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

[Athlonite]

True, that is my bad, second kid came, things changed

excuse accepted thankfully mines 22 and can look after himself

 

[Felix123BU]

excuse accepted thankfully mines 22 and can look after himself

With a 2 year old and a freshly born one its a bit more tricky
Anyway, on topic, I was thinking of doing that for some time, hope to also find the time, it should be some interesting comparisons keeping in mind the amount of results gathered here.

 

[Ibizadr]

This is my best score with 5800x +200mhz pbo +CO per core +2x8gb 3800mhz cl14. Next step try to reach 200fps

 

[Lew Zealand]

CPU re-test after replacing GTX 1080 with RX 6600XT and i7-9700F gets faster. OK then. Makes me want to try the 1080 again and as I'm replacing it's cooler with an aftermaket one this weekend, that does seem to be in the cards.

No that was not a pun, your brain is just broken.

 

[Ibizadr]

CPU re-test after replacing GTX 1080 with RX 6600XT and i7-9700F gets faster. OK then. Makes me want to try the 1080 again and as I'm replacing it's cooler with an aftermaket one this weekend, that does seem to be in the cards.

No that was not a pun, your brain is just broken.

View attachment 243644

Pls use AA in taa its on first page the settings

 

[Lew Zealand]

Pls use AA in taa its on first page the settings

Whoops! I'll re-test this again tonite. Lol when you think you're setting everything to Lowest and forget that it's actually another Lowest...

 

[Ibizadr]

Whoops! I'll re-test this again tonite. Lol when you think you're setting everything to Lowest and forget that it's actually another Lowest...

Its only to be a standard for everyone.

 

[Lew Zealand]

Its only to be a standard for everyone.

OP says to do this with AA off and a lot of the previous posts (including my older ones) also have AA off. The other SotTR thread targetting GPUs uses TAA. But then I see quite a few people in this thread with TAA on as well. Not consistent either way. In any case, here's the same settings as above but with TAA on:

i7-9700F, 2666 CL13 (B360), power limit set to 130W (ie: no limit as this is an 8-core "65W" CPU, lol)
RX 6600XT, 2750 MHz cores, 17600 MHz memory, +20% Power limit

Still a faster CPU score than when tested with the 1080 earlier in the thread. Should be building the 1080's cooler today and we'll see how that changes, if anything.

 

[Ibizadr]

You are right bro sorry for my mistake. Another best result with change trcdwr. This time I put AA off.

 

[lawood]

5600x + 3070 ti

 

[Det0x]

tweaking settings, not finalized scores

Things left to try: disable SMT, enable/disable rebar, new bench-only windows install and giving it a shoot with my old 2x8gb memory sticks.
Main bottleneck is CPU clocks, motherboard without external clockgen -> stuck on 100mhz baseclock = max 4550mhz ST and 4450mhz MT

 

[lawood]

That software looks useful (pbo2 tuner)
Can you share where to download? A quick google search found nothing.

 

[Det0x]

That software looks useful (pbo2 tuner)
Can you share where to download? A quick google search found nothing.

Hope this link for works for you

 

[Athlonite]

Main bottleneck is CPU clocks, motherboard without external clockgen -> stuck on 100mhz baseclock = max 4550mhz ST and 4450mhz MT

What mobo are you running with

also can you fill in your system specs in your profile so we don't have to keep asking what you're running

 

[Det0x]

What mobo are you running with

also can you fill in your system specs in your profile so we don't have to keep asking what you're running

Zentimings also show what memory i use

 

[Athlonite]

You should still be able to push 110~115MHz on your fsb though

 

[IamVoo]

tweaking settings, not finalized scores
View attachment 245109

View attachment 245110

Things left to try: disable SMT, enable/disable rebar, new bench-only windows install and giving it a shoot with my old 2x8gb memory sticks.
Main bottleneck is CPU clocks, motherboard without external clockgen -> stuck on 100mhz baseclock = max 4550mhz ST and 4450mhz MT

I've been trying to figure out where the huge change in results is coming from, if you can help me out. I understand the hardware isnt the same and youve got a more aggressive overclock on the memory and also CO on the cores but I feel like that cant be a 12000 frame difference. I could be wrong but I'd like to know for sure. I see you are running windows 11 while im on 10 so that too is another variable.

I'm on a 6800xt with a mild OC around 2500mhz but in none of the tests I've run have I shown to be GPU bound so im wondering if driver overhead from radeon or just how the 3090 performs could be a factor here or is it specifically due to your memory OC. Even with CO on a few extra megahertz shouldnt play that big of a role. I'm trying to determine if you are just that far ahead in performance, or if I'm running into technical problems somewhere I havent realized yet.

 

[Ibizadr]

I've been trying to figure out where the huge change in results is coming from, if you can help me out. I understand the hardware isnt the same and youve got a more aggressive overclock on the memory and also CO on the cores but I feel like that cant be a 12000 frame difference. I could be wrong but I'd like to know for sure. I see you are running windows 11 while im on 10 so that too is another variable.
View attachment 245363
I'm on a 6800xt with a mild OC around 2500mhz but in none of the tests I've run have I shown to be GPU bound so im wondering if driver overhead from radeon or just how the 3090 performs could be a factor here or is it specifically due to your memory OC. Even with CO on a few extra megahertz shouldnt play that big of a role. I'm trying to determine if you are just that far ahead in performance, or if I'm running into technical problems somewhere I havent realized yet.

My first advice it's to correct some of your timmings. Trc(45)=trp(15)+tras(30) and tFAW(16) 2x trtp(8). And try to running GDM disable even if you won't can't at 1t go to 2t GDM disable. Try it again to see improvement. You can improve ram even more but you need to learn the basics and do some ram test (testmem5, kahru, hci) to see if it's stable.

5600x + 3070 ti

Nice result bro

 

[phanbuey]

I've been trying to figure out where the huge change in results is coming from, if you can help me out. I understand the hardware isnt the same and youve got a more aggressive overclock on the memory and also CO on the cores but I feel like that cant be a 12000 frame difference. I could be wrong but I'd like to know for sure. I see you are running windows 11 while im on 10 so that too is another variable.
View attachment 245363
I'm on a 6800xt with a mild OC around 2500mhz but in none of the tests I've run have I shown to be GPU bound so im wondering if driver overhead from radeon or just how the 3090 performs could be a factor here or is it specifically due to your memory OC. Even with CO on a few extra megahertz shouldnt play that big of a role. I'm trying to determine if you are just that far ahead in performance, or if I'm running into technical problems somewhere I havent realized yet.

Awesome result thx for sharing - that 5800X3D is a monster.

It's way faster than 12600K w DDR5 (itx) 5.2Ghz 1.24v DDR5 6200 32-35-35-72... we do have different game versions tho... do you have the full game or demo?

Edit: the difference in your results is that you're runnin game demo - demo is way way slower. If you get full version of the game you will get better results.

Here's mine:

 

[IamVoo]

Awesome result thx for sharing - that 5800X3D is a monster.

It's way faster than 12600K w DDR5 (itx) 5.2Ghz 1.24v DDR5 6200 32-35-35-72... we do have different game versions tho... do you have the full game or demo?

Edit: the difference in your results is that you're runnin game demo - demo is way way slower. If you get full version of the game you will get better results.

Here's mine:
View attachment 245384

wait is that it? All this time no matter what I couldnt reach similar scores. Demo is just slower? This would alleviate alot of issues man, hope it's true.

I ran earlier with just XMP and also with a tune I've used on my 5800x in the past that was stable 3600cl14 and at best there is a 3fps difference and at worse no change at all. I've heard that the cache can really cover up memory issues and excels in bringing rigs with slower memory back to par with other faster memory rigs. Maybe it's just game specific where I dont see much uplift?

Regardless, knowing the demo runs much slower makes me feel alot better. That said i cant see myself buying a game just for a benchmark.

 

[phanbuey]

wait is that it? All this time no matter what I couldnt reach similar scores. Demo is just slower? This would alleviate alot of issues man, hope it's true.

I ran earlier with just XMP and also with a tune I've used on my 5800x in the past that was stable 3600cl14 and at best there is a 3fps difference and at worse no change at all. I've heard that the cache can really cover up memory issues and excels in bringing rigs with slower memory back to par with other faster memory rigs. Maybe it's just game specific where I dont see much uplift?

Regardless, knowing the demo runs much slower makes me feel alot better. That said i cant see myself buying a game just for a benchmark.

Yeah I wouldn't buy it just for benchmark but is a great game, and yes the demo is super slow compared to the real game, they've optimized it a bunch since the demo release.

Always look at the game version, different game versions are not comparable in the benchmark for SOTTR. It's for sure true - I have both and they're totally different numbers.

 

[Det0x]

Manage to break magical 400fps cpu game limit

5800x3d @ stock 4450mhz
Average CPU Game = 406 fps

If only we could find a way to overclock these cpu, for us without external clockgen for baseclock..
Imagine what this cpu could do @ 4.8 to 5ghz with unlocked multiplier

 

[harm9963]

Manage to break magical 400fps gpu game limit

5800x3d @ stock 4450mhz
Average CPU Game = 406 fps
View attachment 245747

If only we could find a way to overclock these cpu, for us without external clockgen for baseclock..
Imagine what this cpu could do @ 4.8 to 5ghz with unlocked multiplier

ASUS DOCS would do that , only two MB can do that - DARK HERO and EXTREME.

 

[QuietBob]

5800X3D with IF @ 1900, RAM @ CL14, 6600XT @ stock + SAM:

 

[freestaler]

Trial Edition, 5800x3d BLCK 104. -30, 3750 c14 IF 1:1