other game at moment, so even less resources to do some whole game engine rewrite that DX11 requires. I remember in 2013 when first Silverstone version came out, it ran quite poorly and it was justified to criticize rF2 for poor performance back then. After that ISI has only optimized the engine without bringing any new FPS consuming effects, new track versions have improved FPS stability, GPU drivers have improved, and finally average GPU performance according to Moore's law has tripled since then. I think the performance complaints 3 years ago were very justified, but not so much anymore since rF2 hasn't got any heavier to run.

Your description regarding rFactor2: "GPU drivers have improved, and finally average GPU performance according to Moore's law has tripled since then" is placebo. GeForce 700 Series Release date May 2013 and GeForce 1000 Series Release date May 2016. I have my doubts that 1080 is outperforming 3 x 780 and I have not experienced any improvement or performance boost with Nvidia drivers since release of rF2. New ISI tracks are improved as you state but the problem is that there are loads of 3pa tracks which don`t perform well. rF2 is modding platform and we can not expect that all moders are redoing they work. I think that only DX11 can give boost which is needed to 3pa tracks.

 

[Promag]

What about Vulcan? been reading some positive things about it. Is it possible?

 

[blakboks]

Your description regarding rFactor2: "GPU drivers have improved, and finally average GPU performance according to Moore's law has tripled since then" is placebo. GeForce 700 Series Release date May 2013 and GeForce 1000 Series Release date May 2016. I have my doubts that 1080 is outperforming 3 x 780 and I have not experienced any improvement or performance boost with Nvidia drivers since release of rF2. New ISI tracks are improved as you state but the problem is that there are loads of 3pa tracks which don`t perform well. rF2 is modding platform and we can not expect that all moders are redoing they work. I think that only DX11 can give boost which is needed to 3pa tracks.

I just upgraded from the GTX760 to the GTX1080. In texture fill rate alone, it's something like 2.75 times higher. In terms of real-world tests (I did a full computer rebuild): I was getting 30-50 fps in Doom on the old card/system, and now am getting capped at 200 fps (Vulkan).

...just food for thought.

 

[WhiteShadow]

I just upgraded from the GTX760 to the GTX1080. In texture fill rate alone, it's something like 2.75 times higher. In terms of real-world tests (I did a full computer rebuild): I was getting 30-50 fps in Doom on the old card/system, and now am getting capped at 200 fps (Vulkan).

...just food for thought.

King vs king Titan Z is GeForce 700 Series card
Maybe you and many others should focus and look sometimes in ISI technical forums and find out is going on with rF2, perhaps start with this thread : http://isiforums.net/f/showthread.php/29462-High-frame-rates-but-stuttering

 

[Will Mazeo]

Would take too much for rF2 to look like this?

The Silvertone video realy looked like rF2 with some shader work and lens flare to me (but the les part is not important, it sucks lol)

 

[WhiteShadow]

I just upgraded from the GTX760 to the GTX1080. In texture fill rate alone, it's something like 2.75 times higher. In terms of real-world tests (I did a full computer rebuild): I was getting 30-50 fps in Doom on the old card/system, and now am getting capped at 200 fps (Vulkan).

...just food for thought.

I stopped by the store on Friday and I bought two GTX 1080. Triple sceens 120fps game play in rF2 is still not possible with descent in game settings with triple screens, single card or SLI.

SLI GTX 780Ti, whql 353.30, sli bits : 0x02D04005 (LEGO Batman 2,.....) is still undisputed champion of rF2 and is able to deliver locked 120 fps gaming with descent settings.

....rF2 seem to be frozen in time.

 

[Led566]

I stopped by the store on Friday and I bought two GTX 1080. Triple sceens 120fps game play in rF2 is still not possible with descent in game settings with triple screens, single card or SLI.

SLI GTX 780Ti, whql 353.30, sli bits : 0x02D04005 (LEGO Batman 2,.....) is still undisputed champion of rF2 and is able to deliver locked 120 fps gaming with descent settings.

....rF2 seem to be frozen in time.

If I had read yout statement last month, I wouldn't ever bought my new system (i5-6600k + single iChill gtx1070).
Thanks god you wait until today.
I now have 5760x1080 everything maxed to the max, fps never goes below 90, but much of the time sits comfortably above 120. Even in Suzuka.
No stuttering no nothing.
My problems before were much probably caused by amd cpu+gpu.
Are you absolutely sure that your system (cpu+memory+mb) is able to feed the gtx1080 sli?

 

[WhiteShadow]

If I had read yout statement last month, I wouldn't ever bought my new system (i5-6600k + single iChill gtx1070).
Thanks god you wait until today.
I now have 5760x1080 everything maxed to the max, fps never goes below 90, but much of the time sits comfortably above 120. Even in Suzuka.
No stuttering no nothing.
My problems before were much probably caused by amd cpu+gpu.
Are you absolutely sure that your system (cpu+memory+mb) is able to feed the gtx1080 sli?

If you click that four colored icon below my user name you find out how my system is.
What did I wrote in my post # 393? This: "Triple sceens 120fps game play in rF2 is still not possible with descent in game settings(not all maxed out) with triple screens, single card or SLI." I think I made my point 100% clear, what is your point? Perhaps justify lack of rFactor2 missing SLI support or pure gMotor 2.5 graphics performance? Or is it something else?

 

[stonec]

If you click that four colored icon below my user name you find out how my system is.
What did I wrote in my post # 393? This: "Triple sceens 120fps game play in rF2 is still not possible with descent in game settings(not all maxed out) with triple screens, single card or SLI." I think I made my point 100% clear, what is your point? Perhaps justify lack of rFactor2 missing SLI support or pure gMotor 2.5 graphics performance? Or is it something else?

Tim Wheatley wrote recently in comments under this video:

Tim ”twhtly” Wheatley1:
Also, 60fps with a 770 is absolutely fine. 970 manages it with triples. 980Ti gets 120fps with triples and 1080 gets 160fps with triples. 700 series came out in 2013, which means if you wanted the performance of our 2013 release, you needed a 2013 card. The horror...

Yes, he is ISI employee, but would he need to promote their product through lies? Well, what do I know, I can't verify the fact on my 2012 system.

 

[cosimo]

My system with a GTX 780 and decent (your idea of decent may vary though?) settings can do 100-120 FPS with triple screens and multiview.

I usually keep track detail to Medium, shadows to medium/high and reflections low.

For me, having ~100+ FPS is far more important than ultra shadows and reflections.

I don't feel like rF2 is worse than other sims I have. Quite the opposite actually, if you consider the fully dynamic lighting.

Take your favourite car to Suzuka from 6pm with 5x-10x time acceleration if you haven't yet.

 

[WhiteShadow]

Tim Wheatley wrote recently in comments under this video:
Yes, he is ISI employee, but would he need to promote their product through lies? Well, what do I know, I can't verify the fact on my 2012 system.

Nobody is accusing Tim to be a liar, but have you never seen anybody from ISI to give technical support with GFX settings? It is easy to make end to all complains about rF2 graphics performance, it is just to post those amazing settings to technical forums, right? Or perhaps you can do that?

 

[Led566]

If you click that four colored icon below my user name you find out how my system is.
What did I wrote in my post # 393? This: "Triple sceens 120fps game play in rF2 is still not possible with descent in game settings(not all maxed out) with triple screens, single card or SLI." I think I made my point 100% clear, what is your point? Perhaps justify lack of rFactor2 missing SLI support or pure gMotor 2.5 graphics performance? Or is it something else?

My experience shows that triple screen 5760x1080@120 fps everything maxed except shadows, is perfectly possible with a single gtx1070.
Seems strange to me that your top notch system isn't able to do the same (or much, much better fwiw).
Enough clear now?

 

[TheGrunt]

My experience shows that triple screen 5760x1080@120 fps everything maxed except shadows, is perfectly possible with a single gtx1070.
Seems strange to me that your top notch system isn't able to do the same (or much, much better fwiw).
Enough clear now?

Since you are running rF2 so well, may I ask what amount of AI cars you are using?

I have a i7-3930K@4.5GHz wit single GTX980ti and 1080p triples and I definitely can't keep 60FPS locked even with reduced settings (mainly shadows medium, reflections off). That means depending on a car and track used, I can get FPS dips all the way down to around 50FPS with 20+ AI cars out on the track. This happens usually at the start when its crowded and track has most details present with all the grandstands and pits and other parts of the track usually run just fine. This is of course very car and track dependent, some combinations stay 60FPS locked all the time. Genreally, ISI cars and tracks are well optimized and have good performance, while not surprisingly some mods can be FPS hogs. For example Flat6 is notoriously hard on the FPS and when driving it, I usually drop shadows to low and it can keep FPS locked at 60 all the time with around 25 AI cars on every track.

Sure, with just few AI cars and without v-sync I can pull 100-200 FPS, but I like to race against more than 3-5 cars. In the end it is useless to compare our performance if we don't know our system specs or our ingame race/event settings including track and car models used.

All in all, rF2 graphics engine is really heavily CPU bound. On most cases when I get below 60 FPS, my GPU usage stays below 100%. PCIE 2.0 vs 3.0 performance difference tells something about CPU overhead too. That is also the reason, why graphics card upgrades don't necessarily lead to significantly improved performance.

 

[Led566]

Since you are running rF2 so well, may I ask what amount of AI cars you are using?

I have a i7-3930K@4.5GHz wit single GTX980ti and 1080p triples and I definitely can't keep 60FPS locked even with reduced settings (mainly shadows medium, reflections off). That means depending on a car and track used, I can get FPS dips all the way down to around 50FPS with 20+ AI cars out on the track. This happens usually at the start when its crowded and track has most details present with all the grandstands and pits and other parts of the track usually run just fine. This is of course very car and track dependent, some combinations stay 60FPS locked all the time. Genreally, ISI cars and tracks are well optimized and have good performance, while not surprisingly some mods can be FPS hogs. For example Flat6 is notoriously hard on the FPS and when driving it, I usually drop shadows to low and it can keep FPS locked at 60 all the time with around 25 AI cars on every track.

Sure, with just few AI cars and without v-sync I can pull 100-200 FPS, but I like to race against more than 3-5 cars. In the end it is useless to compare our performance if we don't know our system specs or our ingame race/event settings including track and car models used.

All in all, rF2 graphics engine is really heavily CPU bound. On most cases when I get below 60 FPS, my GPU usage stays below 100%. PCIE 2.0 vs 3.0 performance difference tells something about CPU overhead too. That is also the reason, why graphics card upgrades don't necessarily lead to significantly improved performance.

I've try with 15 AI.
Or online racing with 20+ people.
But you have a point, main reason for upgrade my system was to shift to intel cpu plus PCI 3.0.
And not because the graphic thread was too heavy on my amd fx6300 but the physics thread that was unbearable by my old cpu.
Interestingly enough you mentioned the Flat6 mod: that one was an eyeopener for me, because of the physic thread being over 90% the system was "sligthly" but constantly out of real time up to the point I was losing 1 to 2 seconds per lap. This was happening to other people too and they don't even noticed why they were so offpace compared to others...

 

[WhiteShadow]

My experience shows that triple screen 5760x1080@120 fps everything maxed except shadows, is perfectly possible with a single gtx1070.
Seems strange to me that your top notch system isn't able to do the same (or much, much better fwiw).
Enough clear now?

I've try with 15 AI.
Or online racing with 20+ people.
But you have a point, main reason for upgrade my system was to shift to intel cpu plus PCI 3.0.
And not because the graphic thread was too heavy on my amd fx6300 but the physics thread that was unbearable by my old cpu.
Interestingly enough you mentioned the Flat6 mod: that one was an eyeopener for me, because of the physic thread being over 90% the system was "sligthly" but constantly out of real time up to the point I was losing 1 to 2 seconds per lap. This was happening to other people too and they don't even noticed why they were so offpace compared to others...

You make a claim that your experience shows that triple screen 5760x1080@120 fps everything maxed except shadows, is perfectly possible with a single GTX 1070 and you haven't tested your new system with AI`s or online racing with other people. Come on mate, pull yourself together.

 

[Euskotracks]

You make a claim that your experience shows that triple screen 5760x1080@120 fps everything maxed except shadows, is perfectly possible with a single GTX 1070 and you haven't tested your new system with AI`s or online racing with other people. Come on mate, pull yourself together.

He is saying that he has tested with 15 AI and 20+ online. Can't you read?

Enviado desde mi ONE A2001 mediante Tapatalk

 

[Tuttle]

@Whiteshadow,

sorry but this is a first warning (means bypassing this would end in a temporary ban);

I mean ok, you have some issue with your system, but man... it's years you are posting the same story all over again and looks like you really don't care about any kind of reply. You keep going for hours, days, months and years (in the form of Antero). We've been VERY openhanded but now, please, let it go.

I think everybody over here got your point. I'm asking you to take a bit of a rest from that fight, especially now which forums seems way more stable than sometime ago.

Please, do not reply to this warning. If you think you have something to say more, please, contact me via PM.

Cheers.

 

[stonec]

Genreally, ISI cars and tracks are well optimized and have good performance, while not surprisingly some mods can be FPS hogs. For example Flat6 is notoriously hard on the FPS and when driving it, I usually drop shadows to low and it can keep FPS locked at 60 all the time with around 25 AI cars on every track.

In the end it is useless to compare our performance if we don't know our system specs or our ingame race/event settings including track and car models used.

Hence why it's useless to compare rF2 performance with mods, which is what many people do because they only race certain mods, which are not well or at all optimized. You can kill even the best GPU and the best DX11 racing sim engine in the world by making a mod with no optimization. It's enough to make an über high-poly car model, neglect any LOD work, add 4K resolution to all textures, add 4K mirrors. Done. That's why it's of no use to compare performance of ISI engine on most mods.

 

[Will Mazeo]

Hence why it's useless to compare rF2 performance with mods, which is what many people do because they only race certain mods, which are not well or at all optimized. You can kill even the best GPU and the best DX11 racing sim engine in the world by making a mod with no optimization. It's enough to make an über high-poly car model, neglect any LOD work, add 4K resolution to all textures, add 4K mirrors. Done. That's why it's of no use to compare performance of ISI engine on most mods.

Well... rF2 is a modding platform...

And to the other guy there, IMO any test with less than 30 AI is invalid, even on my i3 I'm checking stuff with 40 URD cars on track, come on xD