Benchmarking

[Mugenjynn]

I know you guys are nerds so your gonna be thinking “oh your a noob bla bla bla.” The thing is I don’t understand benchmarking. I mean I understand the results, but I don’t know like what programs you guys are using. For example, I see that one with the dragon in the beginning like I have no idea what that program is. I can understand the difference between frames per second with gpu’s, the other numbers I don’t really understand. Anyone willing to explain these things to me? Sorry if I’m a little hard tk understand (I have a difficult time expressing my thoughts with words).

[Slottr]

3DMark, Unigen Heaven, unigen valley, etc

[vanished]

The one with the dragon is unigine heaven https://benchmark.unigine.com/heaven

 

Other popular ones include 3DMark Firestrike and Time Spy, found here: https://benchmarks.ul.com/3dmark

 

Many popular games include their own benchmarks as well.  You may have seen some from GTA V, The Witcher 3, Crysis 3, Ashes of the Singularity (one could argue this is just a benchmark and not really a game), The Division, various versions of Tomb Raider, Far Cry, and Assassin's Creed as well.

[Mugenjynn]

2 minutes ago, Ryan_Vickers said:

The one with the dragon is unigine heaven https://benchmark.unigine.com/heaven

 

Other popular ones include 3DMark Firestrike and Time Spy, found here: https://benchmarks.ul.com/3dmark

 

 

2 minutes ago, Slottr said:

3DMark, Unigen Heaven, unigen valley, etc

What about the score in the end? Are they different for each program? Also thanks

[i_build_nanosuits]

i like to use unigine superpostion i think the 4K optimized setting is good to test performance, heat and clockpseed and stability etc....it's the one i use these days.

[vanished]

1 minute ago, Mugenjynn said:

 

What about the score in the end? Are they different for each program? Also thanks

Yes, the scores are unique to each program and so scores for one program can be used to compare different hardware, but you cannot compare scores from two different programs.  Each program even has several different modes and settings that can be run, all of which will give a very different result.

[i_build_nanosuits]

2 minutes ago, Mugenjynn said:

 

What about the score in the end? Are they different for each program? Also thanks

yes they are and most of them will let you at least compare your scores with others, and some of them will let you upload it for free...but you can get a 3d mark license for cheap on steam these days and it's pretty nice.

[emosun]

passmark is bae

other programs are kinda crap when you need to compare your results with identical hardware

[Enochian]

I was thinking about the same question.. and I'am actually happy to find an existing thread

Will the 3dmark basic (free) edition be enough to benchmark my whole pc? Will this edition allow me to compare results with others? 

- When it comes to using it, are there any precautions or any preparations i need to make?  
- How does the things with settings work

[vanished]

13 hours ago, Enochian said:

I was thinking about the same question.. and I'am actually happy to find an existing thread

Will the 3dmark basic (free) edition be enough to benchmark my whole pc? Will this edition allow me to compare results with others? 

- When it comes to using it, are there any precautions or any preparations i need to make?  
- How does the things with settings work

It depends what you want to test.  Ideally, if you have one specific application or game you want to test it for, you'd be best just running that.  But if you want a more generic test, any of the ones I've listed above are good for GPU scoring.  If you want to test disk speed, you'd want something like CrystalDiskMark or ATTO.  You can test CPU performance with Cinebench or Passmark (I've actually never run Passmark myself but it's a go-to for me when looking up CPU performance) or even CPU-Z.  Memory can be benchmarked with AIDA64, which is also a great stress test tool for examining stability and temps.  Some of these tools are also able to test other things but those are what I know them for, and what I most often see them referenced for.  There's also UserBenchmark which tests everything!

 

As for precautions to take, all benchmarks are the same:

1. Make sure nothing else is stressing the system while it runs or you will get inaccurate (low) scores
2. Make sure you have adequate cooling for the part being tested.  Benchmarks don't last a super long time, but they are designed to work your hardware to the limit in order to determine maximum performance, so make sure you can handle it.

[Enochian]

Ok, now i can tell "i see what you did back there". I ran Time Spy and FireStrike for the first time and it's quite addicting to do, then to check what you can upgrade and try some more.