UserBenchmark offers explanation for changes to CPU score weights

[Evernessince]

I fail to see what's wrong here. Single core, single threaded ipc performance IS more important for power users. They didn't change much, only increased the weight from 30% (which I'd say it was indeed too low) to 40% - in fact, if it was up to me, I'd put the weighted average of single core performance at 50% at the very least, unless my benchmarks were targeted at corporate consumers for whom multi-core performance might indeed have more weight.
Anyways, their timing for this change couldn't be worse. Of course there would be drama especially from AMD fanboys. This coming from someone who's an opportunistic buyer and doesn't favor any companies.

"Power users" like wendel would vastly prefer to have multi-core performance so it really depends what you mean with that term. Given that a majority of programs and the windows operating system are well multi-threaded, I do not see the sense in increasing the weight of single thread.

50% single thread? Really? So according to you a single core 6 GHz processor is better then a 5GHz 9900K? That thing will choke running firefox and chrome, let alone gaming. I think you ought to reconsider your metrics.

 

[arrowflash]

"Power users" like wendel would vastly prefer to have multi-core performance so it really depends what you mean with that term. Given that a majority of programs and the windows operating system are well multi-threaded, I do not see the sense in increasing the weight of single thread.

50% single thread? Really? So according to you a single core 6 GHz processor is better then a 5GHz 9900K? That thing will choke running firefox and chrome, let alone gaming. I think you ought to reconsider your metrics.

Well, clock speed usually doesn't mean anything when comparing single threaded performance between different CPUs, it's the ipc that matters. But for sake of argument, I'm going to assume you mean a hypothetical single core, single threaded 9900K @6GHz vs a standard 9900K @5GHz, so 20% better theoretical single thread performance (in practice, a 20% increase in clock speed doesn't always mean a 20% performance increase even for single threaded apps/benchmarks).
Maybe I'm missing something but in this scenario, even with 50% weight for single-threaded performance, wouldn't the comparatively abysmal performance of the single threaded CPU in the quad-core and multi-core tests severely weigh down its rating, so it would end up falling behind the standard 9900K in the overall benchmark? In this case the standard multi-core 9900K would still rightfully be considered the better CPU.

 

[Cycloid Torus]

Umm, I like my 6c/12t cpu and it isn't AMD either. Multi-core is where it will be at.

 

[Evernessince]

Well, clock speed usually doesn't mean anything when comparing single threaded performance between different CPUs, it's the ipc that matters. But for sake of argument, I'm going to assume you mean a hypothetical single core, single threaded 9900K @6GHz vs a standard 9900K @5GHz, so 20% better theoretical single thread performance (in practice, a 20% increase in clock speed doesn't always mean a 20% performance increase even for single threaded apps/benchmarks).
Maybe I'm missing something but in this scenario, even with 50% weight for single-threaded performance, wouldn't the comparatively abysmal performance of the single threaded CPU in the quad-core and multi-core tests severely weigh down its rating, so it would end up falling behind the standard 9900K in the overall benchmark? In this case the standard multi-core 9900K would still rightfully be considered the better CPU.

It was a hypothetical situation. Given that high weight is for single core and quad core, the highest clocking quad core will be ranked the highest. Anything over 4 cores is essentially worthless as you only get a maximum of 2 pts. You don't need to nearly go as high as the 9900K as a 9350K will match it in this test.

The point was that cores past 4 do matter in a majority of modern applications and games. You loose significant performance with only 4 cores. Case in point: https://www.techspot.com/article/1803-are-quad-cores-dead/

 

[arrowflash]

It was a hypothetical situation. Given that high weight is for single core and quad core, the highest clocking quad core will be ranked the highest. Anything over 4 cores is essentially worthless as you only get a maximum of 2 pts. You don't need to nearly go as high as the 9900K as a 9350K will match it in this test.

The point was that cores past 4 do matter in a majority of modern applications and games. You loose significant performance with only 4 cores. Case in point: https://www.techspot.com/article/1803-are-quad-cores-dead/

Oh, I definitely agree that multi-core performance should have a higher weight, and bringing it down from 10% to 2% doesn't make sense to me. I'd say it probably should be evenly split with quad core (so 40/30/30). Only didn't mention that because it wasn't the focus of the debate.

 

[Darstuer]

It's 2019 and single core performance is suddenly becoming more important and multi core performance becoming less relevant?.


LMAO. What a joke!.

 

[Markoni35]

It's 2019 and single core performance is suddenly becoming more important and multi core performance becoming less relevant?

Intel being the main sponsor of UserBenchmark would easily explain this decision. Noooo.... I'm not saying that benchmarks are biased and corrupt. Okay... yeah... that's exactly what I'm saying.

 

[Lew Zealand]

If UserBenchmark's test was actually worth a damn they would use a sliding thread count scale, not an arbitrarily proportioned mix of 3 arbitrary thread counts of 1, 4, and CPU Max.

But that would necessitate a toss of all their current arguably useless data and rewrite of their ever more outdated benchmark to do tests of one through N threads. I guess N=64 for now to account for Epyc and Xeon, but should at least be able to scale to 256 threads for at least few years of futureproofing.

Fat chance that'll happen.