getting bad fps in gta v w/ good computer....

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Ketchup

Elite Member
Sep 1, 2002
14,553
248
106
2500k to 5960x (used in the video bench) is a little over a 50% performance increase in this game. Since the 2500k is such an overclock-friendly chip, it is in your best interest to overclock this CPU.

Edit: unfortunately, you are looking at a game that really scales with CPU power, not just GPU power. So while you have a great video card, and a CPU that should have plenty of life left in it for most tasks, the CPU is what's holding you back in this scenario.
 
Last edited:

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
51-60 FPS is with 980 SLI. So 20-40 fps for you is right. That is without AA.

But the 290x isn't bottlenecked in game. I think the post below yours is accurate. The coy is the bottleneck. I ordered a coolermaster 212 Evo. I will install that and see about getting the chip to 4.5ghz
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
2500k to 5960x (used in the video bench) is a little over a 50% performance increase in this game. Since the 2500k is such an overclock-friendly chip, it is in your best interest to overclock this CPU.

Edit: unfortunately, you are looking at a game that really scales with CPU power, not just GPU power. So while you have a great video card, and a CPU that should have plenty of life left in it for most tasks, the CPU is what's holding you back in this scenario.

Thanks. I will do some overclocking and report back.
 

Ketchup

Elite Member
Sep 1, 2002
14,553
248
106
Looking forward to your results. Mine did pretty well, topping at 4.9 (but took more vcore than I was comfortable keeping it at 24/7).
 

MTDEW

Diamond Member
Oct 31, 1999
4,284
37
91
JST0rm, I can give you the info you need pretty easily with my rig in sig.

I can run my 2600k at stock with HT disabed, which will basically match your 2500k.
Give me your 290 Tri-x's clock speeds and i'll match those too.

Then we can both run the Valley benchmark and GTA V benchmark at the same settings and then you'll know if indeed a clean install may be needed.

Let Me know. ;)
 

MTDEW

Diamond Member
Oct 31, 1999
4,284
37
91
Here are my GTA V numbers with the following settings.

CPU - 2600k @ 3.4ghz with HT off (no overclocking).....(your 2500k without overclock is 3.3ghz...so close enough...LOL)
GPU - R9 290 Tri-x 1050/1300
Win 7 64bit
CAT 15.4 beta drivers

Ingame benchmark using 1080p with everything at game defaults except vsync set to OFF.
Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 39.061115, 100.773621, 81.080208
Pass 1, 58.909767, 118.589592, 90.773499
Pass 2, 50.402103, 128.165741, 82.129074
Pass 3, 54.794628, 134.325409, 102.218925
Pass 4, 11.796142, 188.937637, 87.626953

Just to be sure, here is what the in-game defaults set things to on my rig.
Ignore Limits - OFF
DX Ver - DX 11

Screen - Fullscreen
Res - 1920x1080 ( I manually set this to match your resolution)
Aspect ratio - Auto
Refresh Rate - 60
Output Monitor - 1

FXAA - On
MSAA - Off
vsync - OFF (i manually turned this off for benchmarking)
Pause on Focus loss - On

Pop Density - max
Pep Variety - max
Dist Scaling - max

Texture Quality - VH
Shader Quality - VH
Shadow Quality - H
Reflection Quality - VH
Reflection MSAA - off
Water Quality - H
Particle Quality - VH
Grass Quality - H
Soft Shadows - Softer
Motion Blur - min/disabled
DOF - on
Aniso - x16
Ambient Occlusion - H
Tesselation - H

Advanced Graphics - ALL OFF

EDIT:

Here are the same settings as above except with my CPU overclocked to 4.6ghz.
Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 61.815323, 125.744530, 95.309227
Pass 1, 66.751305, 124.719597, 94.259125
Pass 2, 62.633884, 149.198380, 89.133125
Pass 3, 74.478714, 139.051468, 105.054001
Pass 4, 26.818159, 156.217819, 95.149925






Here are my Unigine Valley 1.0 scores.
GPU = 1050/1300
Using Extreme HD preset.
note: It varies from run to run, so as long as its close.
(as you can see by my CPU non-oc'ed max is actually higher than my CPU OC'ed max...LOL)
Its a GPU benchmark, so CPU OCs dont really effect it too much anyway.
I just ran both twice and recorded the best of the two runs for each.

CPU @ 4.6ghz
FPS - 60.5
Score - 2531
Min - 29.0
Max - 114.0

CPU @ 3.4ghz
FPS - 59.9
Score - 2504
Min - 26.6
Max - 115.6
 
Last edited:

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
awesome man. Thank you! I am at work right now but I will try this tomorrow (working like 15 hours on a saturday for reasons).

My 290x is clocked at 1020 but why worry over such a small amount. I can see you are doing way better then mine. lol. Ill match your settings and check back.

Looking at your frames: the lowest you got was on your first pass and that amount is about my highest. No way this is a oc issue already.
 
Last edited:

MTDEW

Diamond Member
Oct 31, 1999
4,284
37
91
awesome man. Thank you! I am at work right now but I will try this tomorrow (working like 15 hours on a saturday for reasons).

My 290x is clocked at 1020 but why worry over such a small amount. I can see you are doing way better then mine. lol. Ill match your settings and check back.

Looking at your frames: the lowest you got was on your first pass and that amount is about my highest. No way this is a oc issue already.
Ok, you say your Tri-x is clocked at 1020mhz core...I can set mine to that.
If you check your memory clocks on your Tri-x also, I'll match those too so we can have a direct comparison at the same GPU clocks.
(GPU core and memory speeds do affect Valley scores, so its best we're the same)

note: My Sapphire 290 Tri-x defaults to 1000/1250 so I didn't know what version of Tri-x you have and your default clocks.
I run mine at 1050/1300 as my daily overclocks because it does that with undervolting (-25mv) from default voltage to keep the temps/wear & tear down...IMO there is no need to stress it with more voltage than it needs just for a few more FPS.
 
Last edited:

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
1350 is the ram speed. Thanks.

Do you have a regular 290? Mine is a 290x
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
Valley pass 1

r7LBqYo.jpg


Valley pass 2

nabUyX8.jpg
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
gta V

Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 10.063225, 37.435627, 28.098972
Pass 1, 18.557266, 45.397552, 32.045891
Pass 2, 15.246042, 47.960159, 33.722572
Pass 3, 19.757921, 47.648823, 34.268497
Pass 4, 2.473881, 54.313091, 30.347734

Time in milliseconds(ms). (Lower is better). Min, Max, Avg
Pass 0, 26.712523, 99.371727, 35.588490
Pass 1, 22.027618, 53.887249, 31.205250
Pass 2, 20.850639, 65.590790, 29.653728
Pass 3, 20.986877, 50.612614, 29.181320
Pass 4, 18.411768, 404.223114, 32.951389

v-sync off
 

escrow4

Diamond Member
Feb 4, 2013
3,339
122
106
I've posted screenshots in other V threads here and I can confirm you are hitting a CPU wall. My 5930K @ 3.7GHz is showing usage on all cores/threads, and not just 10%; I'm talking 30%+ or so. I've seen CPU usage hit 55% all up in total. Vysnc off I can max out (AA disabled except for FXAA and Reflections @ 2x) all standard settings except for grass at Very High, all others are Ultra and Very High at 1200p. Advanced settings I can up the sliders to 60% and the flying option is enabled. That is enough for 70-80FPS in the city, 55FPS roughly outside in the countryside some dips to 30s. Subs underwater for some reason hit 20s but that is hardly part of the game so meh. 30hrs in and coming to the end of the (vastly shorter *snort*) single player campaign.
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
I certainly agree with you but it seems others are posting benches at 200% my benches at the same clock. So I think somethign is wrong. I'm going to do a full system reinstall and bench these 2 again.
 

MTDEW

Diamond Member
Oct 31, 1999
4,284
37
91
1350 is the ram speed. Thanks.

Do you have a regular 290? Mine is a 290x
Yes mine is a SAPPHIRE TRI-X OC R9 290, so your 290x benches should be slightly higher than mine then.

Give me a bit , i have a few things to do first and I'll post the numbers for you.

But taking an initial look, your scores do seem usually low.


EDIT to add benchmarks.

Here is my Valley benchmark, I couldnt get alt+printscreen to work, so I had to use Valleys own built in screenshot, so its fullscreen...sorry.

NOTE: I see you used a custom preset, I used the Extreme HD preset so we can ensure we are both using the exact same settings.

This is with...
CPU - 3.4ghz HT off
GPU - 1020/1350
Valley - Extreme HD preset
CAT 15.4 beta drivers
l2sQHPM.png
 
Last edited:

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
0JTPsC7.jpg


Ok. Its less then yours for sure. So something is off with the system but I dont believe its the card because of how it was sleeping through the benchmarks in gta.
 

MTDEW

Diamond Member
Oct 31, 1999
4,284
37
91
Yeah, the only thing I could think to check is maximizing the power limit setting for the GPU and use something like MSI AB onscreen display so you can see the clockspeeds in realtime while running the benchmark to be sure its not throttling.
Its not that far off, but being a 290x, it should beat my 290 at the same clock speeds.
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
yeah thanks for your help fir sure. I will check back when I can do a reinstall of windows and then again when I do some overclocking.
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
ok fresh install of windows

b2B9cYc.jpg


Clearly an improvement in valley but still low compared to yours. Will check gta now. This is with no overclock yet. Thats in the mail still.
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 19.603251, 38.115330, 32.084393
Pass 1, 23.552719, 46.607922, 34.538612
Pass 2, 23.509518, 55.328621, 35.434128
Pass 3, 9.800073, 46.942535, 35.087711
Pass 4, 12.537157, 58.886543, 34.318260

New gta benches but the install is on the same disk. Im going to redo the valley bench with only the system disk in the system.

old bench below

Frames Per Second (Higher is better) Min, Max, Avg
Pass 0, 10.063225, 37.435627, 28.098972
Pass 1, 18.557266, 45.397552, 32.045891
Pass 2, 15.246042, 47.960159, 33.722572
Pass 3, 19.757921, 47.648823, 34.268497
Pass 4, 2.473881, 54.313091, 30.347734
-----------------------------
Valley was 2343 without the other drives. I will try the overclock and update
 
Last edited:
Aug 11, 2008
10,451
642
126
Where do you see that the 2500k benched at 30fps? Im seeing 72 per my post here:

http://forums.anandtech.com/showpost.php?p=37353682&postcount=3

Im just interested. If I do infact have massive cpu bottlenecking I will probably attempt overclocks (with a new fan) and see if I can get another year out of this cpu.

According to game.gpu, 2500k gives 60 FPS ave and 50 min, VHQ 1080p. That is with a very powerful gpu (980SLI). That is pretty close to the results you posted earlier.

I dont think you should be cpu limited to 30FPS with a 2500k. The game does seem to run quite a bit better on haswell though, and likes hyperthreading.
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
My next bit of trouble shooting will be to uninstall steam and reinstall on my ssd. I will then download gta v and see if being on a ssd matters. During the bench I hear the drive doing a lot of spinning so I am suspect.
 

MTDEW

Diamond Member
Oct 31, 1999
4,284
37
91
My next bit of trouble shooting will be to uninstall steam and reinstall on my ssd. I will then download gta v and see if being on a ssd matters. During the bench I hear the drive doing a lot of spinning so I am suspect.
I have my steam/games on my mechanical HDD and use my SSD only for OS and needed programs.
So in case I didn't mention it before, I'm running GTA V off of my mechanical HDD.
Sure, my game load times are longer than they would be if the game was on my SSD.
But I really doubt it effects in-game performance or benchmarks, just load times.
 

JSt0rm

Lifer
Sep 5, 2000
27,399
3,947
126
Im wondering if the drive is failing and causing slow downs though. Possibly I may need to replace it.