1080TI + 4790k@4.4GHZ, unexpected Frames in demanding games

[TVwazhere]

i7-4790k @4.4ghz stock voltage stable

Gigabyte Z97N-WIFI

G.Skill Sniper DDR3 16GB 1866mhz

EVGA GTX1080Ti FTW3 Slider set to 117% power, no other overclocks

Corsair 380T

1x Samsung EVO 250GB (OS), 1x WD Black 3TB(games), 1x WD Black 5TB(Video only)

Corsair CX550M

Asus VG248QE

Cryorig H7

 

Windows build 1803

Nvidia drivers 416.34  (freshly installed after removal of GTX970 using DDU)

Nvidia Control Power Management mode set to Maximum performance (was originally on a default setting, but enabling it only gave me minimal improvements)

 

***ALL BENCHMARKS ARE IN 1080P UNTIL I GET A NEW MONITOR. I AM AWARE 1080TI IS OVERKILL***

 

Both Firestrike and Heaven Bench 4.0 perform as expected. Same story with moderately demanding games like Overwatch (250+fps) and Rainbow Six Seige. However AAA titles like GTA (60-90fps avg) and Battlefield 1 (90-110) are much lower than expected at 1080p max settings, or with some settings turned down (Thought this was the issue for GTAV, but it turns out it's not) When these games are active I notice both GPU Usage and Max Power do not hit 100% like they do in the benchmarks, in BF1 specifically even rendering at 150% (1440p) doesnt achieve over 110fps, nor does it bring usage and power draw to 100% (usually 85)

Spoiler

 

3Dmark test

 

Percision Xoc settings

Rainbow 6 seige

 

GTA Settings

GPU usage on GTAv

 

Battlefield 1 (tried both dx12 on and off with no result change)

Battlefield GPU usage

 

Checked a few forums but nothing seems to give me a solid answer. I'm not ruling out the dreaded B word, but I haven't seen any game pin the CPU to 100% (though in BF1 it's usually around 88%)

 

As of posting, I'm about to go to bed so please leaves thoughts and I will get to responses in the morning. Thank you in advance for your help.

[Cereal5]

It is most likely the dreaded b-word. However, even on 3440x1440 my 1080 Ti doesn't hit 100% power usage... it usually is around 80-95. (also, 1080p 150% render is 1620p , 1440p would be 133%)

[GamerDude]

I think Heaven benchmark should be run with 8xAA, not 2xAA for the standard High End GPU benchmark run.

[Spotty]

18 minutes ago, TVwazhere said:

Checked a few forums but nothing seems to give me a solid answer. I'm not ruling out the dreaded B word, but I haven't seen any game pin the CPU to 100% (though in BF1 it's usually around 88%)

Show individual core usage with MSI Afterburner. You will likely see that one or more of the cores are being pegged at 100% while other cores may only be at 60-70% - varying depending on what task each core is assigned to do. You'll notice this especially in Battlefield games that are very CPU heavy.

 

FWIW, I haven't played Battlefield 1, but in Battlefield V beta I was averaging 110-130FPS at 1440p high settings, and that was with a CPU bottleneck. So the 110FPS you're getting in Battlefield 1 with your system seems about right to me.

Edited October 18, 2018 by Spotty

[i_build_nanosuits]

im playing at 2560x1440p max details on a 144hz Gsync monitor and in many modern games i run out of CPU before i run out of GPU...clockspeeds can be seen in my signature...

[Crunchy Dragon]

Sounds like you have a bit of a bottleneck there.

[Ashiella]

It's very similar to my issue with my 1080 due to a lacking single core performance and most games lacking support for high thread counts. Probably a bottleneck

[xAcid9]

Try 4K DSR and check if GPU usage go up?

[TVwazhere]

9 hours ago, Cereal5 said:

 the dreaded b-word.

9 hours ago, Spotty said:

CPU bottleneck

9 hours ago, i_build_nanosuits said:

 run out of CPU

9 hours ago, Crunchy Dragon said:

bottleneck

6 hours ago, Ashiella said:

Probably a bottleneck

 

I guess possible for a CPU bottleneck in BF1, but I'm not sure about GTAV. Like Spotty suggested, I'll have to check single core usage in that game because the averages in that game are SUPER low compared to what they should be.

6 hours ago, xAcid9 said:

Try 4K DSR and check if GPU usage go up?

When running BF1 at 200% resolution, the GPU usage does increase to 99%, and the FPS does decrease to about 75. Forgot to check CPU usage but I'm assuming it was slightly lower. 

9 hours ago, GamerDude said:

I think Heaven benchmark should be run with 8xAA, not 2xAA for the standard High End GPU benchmark run.

Probably, though I guess it helps reduce potential GPU bottlenecks in that test. 

 

I just wouldn't have thought the 4790K and something like the 7700K (CPU used in a lot of benchmarks I was seeing) would have differed so much in performance in games considering IPC and core speed between generations has only marginally improved. I might try overclocking the CPU beyond 4.4ghz once I get into 1440p 144fps monitors, perhaps that and the increase resolution will help strike a balance in CPU and GPU usage in games like these. 

[Cereal5]

Lol good montage there. Yeah... Haswell is starting to show its age with higher end hardware. Thankfully it’s still great for medium and lower, so you should be able to get a decent bit if you sell it all as a combo

[TVwazhere]

Just now, Cereal5 said:

Lol good montage there. Yeah... Haswell is starting to show its age with higher end hardware. Thankfully it’s still great for medium and lower, so you should be able to get a decent bit if you sell it all as a combo

I was expecting to one day upgrade but seeing how bad the 9900K was priced, I was hoping to wait till past 2019 (Unless Zen 2 turns out to be a killer CPU)

[Mick Naughty]

I’ll tell you like people tell me, there’s something wrong with your rig. That’s as far as the experts will go in diagnosing. 

 

Get a real monitor and see how it goes, kinda pointless trying to make that card work on 1080p. 

[Spotty]

15 minutes ago, TVwazhere said:

I guess possible for a CPU bottleneck in BF1, but I'm not sure about GTAV. Like Spotty suggested, I'll have to check single core usage in that game because the averages in that game are SUPER low compared to what they should be.

Not sure about GTA V. It should run fine on your system. One thing I've noticed about GTA V is that it favours fast memory, and if you have slower memory - or not enough capacity - you will face quite a lot of stutters in game as you're driving around. I don't have any info to back this up, so take it with a grain of salt.

 

15 minutes ago, TVwazhere said:

I just wouldn't have thought the 4790K and something like the 7700K

There's really not a lot of difference between the 4790k and 6700k that I have, so you should be seeing results pretty similar to what I have.

 

Here's my GTA V benchmark results (1440p Very High settings [I think - It's an older test]). Looking at averages of 80-110FPS depending on the test.

Spoiler

Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 18.898767, 110.227463, 86.310989
Pass 1, 54.117012, 115.622917, 83.790649
Pass 2, 54.962818, 146.968582, 98.115417
Pass 3, 27.706285, 157.825302, 111.777176
Pass 4, 39.899967, 163.079330, 90.927277

Time in milliseconds(ms). (Lower is better). Min, Max, Avg
Pass 0, 9.072149, 52.913502, 11.586010
Pass 1, 8.648805, 18.478477, 11.934505
Pass 2, 6.804175, 18.194118, 10.192079
Pass 3, 6.336120, 36.092892, 8.946370
Pass 4, 6.131985, 25.062677, 10.997800

Frames under 16ms (for 60fps):
Pass 0: 804/808 frames (99.50%)
Pass 1: 767/782 frames (98.08%)
Pass 2: 897/899 frames (99.78%)
Pass 3: 1032/1034 frames (99.81%)
Pass 4: 10048/10227 frames (98.25%)

Frames under 33ms (for 30fps):
Pass 0: 807/808 frames (99.88%)
Pass 1: 782/782 frames (100.00%)
Pass 2: 899/899 frames (100.00%)
Pass 3: 1033/1034 frames (99.90%)
Pass 4: 10227/10227 frames (100.00%)

Percentiles in ms for pass 0
50%,    11.00
75%,    12.00
80%,    12.00
85%,    12.00
90%,    13.00
91%,    13.00
92%,    13.00
93%,    13.00
94%,    13.00
95%,    13.00
96%,    13.00
97%,    14.00
98%,    14.00
99%,    14.00

Percentiles in ms for pass 1
50%,    11.00
75%,    12.00
80%,    13.00
85%,    13.00
90%,    14.00
91%,    14.00
92%,    14.00
93%,    14.00
94%,    14.00
95%,    15.00
96%,    15.00
97%,    15.00
98%,    15.00
99%,    16.00

Percentiles in ms for pass 2
50%,    10.00
75%,    12.00
80%,    12.00
85%,    12.00
90%,    12.00
91%,    13.00
92%,    13.00
93%,    13.00
94%,    13.00
95%,    13.00
96%,    13.00
97%,    13.00
98%,    13.00
99%,    14.00

Percentiles in ms for pass 3
50%,    8.00
75%,    9.00
80%,    10.00
85%,    10.00
90%,    10.00
91%,    11.00
92%,    11.00
93%,    11.00
94%,    11.00
95%,    11.00
96%,    11.00
97%,    11.00
98%,    12.00
99%,    12.00

Percentiles in ms for pass 4
50%,    11.00
75%,    12.00
80%,    13.00
85%,    13.00
90%,    13.00
91%,    14.00
92%,    14.00
93%,    14.00
94%,    14.00
95%,    14.00
96%,    15.00
97%,    15.00
98%,    15.00
99%,    16.00

=== SYSTEM ===
Windows 10 Pro 64-bit (6.2, Build 9200)
DX Feature Level: 11.0
Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz (8 CPUs), ~4.0GHz
16384MB RAM
NVIDIA GeForce GTX 1080 Ti, 11667MB, Driver Version 399.07
Graphics Card Vendor Id 0x10de with Device ID 0x1b06

=== SETTINGS ===
Display: 2560x1440 (FullScreen) @ 143Hz VSync OFF
Tessellation: 3
LodScale: 1.000000
PedLodBias: 0.200000
VehicleLodBias: 0.000000
ShadowQuality: 3
ReflectionQuality: 2
ReflectionMSAA: 4
SSAO: 2
AnisotropicFiltering: 8
MSAA: 2
MSAAFragments: 0
MSAAQuality: 0
SamplingMode: 0
TextureQuality: 2
ParticleQuality: 2
WaterQuality: 2
GrassQuality: 2
ShaderQuality: 2
Shadow_SoftShadows: 5
UltraShadows_Enabled: true
Shadow_ParticleShadows: true
Shadow_Distance: 2.000000
Shadow_LongShadows: true
Shadow_SplitZStart: 0.930000
Shadow_SplitZEnd: 0.890000
Shadow_aircraftExpWeight: 0.990000
Shadow_DisableScreenSizeCheck: false
Reflection_MipBlur: true
FXAA_Enabled: true
TXAA_Enabled: true
Lighting_FogVolumes: true
Shader_SSA: true
DX_Version: 2
CityDensity: 1.000000
PedVarietyMultiplier: 1.000000
VehicleVarietyMultiplier: 1.000000
PostFX: 2
DoF: true
HdStreamingInFlight: true
MaxLodScale: 1.000000
MotionBlurStrength: 0.000000

 

(From memory I was having issues with VOLTAGE LIMIT warning showing when running those benchmarks, and no amount of tweaking the power/voltage settings in MSI would fix it, and from memory I was only at about 80% GPU usage and 40% CPU usage)

[GamerDude]

22 minutes ago, TVwazhere said:

I was expecting to one day upgrade but seeing how bad the 9900K was priced, I was hoping to wait till past 2019 (Unless Zen 2 turns out to be a killer CPU)

You know what? You should just stop stressing out on the benchmarks and just play games, IF you see massive stutter or framerate drops, then yeah, maybe a CPU/mobo/RAM upgrade is in order. I'm more concern about gaming now, and all the games I've played thus far have been smooth, even Forza 4 Horizon Demo benchmark ran pretty nicely on my weakest rig, the 2600K + GTX 1080 system. When I play games on my RTX rig, I see my GPU being pushed to 90-95% depending on the game, but most importantly, games run smooth.....except for Gears of War Ultimate PC. I get some slowdowns here and there.....even on the RTX rig. That game's just messed up I guess....

[TVwazhere]

22 minutes ago, Mick Naughty said:

Get a real monitor and see how it goes, kinda pointless trying to make that card work on 1080p. 

That's definitely in the future. Waiting for Black friday/Newegg's Black "month" since it's so close.

22 minutes ago, Spotty said:

Not sure about GTA V. It should run fine on your system. One thing I've noticed about GTA V is that it favours fast memory, and if you have slower memory - or not enough capacity - you will face quite a lot of stutters in game as you're driving around. I don't have any info to back this up, so take it with a grain of salt.

 

There's really not a lot of difference between the 4790k and 6700k that I have, so you should be seeing results pretty similar to what I have.

 

Here's my GTA V benchmark results (1440p Very High settings [I think - It's an older test]). Looking at averages of 80-110FPS depending on the test.

  Reveal hidden contents

Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 18.898767, 110.227463, 86.310989
Pass 1, 54.117012, 115.622917, 83.790649
Pass 2, 54.962818, 146.968582, 98.115417
Pass 3, 27.706285, 157.825302, 111.777176
Pass 4, 39.899967, 163.079330, 90.927277

Time in milliseconds(ms). (Lower is better). Min, Max, Avg
Pass 0, 9.072149, 52.913502, 11.586010
Pass 1, 8.648805, 18.478477, 11.934505
Pass 2, 6.804175, 18.194118, 10.192079
Pass 3, 6.336120, 36.092892, 8.946370
Pass 4, 6.131985, 25.062677, 10.997800

Frames under 16ms (for 60fps):
Pass 0: 804/808 frames (99.50%)
Pass 1: 767/782 frames (98.08%)
Pass 2: 897/899 frames (99.78%)
Pass 3: 1032/1034 frames (99.81%)
Pass 4: 10048/10227 frames (98.25%)

Frames under 33ms (for 30fps):
Pass 0: 807/808 frames (99.88%)
Pass 1: 782/782 frames (100.00%)
Pass 2: 899/899 frames (100.00%)
Pass 3: 1033/1034 frames (99.90%)
Pass 4: 10227/10227 frames (100.00%)

Percentiles in ms for pass 0
50%,    11.00
75%,    12.00
80%,    12.00
85%,    12.00
90%,    13.00
91%,    13.00
92%,    13.00
93%,    13.00
94%,    13.00
95%,    13.00
96%,    13.00
97%,    14.00
98%,    14.00
99%,    14.00

Percentiles in ms for pass 1
50%,    11.00
75%,    12.00
80%,    13.00
85%,    13.00
90%,    14.00
91%,    14.00
92%,    14.00
93%,    14.00
94%,    14.00
95%,    15.00
96%,    15.00
97%,    15.00
98%,    15.00
99%,    16.00

Percentiles in ms for pass 2
50%,    10.00
75%,    12.00
80%,    12.00
85%,    12.00
90%,    12.00
91%,    13.00
92%,    13.00
93%,    13.00
94%,    13.00
95%,    13.00
96%,    13.00
97%,    13.00
98%,    13.00
99%,    14.00

Percentiles in ms for pass 3
50%,    8.00
75%,    9.00
80%,    10.00
85%,    10.00
90%,    10.00
91%,    11.00
92%,    11.00
93%,    11.00
94%,    11.00
95%,    11.00
96%,    11.00
97%,    11.00
98%,    12.00
99%,    12.00

Percentiles in ms for pass 4
50%,    11.00
75%,    12.00
80%,    13.00
85%,    13.00
90%,    13.00
91%,    14.00
92%,    14.00
93%,    14.00
94%,    14.00
95%,    14.00
96%,    15.00
97%,    15.00
98%,    15.00
99%,    16.00

=== SYSTEM ===
Windows 10 Pro 64-bit (6.2, Build 9200)
DX Feature Level: 11.0
Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz (8 CPUs), ~4.0GHz
16384MB RAM
NVIDIA GeForce GTX 1080 Ti, 11667MB, Driver Version 399.07
Graphics Card Vendor Id 0x10de with Device ID 0x1b06

=== SETTINGS ===
Display: 2560x1440 (FullScreen) @ 143Hz VSync OFF
Tessellation: 3
LodScale: 1.000000
PedLodBias: 0.200000
VehicleLodBias: 0.000000
ShadowQuality: 3
ReflectionQuality: 2
ReflectionMSAA: 4
SSAO: 2
AnisotropicFiltering: 8
MSAA: 2
MSAAFragments: 0
MSAAQuality: 0
SamplingMode: 0
TextureQuality: 2
ParticleQuality: 2
WaterQuality: 2
GrassQuality: 2
ShaderQuality: 2
Shadow_SoftShadows: 5
UltraShadows_Enabled: true
Shadow_ParticleShadows: true
Shadow_Distance: 2.000000
Shadow_LongShadows: true
Shadow_SplitZStart: 0.930000
Shadow_SplitZEnd: 0.890000
Shadow_aircraftExpWeight: 0.990000
Shadow_DisableScreenSizeCheck: false
Reflection_MipBlur: true
FXAA_Enabled: true
TXAA_Enabled: true
Lighting_FogVolumes: true
Shader_SSA: true
DX_Version: 2
CityDensity: 1.000000
PedVarietyMultiplier: 1.000000
VehicleVarietyMultiplier: 1.000000
PostFX: 2
DoF: true
HdStreamingInFlight: true
MaxLodScale: 1.000000
MotionBlurStrength: 0.000000

 

(From memory I was having issues with VOLTAGE LIMIT warning showing when running those benchmarks, and no amount of tweaking the power/voltage settings in MSI would fix it, and from memory I was only at about 80% GPU usage and 40% CPU usage)

I haven't touched my RAM since installing it and enabling XMP. In fact I havent tweaked the CPU settings since 2014  

1 minute ago, GamerDude said:

You know what? You should just stop stressing out on the benchmarks and just play games, IF you see massive stutter or framerate drops, then yeah, maybe a CPU/mobo/RAM upgrade is in order.

You sound just like my friends  (In a good way)

 

I know I should. I just want to maximize my potential, considering I spent good money on the card. It still performs really well and I can now running every thing at max settings until I upgrade. My deeper concern is that I'll be having these same issues at 1440p. I guess it's a good problem to have, but one I was hoping to avoid.

[maizenblue]

I have a 4790k at stock speed and a 1070 ti with 16gb @ 1866mhz. At 1440p i get 90 to 120 fps in bf1 with everything maxed except i have 2x mxaa instead of 4x, so something doesnt seem right. Sure its not a ram issue? Are you using 2 sticks of 8gb?  

 

If by b word u mean bottleneck, how is your 4790k @ 4.4gz underperforming against mine running at 4.0. If i can manage the same framerates at 1440p something doesnt seem right. The 2x mxaa diference shouldn't account for much.  

[TVwazhere]

58 minutes ago, maizenblue said:

Sure its not a ram issue? Are you using 2 sticks of 8gb?   

Yes, 2x 8GB kit

58 minutes ago, maizenblue said:

If by b word u mean bottleneck, how is your 4790k @ 4.4gz underperforming against mine running at 4.0. If i can manage the same framerates at 1440p something doesnt seem right. The 2x mxaa diference shouldn't account for much.  

I really dont know. I'm considering a Win 10 reinstall but I've seen people do that and have no results

[Hayter]

I also notice weird performance in Call of Duty Black Ops 4 Black Out playing at 4K 1aa everything else ultra I get random stutter freezes etc. I hope is not my vram. Vram  sits around 96% frame buffer 44% but I think it does that with every single size of vram doesn’t matter if you have 4 or 6 or 11 it uses all of it. My Cpu sits at 88%-90% and my gpu at 101% usage 

temps on GPU 77c cpu temps 70c

specs

4790k 4.60 oc 1.290v H100iv2

16GB DDR3 2.40  1.65v

ASUS Maximus VII Hero 3201

EVGA RTX 2080 XC reference overclock clocks 1905 core 7000 memory 416.32

860w Corsair AX 

850Evo 500GB latest drivers fresh install of windows.

460x case Corsair with 6 120mm fans

I was thinking of buying 970Evo 500GB maybe removes my stutter and freezes or my components are hold me back with a RTX 2080 or drivers.

 

[Mick Naughty]

1 hour ago, maizenblue said:

I have a 4790k at stock speed and a 1070 ti with 16gb @ 1866mhz. At 1440p i get 90 to 120 fps in bf1 with everything maxed except i have 2x mxaa instead of 4x, so something doesnt seem right. Sure its not a ram issue? Are you using 2 sticks of 8gb?  

 

If by b word u mean bottleneck, how is your 4790k @ 4.4gz underperforming against mine running at 4.0. If i can manage the same framerates at 1440p something doesnt seem right. The 2x mxaa diference shouldn't account for much.  

What usage is the cpu seeing on amines?

[i_build_nanosuits]

 

5 hours ago, TVwazhere said:

I was expecting to one day upgrade but seeing how bad the 9900K was priced, I was hoping to wait till past 2019 (Unless Zen 2 turns out to be a killer CPU)

yeah same here...but i'm using my rig mostly for VR gaming and thankfully this is MUCH more GPU demanding and much less CPU intensive and in this situation my system is plenty to use the 1080ti to it's max potential...and even in 1440p gaming there are only some games where i see GPU usage drop to 80 or 85% at times, BF1 is one of them...but in GTA V at 1440p im entirely maxing out the GPU all the time though...

I was using a GTX 1080 previously and i was maxing it out all the time, 1080ti is that little bit extra powerful where it's a bit over the edge to what the CPU can produce...that said if you're on a 1080p screen then yeah upgrade that ASAP the 1080ti is WAY overpowered for that cause that card can push 1440p at 120FPS+ in most AAA titles...

 

1 hour ago, maizenblue said:

I have a 4790k at stock speed and a 1070 ti with 16gb @ 1866mhz.

well there you go...1070ti is much less powerful and i had a 1080 before and a 980ti and with those cards i was always GPU bound. OP has a 1080ti.

[TVwazhere]

3 minutes ago, i_build_nanosuits said:

if you're on a 1080p screen then yeah upgrade that ASAP the 1080ti is WAY overpowered for that cause that card can push 1440p at 120FPS+ in most AAA titles...

In the works. I saw a 1440p Gsync IPS 144hz display for $500 on new egg but I'm holding out for black Friday

 

The shortlist:

https://pcpartpicker.com/products/monitor/#A=1&r=256001440&sort=price&p=1

[i_build_nanosuits]

Just now, TVwazhere said:

In the works. I saw a 1440p Gsync IPS 144hz display for $500 on new egg but I'm holding out for black Friday

 

The shortlist:

https://pcpartpicker.com/products/monitor/#A=1&r=256001440&sort=price&p=1

For BF1 i now use the DX12 mode...it stutter a bit during the first few minutes, maybe one round or something...and then it clears up and CPU usage is much lower (in the 60% overall i would say) and GPU usage is 90%+ in most maps and i get smooth 120FPS rendering....it's worth the few first minute of stuttery mess to get better performance after IMHO...

[maizenblue]

19 minutes ago, Mick Naughty said:

What usage is the cpu seeing on amines?

Id have to check later on. Its been a while since i paid attention to such things. I did check my fps on multiplayer the other day and i was averaging around 115 fps on verdun heights on a full server, sometimes up to 126 fps. 

 

I seriously doubt a cpu bottleneck is his problem if i can get the same frame rates with a non oc 4790k and a weaker gpu gaming at a much higher resolution. 

[i_build_nanosuits]

Just now, maizenblue said:

I seriously doubt a cpu bottleneck is his problem if i can get the same frame rates with a non oc 4790k and a weaker gpu gaming at a much higher resolution. 

thats basically the definition of a CPU limitation.

[Mick Naughty]

Just now, maizenblue said:

Id have to check later on. Its been a while since i paid attention to such things. I did check my fps on multiplayer the other day and i was averaging around 115 fps on verdun heights on a full server, sometimes up to 126 fps. 

 

I seriously doubt a cpu bottleneck is his problem if i can get the same frame rates with a non oc 4790k and a weaker gpu gaming at a much higher resolution. 

I don’t notice till it effects my gameplay. Slight stutter that will throw off my aim. Or on other games where the gpu won’t be fully utilized. But I’m done with bf1 now so isn’t. A big deal.