Question Reliable CPU Benchmark site that compares laptops vs mobile vs desktop CPUs?

milkygirl

Distinguished
Jun 12, 2013
68
0
18,630
0

USAFRet

Titan
Moderator
Mar 16, 2013
147,288
9,323
175,390
23,020
I used a few of the following and it doesn't seem to give a realistic representation of the performance across platforms. Any suggestions for a cross-platform benchmark check?

https://cpu.userbenchmark.com/Compare/AMD-Ryzen-5-3600-vs-AMD-Ryzen-5-5600G/4040vsm1553183
UBM is worse than useless for this.
Strongly biased against AMD


Desktop vs mobile is just as bad, no mater who is doing the compare.
Totally different software.
 
Reactions: drea.drechsler
I also have to think comparing a cross-comparing notebook/laptop, tablet and desktop computers is an exercise in frustration that should be avoided. Reason is these three platforms are heavily optimized for entirely different use cases where things like battery life and carry weight is increasingly more important as performance compromise is increasingly tolerated as you move from desktop to laptop/notebook to tablet. That makes it nigh on worthless to compare performance.
 

milkygirl

Distinguished
Jun 12, 2013
68
0
18,630
0
This doesn't have mobile (snapdragon) benchmarks and no M1 Pro/Max benchmarks too..

to others: I understand it's not apple-to-apple comparison when comparing against different platforms, but I just want to get a sensing of how close phone chips are to overtaking laptop/desktop variants (E.g. If snapdragon 888 is run on a desktop, what's the desktop equivalent intel/amd model?
 

USAFRet

Titan
Moderator
Mar 16, 2013
147,288
9,323
175,390
23,020
This doesn't have mobile (snapdragon) benchmarks and no M1 Pro/Max benchmarks too..

to others: I understand it's not apple-to-apple comparison when comparing against different platforms, but I just want to get a sensing of how close phone chips are to overtaking laptop/desktop variants (E.g. If snapdragon 888 is run on a desktop, what's the desktop equivalent intel/amd model?
You can't really compare.

Phone software is completely different than desktop, or even laptop.
 
Benchmark software is one thing.

Compiled OS and applications are something else.
The problem I have with "realistic tests" is what you, I, and everyone else uses the computer for is different. For example, I don't care about compiling OSes and applications because I almost never do that. But I will care about how well the CPU does in Handbrake because that's something I'll do. But even then, those have too many variables that cross comparing is kind of a wash

So I'd rather go with benchmarking tools where everyone is likely using the default suite of tests.
 
This doesn't have mobile (snapdragon) benchmarks and no M1 Pro/Max benchmarks too..

to others: I understand it's not apple-to-apple comparison when comparing against different platforms, but I just want to get a sensing of how close phone chips are to overtaking laptop/desktop variants (E.g. If snapdragon 888 is run on a desktop, what's the desktop equivalent intel/amd model?
You are probably not going to find anything even close to what you are looking for, and specially not in one place. What you are asking will require a lot of research and comparing the results of many reviewers online.

If you wana know how good or bad a software may run (and please keep in mind the word "may" is very important) then you will have to check several websites/youtubers, which will all use different approaches and methodologies for benchmarking/testing.

As you wrote it wont be apples to apples, most likely it wont even be apples to elephants.

I believe the devices using M1 Pro/MAX were just announced 3 days ago, you wont see many benchmarks out there, not till later.

As for phone chips overtaking laptop and dekstop variants, no one can tell when it will happend or how close we are. If it ever happend at all. No one have such a powerfull crystal ball. One thing is for sure if it ever happned, it wont happend tomorrow, and probably not anytime soon.
 
Last edited:

USAFRet

Titan
Moderator
Mar 16, 2013
147,288
9,323
175,390
23,020
The problem I have with "realistic tests" is what you, I, and everyone else uses the computer for is different. For example, I don't care about compiling OSes and applications because I almost never do that. But I will care about how well the CPU does in Handbrake because that's something I'll do. But even then, those have too many variables that cross comparing is kind of a wash

So I'd rather go with benchmarking tools where everyone is likely using the default suite of tests.
Its not a case of performance while compiling.

Rather...the theoretical same application is compiled differently for each platform.
The same application for mobile is not the same application for desktop. Even if the name is the same and the buttons are in the same place.

Inspect the "same application" for mobile and desktop at the machine code level, they are different.
 

ASK THE COMMUNITY