How to target the CPU in gaming benchmarks

Everything about displays and monitors. 120Hz, 144Hz, 240Hz, 4K, 1440p, input lag, display shopping, monitor purchase decisions, compare, versus, debate, and more. Questions? Just ask!
Post Reply
daninthemix
Posts: 19
Joined: 05 Sep 2017, 04:36

How to target the CPU in gaming benchmarks

Post by daninthemix » 29 Sep 2022, 01:01

What is the correct way to target the CPU in gaming benchmarks? Set everything to Ultra and run at 720p to eliminate GPU?

I want to do some tests and see whether ECO mode on my 5950x makes any difference to frame-rates, so I thought I'd run a few games with benchmarks.

Kyouki
Posts: 190
Joined: 20 Jul 2022, 04:52

Re: How to target the CPU in gaming benchmarks

Post by Kyouki » 29 Sep 2022, 04:52

Awful link but some good insights:

https://www.capframex.com/blog/post/Abo ... g%20sector...
CPU: AMD R7 5800x3D ~ PBO2Tuner -30 ~ no C states
RAM: Gskill Bdie 2x16gb TridentZ Neo ~ CL16-16-16-36 1T ~ fine tuned latency
GPU: ASUS TUF 3080 10G OC Edition(v1/non-LHR) ~ disabled Pstates ~ max oced
OS: Fine tuned Windows 10 Pro, manual tuned.
Monitor: Alienware AW2521H ~ mix of ULMB/Gsync @ 240hz/360hz
More specs: https://kit.co/Kyouki/the-pc-that-stomps-you

Bloods
Posts: 15
Joined: 08 Nov 2020, 14:02

Re: How to target the CPU in gaming benchmarks

Post by Bloods » 29 Sep 2022, 10:03

Specificity is king in this regard - Run the benchmarks with the EXACT same settings you would otherwise play with, that's the only way you are going to see if you are going to get real life improvements using ECO mode.

720p tests are fun to see for some general idea how the CPU performs with no bottlenecking in that specific game(engine) but only in a general sense.
Acer Nitro XV252QF (390HZ) VRB Off, Overdrive: Normal/Extreme) Max Brightness ON.
Asus VG248QE (144HZ)

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: How to target the CPU in gaming benchmarks

Post by Chief Blur Buster » 29 Sep 2022, 19:58

daninthemix wrote:
29 Sep 2022, 01:01
What is the correct way to target the CPU in gaming benchmarks? Set everything to Ultra and run at 720p to eliminate GPU?

I want to do some tests and see whether ECO mode on my 5950x makes any difference to frame-rates, so I thought I'd run a few games with benchmarks.
Depends on what you're doing -- as a software developer, this is an important optimization tool.

Actually disabling Ultra and using 720p + low detail, in some engines amplifies CPU bottlenecks more than 720p + Ultra (usually, with most game engine workflows). Basically, the immutable minimum processing per frame (geometry updates, physics updates, hitboxes, controller updates, etc) generally scales with increased frame rates, so low-rez + low-detail will classically produce more CPU bottlenecks than low-rez + Ultra-detail, at least in classical game engine workflows.

While best practice recommends playing with real-world settings -- sometimes synthetic tests are necessary to reveal CPU weaknesses in the refresh rate race to retina refresh rates, especially when we're working with geometric-scaling problems. e.g. revealing weaknesses to optimize (optimizing software development, optimizing for VR rendering, etc).

High 720p+Ultra frame rates are useful for DLSS 3.0 type algorithms, so a software developer working on DLSS-less system may need to optimize CPU performance for 720p, which would help other members of the same software team that has DLSS.

It also makes porting the game codebase to Nintendo Switch, Steam Deck, mobile devices, and other 720p systems easier to do. The intentional method of amplify CPU problems by intentionally configuring the game settings to CPU-limiting settings and then optimizing the game source code to reduce CPU utilization.

This is also best concurrently done with artifically lowering of clockspeeds and core-count limiting, to amplify CPU bottlenecks more. -- to assist in coding optimization work.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

Post Reply