The ACC CPU Benchmark Thread

This is a spin-off to "The ACC Benchmark Thread" & "The ACC GPU Benchmark Thread" :

https://www.assettocorsa.net/forum/index.php?threads/the-acc-benchmark-thread.53257/
https://www.assettocorsa.net/forum/index.php?threads/the-acc-gpu-benchmark-thread.53426/

While that one is about optimizing systems and learning about which settings affect which GPU/CPU the most, this one is focussed on pure CPU performance.

You'll need to download FRAPS - it's free, lean and easy to use. You can get it here, for example:

https://www.computerbase.de/downloads/systemtools/grafikkarten/fraps/

Set up a hotkey for benchmarking and tick the checkbox at "Stop benchmark after 60 seconds".



IMPORTANT: Deactivate FPS cap and VSYNC for the benchmark run! It can potentially make your system look worse than it actually is! Adaptive sync like FreeSync and GSync can remain active.

Now, set up Quickrace like this:


Realism:

We want Full Formation Lap for consistency.

Here's a zoom-in on the left red box:



Those are the fixed video settings you need to use:

We want to reduce the GPU Usage to the lowest possible.

Start the session. Hit the button Drive. When you are in cockpit view, hit your benchmark key as fast as possible and drive the formation lap.
You'll notice this by the FPS counter reappearing in the corner of the screen.

Open this thread and post the FRAPS result from the file FRAPSLOG.txt.
Please write down your CPU, GPU and RAM and make sure your signature is up to date.

Here's the preview picture of the result table, still FHD, including all valid test runs. The original table will be attached later to this post.
FHD Benches:

720p Benches:


Looking forward to seeing your results

 

RyZen 1600 3,8Ghz:
2018-12-07 13:13:37 - AC2-Win64-Shipping
Frames: 4884 - Time: 60000ms - Avg: 81.400 - Min: 73 - Max: 99

Core i7 8750H Stock:
2018-12-07 13:52:14 - AC2-Win64-Shipping
Frames: 4865 - Time: 60000ms - Avg: 81.083 - Min: 71 - Max: 97

Laptop i7 is equally to my Ryzen

 

Hi Deatroy, did you have that?
https://www.assettocorsa.net/forum/index.php?threads/the-acc-benchmark-thread.53257/

 

Differenz is now, we test with full grid + low graphics which brings us into full CPU Limit.
Taking also a worst case situation.
When you hit benchmarking directly after spawn behind AI, everyone has 20 FPS more in min.FPS which isnt true for AI Racing or even MP later on.

But yes you are right, the old one was quite similar.
This one is just for CPU. Not combined system or GPU^^

 

2018-12-07 17:31:48 - AC2-Win64-Shipping
Frames: 4656 - Time: 60000ms - Avg: 77.600 - Min: 65 - Max: 90
ryzen7-1700

 

Sorry guys...
But in the end, what's the conclusion?
Is better a new intel i5 (8th gen) or new a amd ryzen 5 (2nd gen)?

A new i7 or a new ryzen 7?

 

definitely better an i7


anyway the game is not optimized yet, will nedd to wait .
after some more testing, i noticed that the global use of the CPU is about 30% (only 2 threads above 50% and below 40% the others) with 15 opponents in multiplayer.
The usage is higher in SP with 20 AI but below 80%.
In both situations the GPU was running at 80% max or below....so clearly lack of optimization

 

Wait for the results. In a few days this thread will surely give you an answer.

 

Not that I don't want to play by the rules of the topic, but I hate fraps. Long story not getting into it.

I very much prefer straight away dinamic benchmarks, so allow me to post two very quick videos (just two laps each) with all statistics live, but recorded with shadowplay (I just installed it to record these two, and have already uninstalled right after. I hate this program too just like fraps, not to mention the gforce experience which I avoid like the plague installing it, but at least serves to record properly at 60 and HD). Let this benchmark be for both cpu and gpu too then, killing two birds with one stone.

i74790k hyperthreading on_970G1_16GB ram ddr3 1600_ssd samsung 850 evo 500Gb. Stock cpu speed, no overclock no mobo core enhancers no nothing. Same for gpu. Windows page file active in ssd where game and steam is installed. Core parking deactivated via registry (w7 user here you know).

First video. Just two laps.., well almost, on bonnet cam which is the one I like to play with. Full grid, I begin last, electric storm:
FPS: min= 46 max = 69 (lol) average= around 52-33


Second video. Just two laps.., well almost too, with the cockpit view you requested with a fixed fov of 45, was it. Full grid, electric storm, I begin last of the pack again.
FPS: min=46 max=71 average: around 54-55


The in-game graphic settings I use you can see it at the end of each video. Same for both anyway.

And yes I'm no youtuber no nothing. You can click on the videos without worrying I earn money or something. Just have that channel to upload few sporadic stuff when requiered, just as it is now. Thought this method would be best.

Hope this helped.

 

Could you repeat these tests also with hyperthreading to off?

It will not be effectively half the cores, since HT is far from doubling the cores but it will be interesting.

Did you applied lowest settings for graphics?
Your gpu is at 99% so your tests are not valid.

Στάλθηκε από το m2 note μου χρησιμοποιώντας Tapatalk

 

@Deatroy we can see from fernadito that this test is still GPU bound.
Check his 970 is always at 100%
We need less resolution or stronger GPUs or he is not using your low settings.

Above test is certainly not valid .

Edit, juSt saw videos. Definitely not lowest settings.

@Fernandito repeat please.

Στάλθηκε από το m2 note μου χρησιμοποιώντας Tapatalk

 

^yup sorry didn't see op's gpu request thing. Anyway why isn't 99% valid?, it is still not maxing out the gpu, so those results should be valid. It even drops to 98% sometimes.

If I go higher presets the gpu marks a fixed 100% of course and there's a loss in performance, but 98-99% for a gpu is still ok without shuttering or other issues that might affect performance, even at a 99-sometimes 100% mark. Exceeding the 85% usage mark-onwards is troublesome and performance loss begins when it comes to cpu and ram usage, specially with cpu, but with gpu it's ok. With those presets I have found the cool spot in my case, and the rest comes down to developers optimization mostly and further polishment of the engine/game, so those results should apply for op's test.

ok i'll repeat without HT tomorrow (results won't be good anyway). I have to reinstall that shadowplay crap again and also already disassembled the wheel already from my desk, as I'm going out now.

 

99% usage at GPU is effectively the same as 100%.
You need to put graphics settings to lowest settings in order to utilize your CPU. It's really simple. GPU needs to be way below usage percentage to let CPU go to 100% instead.

Στάλθηκε από το m2 note μου χρησιμοποιώντας Tapatalk

 

I have to disagree man. 98-99 gpu usage is not effectively 100%. gpus are prepared for max workloads without seeing an impact on performance, specially when you have very low temps like in this case ranging from 45-50º, which is a laughably cool temp for a gpu almost at max speed. Problem at maximum performance on gpus begin only when you exceed a temp treshold, which is pretty high though too.

And I didn't see btw that in the cockpit video there's drops even down to 95%. This shouls apply ok for the test too, as he wanted the cockpit view for everybody.

It's with cpus usages of 85% onwards that you r performance gets a hit. And with ram most of the times too.

but this cpu will never go 100% usage regardless the low graphical settings. The only way is turning off HT as you asked me (I'll do tomorrow, word), but four cores only is not optimal for this game, so the test should not be either convincing or apply at all. I mean, this game needs for an optimal use either a 4 core+6threads cpu (minimum), or a six physical core cpu straight away.

 

Hmmm

Dude, just read what I wrote. You can cross check this also reading on internet.

Also you will realize all these, when you will do the tests as adviced. You will see way higher fps, way higher cpu usage across all cores and way less GPU usage (thankfully away from 99% , which is the same as 100% not as just a figure but practically also).

Then your tests will be valid.

Στάλθηκε από το m2 note μου χρησιμοποιώντας Tapatalk

 

...at least you're hilarious

 

I'll do I'll do, but I have just seen that I have drops down to even 95% in the cockpit video, which should comply with ops test too, as what interested him the most was everybody in cockpit view.

And pardon if I sound just a little bit stubborn, but man.., 95-96-97-98-99% usage on a gpu is perfectly fine without any loss on performance. It is with cpus only when exceeding the 85%mark hits performance. There's many differences between a gpu chip and a cpu chip. Cpu gotta do hundreds of other task simultaneously OS wise than just moving the game, and gpu doesn't. Not to mention how much more powerful a gpu chip is.
And remember that FPS are not the most important stuff as long as you equal the same Hz refresh as you monitor. Frametime is what cuts the cheese. It needs to be steady, and as you see the game still suffers from this regardless the FPS dropping bellow 60. In this case It happens first due to optimization (game needs more polishment, it is an early access), and second due to the tremendous conditions a full grid plus that stormy weather plus those particles+ tyre sim plus AI sim plus many things It poses for your cpu/gpu without fine tuning and ok optimization (early access).

Tomorrow I do the test without ht bro. Word

ah yes you were that one who was unable to read and comprehend that topic about latest w10 update + new nvidia drivers fixing the shuttering problems introduced in last falls update.

Look, you have to prove that you can count from 1 to ten without getting lost in the process if you want to be taken in serious henceforth. Otherwise you may crawl back to your cave

 

I can't exactly follow all your logic. Your are correct that generally we would like to be away from high cpu usage and close to full utilization of gpu or at vsync

But

This is not what this thread wants to highlight and this is CPU performance that we will correctly evaluated only if GPU stays out of the equation as much as possible.

So, please do repeat your tests with exactly same settings as op putted even without fraps. Video will be ok. Then again but with HT off if you like (as option).

Στάλθηκε από το m2 note μου χρησιμοποιώντας Tapatalk

 

Comedy gold

Why do you have to be so obstinate @Fernandito? The OP made a request using certain settings to get benchmark results. If you want to do your own test, that's fine, go right ahead, happy days. There really is no need for your contrary attitude in almost every thread you post in. You challenge people without posting any evidence to your claims, and resort to name calling when somebody disagrees with you.
You are of course entitled to your own opinion and healthy discourse is good for the game and life in general. So please just reign in the attitude a little, it'll do everyone some good. x

Edit - quote fail!..

 

it would be better you to shut off