Best CPU / Rig build for G2

[ACG_PanzerVI]

Ok, I preordered the G2 and am looking for advice on the rig I'm going to build to run it.  I currently have an old I5-4690K pushing 16gb of ram on an Asrock board, Founders RTX 2070 Super, driving a 35" ACER Predator ultrawide monitor at 3440 x 1440.  Been using MSAA and no, spotting is not great, but the sim is beautiful.  Not sure my FPS, but I have few/zero issues with it.  

 

My current plan based on reviewer on MRTV is to carry my 2070S forward.  I'll convert my old rig to business-only, and build a new one to run the VR.  I can go fairly high end here, per others comments on their hobby budgets, but I don't want to buy stupid.  That said, I like buying new hardware as infrequently as possible, so 5 years is my target lifespan for the rig, with a potential GPU upgrade in there.  If I overclock, it will be the simplest, most conservative OC I can figure out.  I like simple, stable, forget-its-on OC solutions.  Would like advice on:

 

CPU selection:  10th Gen Intel I5 / I7 vs. AMD 3700x 3900x

Motherboard and Chipset

Other considerations

 

Plan is for 32GB of ram regardless and a 2TB SSD and a 750w Gold PSU.    I'll put it all in a nice box. Any thoughts, with rationale, appreciated.

[ZachariasX]

Well, if a later CPU upgrade is desired, you’d have to go with a recent AMD rig. Intel Socket 1200 will just see one more upgrade before it is ditched for Socket 1700.

 

Alas, the 10900K would probably give you some of the best performance for the game today. Whether the small lead makes a difference for the G2 with IL2 remains to be seen.

[chiliwili69]

Your question is mostly also applicable to all modern VR headsets. A more powerfull PC will give you more % of time at full native fps in complex scenarios.

 

We have been studying in the past what are the main factors to obtain better VR performance. The last series of tests was conducted with the Remagen becnhmark, but it doesn´t work anymore with latest game updates with referred rendering (recorded flights became unplayable).

 

Ideally the game could have an always playable recorded flight, so we would be always able to use a common track to measure performance over the same baseline, but it seems it is not on the top priorities of the development plans. I created a poll to know if this is important and people has same opinon than me. With a common benchmark it would be easier to compare performance of different headsets, hardware, software and settings. So all tricks and wisdom would be compiled and we will be able to take the best decissions when acquiring new hardware. @Jason_Williams do you think this would be important for the VR (and non-VR) community?

 

After this intro, let see were is wise to invest money for a PC for the Reverb G2. Assuming you only plays IL-2 VR:

 

1. Intel vs AMD. In all the test we have done Intel was always the winner. It is true that new AMD CPUs could change this, but current chips don´t. So forget AMD until someone prove the opposite

2. Intel CPU: Overclock is a must, so go for a K version. Forget the "X" chips since they are more difficult to overclock. Current 10th gen is incresing the number of cores but for IL-2 VR you will not need more than 4 cores. So you can go to 10600K, or 10700K or 10900K., all of them have 2Mb cache per core (we suspect that cache plays also a role).  Personally I would go to 10600K, having the option to go to higher versions if other demostrate that they are better for IL-2 VR.

3. CPU cooling: Go for a good AIO liquid cooling of 240 or 360mm. They are worth the money. Put it at the top of the case. An inlcude 3 or 4 more fans in the case.

4. RAM: IL-2 VR RAM usage is around 7-8 Mb. So please, don´t spend money on size (32GB) and go for speed. Try to go to 16 (2x8) RAM at 4400MHz (or higher) with lowest Cas Latency (CL). So learnt that frequency (the higher the better) is better than CL.

5. Motherboard: Look for a known brand with capability for high RAM frequencies. And with easy software for overclock (like MSI Command Center). But I think any brand would be OK.

6. GPU: Don´t spend a fortune in the lastest top GPU, probably they are not worth the money. For the G2 I would try first with a 1080Ti or 2080S. You can always upgrade GPU easily.

 

my two cents.

 

[I.JG3_CDRSEABEE]

3 hours ago, chiliwili69 said:

6. GPU: Don´t spend a fortune in the lastest top GPU, probably they are not worth the money. For the G2 I would try first with a 1080Ti or 2080S. You can always upgrade GPU easily.

Do you think a 1080ti can drive a G2? I have a I7 7700k not overclocked and currently running a O+ at 130SS and get decent FR's with the 1080ti.

 

I dont think its going to push the G2 with 100SS though. I may be wrong. Im looking at the 3080 but waiting on the benchmarks.

[Remontti]

54 minutes ago, CDRSEABEE said:

Do you think a 1080ti can drive a G2? I have a I7 7700k not overclocked and currently running a O+ at 130SS and get decent FR's with the 1080ti.

 

I dont think its going to push the G2 with 100SS though. I may be wrong. Im looking at the 3080 but waiting on the benchmarks.

I have RTX2080Ti and Reverb G1. My fps drops below 90 from time to time with High settings. SteamVR shows that it is usually the graphics card that can't keep up when getting fps drops.

[dburne]

8 minutes ago, Remontti said:

I have RTX2080Ti and Reverb G1. My fps drops below 90 from time to time with High settings. SteamVR shows that it is usually the graphics card that can't keep up when getting fps drops.

 

That is certainly good to hear. Wondering how the performance might be with my upcoming Reverb G2.

[I.JG3_CDRSEABEE]

1 hour ago, dburne said:

 

That is certainly good to hear. Wondering how the performance might be with my upcoming Reverb G2.

Isnt it the same amount of pixels? G1 = G2 pixel wise? 

1 hour ago, Remontti said:

I have RTX2080Ti and Reverb G1. My fps drops below 90 from time to time with High settings. SteamVR shows that it is usually the graphics card that can't keep up when getting fps drops.

How is plane spotting?

[chiliwili69]

3 hours ago, CDRSEABEE said:

Do you think a 1080ti can drive a G2?

 

About a year ago I was testing the Reverb G1: 

 

 

The performance was quite OK for the settings I was using (9.3 Million pixels). But since then the game has been updated several times and now I think there is more load on GPU.

Probably the 1080Ti is on the limit, but I should see that in a month or so...

1 hour ago, CDRSEABEE said:

Isnt it the same amount of pixels? G1 = G2 pixel wise?

yes is the same

2 hours ago, Remontti said:

High settings

 What settings for clouds, shadows and AA? (They eat a lot GPU)

[I.JG3_CDRSEABEE]

44 minutes ago, chiliwili69 said:

About a year ago I was testing the Reverb G1: 

I get really confused with all that post. lol But thanks for all the work.

 

What do you think about the new cards? Should one just get the 3090 and pay the big price or go for a 3080? I know the benchmarks are not out but would it be beneficial just to get the best available and be set for four years? Or is there overkill ( GPU) for IL2 at this point?   I have a I7 7700k not OC'd yet? The CPU seems to not be working over 35% running my current settings .

 

I plan on upgrading the GPU and want to max out groundwork clouds etc. So just debating which to get.

 

Thanks

[Remontti]

On 8/25/2020 at 5:21 PM, CDRSEABEE said:

How is plane spotting?

I think the spotting is alright. Identifying planes is also good with the new zoom.

 

On 8/25/2020 at 6:50 PM, chiliwili69 said:

 

 

 What settings for clouds, shadows and AA? (They eat a lot GPU)

For now I have High clouds, medium shadow and 2xFXAA. And 60Hz mode for Reverb. 

I was little disappointed that RTX2080Ti can't run it flawless but it is what it is. Mostly it runs ok but in certain missions you get some severe fps drops. Like Sea Dragon campaign where there is fire and smoke everywhere and all kind of action around you.

 

[chiliwili69]

1 hour ago, Remontti said:

For now I have High clouds, medium shadow and 2xFXAA. And 60Hz mode for Reverb. 

I was little disappointed that RTX2080Ti can't run it flawless but it is what it is. Mostly it runs ok but in certain missions you get some severe fps drops. Like Sea Dragon campaign where there is fire and smoke everywhere and all kind of action around you

 

I am really surprised that a 2080Ti can not handle 60fps always with your moderate settings for the Reverb (g1). I assume that you don´t apply any extra Supersampling to the native resolution of reverb.

 

Have you checked with fpsVR when you get fps drops is due the GPU frametimes are below 16.6 ms?

What is the speed of your RAM?

 

It seems that the new referred redenring has loaded the GPUs more than I thought.

[Gomoto]

Best CPU for G2 has yet to be invented. All people jumping on the 3080/3090 bandwagon for VR will have a unpleasant surprise.

 

My 1080Ti can easily drive 200% and more resolution of the G1 in low quality 1.0 scale rendering or ultra quality in 0.7 scale rendering.

Why is this? Because we are totally CPU bound in VR.

 

A 3080/3090 will drive 500% or 600% resolution of the G2, but FPS will still be bound by your CPU performance and for ultra quality at 90 FPS you will have to drop rendering scale until faster CPU's will come available.

[Alonzo]

40 minutes ago, Gomoto said:

Best CPU for G2 has yet to be invented. All people jumping on the 3080/3090 bandwagon for VR will have a unpleasant surprise.

 

My 1080Ti can easily drive 200% and more resolution of the G1 in low quality 1.0 scale rendering or ultra quality in 0.7 scale rendering.

Why is this? Because we are totally CPU bound in VR.

 

A 3080/3090 will drive 500% or 600% resolution of the G2, but FPS will still be bound by your CPU performance and for ultra quality at 90 FPS you will have to drop rendering scale until faster CPU's will come available.

 

CPU is not currently a limitation for most scenarios, including multiplayer, if you have a 5ghz Intel. 600% of the resolution of the G2 would be 2160 x 2160 x 2 (eyes) x 90hz x 6 (hundred percent). That's 5 billion pixels per second.

 

You're not making much sense here. I agree that people expecting a 3080 or 3090 to be magical will be disappointed, but we are not CPU bound in VR. We've measured it and proven it to be true.

[Gomoto]

I know and you will see

[Remontti]

8 hours ago, chiliwili69 said:

 

I am really surprised that a 2080Ti can not handle 60fps always with your moderate settings for the Reverb (g1). I assume that you don´t apply any extra Supersampling to the native resolution of reverb.

 

Have you checked with fpsVR when you get fps drops is due the GPU frametimes are below 16.6 ms?

What is the speed of your RAM?

 

It seems that the new referred redenring has loaded the GPUs more than I thought.

I can't maintain 90fps so I use the 60Hz mode. Mostly it is 60 fps. At least with quick missions and PWCG.

 

Yesterday I was flying on Combat Box and fps was in the 35-50 with a rather quick mission at low altitude. Made a track of it and viewing the track I got 57-60 fps. I suppose I was limited by Cpu there.

 

I changed my settings a bit:

Fxaa off

Clouds medium (no clouds on that Combat box mission)

Disabled 4k textures

 

Now with these settings I was getting 60 fps in the track. I joined back to Combat Box and now I were getting constant 60 fps. MP performance in VR is total mystery to me. Lol.

 

I think I have 16Gb DDR4 2666.

Super Sampling is at 100%

I've not used fpsVR. I've used steamVR setting that shows a colored box in the corner if you are cpu or gpu limited. It was disabled now.

 

 

 

 

[jollyjack]

On 8/25/2020 at 10:02 AM, chiliwili69 said:

Your question is mostly also applicable to all modern VR headsets. A more powerfull PC will give you more % of time at full native fps in complex scenarios.

 

We have been studying in the past what are the main factors to obtain better VR performance. The last series of tests was conducted with the Remagen becnhmark, but it doesn´t work anymore with latest game updates with referred rendering (recorded flights became unplayable).

 

Ideally the game could have an always playable recorded flight, so we would be always able to use a common track to measure performance over the same baseline, but it seems it is not on the top priorities of the development plans. I created a poll to know if this is important and people has same opinon than me. With a common benchmark it would be easier to compare performance of different headsets, hardware, software and settings. So all tricks and wisdom would be compiled and we will be able to take the best decissions when acquiring new hardware. @Jason_Williams do you think this would be important for the VR (and non-VR) community?

 

After this intro, let see were is wise to invest money for a PC for the Reverb G2. Assuming you only plays IL-2 VR:

 

1. Intel vs AMD. In all the test we have done Intel was always the winner. It is true that new AMD CPUs could change this, but current chips don´t. So forget AMD until someone prove the opposite

2. Intel CPU: Overclock is a must, so go for a K version. Forget the "X" chips since they are more difficult to overclock. Current 10th gen is incresing the number of cores but for IL-2 VR you will not need more than 4 cores. So you can go to 10600K, or 10700K or 10900K., all of them have 2Mb cache per core (we suspect that cache plays also a role).  Personally I would go to 10600K, having the option to go to higher versions if other demostrate that they are better for IL-2 VR.

3. CPU cooling: Go for a good AIO liquid cooling of 240 or 360mm. They are worth the money. Put it at the top of the case. An inlcude 3 or 4 more fans in the case.

4. RAM: IL-2 VR RAM usage is around 7-8 Mb. So please, don´t spend money on size (32GB) and go for speed. Try to go to 16 (2x8) RAM at 4400MHz (or higher) with lowest Cas Latency (CL). So learnt that frequency (the higher the better) is better than CL.

5. Motherboard: Look for a known brand with capability for high RAM frequencies. And with easy software for overclock (like MSI Command Center). But I think any brand would be OK.

6. GPU: Don´t spend a fortune in the lastest top GPU, probably they are not worth the money. For the G2 I would try first with a 1080Ti or 2080S. You can always upgrade GPU easily.

 

my two cents.

 

 

Good advice, and i'd go for Intel i7, having (and had) 3 PCs and a laptop using them. Had a burnt AMD once. NoMo AMD CPUs.

I think a watercooler is best choice, and i found that my now 1 year old 9700k produces quite a bit of heat running at 43 Celsius averagely.

Watercooler benefit is also that you can set all fans to blow out, and have one intake with a large (cleanable) dust-filter. I noticed no dust in my case for a year now.

 

Edited August 27, 2020 by jollyjack

[chiliwili69]

5 hours ago, Gomoto said:

I know and you will see

 

This a quite good scientific reason...

As Carl Sagan said: Extraordinary claims require extraordinary evidence

27 minutes ago, Remontti said:

Yesterday I was flying on Combat Box and fps was in the 35-50 with a rather quick mission at low altitude. Made a track of it and viewing the track I got 57-60 fps. I suppose I was limited by Cpu there.

 

With the Remagen test we demostrated that a recorded flight produce almost the same fps than the live game in single player. If I were you I will study first your performance in just SP to know exactly how bottlenecked is your 2080Ti with the Reverb G1. 

SteamVR frametimes window is OK but fpsVR is much handy to use and you have a HUD inside the game where you can see instantly how your CUP or GPU frametimes change when you look at the smoke or at the clean sky, and how far you are from the limit. If you spent that $ in a 2080Ti you can spend 4$ for fpsVR, you will not regret it.

 

My point is that with your 2080Ti and your settings you should not be bottlenecked by GPU at 60Hz. If you are below that when there is a lot of smoke or fire it could be due to CPU (to handle all those particles). You can test that in Stalingrad city flying low inside all the smoke of the burning oil tanks.

[chiliwili69]

4 hours ago, Remontti said:

16Gb DDR4 2666

 

This could be also a weak point in your PC. You have the current top GPU but in memory you are not at that level.

See what frequencies support your MoBo. If they support 3200 or more it could be worth to upgrade to gain some performance. The higher the better.

[Remontti]

1 hour ago, chiliwili69 said:

 

This could be also a weak point in your PC. You have the current top GPU but in memory you are not at that level.

See what frequencies support your MoBo. If they support 3200 or more it could be worth to upgrade to gain some performance. The higher the better.

I might just do that since memory isn't that expensive.

 

My biggest problem with VR is that I actually like my monitor more. I tend to fly in VR for week or two and then go back to monitor. 

[JG7_X-Man]

I usually build once every 2-3 yrs, so I have to go for the top shelf performance because my box has to last me awhile. This was going to be a build year but I am postponing for MS 2020, Intel motherboards that run PCIe 4.0, the supply chain to pick back up as COVID-19 impact subsides and we figure out what the new normal is and to top it all off, I moved.

 

So as you can see from my specs below, going high-end on a build gives you the advantage of being able to hold out for a longer time without sacrificing performance.

 

 

 

 

 

[dburne]

Yeah with the exception of video card and possibly ram, I normally get a build to last around 4-5 years.

This rig I built in Jan 2019 replaced a previous build from 2013.

[ACG_PanzerVI]

Thanks very much for all of this.  Very valuable input that I will heed in my process.

 

 

Your question is mostly also applicable to all modern VR headsets. A more powerfull PC will give you more % of time at full native fps in complex scenarios.

 

We have been studying in the past what are the main factors to obtain better VR performance. The last series of tests was conducted with the Remagen becnhmark, but it doesn´t work anymore with latest game updates with referred rendering (recorded flights became unplayable).

 

Ideally the game could have an always playable recorded flight, so we would be always able to use a common track to measure performance over the same baseline, but it seems it is not on the top priorities of the development plans. I created a poll to know if this is important and people has same opinon than me. With a common benchmark it would be easier to compare performance of different headsets, hardware, software and settings. So all tricks and wisdom would be compiled and we will be able to take the best decissions when acquiring new hardware. @Jason_Williams do you think this would be important for the VR (and non-VR) community?

 

After this intro, let see were is wise to invest money for a PC for the Reverb G2. Assuming you only plays IL-2 VR:

 

1. Intel vs AMD. In all the test we have done Intel was always the winner. It is true that new AMD CPUs could change this, but current chips don´t. So forget AMD until someone prove the opposite

2. Intel CPU: Overclock is a must, so go for a K version. Forget the "X" chips since they are more difficult to overclock. Current 10th gen is incresing the number of cores but for IL-2 VR you will not need more than 4 cores. So you can go to 10600K, or 10700K or 10900K., all of them have 2Mb cache per core (we suspect that cache plays also a role).  Personally I would go to 10600K, having the option to go to higher versions if other demostrate that they are better for IL-2 VR.

3. CPU cooling: Go for a good AIO liquid cooling of 240 or 360mm. They are worth the money. Put it at the top of the case. An inlcude 3 or 4 more fans in the case.

4. RAM: IL-2 VR RAM usage is around 7-8 Mb. So please, don´t spend money on size (32GB) and go for speed. Try to go to 16 (2x8) RAM at 4400MHz (or higher) with lowest Cas Latency (CL). So learnt that frequency (the higher the better) is better than CL.

5. Motherboard: Look for a known brand with capability for high RAM frequencies. And with easy software for overclock (like MSI Command Center). But I think any brand would be OK.

6. GPU: Don´t spend a fortune in the lastest top GPU, probably they are not worth the money. For the G2 I would try first with a 1080Ti or 2080S. You can always upgrade GPU easily.

 

my two cents.

Edited August 28, 2020 by ACG_PanzerV
Add Quote.

[dburne]

I would not build a system today with 16 GB of ram.

Granted if IL-2 is the only game you will ever play, that will work at least for the near future. There may be a time in the future that it will use more.

However other games can and will certainly make use of more, for instance DCS from what I have seen does.

It does not hurt to have more ram than needed, well except for the pocketbook. 

 

When I built my current rig in Jan 2019, I put 32 GB in to be safe.

[chiliwili69]

I would not build a system today with 32Gb of ram if the purpose is IL-2 VR. I also tested other VR games like RoboRecall, EliteDangerus, etc and they run ok.

My current system has 16Gb and my son can ran flawlessly any new game he is playing (Battlefield V, Apex, Overwatch, Unturned, etc).

I also doubt IL-2 VR would require more than 16Gb in the near future since it will reduce the number of potential users, since only 8.6% (according to Steam surverys) has more than 16Gb:

 

So for a given budget, talking about RAM today it is better to invest in frequency rather than in size if your interest is IL-2 VR.

 

If I will rig a rig today I would go to something like (check the Mobo support 4800 MHz):

https://pcpartpicker.com/product/jhDJ7P/gskill-trident-z-royal-16-gb-2-x-8-gb-ddr4-4800-memory-f4-4800c18d-16gtrg

 

If one day (in two years) you need more RAM you can easily add 2 more stick of RAM (2x8Gb) to your MoBo. But at that time probably DDR5 would the standard.

 

If you are going to use the PC for other RAM heavy duties (virtual machines, physical simulations, complex video editing, etc) then it is another story.

Edited August 30, 2020 by chiliwili69

[Dutch2]

On 8/30/2020 at 11:02 AM, chiliwili69 said:

I would not build a system today with 32Gb of ram if the purpose is IL-2 VR. I also tested other VR games like RoboRecall, EliteDangerus, etc and they run ok.

My current system has 16Gb and my son can ran flawlessly any new game he is playing (Battlefield V, Apex, Overwatch, Unturned, etc).

I also doubt IL-2 VR would require more than 16Gb in the near future since it will reduce the number of potential users, since only 8.6% (according to Steam surverys) has more than 16Gb:

 

So for a given budget, talking about RAM today it is better to invest in frequency rather than in size if your interest is IL-2 VR.

 

If I will rig a rig today I would go to something like (check the Mobo support 4800 MHz):

https://pcpartpicker.com/product/jhDJ7P/gskill-trident-z-royal-16-gb-2-x-8-gb-ddr4-4800-memory-f4-4800c18d-16gtrg

 

If one day (in two years) you need more RAM you can easily add 2 more stick of RAM (2x8Gb) to your MoBo. But at that time probably DDR5 would the standard.

 

If you are going to use the PC for other RAM heavy duties (virtual machines, physical simulations, complex video editing, etc) then it is another story.

Any proof this high speed RAM does any impact, because lots of other test I did read on the internet, did show no effect. 

Further as it seems some will go for MS FS2020 and in combination with the later VR update, the extra 32Gb is needed. 

 

About the benchmark I do hope mr Jason, bring us an decent benchmark or nothing. Because benchmarking on a video recordings, sorry, despite all your efforts, is rather useless. 

 

 

On 8/29/2020 at 1:51 AM, ACG_PanzerV said:

Thanks very much for all of this.  Very valuable input that I will heed in my process.

 

 

Your question is mostly also applicable to all modern VR headsets. A more powerfull PC will give you more % of time at full native fps in complex scenarios.

 

We have been studying in the past what are the main factors to obtain better VR performance. The last series of tests was conducted with the Remagen becnhmark, but it doesn´t work anymore with latest game updates with referred rendering (recorded flights became unplayable).

 

Ideally the game could have an always playable recorded flight, so we would be always able to use a common track to measure performance over the same baseline, but it seems it is not on the top priorities of the development plans. I created a poll to know if this is important and people has same opinon than me. With a common benchmark it would be easier to compare performance of different headsets, hardware, software and settings. So all tricks and wisdom would be compiled and we will be able to take the best decissions when acquiring new hardware. @Jason_Williams do you think this would be important for the VR (and non-VR) community?

 

After this intro, let see were is wise to invest money for a PC for the Reverb G2. Assuming you only plays IL-2 VR:

 

1. Intel vs AMD. In all the test we have done Intel was always the winner. It is true that new AMD CPUs could change this, but current chips don´t. So forget AMD until someone prove the opposite

2. Intel CPU: Overclock is a must, so go for a K version. Forget the "X" chips since they are more difficult to overclock. Current 10th gen is incresing the number of cores but for IL-2 VR you will not need more than 4 cores. So you can go to 10600K, or 10700K or 10900K., all of them have 2Mb cache per core (we suspect that cache plays also a role).  Personally I would go to 10600K, having the option to go to higher versions if other demostrate that they are better for IL-2 VR.

3. CPU cooling: Go for a good AIO liquid cooling of 240 or 360mm. They are worth the money. Put it at the top of the case. An inlcude 3 or 4 more fans in the case.

4. RAM: IL-2 VR RAM usage is around 7-8 Mb. So please, don´t spend money on size (32GB) and go for speed. Try to go to 16 (2x8) RAM at 4400MHz (or higher) with lowest Cas Latency (CL). So learnt that frequency (the higher the better) is better than CL.

5. Motherboard: Look for a known brand with capability for high RAM frequencies. And with easy software for overclock (like MSI Command Center). But I think any brand would be OK.

6. GPU: Don´t spend a fortune in the lastest top GPU, probably they are not worth the money. For the G2 I would try first with a 1080Ti or 2080S. You can always upgrade GPU easily.

 

my two cents.

 

For the motherboard and you are not afraid to be an overclocker, get the one that does have the lowest VRM temperatures. About AMD vs Intel, if not being an overclocker do not skip AMD. Read the tests and get whats best for your budget.

 

RAM always get single rank RAM if going for 2x8Gb first and later for the extra 2x8Gb. I would get something like 3333mhz/cl16

 But most important get the RAM that is tested in the motherboard of your choice, see the RAM recommendation listing that is included in the motherboard site. 

 

AIO, the best bang for bucks is the Arctic Liquid FreezerII 360. I did get mine for $100,-  Its very big but check if its fit in your case, alternatively is the more a less same performer and cheaper, the Artic Liquid FreezerII 280. Check GemersNexus for the review as he is the only one who does a decent test btw.

 

Computer Case, If not have chosen any case I would sure select one that can handle an 360x120 & 240x140 radiator on front and top. 

 

Get your info from the recourses like Gamernexus, Jazztwocent, Linus, Der8auer and Hardware unboxed on Youtube. 

Edited August 31, 2020 by Dutch2

[chiliwili69]

1 hour ago, Dutch2 said:

Any proof this high speed RAM does any impact, because lots of other test I did read on the internet, did show no effect. 

 

If you read the post of Greif in the Remagen Tests thread you will see that RAM speed was one of the main influencing variables to achieve high fps in monitor 2D (for a PC with no bottlenecked GPU).

1 hour ago, Dutch2 said:

About the benchmark I do hope mr Jason, bring us an decent benchmark or nothing. Because benchmarking on a video recordings, sorry, despite all your efforts, is rather useless

 

Well , I would not say that the tests performed are useless. They helped us to determine what are the main factor which influence performance in IL-2 in monitor, and therefore in VR for a not bottlenecked GPU.

 

Any news about what Mr Jason thinks about this?

[Dutch2]

51 minutes ago, chiliwili69 said:

 

If you read the post of Greif in the Remagen Tests thread you will see that RAM speed was one of the main influencing variables to achieve high fps in monitor 2D (for a PC with no bottlenecked GPU).

 

Well , I would not say that the tests performed are useless. They helped us to determine what are the main factor which influence performance in IL-2 in monitor, and therefore in VR for a not bottlenecked GPU.

 

Any news about what Mr Jason thinks about this?

 

for checking what is the bottleneck at you rig, there is no need for a benchmark test, you can use something like fpsVR or for 2D MSI afterburner. 

If Jason is considering an ingame benchmark, I do hope @Jason_Williams  first check this video before even think about introducing an benchmark that is missing every game specific reality. 

 

[ACG_PanzerVI]

On 8/30/2020 at 5:02 AM, chiliwili69 said:

 

So for a given budget, talking about RAM today it is better to invest in frequency rather than in size if your interest is IL-2 VR.

 

If I will rig a rig today I would go to something like (check the Mobo support 4800 MHz):

https://pcpartpicker.com/product/jhDJ7P/gskill-trident-z-royal-16-gb-2-x-8-gb-ddr4-4800-memory-f4-4800c18d-16gtrg

 

If one day (in two years) you need more RAM you can easily add 2 more stick of RAM (2x8Gb) to your MoBo. But at that time probably DDR5 would the standard.

 

If you are going to use the PC for other RAM heavy duties (virtual machines, physical simulations, complex video editing, etc) then it is another story.

Edited yesterday at 05:04 AM by chiliwili69

 

I too would like to know the real performance bennie here.  The price delta between 4000 and 4800 is roughly x 2+.  Would want to appreciate a significant performance difference.  I could get 32gb of 4200 speed for the price of 16gb of 4800.

[chiliwili69]

12 hours ago, Dutch2 said:

or checking what is the bottleneck at you rig, there is no need for a benchmark test, you can use something like fpsVR or for 2D MSI afterburner. 

 

fpsVR is useful too to know your GPU/CPU frametimes and how bottlenecked is your PC. But the purpose of the benchmark is quite beyond that:

 

- To know how every hardware option (CPU, GPU, RAM,etc) influences the game performance, based on the multiple tests performed by users

- To know how every single game setting influences the game performance, this is what I did here. So you can better know the fps versus detail trade off.

- To know how other VR systems/Software perform for IL-2

- To know if your rig is performing how it should be according to the test of your peers or the expected performance of the greif correlation model.

 

With the test we performed along the years we gain a significant knowledge about the role that every piece of the PC was playing.It was a collective small effort that benefit many of us.

12 hours ago, Dutch2 said:

If Jason is considering an ingame benchmark, I do hope @Jason_Williams  first check this video before even think about introducing an benchmark that is missing every game specific reality.

 

That´s a nice video. It points out many of the items which are desirable in a game benchmark.

Even an always playable track record of 2 minutes in a complex map/scenario (wich would not require a lot of devlopment effort from the IL-2 Team) would be a good thing to have.

[Dutch2]

I’m only saying: for only checking you bottleneck you can use fpsVR or afterburner, I do know these are not benchmarktools. 

 

I’m not that good in the English language and at that point weak for some funny guys who are trying to make ridicules remarks on my reply,  but this video from digital foundry guy, does tell you everything I always wanting to say about what your test methodology is missing,  besides, you need to rerun a couple of times and its about frame-times you wanting to know, even more specific the straightness then an average FPS, but that is another discussion.  

 

I know it was the only method available, and for that a big thank on al your efforts you did made, it also did show me the way of direction. Only its not telling the right story as being clearly explained in the video.

 

About your last remark I’m disagree on that.   I do hope if @Jason_Williams put inn a ingame benchmark, he do it on the right way or NOT, there is no compromise, like an track from a video recording.

[chiliwili69]

8 hours ago, ACG_PanzerV said:

I too would like to know the real performance bennie here.  The price delta between 4000 and 4800 is roughly x 2+.  Would want to appreciate a significant performance difference.  I could get 32gb of 4200 speed for the price of 16gb of 4800.

 

The theoretical gain you should get by every +100MHz of RAM speed is about +1.6fps in monitor, so about +0.8fps in VR. This is only an statistical correlation based on tests of not constrained GPU and using the previous game version. So you night not achive exactly those numbers. See this.

 

Initially I thought that the important number to correlate was going to be the first word latency (it is 2000*CL/Freq). The lower the better. But Greif saw that frequency was correlating better than first word latency.

 

I think you refer to 4000Mhz, not to 4200. This are the two choices for almost same price with same First word latency 7.5 ns, right?

 

https://pcpartpicker.com/product/jhDJ7P/gskill-trident-z-royal-16-gb-2-x-8-gb-ddr4-4800-memory-f4-4800c18d-16gtrg

 

https://pcpartpicker.com/product/TMmFf7/gskill-ripjaws-v-32-gb-4-x-8-gb-ddr4-4000-memory-f4-4000c15q-32gvk

 

So for +800MHz you will get about +6.4fps in VR.

 

If you just want the PC for IL-2 VR (like in my case) I will go to 16Gb at 4800.

If you think you will be also a frequent player of FS2020 or DCS or other heavy apps then 32Gb at 4000 would be also quite good.

 

Remember to check that your Mobo support that freq and RAM model.

 

[Alonzo]

On 8/31/2020 at 8:37 AM, Dutch2 said:

Any proof this high speed RAM does any impact, because lots of other test I did read on the internet, did show no effect.

 

The VR community has done plenty of testing and benchmarking. Your assertion that other information on the internet is more correct leads me to think you have not properly studied the IL2 VR research and methods. You should at least read those things and understand what they are before making such an assertion.

 

Chili has kindly explained the existing research. As part of that research I did some benchmarking where the only thing I changed about my rig was the RAM speed. I demonstrated a significant increase in FPS by increasing the RAM frequency.

 

Please, when someone like Chili or Fenris states something about VR performance, give them the respect they deserve -- they have put significant time into the testing. Yes, we are without a standardized benchmark, but Chili's method isolates as much as possible and does give reproducible numbers.

 

Information from non-IL2 testing should be treated as suspect. "Rules of thumb" or results for other games do not apply here. The recipe has been the same for my entire time playing IL2, more than two years now: Fastest clock speed Intel chip you can find, good motherboard, fast RAM, fast GPU.

 

Everyone falls back to "where is the proof?" or "got a link for that?" all the time, even when an expert is speaking. It's very tiring. We have actual experts who have poured hundreds of hours into testing, we have spreadsheets full of numbers, and we have solid conclusions. Please listen to those.

[Georgio]

Regarding the G2, remember that you can always run it on half resolution initially and it will still visually blow away every other headset out there.

This is what I intend to do until the 3000 series are at an affordable price without having to sell my gran...?

 

I currently have an i7 8700k / 1080Ti  / Rift S setup.

[Charlo-VR]

I'm in a similar boat - my EVGA Titan X Hybrid only has DisplayPort 1.2, not the DisplayPort 1.3 listed in the G2 specs. I'm hoping I can at least get some enjoyment from the G2 until I can see some benchmarks and GPU AIO cooling solutions before committing to my much-needed upgrades to my current 5-year-old rig

[SCG_Fenris_Wolf]

You have more time, our HP Reverb G2 was just delayed to 3rd week of October by HP. Went under the carpet in the hypestorm the new RTX-gen created.

 

 

[chiliwili69]

On 9/1/2020 at 4:54 PM, Alonzo said:

experts

Thanks for your comments. 

It is understandable that not all people have time to read all the info of this forum, so it logical that they ask. So we answer. No problem at all.

 

Also, I would not consider myself an expert in almost anything. Just a man who like VR and planes, and want to get the most of it by understanding how things work. And thanks to the collective effort we have learnt something.

[Charlo-VR]

On 9/1/2020 at 12:24 AM, chiliwili69 said:

 

The theoretical gain you should get by every +100MHz of RAM speed is about +1.6fps in monitor, so about +0.8fps in VR. This is only an statistical correlation based on tests of not constrained GPU and using the previous game version. So you night not achive exactly those numbers. See this.

 

Initially I thought that the important number to correlate was going to be the first word latency (it is 2000*CL/Freq). The lower the better. But Greif saw that frequency was correlating better than first word latency.

 

I think you refer to 4000Mhz, not to 4200. This are the two choices for almost same price with same First word latency 7.5 ns, right?

 

https://pcpartpicker.com/product/jhDJ7P/gskill-trident-z-royal-16-gb-2-x-8-gb-ddr4-4800-memory-f4-4800c18d-16gtrg

 

https://pcpartpicker.com/product/TMmFf7/gskill-ripjaws-v-32-gb-4-x-8-gb-ddr4-4000-memory-f4-4000c15q-32gvk

 

So for +800MHz you will get about +6.4fps in VR.

 

If you just want the PC for IL-2 VR (like in my case) I will go to 16Gb at 4800.

If you think you will be also a frequent player of FS2020 or DCS or other heavy apps then 32Gb at 4000 would be also quite good.

 

Remember to check that your Mobo support that freq and RAM model.

 

I’m about to upgrade my rig to an i7 10700K on an MSI MEG Z490 Unify. And of course, I’ll get an RTX 3000 series GPU whenever the supply chain provides them after I see some benchmarks.
 

When I look at 32GB RAM at 4000MHz I see most run at 1.5v, but according to pcpartpicker Intel warns against running RAM higher than at 1.35v. Amongst the RAM listed on the MSI QVL for that board, though, the latencies both go higher if I limit myself to 1.35v RAM. I’m considering this 1.35v RAM:

 

https://pcpartpicker.com/product/gXKcCJ/gskill-trident-z-royal-32-gb-4-x-8-gb-ddr4-4000-memory-f4-4000c17q-32gtrs

[dburne]

When talking ram I personally think the sweet spot is in the 1.35v ram, 3200-3600 CL14 or CL15.

I run 32 GB of G Skill 3200 MHz CL14 and am very happy with it.

It is a bit of balancing act with speed and latency both taken into consideration.

[Bernard_IV]

I got solid performance in my G1 with a 7700k 5.2ghz, 3200 cl14 ram sticks, and a 1080ti at 2000mhz.  I had to turn the clouds to low, AA off, and grass off to keep above 90htz most of the time.  I tried to run faster ram but my system didn't like it, the 3200 was the most it would take.  I'm looking to get the 3080 or 3090 card depending on if the 3090 is worth it or not along with the new G2.  Looking forward to high settings!  I'm going to keep the 7700k for now.  I'll see how it does in other games, I know for IL2 it is as good as it will get.

[chiliwili69]

14 hours ago, Charlo-VR said:

but according to pcpartpicker Intel warns against running RAM higher than at 1.35v

 

That´s strange.

According to Intel offical webpage there three levels of RAM overclock: Beginners (just Intel XMP profiles), intermediate (tweak the XMP profile) and advanced (on your own).

https://www.intel.com/content/www/us/en/gaming/resources/overclock-ram.html

 

I am just talking about the Beginners, which is the one I currently use, just a default XMP profile.

 

There are standard Intel XMP profiles which uses voltages higher than 1.35volts, look this table.

 

Perhaps it would be better to ask to people like @WallterScott or @jarg1 who were running at higher than 4000MHz and achieving the top performance of the Remagen tests. What they did?

I have never done RAM OC apart from XMP profiles. So, I am just a beginner here with no experience.