GTX1080Ti being hold back?

[CTR640]

After slightly more than 4 years it was time to upgrade the GPU, GTX780 to 1080Ti. It got installed all nicely and no errors so far. However, I've seen other 1080Ti owners they can run GTAV on max 4K60 and others can run the game easily with extended distance scaling on max while mine will get its fps crapped to the 40's. A friend on GTAO got a 1070Ti and he runs it maxed, 1080p60 and also with extended distance scaling at max... So something is not right and I'm not sure what. Is the 1080Ti being hold back by this shit CPU Intel i7-4770? GTAV even runs like shit with map mods installed and gameplay was much stable and smooth when I used GTX780 while 1080Ti struggles with the map mods. I tried different settings and resolutions like 1080p, 1440p and 2160p. Same results. Temps are fine, fans are fine, no problems and checked the temps and usage every time with MSI Afterburner.

 

My system:

W7 64bit Home

i7-4770

16GB RAM

PSU: CoolerMaster V650S

 

No BSOD, no errors, no artifacts and any other trouble that asks for attention. So, is the culprit the CPU? Should I upgrade to a better CPU?

Not only GTAV, also NFS Payback is getting shitty FPS, especially in the city and settings are Auto and Custom and resolution is 1440p. 2160p is out of question.

 

Not to mention those with better performance got 6core CPU's, the 5820K and 8600K.

[TheGlenlivet]

In Afterburner, what is your CPU usage while gaming?

[Fasauceome]

Even if the 1080 ti is being "held back," the performance should not be worse. If, as you said, 

7 minutes ago, CTR640 said:

gameplay was much stable and smooth when I used GTX780 while 1080Ti struggles with the map mods

then something is amiss

[Zyndo]

16 minutes ago, CTR640 said:

Snip

Did your performance go DOWN after the upgrade? If it did, a CPU bottleneck can't be the culprit here. Bottlenecks don't reduce performance, rather, they limit performance increases.

If your performance went down, I would recommend uninstalling/reinstalling your drivers as it is something you didn't state that you did and it is HIGHLY recommended when changing GPU's (and i'm not saying update your drivers, i'm saying get rid of them and then install them again as previous driver into from your old GPU can be interfering). A 4770 is unlikely to be a "bottleneck" in 4k gaming, even with a 1080ti, because the graphical load is so extreme you aren't getting high enough FPS to really stress the CPU. If we were talking 1440p or 1080p then you can certainly be CPU limited in many titles when paired with a 1080ti, but at 4k the CPU load is relatively quite low.

[CTR640]

18 minutes ago, Zyndo said:

Did your performance go DOWN after the upgrade? If it did, a CPU bottleneck can't be the culprit here. Bottlenecks don't reduce performance, rather, they limit performance increases.

If your performance went down, I would recommend uninstalling/reinstalling your drivers as it is something you didn't state that you did and it is HIGHLY recommended when changing GPU's (and i'm not saying update your drivers, i'm saying get rid of them and then install them again as previous driver into from your old GPU can be interfering). A 4770 is unlikely to be a "bottleneck" in 4k gaming, even with a 1080ti, because the graphical load is so extreme you aren't getting high enough FPS to really stress the CPU. If we were talking 1440p or 1080p then you can certainly be CPU limited in many titles when paired with a 1080ti, but at 4k the CPU load is relatively quite low.

Yes, before starting Windows in normal mode, i started it in safe mode the very first time with DDU. All went fine.

And now also uninstalled the drivers again with DDU and installed older driver, same results. And that is the problem, 1080Ti should be able to handle it with ease on GTAV at 4K or 1440p but performance is just shit while others with 1080Ti/8600K and 1070Ti/5820K got much better performance than I have. So it gives the impression GTAV likes better CPU's, most likely 6core.

 

34 minutes ago, TheGlenlivet said:

In Afterburner, what is your CPU usage while gaming?

The max was 81% on one core but overall the usage are always between 39%-75% between the cores and threads, there is not one core that's 100%.

[Fasauceome]

3 minutes ago, CTR640 said:

1080Ti/8600K and 1070Ti/5820K got much better performance

Those CPUs also have faster single cores instead of just more cores. Your CPU should be able to handle itself though, I used to have an i7 4770 and I have nothing but good things to say about it.

[TheGlenlivet]

3 minutes ago, CTR640 said:

The max was 81% on one core but overall the usage are always between 39%-75% between the cores and threads, there is not one core that's 100%.

Doesn't sound like the CPU is the bottleneck with that kind of usage.  RAM is dual channel/2 8GB DIMMS correct?

 

[Zyndo]

21 minutes ago, CTR640 said:

1080Ti should be able to handle it with ease on GTAV at 4K or 1440p

Stop saying "4k or 1440" like they're somehow close to the same thing. at 4k you will not be CPU limited. at 1440p you will almost certainly run the risk of being CPU limited.

 

21 minutes ago, CTR640 said:

The max was 81% on one core but overall the usage are always between 39%-75% between the cores and threads, there is not one core that's 100%.

This is definitive evidence you are not being CPU limited

 

 

Look at this video. Make sure your game is set to the same settings as it is, then run your game benchmark and see if your performance is close to it. Your fps should be SLIGHTLY lower because he's using a well clocked 7700k with better IPC, but it should be very close. you should be getting easily 95% of the performance of him, with only the occasional frametime being too high.

 

 

[FaintLocket424]

I'm running an i5 750 (1st gen) and I'm running 1080p60fps easily with a GTX 980 in GTAO.

 

What are your temps? You may be throttling, however unlikely it is.

 

If nothing works to get it working then you may have a defective GPU. If so you should consider getting an RMA.

[CTR640]

10 minutes ago, TheGlenlivet said:

Doesn't sound like the CPU is the bottleneck with that kind of usage.  RAM is dual channel/2 8GB DIMMS correct?

 

Forgot to mention that, no, the 16GB RAM are not dual channel, got 4 sticks of them.

 

7 minutes ago, Zyndo said:

Stop saying "4k or 1440" like they're somehow close to the same thing. at 4k you will not be CPU limited. at 1440p you will almost certainly run the risk of being CPU limited.

 

This is definitive evidence you are not being CPU limited

 

 

Look at this video. Make sure your game is set to the same settings as it is, then run your game in similar instances and see if your performance is close to it. Your fps should be SLIGHTLY lower because he's using a well clocked 7700k with better IPC, but it should be very close. you should be getting easily 95% of the performance of him, with only the occasional frametime being too high.

 

 

That kind of benchmarks are not that really good indications. And even increasing the 'Extended Distance Scaling' to 50% already makes the game stutter and that's only on lower resolutions like the 1080p and 1440p. Everything in 'Advanced Graphics' must be off to keep the smooth performance while in the video he got all the Advanced Graphics cranked up and 4K. I'll try it again to make sure everything to be equal.

[CTR640]

1 minute ago, FaintLocket424 said:

I'm running an i5 750 (1st gen) and I'm running 1080p60fps easily with a GTX 980 in GTAO.

 

What are your temps? You may be throttling, however unlikely it is.

 

If nothing works to get it working then you may have a defective GPU. If so you should consider getting an RMA.

Nah, temps have been very fine, the max was 72°C while fans are at 55%.

[FaintLocket424]

Ok. I think you may have a defective GPU.

[Zyndo]

1 minute ago, CTR640 said:

and that's only on lower resolutions like the 1080p and 1440p

stop saying 1080p and 1440p like they are the same thing as 4k. You only need a more powerful GPU with each resolution increase. the lower framerates of 4k make it significantly easier for a weaker CPU to keep up. You started this topic asking about 4k. Are you trying to run the game at 1080p or at 4k? because our recommendations are going to be VERY different based on this information.

 

3 minutes ago, CTR640 said:

That kind of benchmarks are not that really good indications.

No. A benchmark like this is the BEST indication. It's a real world, repeatable benchmark, not a synthetic benchmark. It runs the game in game scenarios the exact way the game would be played. It runs the camera on a rail. I'm not asking you to run something like firestrike/timespy/superposition because that would be useless, but a benchmark like the GTA V benchmark is an extremely useful tool for troubleshooting performance issues.

[CTR640]

3 minutes ago, Zyndo said:

stop saying 1080p and 1440p like they are the same thing as 4k. You only need a more powerful GPU with each resolution increase. the lower framerates of 4k make it significantly easier for a weaker CPU to keep up. You started this topic asking about 4k. Are you trying to run the game at 1080p or at 4k? because our recommendations are going to be VERY different based on this information.

 

No. A benchmark like this is the BEST indication. It's a real world, repeatable benchmark, not a synthetic benchmark. It runs the game in game scenarios the exact way the game would be played. It runs the camera on a rail. I'm not asking you to run something like firestrike/timespy/superposition because that would be useless, but a benchmark like the GTA V benchmark is an extremely useful tool for troubleshooting performance issues.

Yes, I asked about 4K because I thougt there is something problematic about my system, especially about the extended distance scaling and others have much better performance than I have. That's why I created this topic. And about the resolutions, maybe I sound like they are the same thing as 4K but it's not. I'm trying the game at 4K.

[Zyndo]

Just now, CTR640 said:

Yes, I asked about 4K because I thougt there is something problematic about my system, especially about the extended distance scaling and others have much better performance than I have. That's why I created this topic. And about the resolutions, maybe I sound like they are the same thing as 4K but it's not. I'm trying the game at 4K.

then do your tests in 4k. stop looking at 1080p and 1440p and stop telling us: "well it can't even do 1080 therefore it can't do 4k". at 1080p your 4770 will very probably keep your 1080ti down in a large majority of titles out there. at 4k that will almost never be the case.

 

It's also worth noting that many people have different definitions of "max" or "ultra" settings. most people will tell you they have everything maxed out, but they aren't including the "extra" sliders like AA, motion blur, or some specialized game features such as the Purehair effects on the Tomb Raider games. These things can make a pretty big difference.  For example, here is a guy playing GTA V with a 1080ti at 1080p and getting less than 60fps in some instances and he is GPU limited. It's all about your settings man.

 

[CTR640]

1 hour ago, Zyndo said:

then do your tests in 4k. stop looking at 1080p and 1440p and stop telling us: "well it can't even do 1080 therefore it can't do 4k". at 1080p your 4770 will very probably keep your 1080ti down in a large majority of titles out there. at 4k that will almost never be the case.

 

It's also worth noting that many people have different definitions of "max" or "ultra" settings. most people will tell you they have everything maxed out, but they aren't including the "extra" sliders like AA, motion blur, or some specialized game features such as the Purehair effects on the Tomb Raider games. These things can make a pretty big difference.  For example, here is a guy playing GTA V with a 1080ti at 1080p and getting less than 60fps in some instances and he is GPU limited. It's all about your settings man.

 

I was very well aware the 4770 would keep the 1080Ti down at 1080p and FHD is not the plan on such GPU. I was only curious why the 780 could handle the map mods I optimized for the 780 easily while 1080Ti struggles with the exact same map mods. And that's not on 1080p. Rest assured, that's not important.

 

And as expected, I do not get the performance you've shown in the first video. At the Vinewood's sign at night, FPS goes all the way down to 29FPS and that's tested on unmodded game and the exact same settings.

[Zyndo]

1 hour ago, CTR640 said:

I was very well aware the 4770 would keep the 1080Ti down at 1080p and FHD is not the plan on such GPU. I was only curious why the 780 could handle the map mods I optimized for the 780 easily while 1080Ti struggles with the exact same map mods. And that's not on 1080p. Rest assured, that's not important.

 

And as expected, I do not get the performance you've shown in the first video. At the Vinewood's sign at night, FPS goes all the way down to 29FPS and that's tested on unmodded game and the exact same settings.

good. now we can move forward. the next thing I would suggest would be to pay attention to your GPU's core frequency and memory frequency during the test and see if it matches the frequencies of the numbers displayed in any of the videos above. 

[CTR640]

9 minutes ago, Zyndo said:

good. now we can move forward. the next thing I would suggest would be to pay attention to your GPU's core frequency and memory frequency during the test and see if it matches the frequencies of the numbers displayed in any of the videos above. 

Sometimes when I start to write, I go a little bit too much.

 

But yes, the GPU's core frequency is identical to mine, mine's even gets to 1999MHz max. I'll be testing the memory frenquency.

[Zyndo]

1 minute ago, CTR640 said:

Sometimes when I start to write, I go a little bit too much.

 

But yes, the GPU's core frequency is identical to mine, mine's even gets to 1999MHz max. I'll be testing the memory frenquency.

monitor the GPU load while you're at it.

[CTR640]

3 minutes ago, Zyndo said:

monitor the GPU load while you're at it.

All are identical to YT, aswell as the memory frequency as the GPU load.

[Zyndo]

7 hours ago, CTR640 said:

All are identical to YT, aswell as the memory frequency as the GPU load.

hrmmmm... other than drivers/software issues, I don't know how you can have 100% load with 100% identical settings and identical (or near identical) clockspeeds and not have the same amount of performance resulting. One of those variables needs to change for your results to change. Any chance you have another system you could plug this GPU into and see if your results are the same there? or maybe an extra HDD laying around which you could install a fresh windows onto, a fresh GTA onto, and fresh drivers onto to see if there is some kind of glitch in your current install?

[CTR640]

4 hours ago, Zyndo said:

hrmmmm... other than drivers/software issues, I don't know how you can have 100% load with 100% identical settings and identical (or near identical) clockspeeds and not have the same amount of performance resulting. One of those variables needs to change for your results to change. Any chance you have another system you could plug this GPU into and see if your results are the same there? or maybe an extra HDD laying around which you could install a fresh windows onto, a fresh GTA onto, and fresh drivers onto to see if there is some kind of glitch in your current install?

Yeah, that's what bugged me and it cannot be the CPU? No need to upgrade to 8700 or even 9700K?

And HDD can also not affect the performance, right? It's a Seagate Barracuda 1TB 7200rpm.

[Zyndo]

5 hours ago, CTR640 said:

Yeah, that's what bugged me and it cannot be the CPU? No need to upgrade to 8700 or even 9700K?

And HDD can also not affect the performance, right? It's a Seagate Barracuda 1TB 7200rpm.

HDD's CAN affect performance in a game like GTA V in some situations if they're slow enough or if you run out of RAM.

As far as the CPU being the problem you stated earlier that you are not ever experiencing 100% load on your CPU, therefore upgrading your CPU which is already being underutilized to something better is going to give you very marginal performance increase at best. Do you have your GPU plugged into your motherboard's x16 PCIe 3.0 slot? Do you have Vsync on? (Vsync will limit your fps to some multiple of your screens refresh rate in order to prevent tearing)

[FaintLocket424]

Do you have any other PCI-e cards uring up all the PCI-e lanes on your CPU? Is your GPU in the top slot?

 

If not I think you may have a defective GPU.

 

[CTR640]

21 hours ago, Zyndo said:

HDD's CAN affect performance in a game like GTA V in some situations if they're slow enough or if you run out of RAM.

As far as the CPU being the problem you stated earlier that you are not ever experiencing 100% load on your CPU, therefore upgrading your CPU which is already being underutilized to something better is going to give you very marginal performance increase at best. Do you have your GPU plugged into your motherboard's x16 PCIe 3.0 slot? Do you have Vsync on? (Vsync will limit your fps to some multiple of your screens refresh rate in order to prevent tearing)

V-sync has been tested on and off. With V-sync off, it goes higher than 60FPS but screen-tearing...

Motherboard is MSI Z87 and GPU is placed in the first PCIe, just like where GTX780 was.

So upgrading the CPU is pretty much pointless?

12 hours ago, FaintLocket424 said:

Do you have any other PCI-e cards uring up all the PCI-e lanes on your CPU? Is your GPU in the top slot?

 

If not I think you may have a defective GPU.

 

What means uring?

There is only one GPU. There are less cables, well, no cables all over the place because the PSU is semi-modulair, CoolerMaster V650S.