Question If I score good in tests like timespy (3D mark) and passmark can I be confident my system is good if I compare scores?

Jul 24, 2020
20
0
10
0
Im also curious if the percentile rating is a good way too. I use passmark and 3dmmark and get a different percentile but theyre both good.
 

Math Geek

Champion
Ambassador
it's an awesome thing to know if you're wanting to compare your score to others. the real question is "does it perform well in what i need it to do for me?"

do you get decent fps for the system in games you play? does it run your massive database and spreadsheet like you need it to? does it do ___ good enough like i need it to?

that's the true measure of the system in my opinion. i know people live and die by these numbers but real world performance in what i ask it to do is what really matters.
 
Jul 24, 2020
20
0
10
0
it's an awesome thing to know if you're wanting to compare your score to others. the real question is "does it perform well in what i need it to do for me?"

do you get decent fps for the system in games you play? does it run your massive database and spreadsheet like you need it to? does it do ___ good enough like i need it to?

that's the true measure of the system in my opinion. i know people live and die by these numbers but real world performance in what i ask it to do is what really matters.
I think i get reasonable fps but i just want to make sure its actually a good fps. Im hoping benchmarks will solve that. Will a benchmark test like 3dmark prove im performing as expected when i compare to others?
 

Math Geek

Champion
Ambassador
to a certain extent yes. but you have no idea what they did to get the score you're comparing with. did they optimize windows real well? or have a bunch of stuff running so their score is probably lower than it should be?

i'd say so long as you're top 25% or so, you're probably getting what you should. you can also look for others get in the same game with similar systems to see how it looks.
 
Jul 24, 2020
20
0
10
0
to a certain extent yes. but you have no idea what they did to get the score you're comparing with. did they optimize windows real well? or have a bunch of stuff running so their score is probably lower than it should be?

i'd say so long as you're top 25% or so, you're probably getting what you should. you can also look for others get in the same game with similar systems to see how it looks.
but if i compare scores using a trusyworthy benchmark i should be good
 

Math Geek

Champion
Ambassador
sure. it'll tell you how you compare to others using the same benchmark. will give some idea how it performs vs others. but i'd not spend too much time stressing over the results. someone may get a bit higher but that does not mean you are "underperforming", it just meas they did something different to their system that changes the results.
 

madmatt30

Titan
Ambassador
3dmark is fairly accurate for benchmark comparisons because you can choose variables related to overclocking etc.

Passmark and userbenchmark not so much because it doesn't allow for these variables.

Ie - if youre running an i7 at stock and 30% of the others on there with the same cpu are overclocked to 5ghz+ (which they absolutely are) it lowers your percentile rating drastically as an overall comparison.
 
Jul 24, 2020
20
0
10
0
sure. it'll tell you how you compare to others using the same benchmark. will give some idea how it performs vs others. but i'd not spend too much time stressing over the results. someone may get a bit higher but that does not mean you are "underperforming", it just meas they did something different to their system that changes the results.
ah ok so if i get aaround the same score im fine?
 
Jul 24, 2020
20
0
10
0
3dmark is fairly accurate for benchmark comparisons because you can choose variables related to overclocking etc.

Passmark and userbenchmark not so much because it doesn't allow for these variables.

Ie - if youre running an i7 at stock and 30% of the others on there with the same cpu are overclocked to 5ghz+ (which they absolutely are) it lowers your percentile rating drastically as an overall comparison.
does userbenchmark purposefully rank AMD lower? If i use 3dmark and get around the same score I can be confident im performing alright?
 

madmatt30

Titan
Ambassador
does userbenchmark purposefully rank AMD lower? If i use 3dmark and get around the same score I can be confident im performing alright?
No it doesn't rank anything lower.
If you open up the detailed results in comparisons you can see what clockspeeds people were running at.

3dmark is kind of a bragging tool amongst the overclocking community so a big percentage on there who post results will have heavy overclocks on both cou and gpu.

3dmark allows you to see those details so you can make a direct comparison to people who are running stock speeds same as you.

With passmark/userbenchmark you cant see overclocking details so when it gives you a percentile ornranking on your hardware its skewed because many users will be running much faster clock speeds than you.
 
Jul 24, 2020
20
0
10
0
No it doesn't rank anything lower.
If you open up the detailed results in comparisons you can see what clockspeeds people were running at.

3dmark is kind of a bragging tool amongst the overclocking community so a big percentage on there who post results will have heavy overclocks on both cou and gpu.

3dmark allows you to see those details so you can make a direct comparison to people who are running stock speeds same as you.

With passmark/userbenchmark you cant see overclocking details so when it gives you a percentile ornranking on your hardware its skewed because many users will be running much faster clock speeds than you.
ah ok. what if i compare passmark scores with non overclocked people, like review sites? and basically if i compare scores on timespy with non overclocked parts im good?
 

beorn

Distinguished
Nov 23, 2010
69
13
18,535
0
No it doesn't rank anything lower.
If you open up the detailed results in comparisons you can see what clockspeeds people were running at.

3dmark is kind of a bragging tool amongst the overclocking community so a big percentage on there who post results will have heavy overclocks on both cou and gpu.

3dmark allows you to see those details so you can make a direct comparison to people who are running stock speeds same as you.

With passmark/userbenchmark you cant see overclocking details so when it gives you a percentile ornranking on your hardware its skewed because many users will be running much faster clock speeds than you.
I thought the percentile rank Passmark gives you after you run it is based on your personal score, not the average of all the different CPUs of your type. Their list is no doubt an average, but your personal score isn't
 

madmatt30

Titan
Ambassador
I thought the percentile rank Passmark gives you after you run it is based on your personal score, not the average of all the different CPUs of your type. Their list is no doubt an average, but your personal score isn't
Well you're right, your score is your score but there's no point in a scoring system without comparing that score to something else?

3dmark is easily the best comparison tool because of the detail it goes into regarding cpu clocks, gpu clocks and ram speed.

My ryzen 7 with 5700xt scores right down on 3dmark (bottom 20%) because the cpu is only overclocked slightly to suit my needs and I've actually downclocked/undervolted the GPU slightly for silent running.
 
Jul 24, 2020
20
0
10
0
Well you're right, your score is your score but there's no point in a scoring system without comparing that score to something else?

3dmark is easily the best comparison tool because of the detail it goes into regarding cpu clocks, gpu clocks and ram speed.

My ryzen 7 with 5700xt scores right down on 3dmark (bottom 20%) because the cpu is only overclocked slightly to suit my needs and I've actually downclocked/undervolted the GPU slightly for silent running.
true. so if i compare im good and not underperforming if my score matches up or is around with others who are non overclocked like me?
 

madmatt30

Titan
Ambassador
true. so if i compare im good and not underperforming if my score matches up or is around with others who are non overclocked like me?
Yes, use the comparison option on the 3d mark results site then work down the table roll you get to similar scores s your own.

You can then directly compare to see what clocks etc they were running at and if your perfoemance is comparable.
 

Gam3r01

Titan
Moderator
You still havent answered the big question though. Why?
Benchmarks can be useful to see if your system is lacking in some way, say overheating causing thermal throttles, etc compared to generally similar systems.

Beyond that theres no real reason to compare scores.
The only thing running a benchmark really does is tell you how good you are at running that benchmark, it says nothing much of your overall system experience.
 
Jul 24, 2020
20
0
10
0
You still havent answered the big question though. Why?
Benchmarks can be useful to see if your system is lacking in some way, say overheating causing thermal throttles, etc compared to generally similar systems.

Beyond that theres no real reason to compare scores.
The only thing running a benchmark really does is tell you how good you are at running that benchmark, it says nothing much of your overall system experience.
I just want to know if my system is performing as expected. If it scores well then i can be confident my system is doing alright,
 

ASK THE COMMUNITY