Performance Valve Index with different SS ratios

[chiliwili69]

One of the forum colleagues asked me to run the Index at different SS ratios and show the frametimes.

The testing conditions was Spitfire IX in summer Kuban map around Novorossiysk mountains at 600ft at 8:00 with no clouds.

Game settings : High Settings, Shadows High, Mirrors off, distant landscape x3, Horizon draw 100Km, Landscape filter off, Grass Normal, Clouds Medium, DRF=1, AA x2, gamma 0.8, 4K texttures, HDR&SSAO off.

My CPU is a 4790K running at 4.8GhZ with 1.3 Vcore. Fastest DDR3 RAM and 1080Ti.

 

I chose four SS levels (58%, 106%, 164%, 264%), just to test subsampling and extreme supersampling. Normally I use 140%.

There is a great visual gain going from 100% to 140%, but I have not taken through lens pictures, just frametimes:

 

The frametimes are shown using "Advanced Frame Timing" option in the Video settings of SteamVR.

The Valve Index was running at 80Hz, which quite OK for me, and with no "Motion Smoothing". So to be at 80fps you need 12.5ms frametimes (is 1/80)

 

You can see that the CPU is around 6ms since there is not too much action when taking snapshots (Maybe with 164% I was lower and a plane around).

But the important aspect to show here is the GPU frametimes, as expected the more SS the closer to 12.5. So you can see that at 264% I am at the very limit of 12.5 for an straight simple flight.

So based on this one can pick a SS ratio around 140%-160% (SteamVR recommend 150%).

But you need to take into account that there is no clouds here, which are the biggest influence in frametimes.

 

 

Edited October 20, 2019 by chiliwili69

[lurch9494]

On 10/20/2019 at 7:57 PM, chiliwili69 said:

One of the forum colleagues asked me to run the Index at different SS ratios and show the frametimes.

The testing conditions was Spitfire IX in summer Kuban map around Novorossiysk mountains at 600ft at 8:00 with no clouds.

Game settings : High Settings, Shadows High, Mirrors off, distant landscape x3, Horizon draw 100Km, Landscape filter off, Grass Normal, Clouds Medium, DRF=1, AA x2, gamma 0.8, 4K texttures, HDR&SSAO off.

My CPU is a 4790K running at 4.8GhZ with 1.3 Vcore. Fastest DDR3 RAM and 1080Ti.

 

I chose four SS levels (58%, 106%, 164%, 264%), just to test subsampling and extreme supersampling. Normally I use 140%.

There is a great visual gain going from 100% to 140%, but I have not taken through lens pictures, just frametimes:

 

The frametimes are shown using "Advanced Frame Timing" option in the Video settings of SteamVR.

The Valve Index was running at 80Hz, which quite OK for me, and with no "Motion Smoothing". So to be at 80fps you need 12.5ms frametimes (is 1/80)

 

You can see that the CPU is around 6ms since there is not too much action when taking snapshots (Maybe with 164% I was lower and a plane around).

But the important aspect to show here is the GPU frametimes, as expected the more SS the closer to 12.5. So you can see that at 264% I am at the very limit of 12.5 for an straight simple flight.

So based on this one can pick a SS ratio around 140%-160% (SteamVR recommend 150%).

But you need to take into account that there is no clouds here, which are the biggest influence in frametimes.

 

 

 

 

I have just ordered a RTX 2080 Ti (overclocked EVGA XC ULTRA Gaming) so I hope to be able to push the settings a little higher than with my 1080 (non Ti). The i9 9900k@5GHz is not the bottle neck now, but the 1080... It should arrive tomorrow, so lots of testing at the weekend ?

Edited October 24, 2019 by lurch9494
speeeling mishtake :)

[skarecrow]

I have the RTX2080 and run everything on ultra with no problem. Always 75 fps.  I7-8700 3.20 GHz

[lurch9494]

4 minutes ago, skarecrow said:

I have the RTX2080 and run everything on ultra with no problem. Always 75 fps.  I7-8700 3.20 GHz

 

Thats good to know  I just dropped a lot of cash on the 2080 Ti so hope it does well lol. The 1080 is struggling, I can get a solid 80fps but most settings are fairly low and Steam SS is only at 100% with the Index set at 80Hz refresh...

[dburne]

31 minutes ago, lurch9494 said:

 

 

I have just ordered a RTX 2080 Ti (overclocked EVGA XC ULTRA Gaming) so I hope to be able to push the settings a little higher than with my 1080 (non Ti). The i9 9900k@5GHz is not the bottle neck now, but the 1080... It should arrive tomorrow, so lots of testing at the weekend ?

 

Running an EVGA RTX 2080 Ti FTW3 with my i9 9900k at 5.1 GHz all cores, and get great results in IL-2 - 80 fps with Rift S practically all the time.

[lurch9494]

1 minute ago, dburne said:

 

Running an EVGA RTX 2080 Ti FTW3 with my i9 9900k at 5.1 GHz all cores, and get great results in IL-2 - 80 fps with Rift S practically all the time.

 

 

Thanks!  What settings are you running? Pretty high? SS?

[skarecrow]

I think you'll be happy with your investment.  I was using the 1080ti (which is a great GPU for sure) and just got tired of constant situational tweaking on flight sims. Like you, bit the bullet on the upgrade (cpu+gpu) and was the best choice I ever made.  Comes overclocked and I haven't bothered pushing with any "partner algorithms".  No need.  Sim runs smoothly when I see others struggling with new maps, multiplayer etc.  I've never had a stutter yet and i fly every day.  Maybe I'm just lucky. I also use a reshade tool as well with no problems.   This is all @ 1080 /75 refresh rate tho...not 1440.

[SCG_Fenris_Wolf]

1 hour ago, skarecrow said:

I have the RTX2080 and run everything on ultra with no problem. Always 75 fps.  I7-8700 3.20 GHz

 

What's the resolution please? Can you drop your settings quickly? What VR headset are you using?

[dburne]

3 hours ago, lurch9494 said:

 

 

Thanks!  What settings are you running? Pretty high? SS?

 

Yes high settings with SS at 1.4.

[chiliwili69]

10 hours ago, skarecrow said:

This is all @ 1080 /75 refresh rate tho...not 1440.

 

Do you mean monitor or VR?

11 hours ago, lurch9494 said:

I can get a solid 80fps but most settings are fairly low and Steam SS is only at 100% with the Index set at 80Hz refresh.

 

The in-game settings (except clouds perhaps) affects basically to CPU load. Basically the higher the more polygons/particles that the CPU has to build the scene.

The SS affects basically the GPU load. The more SS the more pixels that need to be rendered.

 

So with your 9900K @5.0Ghz you can already raise the game settings quite well.

With your new 2080Ti you will then be able to raise the SS bar a bit more.

 

Use the tool fpsVR to see the CPU and GPU frametimes while you are playing. So you can track both loads and adjust settings accordingly.

7 hours ago, dburne said:

Yes high settings with SS at 1.4

 

@lurch9494 please note that Dburne refers here to 1.4 PixelDensity in OculusTrayTool with RiftS at 80Hz. This is 11.7 Million pixels at 80Hz.

 

If you want the same GPU load at Index then you need to put 130% SS at SteamVR for the Index at 80Hz. This is 11.7 million pixels at 80Hz.

 

 

The cals are in the "SS per device" tab at this spreadsheet:

https://docs.google.com/spreadsheets/d/1gJmnz_nVxI6_dG_UYNCCpZVK2-f8NBy-y1gia77Hu_k

 

But my current 1080Ti handle well the 140% SS at SteamVR for Index at 80Hz as demonstrated in above pictures (but with no clouds). So I think your 2080Ti will be able to run 160%SS at SteamVR with Index at 80Hz.

Edited October 25, 2019 by chiliwili69

[lurch9494]

On 10/25/2019 at 8:15 AM, chiliwili69 said:

 

Do you mean monitor or VR?

 

The in-game settings (except clouds perhaps) affects basically to CPU load. Basically the higher the more polygons/particles that the CPU has to build the scene.

The SS affects basically the GPU load. The more SS the more pixels that need to be rendered.

 

So with your 9900K @5.0Ghz you can already raise the game settings quite well.

With your new 2080Ti you will then be able to raise the SS bar a bit more.

 

Use the tool fpsVR to see the CPU and GPU frametimes while you are playing. So you can track both loads and adjust settings accordingly.

 

@lurch9494 please note that Dburne refers here to 1.4 PixelDensity in OculusTrayTool with RiftS at 80Hz. This is 11.7 Million pixels at 80Hz.

 

If you want the same GPU load at Index then you need to put 130% SS at SteamVR for the Index at 80Hz. This is 11.7 million pixels at 80Hz.

 

 

The cals are in the "SS per device" tab at this spreadsheet:

https://docs.google.com/spreadsheets/d/1gJmnz_nVxI6_dG_UYNCCpZVK2-f8NBy-y1gia77Hu_k

 

But my current 1080Ti handle well the 140% SS at SteamVR for Index at 80Hz as demonstrated in above pictures (but with no clouds). So I think your 2080Ti will be able to run 160%SS at SteamVR with Index at 80Hz.

 

 

 

Thanks for this, I have spent most of today 'tweeking' my settings for the 2080 Ti. I used fpsVR for all the testing and flew over areas I knew would cause the highest 'strain' on the system inc very low level over the citys on the new BoBP map.

 

I am now at a point that with the 2080ti and the i9 9900k @ 5GHz they are both about equal and no actual bottle neck. The settings below are for the Index @ 80Hz and basically I dont see drops below 80 fps and it is buttery smooth. I could lower the settings a bit and run the Index higher (120 Hz is quite achievable) but 80 works flawlessly and I can have the eye candy almost maxed!

 

The image through the Index at these settings is close to stunning! not quite 1080p monitor, but really not that far off and a massive improvement over my old seup with the CV1

 

Steam VR SS @ 148

 

ULTRA
Screen Res 640x480

UI Scale Auto

Shadows LOW

Mirrors Simple

Distant Landscape Detail x4

Horizon Draw Distance 150km

Landscape Filter Sharp

Grass Quality Normal

Clouds Quality Low

Dynamic res Full

Antialiasing @2

SSAO unchecked

HDR Unchecked

Sharpen Checked

Use 4K Tax Checked

 

The end result is excellent and the 2080 Ti made a MUCH bigger improvement than I had expected!

Edited October 26, 2019 by lurch9494
speeeling mishtake :)

[chiliwili69]

On 10/26/2019 at 5:30 PM, lurch9494 said:

The image through the Index at these settings is close to stunning! not quite 1080p monitor, but really not that far off

 

I am glad you get a solid 80fps experience with those settings.

Regarding the 1080p monitor, I made some screenshoots comparison in another thread.

 

[lurch9494]

11 hours ago, chiliwili69 said:

 

I am glad you get a solid 80fps experience with those settings.

Regarding the 1080p monitor, I made some screenshoots comparison in another thread.

 

 

 

Very interesting! Your efforts are much appreciated chiliwili69 !! I must say, I am very pleased with the Index now I have fully upgraded my rig

[Alonzo]

Low clouds and shadows but Ultra overall settings? Seems a weird combination, I'm curious how you settled on that. The clouds previously looked like junk on anything less than High, but I haven't tested recently.

[WallterScott]

28.10.2019 в 11:11, chiliwili69 сказал:

I am glad you get a solid 80fps experience with those settings.

Does anyone get a solid 80Hz online? For example on in berloga? I have a very powerful system, but the game is not very smooth. there are settings for good online, any help.

[SCG_Fenris_Wolf]

On 11/1/2019 at 11:18 PM, WallterScott said:

Does anyone get a solid 80Hz online? For example on in berloga? I have a very powerful system, but the game is not very smooth. there are settings for good online, any help.

It's not you, everyone is suffering equally from it.

 

@lurch9494 How do the clouds look like on that setting? Aren't they creating any steps? What about the shadows, do they not flicker? If no, got to try this.

[lurch9494]

On 11/3/2019 at 12:07 AM, SCG_Fenris_Wolf said:

It's not you, everyone is suffering equally from it.

 

@lurch9494 How do the clouds look like on that setting? Aren't they creating any steps? What about the shadows, do they not flicker? If no, got to try this.

 

Generally I get a very smooth, solid 80 fps, but from time to time I can get a slight loss or minor stutter... Clouds are ok (not great) and no problems with shadows at all  Hope it helps. I have now overclocked my GPU (7500 MHz Ram and GPU at max 2100MHz) and gone to full custom 360 open loop water cooling for the CPU ( now @ 5.2 GHz with no AVX offset ) So my system is pretty beefy...

Edited November 4, 2019 by lurch9494
speeeling mishtake :)

[chiliwili69]

On 10/25/2019 at 9:15 AM, chiliwili69 said:

The in-game settings (except clouds perhaps) affects basically to CPU load

 

I have been experimenting this afternoon with shadows, mirrors and clouds to see how they load the GPU or the CPU.

All tests are in SinglePlayer QuickMission at Kuban map with sun, 2vs2 scenario, flying Spit IX at High Settings with SteamVR-SS 140%. Everything else like OP.

 

With no clouds, either Shadows at High or/and Mirror at complex the GPU load is not affected and I get 80fps 99% of the time (CPU frametimes 6-8 ms, GPU frametimes 8-11ms)

 

Then I put mirror off and shadows off and put clouds at High and wheater on average. Then the CPU load is basically the same (frametimes around 8ms), but the GPU load is at 70-80% and massive frame drops with GPU frametimes well above the 12.5ms.

 

So today I have understood that clouds are totally dependent on GPU performance. So if you want great clouds get a nice GPU and decrease your SS. At least before the promised update...!

[WallterScott]

03.11.2019 в 18:48, lurch9494 сказал:

Generally I get a very smooth, solid 80 fps, but from time to time I can get a slight loss or minor stutter... Clouds are ok (not great) and no problems with shadows at all  Hope it helps. I have now overclocked my GPU (7500 MHz Ram and GPU at max 2100MHz) and gone to full custom 360 open loop water cooling for the CPU ( now @ 5.2 GHz with no AVX offset ) So my system is pretty beefy...

Smooth, solid fps offline or online? In fast missions and campaign I have no problem.. But it is necessary to go online (fps subsides to 50-60 (berloga).

I have already built a PC that it hardly fits in my room. I organized two separate water circuits.. I went crazy and made a waterblock on the processor of silver 999)) i9 9900K 5.3 avx0 GHz, 2080Ti, 16 GB DDR4 4400, Samsung 960 Pro (M2) But this infection still stutters. The monitoring program fpsVR bottleneck - the time frame of the processor.

Edited November 4, 2019 by WallterScott

[lurch9494]

7 hours ago, WallterScott said:

Smooth, solid fps offline or online? In fast missions and campaign I have no problem.. But it is necessary to go online (fps subsides to 50-60 (berloga).

I have already built a PC that it hardly fits in my room. I organized two separate water circuits.. I went crazy and made a waterblock on the processor of silver 999)) i9 9900K 5.3 avx0 GHz, 2080Ti, 16 GB DDR4 4400, Samsung 960 Pro (M2) But this infection still stutters. The monitoring program fpsVR bottleneck - the time frame of the processor.

 

 

Im OK online too, I did get the odd stutter (fairly infrequent to be honest) and using fpsVR showed that the 2080 Ti and i9 were fairly similar and no big bottle neck from either. The CPU was always working a little harder, but not enough to call it a bottle neck. Next time I go online to play, I will try and run fpsVR and check things. It may be fps is dropping, but doing it smoothly and I am not noticing it during game play?? But I cant think it is dropping a lot or I would notice it...

 

Are you shutting down all un-needed software etc, I run a pretty stripped down OS (usually showing 0 to 1% occupancy at idle)...

[lurch9494]

Ok, just did a quick test with fpsVR again, offline both the CPU and GPU times are around 8ms each with occasional spikes (very few above 12) Online is a little higher, but still around 9-ish for both, but the spikes can be more numerous and slightly more break the 12ms point, but if you are playing and not checking the figures, its very smooth and nice, just the very rare micro stutter...

[Dirt_Merchant]

goodness @lurch9494, if you're getting stutters we're quite some time away from solving that problem entirely lol.

[lurch9494]

3 minutes ago, [_FLAPS_]Dirt_Merchant said:

goodness @lurch9494, if you're getting stutters we're quite some time away from solving that problem entirely lol.

 

I just managed to run my overclock to 5.4GHz @ 1.42v Vcore and did a quick flight!  It was superb! However, temps were NOT and would not run my cpu at that voltage for long anyway lol. But it did prove that CPU speed is still lacking on my system as I got no stutters @ 5.4 GHz... Possibly a fluke (but I am not running my CPU that hard again to find out LOL) I seem to have gotten a bit of a Golden CPU here

[chiliwili69]

11 hours ago, lurch9494 said:

my overclock to 5.4GHz @ 1.42v Vcore and did a quick flight!  It was superb! However, temps were NOT

 

I assumed you reach that with all cores and no AVX offset (that´s indeed a great CPU), but as you said, 1.42V is a high voltage so high power so high temp.

If you want to reach that 5.4 Ghz with less voltage try to disable some cores in the BIOS. Most likely with only 4 cores active (and Hyperthreading off) you would reach 5.4 with just 1.32V and lower temps. At work we were able to reach 5.3Ghz with just 1.26V with two cores and CPU air cooling: https://forum.il2sturmovik.com/topic/54687-disabling-some-cores-to-get-more-overclock/

 

 

11 hours ago, lurch9494 said:

But it did prove that CPU speed is still lacking on my system as I got no stutters @ 5.4 GHz..

 

If you really solved the strange sttuters issues online you could report your experience in these threads that Jason created:

https://forum.il2sturmovik.com/topic/55320-stuttering-in-single-player-research/

https://forum.il2sturmovik.com/topic/55322-stutters-in-multiplayer-research/

 

Perhaps reaching 5.4Ghz is not an option for everyone, but if it solves the issue it could give clues about the origin of the problem.

[dburne]

14 hours ago, lurch9494 said:

 

I just managed to run my overclock to 5.4GHz @ 1.42v Vcore and did a quick flight!  It was superb! However, temps were NOT and would not run my cpu at that voltage for long anyway lol. But it did prove that CPU speed is still lacking on my system as I got no stutters @ 5.4 GHz... Possibly a fluke (but I am not running my CPU that hard again to find out LOL) I seem to have gotten a bit of a Golden CPU here

 

I would be very cautious and watch what vcore is being applied under load.

5.1 GHz on all cores no AVX seems to be the sweet spot for my 9900k, yeah I can get it to 5.2 but am uncomfortable

with both the vcore required and core temps under load.

 

Edited November 5, 2019 by dburne

[lurch9494]

2 hours ago, dburne said:

 

I would be very cautious and watch what vcore is being applied under load.

5.1 GHz on all cores no AVX seems to be the sweet spot for my 9900k, yeah I can get it to 5.2 but am uncomfortable

with both the vcore required and core temps under load.

 

 

I think I am pretty lucky, I can get a solid, no AVX offset  5.2GHz at a Vcore of 1.305 with maximum peak temps of 62 degs with my new custom 360 loop. But beyond 5.2GHz it needs silly volts to be stable. At 1.305v the vdroop is about 1.285v under full synthetic load (AVX) . Strangely, I need a Vcore of 1.295 for 5.1GHz with similar temps, so 5.2 seems to be my sweet spot. I have a second 360 radiator and second pump to fit as well as a better waterblock and thermal compound, so I may be able to drop the temps a little more yet too

4 hours ago, chiliwili69 said:

 

I assumed you reach that with all cores and no AVX offset (that´s indeed a great CPU), but as you said, 1.42V is a high voltage so high power so high temp.

If you want to reach that 5.4 Ghz with less voltage try to disable some cores in the BIOS. Most likely with only 4 cores active (and Hyperthreading off) you would reach 5.4 with just 1.32V and lower temps. At work we were able to reach 5.3Ghz with just 1.26V with two cores and CPU air cooling: https://forum.il2sturmovik.com/topic/54687-disabling-some-cores-to-get-more-overclock/

 

 

 

If you really solved the strange sttuters issues online you could report your experience in these threads that Jason created:

https://forum.il2sturmovik.com/topic/55320-stuttering-in-single-player-research/

https://forum.il2sturmovik.com/topic/55322-stutters-in-multiplayer-research/

 

Perhaps reaching 5.4Ghz is not an option for everyone, but if it solves the issue it could give clues about the origin of the problem.

 

 

To be honest, 5.4GHz is not a serious option for me. I am pretty happy with the 5.2GHz all cores / 16 threads no AVX offset with Vcore @ 1.305 and temps maxing at 62 (highest peak temp on an hour long AVX enabled synthetic test). The trouble with my test at 5.4 is I dont want to risk my CPU by repeating it, so the lack of stutters on a short flight does not mean much... But I will reprt it tonight when I get home

 

 

Edited November 5, 2019 by lurch9494

[dburne]

18 minutes ago, lurch9494 said:

 

I think I am pretty lucky, I can get a solid, no AVX offset  5.2GHz at a Vcore of 1.305 with maximum peak temps of 62 degs with my new custom 360 loop. But beyond 5.2GHz it needs silly volts to be stable. At 1.305v the vdroop is about 1.285v under full synthetic load (AVX) . Strangely, I need a Vcore of 1.295 for 5.1GHz with similar temps, so 5.2 seems to be my sweet spot. I have a second 360 radiator and second pump to fit as well as a better waterblock and thermal compound, so I may be able to drop the temps a little more yet 

 

To be honest, 5.4GHz is not a serious option for me. I am pretty happy with the 5.2GHz all cores / 16 threads no AVX offset with Vcore @ 1.305 and temps maxing at 62 (highest peak temp on an hour long AVX enabled synthetic test). The trouble with my test at 5.4 is I dont want to risk my CPU by repeating it, so the lack of stutters on a short flight does not mean much... But I will reprt it tonight when I get home

 

 

 

Yeah sounds like you got a very sweet chip there.

[lurch9494]

Just now, dburne said:

 

Yeah sounds like you got a very sweet chip there.

 

 

It makes a nice change, my previous CPU ( 6600k ) was a turnip and terrible for O/Clocking  LOL

[chiliwili69]

19 hours ago, lurch9494 said:

I have a second 360 radiator and second pump to fit as well as a better waterblock and thermal compound, so I may be able to drop the temps a little more yet too

Hey, what a beautiful open case and water circuit! it is as big as the whole PC. I didn´t know about open case PCs. Isn´t the dust a problem?

19 hours ago, lurch9494 said:

The trouble with my test at 5.4 is I dont want to risk my CPU by repeating it,

With that custom loop, you can really try many things. Did you try with only with two or four cores active at 5.4? (it will be safe I think)

[lurch9494]

1 hour ago, chiliwili69 said:

Hey, what a beautiful open case and water circuit! it is as big as the whole PC. I didn´t know about open case PCs. Isn´t the dust a problem?

With that custom loop, you can really try many things. Did you try with only with two or four cores active at 5.4? (it will be safe I think)

 

 

Not yet, but I was thinking about dropping the HT as 8 cores on their own is plenty to be honest. I have a second 360 radiator and pump to add to the loop also, plus a better (possibly better lol) water block. I also used some cheap thermal paste I had hanging around that was pretty old, so I am going to get some Thermal Grizzly Cryonaught paste to try. I hope to lower the temps another 5+ degs from an average peak of 60 to 55 or lower. I might then up the vCore to a max of 1.35v (my personal sfe limit for 24/7 oc'ing)

 

All that may alow me to run 5.3 or higher... or not lol

[WallterScott]

04.11.2019 в 23:07, lurch9494 сказал:

Ok, just did a quick test with fpsVR again, offline both the CPU and GPU times are around 8ms each with occasional spikes (very few above 12) Online is a little higher, but still around 9-ish for both, but the spikes can be more numerous and slightly more break the 12ms point, but if you are playing and not checking the figures, its very smooth and nice, just the very rare micro stutter...

I specifically said about the berloga server, as the servers are different. In the clear sky and on an empty airfield there are no problems, everything is really smooth. But it is necessary to arrange dogfight, which will track shots, debris, smoke fps drops to 50 and even help save 5.3 GHz (I think 5.4 is not a big difference). In berloga server you can quickly fly into a mass battle, there quickly check the performance of online mode. If you don't mind, try it there

Today we will take a screenshot of the fpsVR  and it can be seen that the toughest fights are not rare spikes (cpu fr ) and is very frequent.

P S If your 5.4 GHz processor holds a stresstest at least Aida FPU (enough for games) then it is truly Golden

[lurch9494]

50 minutes ago, WallterScott said:

I specifically said about the berloga server, as the servers are different. In the clear sky and on an empty airfield there are no problems, everything is really smooth. But it is necessary to arrange dogfight, which will track shots, debris, smoke fps drops to 50 and even help save 5.3 GHz (I think 5.4 is not a big difference). In berloga server you can quickly fly into a mass battle, there quickly check the performance of online mode. If you don't mind, try it there

Today we will take a screenshot of the fpsVR  and it can be seen that the toughest fights are not rare spikes (cpu fr ) and is very frequent.

 

 

Hi, OK I will try and get on tonight for a quick test if I have time, if not asap

2 hours ago, chiliwili69 said:

Hey, what a beautiful open case and water circuit! it is as big as the whole PC. I didn´t know about open case PCs. Isn´t the dust a problem?

 

 

 

The case is great, so easy to work on the PC and very versatile, lots of mounting options and space to play with  Dust isnt much of an issue (though only been using it a few days, but my old NZXT (normal case) was always getting dust in it and looked bad withing a few days due to the high pressure fans etc...

 

At least this case is fast n easy to clean

[Alonzo]

8 hours ago, lurch9494 said:

Not yet, but I was thinking about dropping the HT as 8 cores on their own is plenty to be honest. I have a second 360 radiator and pump to add to the loop also, plus a better (possibly better lol) water block. I also used some cheap thermal paste I had hanging around that was pretty old, so I am going to get some Thermal Grizzly Cryonaught paste to try. I hope to lower the temps another 5+ degs from an average peak of 60 to 55 or lower. I might then up the vCore to a max of 1.35v (my personal sfe limit for 24/7 oc'ing)

 

All that may alow me to run 5.3 or higher... or not lol

 

You definitely do not need hyper threading for IL2. It doesn't hurt except that you might get a higher overclock with it disabled. 

 

For me I was getting stutters but then upgraded to whatever that recent Windows patch was (one of the numbered ones) and it fixed a lot of problems. I might be getting to the point where I need to do a clean O/S install, it's been well over a year since I last did it.

[lurch9494]

1 hour ago, Alonzo said:

 

You definitely do not need hyper threading for IL2. It doesn't hurt except that you might get a higher overclock with it disabled. 

 

For me I was getting stutters but then upgraded to whatever that recent Windows patch was (one of the numbered ones) and it fixed a lot of problems. I might be getting to the point where I need to do a clean O/S install, it's been well over a year since I last did it.

 

Yeah, I think HT is pretty useless for me, I dont have any software that really needs more than the 8 physical cores...

[lurch9494]

12 hours ago, WallterScott said:

I specifically said about the berloga server, as the servers are different. In the clear sky and on an empty airfield there are no problems, everything is really smooth. But it is necessary to arrange dogfight, which will track shots, debris, smoke fps drops to 50 and even help save 5.3 GHz (I think 5.4 is not a big difference). In berloga server you can quickly fly into a mass battle, there quickly check the performance of online mode. If you don't mind, try it there

Today we will take a screenshot of the fpsVR  and it can be seen that the toughest fights are not rare spikes (cpu fr ) and is very frequent.

P S If your 5.4 GHz processor holds a stresstest at least Aida FPU (enough for games) then it is truly Golden

 

Hi  Just did a 20 min session on Berloga server, got 3 kills and mixed it up a bit, but I had no stutters at all? (maybe one, but I was in a tight turn fight and wasnt sure if I had a stutter or not) This was at 5.2 GHz. I will probably have a go at disabling HT and 4 cores and try 5.4 or 5.5GHz if possible at the weekend when I have more time to stress test and check temps properly

 

Edited November 6, 2019 by lurch9494
speeeling mishtake :)

[dburne]

I would keep a close eye on voltages/core temps under load.

[WallterScott]

57 минут назад, lurch9494 сказал:

Hi  Just did a 20 min session on Berloga server, got 3 kills and mixed it up a bit, but I had no stutters at all

Wow.. And what settings (in steam and in the game)?

What would overclocking to 5.4-5.5 remove the CPU cover. Put the waterblock right on top of crystal. Instead of a paste - liquid metal. Minus 5-6 degrees.

Edited November 6, 2019 by WallterScott

[lurch9494]

7 hours ago, WallterScott said:

Wow.. And what settings (in steam and in the game)?

What would overclocking to 5.4-5.5 remove the CPU cover. Put the waterblock right on top of crystal. Instead of a paste - liquid metal. Minus 5-6 degrees.

 

Hi, I used the same settings as before (below)

 

ULTRA
Screen Res 640x480

UI Scale Auto

Shadows LOW

Mirrors Simple

Distant Landscape Detail x4

Horizon Draw Distance 150km

Landscape Filter Sharp

Grass Quality Normal

Clouds Quality Low

Dynamic res Full

Antialiasing @2

SSAO unchecked

HDR Unchecked

Sharpen Checked

Use 4K Tex Checked

 

Steam SS 150%, motion smoothing off

 

I am going to try what @chiliwilly suggested, turning of HT and a couple of cores (six physical cores is still plenty for iL2) That should enable me to drop the voltage or at least leave it the same and have extra headroom to up the multiplier higher  It may or may not work lol

8 hours ago, dburne said:

I would keep a close eye on voltages/core temps under load.

 

Hi dburne

 

At 5.2 GHz (No AVX offset) I am running vCore @ 1.305v, vdroop is 1.285v and max peak temps is 62 deg under full sysnthetic load (AVX). So pretty happy with that. Once I install the second rad an pump along with some better thermal paste, I hope to lower the temps further...

Edited November 7, 2019 by lurch9494
speeeling mishtake :)

[chiliwili69]

1 hour ago, lurch9494 said:

six physical cores is still plenty for iL2

 4 cores is also plenty for IL-2 VR.

My 4790K has only 4 physical cores and HT off. No problem at all with IL-2 running at 4.8GHz in SP mode and my settings.

[lurch9494]

2 minutes ago, chiliwili69 said:

 4 cores is also plenty for IL-2 VR.

My 4790K has only 4 physical cores and HT off. No problem at all with IL-2 running at 4.8GHz in SP mode and my settings.

 

Indeed, but I run a few other applications / games that would benefit from having 6 cores, so I am hoping to find a setup that will work well on all so I dont have to keep loading a different bios  lol