Shadow of the Tomb Raider benchmark

[Det0x]

No point to discuss this further, like with many millennials, feelings are more important than facts.
And i won't stoop to that level with namecalling and/or telling others what they should or shouldn't do.

In my view we are not sheep and such should not follow just because it feels comforting, when faced with a better option..
Anyway, this is the fastest rocketlake ive seen to date:
250fps average CPU Game

5250mhz core
5000mhz singlerank memory

*edit typo*

 

[Felix123BU]

No point to discuss this further, like with many millenniums, feelings are more important than facts.
And i won't stoop to that level with namecalling and/or telling others what they should do.

In my view we are not sheep and such should not follow just because it feels comforting, when faced with a better option..
Anyway, this is the fastest rocketlake ive seen to date:
250fps average CPU Game
View attachment 195120

This is not about feelings but respect towards the posters in this thread. Using the "millenniums" argument, which I suppose is millennials, is just pure lame.
But hey, feel free to continue trolling with "facts", but disregarding the fact that this is a 1080 Highest thread, you non-millennial you

 

[DemonicRyzen666]

In my view we are not sheep and such should not follow just because it feels comforting, when faced with a better option..

This quote just makes me think about how everyone reads reviews.
I'm wondering no one has test mGPU on this benchmark ?
It does support it does it not ? ¯\_(ツ)_/¯

 

[Felix123BU]

This quote just makes me think about how everyone reads reviews.
I'm wondering no one has test mGPU on this benchmark ?
It does support it does it not ? ¯\_(ツ)_/¯

In a way it supports it, in another it does not. It supports it in the way that you can always use "better" settings to get "better" results. In where it does not is that the point of having a meaningful review is to test everything at a set in stone standard, aka settings across the board, else the results have no meaning.
That's the point some don't get or ignore, this thread was meant for people comparing FPS at a very specific setting. The moment you change settings for whatever reason, the whole comparisons done here since last year lose their meaning and validity.
Not that having a wide variety of hardware configuration here gives a super accurate picture, but its still interesting to see how stuff stacks in the grand scheme of things.


This is the church of "1080 Highest", all who transgress and wish to receive absolution must post a "1080 Highest" confession and all their sins shall be forgiven

 

[Felix123BU]

@ Felix123BU

How about this then, you maintain the GPU highest standing, and i maintain the CPU standing.
Or maybe its much more fun to argue in the thread instead of actually contributing ?


Current CPU Game standings:
(Have highlighted those submissions i think that are punching above their weight class)

284 average cpu fps, me @ 5950x
270 average cpu fps, mrthanhnguyen @ 10900k
263 average cpu fps, professor dumb dumb @ I9 10980XE
253 average cpu fps, me @ 8core Zen3 cpu (1CCD 5950x)
250 average cpu fps, NIZZEN @ 11900k
229 average cpu fps, harm9963 @ 5950x
226 average cpu fps, harm9963 @ 5800x
216 average cpu fps, Felix123BU @ 5800x
212 average cpu fps, mouacyk @ 9900k
207 average cpu fps, thesmokingman @ 5900x
200 average cpu fps, Felix123BU @ 5600x
198 average cpu fps, jonRock3 @ 5800x
197 average cpu fps, nguyen @ 9900k
196 average cpu fps, lsevald @ 9900k
193 average cpu fps, weekendgeek @ 10850k
192 average cpu fps, phanbuey @ 10850k
181 average cpu fps, jlewis02 @ 10850k
180 average cpu fps, lsevald @ 9900k
169 average cpu fps, Hardi @ 3900x
168 average cpu fps, UncleFuzz @ 9900k
167 average cpu fps, phanbuey @ 8700k
167 average cpu fps, Duvar @ 3600
166 average cpu fps, Felix123BU @ 3800x
165 average cpu fps, Lindatje @ 3600
164 average cpu fps, dOBER @ 9900k
164 average cpu fps, n1mmel @ 8700k
161 average cpu fps, mtrai @ 2700x
156 average cpu fps, Splinterdog @ 5600x
156 average cpu fps, IchigoQc @ 10700k
155 average cpu fps, phanbuey @ I7 7820x
152 average cpu fps, Fizban @ 10875H
151 average cpu fps, augustben @ 9900k
149 average cpu fps, phanbuey @ I7 7820X
148 average cpu fps, weekendgeek @ 10400
148 average cpu fps, bottleofrum @ 3600
146 average cpu fps, championdude1 @ 9900k
145 average cpu fps, NoJuan999 @ R7 3700x
145 average cpu fps, Taraquin @ R5 3600
144 average cpu fps, Vayra86 @ 8700k
143 average cpu fps, Brunsas @ 8086k
142 average cpu fps, harm9963 @ 2700x
142 average cpu fps, Taraquin @ 8400
141 average cpu fps, Lew Zealand @ 8400
141 average cpu fps, cRs @8700k
140 average cpu fps, anachron @ 8600k
138 average cpu fps, Fizban @ 10875H
138 average cpu fps, RandallFlagg @ 10400
138 average cpu fps, lsevald @ 4770k
137 average cpu fps, scope54 @ 6600k
137 average cpu fps, Taraquin @ I5 8400
134 average cpu fps, jboydgolfer @ 8600k
134 average cpu fps, Solaris17 @ 7980XE
133 average cpu fps, QuietBob @ 3300x
133 average cpu fps, Pumper @ 4770k
133 average cpu fps, biffzinker @ 2600x
133 average cpu fps, Athlonite @ 3700x
130 average cpu fps, Hugis @ 4770k
130 average cpu fps, Lindatje @ 1600x
128 average cpu fps, Hardi @ 2700x
127 average cpu fps, Chloe Price @ R5 3600
126 average cpu fps, Vendor @ I5 7500
126 average cpu fps, Lindatje @ r5 1600x
126 average cpu fps, Hockster @ 2700x
125 average cpu fps, jlewis02 @ 7700k
125 average cpu fps, Athlonite @ R7 3700x
123 average cpu fps, advanced3 @ 3700x
123 average cpu fps, Melvis @ 2700x
122 average cpu fps, jormungand @ 7700k
122 average cpu fps, CGi-Quality @ I9 7920x
121 average cpu fps, CrAsHnBuRnXp @ 9900k
121 average cpu fps, droopyRO @ 8600k
120 average cpu fps, Warsaw @ 2700X
118 average cpu fps, Mac2580 @ 7700k
118 average cpu fps, TewChainsaw @ I5 4690
117 average cpu fps, Hyderz @ 6700k
117 average cpu fps, blued @ 7700k
113 average cpu fps, jlewis02 @ 7700k
113 average cpu fps, mozo @ 9900k
110 average cpu fps, Splinterdog @ 2600x
109 average cpu fps, JrRacinFan @ 3600
108 average cpu fps, blobster21 @ 4790k
107 average cpu fps, Chloe Price @ r5 2600
106 average cpu fps, MrGenius @ 3770k
105 average cpu fps, reavl @ 5820k
105 average cpu fps, Athlonite @ R7 2700
104 average cpu fps, mb194dc @ R5 1600
104 average cpu fps, uco73 @ E5 2690
104 average cpu fps, tvamos @ R5 1600
102 average cpu fps, Tomgang @ I7 980
102 average cpu fps, Lew Zealand @ 8259U
98 average cpu fps, Fizban @ 9750H
98 average cpu fps, potato580+ @ 2200G
96 average cpu fps, Hardcore Games @ 3600
89 average cpu fps, KIKOFREDY @ R5 3600
85 average cpu fps, Taraquin @ 6700HQ
84 average cpu fps, masterdeejay @ X5675
73 average cpu fps, Chloe Price @ Phenom 2 x4 B55
70 average cpu fps, Taraquin @ 8400
60 average cpu fps, StefanM @ 7700hq
55 average cpu fps, Edwired @ E5450

I see you can not let go of the CPU obsession
But damn, it must have take you some time to extract these results, see, arguing sometimes gets interesting results
What about you open a thread for using this game as a CPU bench, and I can chip in, as in we settle on a fixed custom resolution (lowest possible to show off CPU), I follow that, you follow the CPU scores, this way we leave this thread to what it was meant from the start and you are happy, I am happy, the universe is a bit happier. I mean it, I am up for it as much as time allows me.
There could be some interesting discussion going on there.

 

[Taraquin]

Can contribute with a higher 5600X score, medium though and my 3060ti at mining uv/uc. 4.8GHz, 4000cl16 B-die tuned.

Det0x: Good job compiling! Think I recognize you from diskusjon.no

A bit interesting that a 5600X close to matches a 10900K@5.5GHz, 4cores extra with 4800 ram-ish. Amd has come a long way since Ryzen gen 1.

 

[Felix123BU]

Look @Det0x , and all who might be interested in this topic, CPU powah in SOTR benchmarks!

Shadow Of The Tomb Raider - CPU Performance and general game benchmark discussions | TechPowerUp Forums

Lets not spam this thread anymore with off-topics, shall we?

 

[nguyen]

Now this is getting old @Det0x , if the CPU results don't change between Medium and Highest, you should have post Highest in the first place. It seems your ego got hurt when others just kindly reminded you.
Now you are just spamming the thread with results nobody cares but you?
What else? post some synthetic CPU benchmark to show off your 5950X?

 

[Felix123BU]

Now this is getting old @Det0x , if the CPU results don't change between Medium and Highest, you should have post Highest in the first place. It seems your ego got hurt when others just kindly reminded you.
Now you are just spamming the thread with results nobody cares but you?
What else? post some synthetic CPU benchmark to show off your 5950X?

"What else? post some synthetic CPU benchmark to show off your 5950X? " ..... don't goad him, he just might because as he said "The rest of the world use this benchmark as a cpu/memeory bench"
Staying on topic and respecting others is for conformists and millennials (not that I see what millennials have to do with staying on topic, but hey, if @Det0x says it, must be true)

 

[the54thvoid]

Okay folks, no more name-calling, distractions or thread derailing.

Another thread has been created by @Felix123BU for CPU comparisons.

Shadow Of The Tomb Raider - CPU Performance and general game benchmark discussions

Hello World! Since Shadow Of The Tomb Raider is a particularly CPU intensive game, or better said, CPU performance has a major impact on FPS, it would be interesting to see how our setups handle the game. For that, lets settle on very specific in-game settings: Fullscreen Exclusive Fullscreen...

www.techpowerup.com

As far as this thread is concerned, it's a mess. So many variables it doesn't seem to have much of a point. I'll keep it open for now as it's quite active but it'd be nice to see some structured guidelines and folk following them.

Also - bench scores taken from other forums don't count. TPU member scores only.

 

[nguyen]

10875H + 2070 Super MaxQ (90W)

Seems like my laptop perform real close to newer high-end laptop with 5900HX + 3080 (115-130W)

 

[Space Lynx]

this is stock gpu/stock cpu and SAM off in BIOS. prob could hit 200 fps avg if i did mild oc's and turned on SAM. not sure

 

[Taraquin]

BTW: CPU game score is a bit lower running highest vs lowest. I get 20 fps more running lowest consistently

In my testing with Zen3 + Ampere SAM lowers the cpu game performance by ~ 10-15 fps.
Would be interesting if you could check if the same was true for Zen3 + Navi21

Since SAM loads CPU more it's not a surprise. I think the results on both Ampere and RDNA2 will be similar.

 

[Det0x]

BTW: CPU game score is a bit lower running highest vs lowest. I get 20 fps more running lowest consistently


Since SAM loads CPU more it's not a surprise. I think the results on both Ampere and RDNA2 will be similar.

*edit*
You are correct

 

[Taraquin]

Maybe due to 5950X have more work capasity? My 3600 experienced the same. I was able to get 160 cpu game medium 1080p, but at 720p lowest I could get 180fps.

 

[jboydgolfer]

Your 8600K seems to be underperforming. Have you tweaked ram/oc cpu?

no, i wasnt running any OC on that CPU. i had RAM & CPU/GPU at stock speed

 

[Taraquin]

You are correct and i am mistaken.
Lowest quality do affects CPU Game numbers

Used the following settings to test this:
Allcore 4.5ghz
SMT and SAM disabled
View attachment 195235

1440P
Highest =256fps average
Medium = 249 fps average
Low = 252 fps average
Lowest = 268 fps average

1080p
Highest =254fps average
Medium = 247 fps average
Low = 254 fps average
Lowest = 275 fps average

720p
Highest =247fps average
Medium = 252 fps average
Low = 247 fps average
Lowest = 276 fps average

Those Russians.. so that's why they are using lowest settings
I will remove all custom and lowest quality entries on my chart, thanks for sharing

Looking for a new 9900k highscore then and the 11900 goes from 250fps to 234 fps, same as 5600x lol

Interessting! Highest would be more of a gpu test while lowest should be used for cpu test Medium is kinda pointless as it pushes neither very hard.

 

[jboydgolfer]

No point to discuss this further, like with many millennials, feelings are more important than facts.
And i won't stoop to that level with namecalling and/or telling others what they should or shouldn't do.

In my view we are not sheep and such should not follow just because it feels comforting, when faced with a better option..
Anyway, this is the fastest rocketlake ive seen to date:
250fps average CPU Game
View attachment 195120


*edit typo*

I was trying to figure out why this dude score was so high I didn’t realize he was using whatever settings got the highest results.

i was asking myself, how with that hardware, is his score so high?
theres no point in just using whatever settings give favorable result, as it renders the standardized benchmark moot.

EDIt
many of these results are using non standard settings. pointless.

Settings: DX12, 1080p, ultra high

 

[thesmokingman]

I was trying to figure out why this dude score was so high I didn’t realize he was using whatever settings got the highest results.

i was asking myself, how with that hardware, is his score so high?
theres no point in just using whatever settings give favorable result, as it renders the standardized benchmark moot.

EDIt
many of these results are using non standard settings. pointless.

View attachment 195252

And lol he apparently didn't even know the difference and back to spamming again.

 

[Felix123BU]

@thesmokingman
The rules of the internet say that at least one special person must pop up from time to time . They go by many names, trolls, shills, fools etc. But this guy was right on one point, its not really mine or another users right to moderate this, this is the role of a moderator. If this crap continues, this thread will die, classical thread hijacking if I have ever seen one

 

[Lew Zealand]

Hey, let's SotTR bench something stupid because it's what I got. Otherwise known as...

A shoddy bomb casing full of used pinball machine parts:

R3 1200 OC to 3.7GHz (@1.3v), 5600XT, 8GB 2800 MHz CL16 DDR4

I had most of the parts laying around so bought an ultra-cheap case, decent PS and B450 and Bob's Your Uncle, you have a CPU-bound gaming machine waiting for AMD CPUs to become affordable in 2023. Frequent frame drops to the mid 20s w/CPU pegged @100% when playing Horizon ZD @1440p Prefer Quality (High equiv.).

Living the dream.

 

[Felix123BU]

Hey, let's SotTR bench something stupid because it's what I got. Otherwise known as...

A shoddy bomb casing full of used pinball machine parts:

View attachment 196484

R3 1200 OC to 3.7GHz (@1.3v), 5600XT, 8GB 2800 MHz CL16 DDR4

I had most of the parts laying around so bought an ultra-cheap case, decent PS and B450 and Bob's Your Uncle, you have a CPU-bound gaming machine waiting for AMD CPUs to become affordable in 2023. Frequent frame drops to the mid 20s w/CPU pegged @100% when playing Horizon ZD @1440p Prefer Quality (High equiv.).

Living the dream.

Not bad, at least in regards to SOTTR, hey, at least you have a decent GPU and a B450, lots of options for an CPU upgrade on the table in the near future that would drastically change you performance.

 

[QuietBob]

R3 1200 OC to 3.7GHz (@1.3v), 5600XT, 8GB 2800 MHz CL16 DDR4

Compared with my result of a 3300X @ 4.5 all core there's barely any difference in the "CPU game" metric. Seems odd.

 

[Lew Zealand]

Duh, I didn't look at your numbers before typing below, but my suspicion of funny business in my benchmark run stands. Your CPU minimums are way higher than mine at 84 FPS where mine is 42. However I'm not sure I'm comparing the correct thing, is CPU Game the only thing that matters? And if so, what's the point of CPU Render? I really dunno what these 2 mean.

Honestly, while watching the benchmark which I've done on 6 or 7 different setups, it looked like it was doing something wrong early in the third stage before the camera descends into the village and the NPC count tanks the FPS thanks to 4c4t. But that would only affect the top end FPS, but not the minimum FPS or averages much, which are visible as high frametimes early and late in the overall benchmark. I'll do it again or a bunch of times and see if I get something drastically different.

BTW I just came out of playing about an hour of RiseotTR and this setup was actually doing rather better than I'd expected at 1440p Very High (like Highest in SotTR) but Textures one step down to High (65fps with dips into the 40s). It's very clear that there's a performance bottleneck with this setup loading or streaming the max textures as FPS is notably smoother with those down 1 notch. Lots of dips to the 20s with Very High textures. VRAM usage hovers around 5800MB with Very High textures and 3100MB with High textures.
_______________

Update:

@QuietBob gets the Brass Ring.

There are a number of missing NPCs in the Market section right at the end of the SotTR benchmark when I test it with the R3 1200 and 5600 XT. I've run this benchmark on a number of other systems and they are always there, even on a 4c8t NUC with the same GPU in an TB3 eGPU case, but here they are missing. Here's who's apparently in the loo during the benchmark:

* Lady on her knees in the central stall who usually sits up on her knees to greet a passerby. The passerby is there doing something else.
* All 3-4 people usually at the end stall near Lara: the kid and his dad, the butcher, and sometimes there's someone else next to him? I forget if there's a 4th person there.

Other notes: at least 2 other NPCs are doing something else, walking around the street and getting right in front of the camera, which you usually don't see in the benchmark. Lara's creeping on looking the opposite way as there's nobody at the butcher's.

This is weird and could be an enforced thing (??) on very low end CPUs as it's repeatable on this one. Ugh, I have a 2c4t NUC that I can eGPU with this to go even lower end but maybe I'll just kick one of the kids off their R5 2600s and knock their setup down to 2c4t in Ryzen Master and try it there. It sucks being curious but too lazy to satisfy it.

Maybe kills this as a good benchmark, I wonder if people with 10900Ks or 59x0s actually get more NPCs?

 

[QuietBob]

Duh, I didn't look at your numbers before typing below, but my suspicion of funny business in my benchmark run stands. Your CPU minimums are way higher than mine at 84 FPS where mine is 42. However I'm not sure I'm comparing the correct thing, is CPU Game the only thing that matters? And if so, what's the point of CPU Render? I really dunno what these 2 mean.

Honestly, while watching the benchmark which I've done on 6 or 7 different setups, it looked like it was doing something wrong early in the third stage before the camera descends into the village and the NPC count tanks the FPS thanks to 4c4t. But that would only affect the top end FPS, but not the minimum FPS or averages much, which are visible as high frametimes early and late in the overall benchmark. I'll do it again or a bunch of times and see if I get something drastically different.

BTW I just came out of playing about an hour of RiseotTR and this setup was actually doing rather better than I'd expected at 1440p Very High (like Highest in SotTR) but Textures one step down to High (65fps with dips into the 40s). It's very clear that there's a performance bottleneck with this setup loading or streaming the max textures as FPS is notably smoother with those down 1 notch. Lots of dips to the 20s with Very High textures. VRAM usage hovers around 5800MB with Very High textures and 3100MB with High textures.
_______________

Update:

@QuietBob gets the Brass Ring.

There are a number of missing NPCs in the Market section right at the end of the SotTR benchmark when I test it with the R3 1200 and 5600 XT. I've run this benchmark on a number of other systems and they are always there, even on a 4c8t NUC with the same GPU in an TB3 eGPU case, but here they are missing. Here's who's apparently in the loo during the benchmark:

* Lady on her knees in the central stall who usually sits up on her knees to greet a passerby. The passerby is there doing something else.
* All 3-4 people usually at the end stall near Lara: the kid and his dad, the butcher, and sometimes there's someone else next to him? I forget if there's a 4th person there.

Other notes: at least 2 other NPCs are doing something else, walking around the street and getting right in front of the camera, which you usually don't see in the benchmark. Lara's creeping on looking the opposite way as there's nobody at the butcher's.

This is weird and could be an enforced thing (??) on very low end CPUs as it's repeatable on this one. Ugh, I have a 2c4t NUC that I can eGPU with this to go even lower end but maybe I'll just kick one of the kids off their R5 2600s and knock their setup down to 2c4t in Ryzen Master and try it there. It sucks being curious but too lazy to satisfy it.

Maybe kills this as a good benchmark, I wonder if people with 10900Ks or 59x0s actually get more NPCs?

Those are some interesting findings! I just ran the benchmark on a 4C/4T Phenom II. After multiple runs I can say for sure that the "missing" NPCs are there: the kneeling vendor, the butcher and the cook, and their two customers. I have noticed, though, that some of the walking customers will appear in a different place in subsequent runs.

Now I'm confused. Could it be that I'm running the demo (build 505) and you're benching with the full version (build 298)?

The way I understand this benchmark, "CPU Game" measures how effectively the CPU executes general game logic (physics simulation, NPC AI, collision detection, sound propagation, and all the underlying game mechanics). "CPU Render" shows how fast the processor can feed the necessary data to the graphics card, and "GPU" indicates how fast the card can present the final image on the screen.

"GPU Bound" at 100% would mean that the GPU cannot process all the information it receives from the processor in real time, resulting in a "GPU bottleneck". Conversely, at 0% the GPU would have to wait to be fed by the processor most of the time, aka "CPU bottleneck". In both cases we would see lower than optimal frame rates. The bigger the discrepancy in processing power between the two components, the lower the ultimate FPS.