It's a shame tbh to side with one team. I never going to trust or use their site. Nor i will ever going to buy intel, as i would never ever would go back to 4 core, world is moving on, but intel is trying to stop it, and some of us is trying to defend poor practices.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
User Benchmark = Fake Benchmark
- Thread starter humbug
- Start date
Somethings very wrong there.
I don't disagree but there are many sites like tech advisor that don't control the active ad content and as such cpu gets hammered whilst you're doing basic browsing, on my 8t lappy such a site wavers around 40-60% cpu.
Yes I could ad block and all sorts but once you start doing that you don't always see what is expected and can break other parts of the browsing experience, years ago 1 or 2 cores was fine and now 4 is tolerable but more threads rather than outright single core pace is generally leading to a better usage experience and it's only going one way seeing as nearly every consumption device these days have 4+ threads.
Jesus, man. Install uBlock Origin and tick all the uBlock filter lists, Easylist, EasyPrivacy Adguard, and Adguard annoyances. Using 60% CPU to render adverts on an 8t CPU? Just lol. I can't remember the last time uBlock Origin broke a website.
I’m just suggesting their benchmark is reflecting real world applications/usage. I know that’s an unpopular view and it doesn’t support the ‘moar core’ majority on these boards but these boards are a microcosm.
Goes off to search for 3900x....:
The problem with it is, and i'll use the quote "you only need 4 cores" Intel's defence for a decade of 4 core CPU's, we now know how overpriced those "you only need 4 cores" CPU's were, not that all of us didn't at the time, this is Intel charging high end money for low end CPU's.
And they don't like having to compete again, this is why User Bench now prioritises "4 core CPU's" in the rankings, FFS stop singing to Intel's tunes.
humbug your 2 example apps are content creation, and its important to state even tho media reviewers and a fair part of the active OCUK base are content creators, the vast vast majority of global PC users are not.
Now is their new balance a good accurate reflection, no idea. But I do feel that the important of super heavy core counts has been hugely overstated by the media recently vs reality.
The majority of desktop software is still single threaded even in 2019.
Personally I have never used their cpu ranking charts, I just use them to help me look at specific comparisons (without weighting been applied to the tests).
Maybe they should start a serverbench website where its reversed with single core been 2% as that would reflect the server world and appease your desire for more cores.
Now is their new balance a good accurate reflection, no idea. But I do feel that the important of super heavy core counts has been hugely overstated by the media recently vs reality.
The majority of desktop software is still single threaded even in 2019.
Personally I have never used their cpu ranking charts, I just use them to help me look at specific comparisons (without weighting been applied to the tests).
Maybe they should start a serverbench website where its reversed with single core been 2% as that would reflect the server world and appease your desire for more cores.
So I have an i7 930. I don't remember what a 9900K was in effective speed but it's now 124%. A 3900X was 150% in effective speed but is now 111%. A whopping 39% decrease.
I have used this site for 5 years in determining 'how good' of an upgrade something will be. I assumed it was an unfiltered aggregate of real world performance.
I have used this site for 5 years in determining 'how good' of an upgrade something will be. I assumed it was an unfiltered aggregate of real world performance.
humbug your 2 example apps are content creation, and its important to state even tho media reviewers and a fair part of the active OCUK base are content creators, the vast vast majority of global PC users are not.
Now is their new balance a good accurate reflection, no idea. But I do feel that the important of super heavy core counts has been hugely overstated by the media recently vs reality.
The majority of desktop software is still single threaded even in 2019.
Personally I have never used their cpu ranking charts, I just use them to help me look at specific comparisons (without weighting been applied to the tests).
Maybe they should start a serverbench website where its reversed with single core been 2% as that would reflect the server world and appease your desire for more cores.
There are far more content creators than gamers, even then we know games now use 8, some even more.
Let me ask you this, who is this benchmark for?
Associate
- Joined
- 7 Sep 2012
- Posts
- 956
- Location
- Lancashire, England
I was wondering why my Ryzen 3600 score went from 109% down to 93% with Userbenchmark.
penalised for having 6 cores, you only need 4.
It's now a comparison for people that don't need a powerful processor.
Yeah, we're all doing it wrong guys, no reason to buy anything but quad cores at this point! There goes Intel's most expensive CPUs!
I dont know their intended viewers for this benchmark on the website.
There is not more content creators than gamers, think about it, fortnite has 10s of millions of players alone.
Did userbench provide a reason for the change?
To me I have never used their league table of CPU's I dont care for it.
I just use it to compare 2 cpu's and as far as I know thats still the same.
e.g.
https://cpu.userbenchmark.com/Compare/Intel-Core-i9-9900K-vs-AMD-Ryzen-9-3900X/4028vs4044
I am still presented with the same data.
single core
quad core
multi core , as well as data for diff types of tests such as sc int, sc float etc
So yeah just use the compare page and ignore the cpu rankings, I never used the cpu rankings anyway LOL. I am not a sheep I can form my own opinion of what cpu is right for me not userbench or any other site, so dont stress yourself over how they rank cpu's. Site's like userbench I have only ever used for their data "not" their opinions, as an opinion is just an opinion, not fact. Personally I dont agree with the 2% weighting, it is too low, but I also dont care. I dont work for AMD, I dont have AMD shares, why would I care?
--edit--
I see they added effective speed metric to compare page, my answer to that is I will just ignore that rating. As its formed on an opinion of which tests are more important.
https://cpu.userbenchmark.com/Faq/What-is-the-effective-CPU-speed-index/55
Last edited:
No point all whining here tbh.
[email protected] is where to direct your complaints.
I've sent them my thoughts.
[email protected] is where to direct your complaints.
I've sent them my thoughts.
I think the problem is some people worry too much about core numbers rather than how their software is performing. Does it actually matter if software is designed for four cores and we only have four cores or if it’s designed for eight and we have eight? As long as it works and it has been working. Just because there has been a space race in the core count doesn’t mean that what was working fine yesterday suddenly doesn’t work at all.
I suspect the vast majority of office workers would kill for a two core let alone a four core machine rather than some useless vm hub! The speed increase would be phenomenal.
Another thingy I consider which may be more relevant than core count is what platform the software was designed for, for example some adobe products just work better on intel (I’m sur s there’s a vice versa too!)
This is a completely meaningless and empty statement given that most software will use how many cores the CPU has to offer.
Well the point is more cores = higher performance = faster work.
Not anymore, Ryzen 3000 also does very well in those applications.
They should just add a slider control within limits of old bench to new bench or wider and allow the 'User' to control 'User benchmark' as it suits them.
Thats a good idea although it will only affect "effective speed" the rest of the data is still as it was.