Is the i5 2500k generally weak with this game?

[millerman101]

Hi guys, I just bought battle of stalingrad to fly with VR but I'm getting terrible stuttering. I'm using the oculus rift.
 

I'm using an i5 2500k with a GTX 1060 6GB and 8gb RAM. Not sure if the i5 2500k is too slow with this game which i've heard is pretty single core intensive.

 

I often fall under 45 FPS on the ground or in combat which leads to nauseating stuttering

 

I can run Elite Dangerous pretty well in VR, as well as many other VR titles including Robo Recall and War Thunder, so I'm not sure what's going on here. 

I plan to get a cooler and OC to 4.5Ghz which may help, but wondering if anyone else is having these types of issues

Thanks!

[A_radek]

Welcome millerman.

A 4.5ghz OC will give you a nice boost in performance at the cost of a good cooler and some time spent tweaking. If you can up ram frequency without spending too much that will also help, how much depending on your current ram speed.

Forgot:
If your new to this sim you might not be aware that pressing the "H" key to hide the hud, gives a good performance increase.

Edited October 19, 2017 by a_radek

[Dutch2]

Check your CPU load to confirm if it is the 2500k.

I never did have any CPU performance problems running the 2500k on 4,8ghz OC. Only I do not have an VRset.

If wanting to go for an aircooler get the Noctua NH D15 the best in town and if the price does count, the new Scythe Mugen5 PCGH edition is the way to go.

[Kraft]

It's not bad with this CPU, I have about 80-90 fps average. But only on "Low" graphics preset. I do not know what the Presets are changing, but already with "Normal" preset, we will see a drop in performance.

Edited October 19, 2017 by RR_Kraft

[BlznSaddles]

The game really uses a lot of resources in VR.  I have a 1070 and an i5 6500 and I still wish I could get some more performance.  I generally run in "balanced" although I do run 2.0 SS in Oculus Tray Tool.

 

Try with everything as low as possible and use OTT (if you're running an Oculus) or Steam VR to run as much SS as you can get away with to get rid of some of the jaggies, and if you have some overhead work up from there.

 

As others have said, turn the HUD off as much as possible.  I have it bound to a HOTAS button so I can glance at it quickly for mission info and then get rid of it.  Icons don't have as much of a performance hit so you can run HUD off and Icons on.  I've heard that the frame counter itself can cost a few frames though I haven't checked on that myself.  Turning clouds off will net a little performance although obviously not an ideal solution.  There is a way in the startup cfg to turn grass almost completely off, that can help and isn't too big a loss IMO.  

[WWGriphos]

The 1060 is a marginal card for VR, and even for regular PC graphics, actually.  The *60 cards are cheap for a reason.

 

I don't use VR, but I run this game with an i5 2500K at 4.4 with a 970 and it runs very well on high settings.  

 

By the way, I have a spare Noctua cooler sitting around for the i5 if you want to try overclocking yours.  PM me.  

Edited October 19, 2017 by Griphos

[chiliwili69]

 

 

I plan to get a cooler and OC to 4.5Ghz which may help, but wondering if anyone else is having these types of issues

 

You will definitely earn fps by OC. So yes, go for a good CPU cooler.

 

You can run the VR performance test before and after your upgrade, so it will helps others to improve their IL2 VR experience.

 

The instructions for the VR test and VR test results are here:

https://forum.il2sturmovik.com/topic/29322-measuring-rig-performance-common-baseline/

[chiliwili69]

 

 

I generally run in "balanced" although I do run 2.0 SS in Oculus Tray Tool.

 

My god!! you use 2.0 SS in OTT, this is equivalent to 4.0 SteamVR_SS

 

A 1070 can´t go beyond 2.0 SteamVR_SS, so try to reduce it to 1.3 in OTT

 

In this post you will find the facts:

https://forum.il2sturmovik.com/topic/29322-measuring-rig-performance-common-baseline/?p=502368

[chiliwili69]

 

 

The 1060 is a marginal card for VR, and even for regular PC graphics, actually.  The *60 cards are cheap for a reason

 

I don´t think so. The 1060 should be OK for BOS VR, it will allow a bit less SS than 1070 or 1080 but it is very valid for IL2 VR. The important thing is the CPU, the OC and the RAM.

Just read the tests:

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

[BlznSaddles]

My god!! you use 2.0 SS in OTT, this is equivalent to 4.0 SteamVR_SS

 

A 1070 can´t go beyond 2.0 SteamVR_SS, so try to reduce it to 1.3 in OTT

 

In this post you will find the facts:

https://forum.il2sturmovik.com/topic/29322-measuring-rig-performance-common-baseline/?p=502368

I think you have the conversion backwards... OTT 2.0 is more like 1.5 in Steam. Anyway as far as frames go I keep above 45 at all times, usually 60-85 occasionally it will peg out at 90 so I’d say no, the 1070 can’t handle that if you’re a stickler for constant 90+ but if your goal is never to drop below 45 then it can do it.

[chiliwili69]

 

 

I think you have the conversion backwards... OTT 2.0 is more like 1.5 in Steam.

 

I already analyzed exactly that in detail in another post:

 

https://forum.il2sturmovik.com/topic/30436-pixel-density-supersampling-steamvr-oculustraytool-and-oculu/ 

 

SS_OTT is the square root of SteamVR_SS. If you have SS_OTT=2, then it is equivalent to put SS_SteamVR=4.

 

Ideally we should be 90 fps all the time and have ASW OFF.

 

Unless you have a very bad system, it is really difficult to go below 45 fps, since 45 is a discontinuity region as explained in this post:

https://forum.il2sturmovik.com/topic/29322-measuring-rig-performance-common-baseline/?p=506112

[dburne]

 Anyway as far as frames go I keep above 45 at all times, usually 60-85 occasionally it will peg out at 90 so I’d say no, the 1070 can’t handle that if you’re a stickler for constant 90+ but if your goal is never to drop below 45 then it can do it.

 

 

 

What counts is how happy you are with your performance in VR, long as you are happy you are golden!

[Retnek]

I use an overclocked i5, too, with a 2 GB Nvdia 960 - so my system is much weaker. If you deactivate most the "Hollywood"-stuff within the graphics settings, choose settings "high" and "range 3" you nearly never sink below 60 FPS with a 1920x1080 screen. And a second 2D-screen for maps etc is running parallel, too. Most of the time it's between 80 and 90 FPS. I won't call that "weak", it's just not 100%, most people won't see the difference. (Not talking about virtual reality, that's another class you won't earn much honour there with an i5)

Edited October 21, 2017 by 216th_Retnek

[BlznSaddles]

I already analyzed exactly that in detail in another post:

 

https://forum.il2sturmovik.com/topic/30436-pixel-density-supersampling-steamvr-oculustraytool-and-oculu/

 

SS_OTT is the square root of SteamVR_SS. If you have SS_OTT=2, then it is equivalent to put SS_SteamVR=4.

 

Ideally we should be 90 fps all the time and have ASW OFF.

 

Unless you have a very bad system, it is really difficult to go below 45 fps, since 45 is a discontinuity region as explained in this post:

https://forum.il2sturmovik.com/topic/29322-measuring-rig-performance-common-baseline/?p=506112

Gotcha. I run ASW off as well, can’t stand it on. Maybe the dynamic resolution factor I use is counteracting the SS in a way? I have that at 0.6. Is it possible as performance slows due to high oversampling that dynamic resolution is downsampling? Migh have to mess around some more but generally I’m happy with the performance as long as the HUD is off.

[WWGriphos]

I don´t think so. The 1060 should be OK for BOS VR, it will allow a bit less SS than 1070 or 1080 but it is very valid for IL2 VR. The important thing is the CPU, the OC and the RAM.

Just read the tests:

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

 

There are no 1060s in that data that I see.  The 60 series are very underpowered GPUs, and GPUs certainly do matter for VR, at least in any VR game testing I've ever seen.  

 

The OP has a 1060 and is having problems in VR.  

[chiliwili69]

 

 

Maybe the dynamic resolution factor I use is counteracting the SS in a way? I have that at 0.6. Is it possible as performance slows due to high oversampling that dynamic resolution is downsampling? Migh have to mess around some more but generally I’m happy with the performance as long as the HUD is off

 

I have not experimented with the Dynamic Resolution factor, but this factor is there to downsampling when you are below your fps target.

Basically I don´t want a system which sacrifice resolution to mantain fps, I prefer the opposite, that resolution is maintained but fps drops from 90 in some scenes. 

Current Rift/Vive resolution is OK to enjoy the game, but more resolution will be awesome. That´s why we use SS. So I prefer to run at fix resolution but variable fps.

 

Here I was doing some objective experiments with SS:

https://forum.il2sturmovik.com/topic/30771-how-much-ss-desirable-image-quality-samples/

 

The difference between SS=0 and SS=1.7 (SteamVR) is notable, if you increase SS more then you have more detail but pay a high price in performance.

 

So, depending of your fps target, you might never run at true SS=2 (OTT). To check exactly the rendered resolution used by BOS you can use Oculus Mirror and activate "Visible HUD layer" in oculus debug tool. 

 

In any case, the important thing is that your personal feeling is fine.

But it is human nature to try to get the most of our hardware and settings.

[chiliwili69]

 

 

There are no 1060s in that data that I see.  The 60 series are very underpowered GPUs, and GPUs certainly do matter for VR, at least in any VR game testing I've ever seen.     The OP has a 1060 and is having problems in VR.

 

There are no test in the first tab, but in the "Only Monitor" tab TUS_SAmuel reported the test at the monitor with 1080Ti, 1080, 1070 and 1060 and the differences were very small for the avg fps.

 

We also know that when you go from FullHD to VR your fps are halved. So, the 1060 will be OK for BOS VR.

 

The problems reported by the OP are most likely not related to the 1060, but to the no-OC of 2500K and probably RAM.

 

If the OP person run the VR test and report his results we will be able to have more data to help him more. 

[millerman101]

There are no test in the first tab, but in the "Only Monitor" tab TUS_SAmuel reported the test at the monitor with 1080Ti, 1080, 1070 and 1060 and the differences were very small for the avg fps.

 

We also know that when you go from FullHD to VR your fps are halved. So, the 1060 will be OK for BOS VR.

 

The problems reported by the OP are most likely not related to the 1060, but to the no-OC of 2500K and probably RAM.

 

If the OP person run the VR test and report his results we will be able to have more data to help him more.

I’ll run the test and post back results

[dingdingchavez]

I used to have an i5 2500k OC to 4.2 and later 4.5, with a 970 and later a 1080.  ( 8GB DDR3 @2133Mhz / Samsung 840 PRO / Win7 )

 

Never tried Elite or War Thunder, but Robo Recall ran with no stutter for me too. The Oculus First Contact tutorial however would turn into a slideshow after five minutes even at 4.5OC with a 1080. Same with games like Onward. IL2 was playable in VR, but I had the same issues as the OP. I can't find my test results for the scenario but the OC didn't stop the stuttering for me.

[chiliwili69]

 

 

urn into a slideshow after five minutes even at 4.5OC with a 1080

 

If the problem appears after 5 minutes it could be due to CPU thermal throttling or GPU thermal throttling. To verify that you will need to monitor CPU and GPU temperatures along the time (free MSI Afterburner allows that for any GPU/CPU brand), and also the GPU/CPU frequencies.

 

MSI Afterburner allows to monitor GPU freq, but not the CPU freq. You can monitor the freq with CPU-Z tool but it doesn´t have trends.

 

The Intel Extreme Tunning Utility allows to trend everything but it is too complicated for me.  Does anybody know a simple free tool to trend (not just show) CPU freq? or an easy way to detect CPU thermal throttling?

[ZachariasX]

The Intel Extreme Tunning Utility allows to trend everything but it is too complicated for me. Does anybody know a simple free tool to trend (not just show) CPU freq? or an easy way to detect CPU thermal throttling?

 

 

Try HWMONITOR

https://www.cpuid.com/softwares/hwmonitor.html

 

Slideshow after a couple of minutes most certainly is due to core overheat and resulting thermal throttling.

 

Best way to find out what your cooling can do is (albeit tedious) run Prime95. What you do is you run the bench at "standard" clocks (etc.) and then re-run the test with each time increased frequencies. At some point, you will see that the gained CPU power is topped off and not increasing anymore with increasing CPU frequency. This is where thermal throttling sets in. Especially with "big" CPUs as the 7900X, this is much before you bluescreen.

 

To make representative measurements for this, you must make sure that your cooling system is staturated with as much heat comes from the CPU. This means you have to run a cycle forst for ~20 minutes, then restart with the run you are actually metering. Else, you have a cooler that first can absorb some heat before it is saturated. You will get too good of a measurement is you count in the thermal capacity of the cooler.

[SeaW0lf]

Does anybody know a simple free tool to trend (not just show) CPU freq? or an easy way to detect CPU thermal throttling?

 

What I think is: if you need a tool to detect CPU thermal throttling is because you are already well beyond safe temperatures. My advice is to always keep your CPU below 80ºC. You will never have to worry about thermal throttling or chip degradation.

 

You can also find the Tj Max (Temperature Junction Maximum / the thermal specification that defines the core temperature limit at which the processor will throttle) on the internet. For example, processors Intel 115X from the 6th to the 8th generation have Tj Max of 100ºC IIRC. So you just need a free tool to see if you are getting there (HWmonitor, Real Temp, HWiNFO64, etc.).

 

Some people accept temperatures up to 85ºC, but I personally don't like to go over 75ºC. Also remember that overclocking changes everything because of the increase in clock and voltage. It varies a lot from person to person, but from what I see, true overclockers do not like to go above 80ºC with an Intel chip. Or you can get to some high 80s on a stress test to stabilize your OC but knowing that in RL you will be below the safe zone.

 

If you do not have the money to invest in cooling, it is much better to lower the clock (and voltage) than to linger near throttling temperatures.

Edited October 30, 2017 by SeaW0lf

[dingdingchavez]

I had CPU-Z and Speccy open whenever the really bad stuttering would start in either First Contact or Onward, and the CPU temps never went beyond 60c. With the 212evo the temps averaged at about 55c. The stuttering with those games would even happen at the base clock (with turbo boost to 3.7).

[ZachariasX]

The idea is to get an idea of the performance of your cooling solution. With average quad cores, you may well be restricted by other factors than thermal issues. Go for eight cores and soon the cooling will define to where you can crank your frequencies.

 

Also just running your game for some minutes and then alt-tab to your temp monitor will give you very little indication of what is actually happening. Checking once and saying, „oh, great, 70 degrees centigrades. I’m safe.“ means not much. You think you are safe and as soon as different loads are on the system you might well be in thermal throttling realms without even noticing it. Often, you just use some FPS, but it can come unnoticed. And you still think „I‘m at 70.“

[SeaW0lf]

I had CPU-Z and Speccy open whenever the really bad stuttering would start in either First Contact or Onward, and the CPU temps never went beyond 60c. With the 212evo the temps averaged at about 55c. The stuttering with those games would even happen at the base clock (with turbo boost to 3.7).

 

Then it is not related to temperatures.

 

Just to make sure it is always good to check with more than one tool to see if the readings match (although Intel is so staple nowadays that these apps usually aren't bugged). A simple one to use as a comparison is Real Temp. Or HWmonitor, which gives you more parameters than just temperatures. 

[dingdingchavez]

Then it is not related to temperatures.

 

Just to make sure it is always good to check with more than one tool to see if the readings match (although Intel is so staple nowadays that these apps usually aren't bugged). A simple one to use as a comparison is Real Temp. Or HWmonitor, which gives you more parameters than just temperatures. 

 

Yeah, I use Hwmonitor/CPUz/Speccy and then MSI Afterburner for on-screen display.

[SeaW0lf]

Yeah, I use Hwmonitor/CPUz/Speccy and then MSI Afterburner for on-screen display.

 

If the temps are matching the numbers on HWmonitor you are fine.

[chiliwili69]

 

 

What I think is: if you need a tool to detect CPU thermal throttling is because you are already well beyond safe temperatures. My advice is to always keep your CPU below 80ºC.

 

I thought there were other variables that could trigger CPU throttling apart from temperatures, like CPU amperage (A), CPU consumed power(W) or Voltage Regulator Temp ©

 

In the BIOS there are some parameter that can be configured to avoid that throtling, look at my BIOS profile. I marked in cyan those parameters:

 

https://docs.google....dit?usp=sharing

 

My temps are around 60 when playing BOS, and that´s why I wanted to monitor also the real CPU freq every 100ms to be sure it is not decreased due to other constraints.

[SeaW0lf]

I thought there were other variables that could trigger CPU throttling apart from temperatures, like CPU amperage (A), CPU consumed power(W) or Voltage Regulator Temp ©

 

In the BIOS there are some parameter that can be configured to avoid that throtling, look at my BIOS profile. I marked in cyan those parameters:

 

https://docs.google....dit?usp=sharing

 

My temps are around 60 when playing BOS, and that´s why I wanted to monitor also the real CPU freq every 100ms to be sure it is not decreased due to other constraints.

 

No, throttling is basically temperature related. Other variables, like insufficient core voltage for example will cause a BSOD or even the computer to don't boot (and you have to clear the CMOS). So either you have temperature throttling or instability issues / screen freezes due to other BIOS parameters (if you are tweaking them. It won't happen in stock BIOS settings). Or when doing overclock the temperature could trigger a BSOD even not near throttling temperatures because it affects how core voltage interacts with the chip for example (possibly expansion of the circuit threads? I have no idea). My i7 over 85ºC is way less stable in stress test at higher clocks (4.6Ghz) and can reboot. If you overdo on the clock / voltage you can damage the chip for higher frequencies. It is common for these website reviews to mention that they forced a given frequency (clock) and voltage in the first couple days but they can't replicate it anymore after a while. So the chip gets wasted for that clock. Usually they have to decease 100-200Mhz to get stable again.

 

Temperature is really a bitch for the chip when doing OC especially. No so much for GPUs, although the current GTX 1060 apparently throttles at 82-83ºC, which is too low for GPUs.

Edited October 31, 2017 by SeaW0lf