RTX 3080 performing as poor as my old GTX 1070 Ti

[GoldMichi]

Hey guys,

 

Today was the day, where I received my long awaited RTX 3080. I went and installed it immediately. After installing I ran a benchmark using the benchmark provided by userbenchmark.com. The results here were as expected. 168% gaming, 92% desktop and 149 workstation. Then I went on and tried the benchmark from Shadow of the Tomb Raider. Here my average fps were around 107. I monitored my temp and the load for the cpu. Temp went up as far as 76 °C on the gpu and 55 °C on the cpu. Load on cpu was always between 50 and 70 percent. Sadly, I wasn't able to monitor the load on my gpu for whatever reason, but given the temps it should have been be maxed out. I've compared the results to other benchmarks for Shadow of the Tomb Raider and realised that my bench was performing around 30% below theirs. Also when just casually gaming, I don't see any big increase in fps. Monitoring my cpu and gpu both linger at around 40% load, when playing WoW Classic (Yes, I know. Using the card to its fullest potential ^^). I play on full HD, 144hz.

I've updated my Bios, checked boost on my ram and uninstalled old gpu drivers with DDU and reinstalled them. I also don't think that my cpu poses that bad of a bottleneck, but maybe you guys think otherwhise.

One thing to mention though; I've installed my gpu vertically using the vertical gpu mounting bracket by cable mods. But I wouldn't figure why that should cause any problems.

My specs are:

cpu: Intel core i7 8700 (non k)

gpu: Palit Gaming Pro RTX 3080

ram: Corsair Vengeance RGB Pro 32gb, 3.6Ghz

motherboard: msi z370 pc pro

psu: Corsair TX750M (750w)

cpu cooling: NZXT Kraken z73 aio

 

I've included the benchmark-results below. (It's in german, but the important stuff should be understandable, sorry for that)

 

Maybe some of you are having the same problem or maybe even managed to solve it. Maybe it's just totally simple and I just don't get it, since I'm fairly new to everything regarding tech and pc.

 

Cheers

Michel

 

[-rascal-]

3 minutes ago, GoldMichi said:

Hey guys,

 

Today was the day, where I received my long awaited RTX 3080. I went and installed it immediately. After installing I ran a benchmark using the benchmark provided by userbenchmark.com. The results here were as expected. 168% gaming, 92% desktop and 149 workstation. Then I went on and tried the benchmark from Shadow of the Tomb Raider. Here my average fps were around 107. I monitored my temp and the load for the cpu. Temp went up as far as 76 °C on the gpu and 55 °C on the cpu. Load on cpu was always between 50 and 70 percent. Sadly, I wasn't able to measure the load on my gpu for whatever reason, but given the temps it should be maxed out. I've compared the results to other benchmarks and realised that my result was performing around 30% below theirs. Also when just casually gaming, I don't see any big increase in fps. Monitoring my cpu and gpu both linger aroung at around 40% load. I play on full HD, 144hz.

I've updated my Bios, checked boost on my ram and uninstalled old gpu drivers with DDU and reinstalled them. I also don't think that my cpu poses that bad of a bottleneck, but maybe you guys think otherwhise.

One thing to mention though; I've installed my gpu vertically using the vertical gpu mounting bracked by cable mods. But I wouldn't figure why that should cause any problems.

My specs are:

cpu: Intel core i7 8700 (non k)

gpu: Palit Gaming Pro RTX 3080

ram: Corsair Vengeance RGB Pro 32gb, 3.6Ghz

motherboard: msi z370 pc pro

psu: Corsair TX750M (750w)

cpu cooling: NZXT Kraken z73 aio

 

I've included the benchmark-results below. (It's in german, but the important stuff should be understandable, sorry for that)

 

Maybe some of you are having the same problem or maybe even managed to solve it. Maybe it's just totally simple and I just don't get it, since I'm fairly new to everything regarding tech and pc.

 

Cheers

Michel

 

 

What are you comparing YOUR Shadow of the Tomb Raider benchmark results to?

In most cases, reviewers have CPU is overclocked, rather than left at stock...or using a non-K SKU chip.

 

You'll see a noticeable difference between ... say a ... i7-8700 vs a i7-8700K @ 5.1 GHz

[GoldMichi]

1 minute ago, -rascal- said:

 

What are you comparing YOUR Shadow of the Tomb Raider benchmark results to?

In most cases, reviewers have CPU is overclocked, rather than left at stock...or using a non-K SKU chip.

 

You'll see a noticeable difference between ... say a ... i7-8700 vs a i7-8700K @ 5.1 GHz

The reviewer used an i9 10900k. So yeah, a significantly better cpu. ^^ But still, shouldn't I be getting better results anyway, or does my i7 8700 pose that bad of a bottleneck, that it was one even when using the gtx 1070Ti?

[AndreiArgeanu]

5 minutes ago, -rascal- said:

 

What are you comparing YOUR Shadow of the Tomb Raider benchmark results to?

In most cases, reviewers have CPU is overclocked, rather than left at stock...or using a non-K SKU chip.

 

You'll see a noticeable difference between ... say a ... i7-8700 vs a i7-8700K @ 5.1 GHz

Linus used a 3950X, bone stock to my knowledge, and it only has 100mhz higher boost clock, the extra 12 cores shouldn't make a difference. I doubt it's the cpu.

[Moonzy]

4 minutes ago, GoldMichi said:

But still, shouldn't I be getting better results anyway, or does my i7 8700 pose that bad of a bottleneck, that it was one even when using the gtx 1070Ti?

SOTR benchmark lets you know that

you're entirely CPU bound

 

edit: GPU-gebunden = GPU Bound, 0% means 100% cpu/other system limited (most probably CPU)

my 3900x can run the game at 144fps when CPU bound, so yours running at 110 is expected

[AndreiArgeanu]

Try running the benchmark at higher resolutions to see whether it's actually the CPU holding you back, but at 50-70% usage I kind of doubt it.

[GoldMichi]

Just now, Moonzy said:

SOTR benchmark lets you know that

you're entirely CPU bound

Oh, totally missed that. lol... that's maybe the reason why I wasn't monitoring any load on my gpu then ^^ Let me try again then. I'll let you know 

[AndreiArgeanu]

Just now, Moonzy said:

SOTR benchmark lets you know that

you're entirely CPU bound

I kind of doubt the idea that the 8700 would hold back the 3080 that much though, not to mention 0% usage is completely unrealistic, especially for an 8700.

[VeganJoy]

most benchmarks were using 9900k/10900ks with overclocks, or stock 3950xs which are still better than a locked 8700. but youre at 1080p so youll be heavily cpu bound no matter what; driver support is such that the 3080 is actually slower than the 2080ti in some titles at that resolution, even with a 10900k

[Moonzy]

21 minutes ago, GoldMichi said:

Oh, totally missed that. lol... that's maybe the reason why I wasn't monitoring any load on my gpu then ^^ Let me try again then. I'll let you know 

21 minutes ago, AndreiArgeanu said:

I kind of doubt the idea that the 8700 would hold back the 3080 that much though, not to mention 0% usage is completely unrealistic, especially for an 8700.

 

23 minutes ago, Moonzy said:

my 3900x can run the game at 144fps when CPU bound, so yours running at 110 is expected

 

at ultra setting that OP used, i drop to about 130-ish

[Moonzy]

4 minutes ago, AndreiArgeanu said:

not to mention 0% usage is completely unrealistic

not 0% usage, but 0% of the time limited by GPU (meaning GPU finish processing before CPU feeds it next frame)

you can see from the graph, each part's processing time, GPU is constantly faster (lower time)

[-rascal-]

5 minutes ago, AndreiArgeanu said:

I kind of doubt the idea that the 8700 would hold back the 3080 that much though, not to mention 0% usage is completely unrealistic, especially for an 8700.

 

The i7-8700 runs with a Base Clock of 3.2 GHz, and Turbo Boost 2.0 table of...

UP TO

• 4.6 GHz when 1/6 cores are active
• 4.5 GHz when 2/6 cores active
• 4.4 GHz when 3/6 cores active
• 4.3 GHz when 4 ~ 6 / 6 cores active

 

If the CPU is running somewhere in between 3.2 GHz ~ 4.3 GHz, after PL2 time has expired, I can see that being some sort of a bottleneck.

Compared to a i7-8700K that is running at 5.0 GHz full time, or a i9-10900K running at constant 5.2 GHz.

[GoldMichi]

1 minute ago, Moonzy said:

not 0% usage, but 0% of the time limited by GPU (meaning GPU finish processing before CPU feeds it next frame)

you can see from the graph, each part's processing time, GPU is constantly faster (lower time)

Ah, I see. So after all my cpu poses a bigger bottleneck than I thought?

[Moonzy]

Just now, GoldMichi said:

Ah, I see. So after all my cpu poses a bigger bottleneck than I thought?

for that game, yes

at 1080p right? think 1440p should see some GPU bound at ultra, never tried though since i only have 1080p

[GoldMichi]

1 minute ago, Moonzy said:

for that game, yes

at 1080p right? think 1440p should see some GPU bound at ultra, never tried though since i only have 1080p

yeah, same. So I'll either have to upgrade my cpu or my monitor ^^ 

[Moonzy]

1 minute ago, GoldMichi said:

yeah, same. So I'll either have to upgrade my cpu or my monitor ^^ 

why? 100 fps is completely playable

upping resolution wont increase fps

https://linustechtips.com/main/profile/187465-moonzy/?status=276612&type=status

[GoldMichi]

Just now, Moonzy said:

why? 100 fps is completely playable

it is. I was just quite disappointed that I'm not seing any improvement over replacing my old card. I'm somewhat of an FPS-whore ^^

[Moonzy]

2 minutes ago, GoldMichi said:

it is. I was just quite disappointed that I'm not seing any improvement over replacing my old card. I'm somewhat of an FPS-whore ^^

you wont see fps increase when you go to 1440p

https://linustechtips.com/main/profile/187465-moonzy/?status=276612&type=status

lowering setting does yield ~10% fps increase though

[GoldMichi]

Just now, Moonzy said:

you wont see fps increase when you go to 1440p

https://linustechtips.com/main/profile/187465-moonzy/?status=276612&type=status

lowering setting does yield ~10 fps increase though

yeah, you're right, but at least I'll be able to play on 1440p and up with 100 fps ^^

[Moonzy]

11 minutes ago, GoldMichi said:

yeah, you're right, but at least I'll be able to play on 1440p and up with 100 fps ^^

if you maintain ultra setting, you'll see lower avg if you also hit gpu bound

 

if you want high fps, intel 10600k/10700k is your go to, or 3600/3700x

not a 1440p panel

 

you can use this to extrapolate, ryzen can refer to 3950x, intel can refer to 10900k

but lower the expected fps slightly because lower tier chip have lower boost

 

but take note, you're expecting a ~35% increase in fps if you choose 10600/10700

and a ~30% fps increase if you choose 3600/3700x

[Mister Woof]

8700 is still a solid chip and generally performs as good or better than any ryzen chip in most games on a 2080ti (it performs exactly the same as a stick 8700k) and you can expect this to be the same trend with a 3080.

 

 

That says, SotTR isn't really the best case scenario for Intel chips to begin with.

 

What resolution is that benchmark your running on?

[Moonzy]

1 minute ago, Mister Woof said:

What resolution is that benchmark your running on?

1080p

[VeganJoy]

13 minutes ago, Moonzy said:

you can use this to extrapolate, ryzen can refer to 3950x, intel can refer to 10900k

but lower the expected fps slightly because lower tier chip have lower boost

the difference in games between a 3700x and a 3950x is extremely minimal since zen 2 boost clocks are a bit weird, right?

 

 

[Moonzy]

Just now, VeganJoy said:

the difference in games between a 3700x and a 3950x is extremely minimal since zen 2 boost clocks are a bit weird, right?

yes, if 3900x loaded all core, it boost to 4ghz (assuming temp is ~70c)

while a 3950x brags 0.1ghz higher boost clock on the box, it actually boost to ~3.7-3.8ghz under all core load

 

that includes most games, take a look at my sig, 3900x clock issue

[GoldMichi]

16 minutes ago, Moonzy said:

if you maintain ultra setting, you'll see lower avg if you also hit gpu bound

 

if you want high fps, intel 10600k/10700k is your go to, or 3600/3700x

not a 1440p panel

 

you can use this to extrapolate, ryzen can refer to 3950x, intel can refer to 10900k

but lower the expected fps slightly because lower tier chip have lower boost

 

but take note, you're expecting a ~35% increase in fps if you choose 10600/10700

and a ~30% fps increase if you choose 3600/3700x

great info, thank you so much!