Ryzen 3000: userbenchmark changed the score weights in favor of intel

[Kedas]

As predicted benchmark tool(s) are changed after Ryzen 3000 (to favor intel)

The first one, and there is always a reason why they needed to change it now.
https://wccftech.com/userbenchmark-...ulti-core-importance-and-raising-single-core/

 

[PotatoWithEarsOnSide]

It is great to know that a Pentium is a better CPU than a 3900X...except if you actually use it.

PS, I love how the image says "beware the army of shills that will sell ice to Eskimos."
Not sure if it was meant to be ironic...

 

[maddogmcgee]

The idiots who bought a Core i9-7960X for $2000 a year and a half ago must be really upset to find out the Core i3-9350KF is a way better CPU.

 

[Abwx]

Cinebench FTW....

 

[amrnuke]

CPU "Avg bench" is now way off. 3800X rated higher than 3900X despite losing to it in almost all benchmarks I see.

They say "UBM effective speed measures performance that's relevant to typical consumer use" but on the actual CPU info page they say the scores "best represent typical gaming fps performance with a single number."

Typical consumer use is not gaming fps performance. It's a combination.

But they don't care, they have an agenda.

Some of the stupidest anomalies:

1) The 3900X beats the 9900K in productivity (by 13% according to TechPowerUp), the 9900K beats the 3900X in gaming, and somehow they have the 3900X 5% lower.

2) They have the 8600K behind the 3600X, which is also weird.

3) Their pricing on the 9900K and 3900X.

It's not just flawed. It's clearly biased.

Just waiting for the GPU update. It won't be long before the 5700XT falls below the 2070S

 

[deathBOB]

Why do you care how some random company rates your favored CPU?

 

[Kenmitch]

Why do you care how some random company rates your favored CPU?

One should never shift the goal posts during the game.

The uproar is over the timing as it makes it look suspicious as to why it was done.

I wouldn't use their data to select a cpu, but the less informed might do so.

 

[Markfw]

Why do you care how some random company rates your favored CPU?

First, even Intel is hurt as we all know that for gaming (the reason they said they did it) needs 6 cores, sometimes 8 cores, and they are trying to discount anything over 4. But AMD is hurt worse.
I also agree with @Kenmitch , that the uninformed may use this data, and its suspicious as to the timing.

 

[rbk123]

Why do you care how some random company rates your favored CPU?

Not really rocket science. It all depends on how prevalent this random company's scoring is used out there. If no one uses it, then no one should care. If many people use it....

 

[Abwx]

What is ridiculous is that 5% better SC or 4C perf count as much as 2x the core count even when scaling from 4 to as low as 8C, while they re at it they should have renamed their test UselessBenchmark...

 

[Zucker2k]

Userbenchmark trolling AMD for that no-show 5GHz chip...

 

[.vodka]

https://twitter.com/x/status/1154315570009071616

Just lol. This site is now CPUboss levels of fail.

I wonder how much did Intel pay them to skew results so much towards ST and QC just to tarnish AMD and hurt their own high end products in the process. 2% MT is ridiculous.

2% is something valid for I don't know, Q6600 launch back in 2007, not 2019.

 

[amrnuke]

https://twitter.com/x/status/1154315570009071616

Just lol. This site is now CPUboss levels of fail.

I wonder how much did Intel pay them to skew results so much towards ST and QC just to tarnish AMD and hurt their own high end products in the process. 2% MT is ridiculous.

2% is something valid for I don't know, Q6600 launch back in 2007, not 2019.

It's crappy.

UBM's response is that their CPU rankings are supposed to reflect gaming. Which is misleading because "Avg Bench" says nothing about "Gaming". One might be misled to assume the 9350KF is better than the 9980XE all around.

Regardless, fortunately their default sort is still by user rating.

 

[rbk123]

Userbenchmark trolling AMD for that no-show 5GHz chip...

Now this is funny.

 

[Ottonomous]

Userbenchmark trolling AMD for that no-show 5GHz chip...

I guess Intel's HEDT and hyperthreaded offerings are being trolled too, what a laugh, totally worth sacrificing the little credibility your site has

 

[DrMrLordX]

The idiots who bought a Core i9-7960X for $2000 a year and a half ago must be really upset to find out the Core i3-9350KF is a way better CPU.

I guess Intel's HEDT and hyperthreaded offerings are being trolled too, what a laugh, totally worth sacrificing the little credibility your site has

I figured the userbenchmark changes would lower scores for some Intel chips as well. How ridiculous.

 

[DominionSeraph]

The amount of estrogen in this thread is overwhelming.

I seem to be the current King of Cores doing actual work on this board (as Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters), and yet even I can't arsed to get off an i7 4790 as my main as 4 cores are all I need for anything except the final crunch. The fact is desktops are losing market share to dual-core laptops and phones as those are apparently all the majority of users need. The idea that the difference between having 99 and 100 cores on a workstation CPU is somehow a life-or-death issue to the general populace's computing needs is quite farcical. Nearly no personal computing workloads are embarrassingly parallel and adding more cores past 4 does nothing for the average user but waste electricity on silicon that they will never use. You AMD fanbois may be desperate to recommend the 2990WX to every grandmother whose Pentium M laptop just died, but the fact is their user experience simply isn't going to be any better for the difference between that and a quad. The performance simply isn't there for their use and so any comparison chart seeks to condense the information for them which sits there and ranks by embarrassingly parallel performance is nothing but BS.
I am using my i7 4790 over my dual X5670 workstations because four fast cores beats 12 slower ones for general use. Userbenchmark has a multicore score which would clearly show the dual X5670's have better multicore performance -- something which anybody who runs these types of tasks can easily scroll down to see. Their overall ranking, however, mirrors my own, and so I take no issue with it whatsoever.
You can pick up four 10-core Xeon E7-4870's for the price of a new quad and get a 4P motherboard for cheaper, too. That doesn't make them a better choice for the majority of people, and a recommendation chart that is aimed at that majority rather than the tiny minority who can fill out those 40 cores should probably reflect this.



Trolling and calling AMD enthusiasts, fanboys, is not allowed.


esquared
Anandtech Forum Director

 

[Fanatical Meat]

The amount of estrogen in this thread is overwhelming.

I seem to be the current King of Cores doing actual work on this board (as Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters), and yet even I can't arsed to get off an i7 4790 as my main as 4 cores are all I need for anything except the final crunch. The fact is desktops are losing market share to dual-core laptops and phones as those are apparently all the majority of users need. The idea that the difference between having 99 and 100 cores on a workstation CPU is somehow a life-or-death issue to the general populace's computing needs is quite farcical. Nearly no personal computing workloads are embarrassingly parallel and adding more cores past 4 does nothing for the average user but waste electricity on silicon that they will never use. You AMD fanbois may be desperate to recommend the 2990WX to every grandmother whose Pentium M laptop just died, but the fact is their user experience simply isn't going to be any better for the difference between that and a quad. The performance simply isn't there for their use and so any comparison chart seeks to condense the information for them which sits there and ranks by embarrassingly parallel performance is nothing but BS.
I am using my i7 4790 over my dual X5670 workstations because four fast cores beats 12 slower ones for general use. Userbenchmark has a multicore score which would clearly show the dual X5670's have better multicore performance -- something which anybody who runs these types of tasks can easily scroll down to see. Their overall ranking, however, mirrors my own, and so I take no issue with it whatsoever.
You can pick up four 10-core Xeon E7-4870's for the price of a new quad and get a 4P motherboard for cheaper, too. That doesn't make them a better choice for the majority of people, and a recommendation chart that is aimed at that majority rather than the tiny minority who can fill out those 40 cores should probably reflect this.

Marks trying to cure F’ing cancer with his set up.

 

[maddie]

The amount of estrogen in this thread is overwhelming.

I seem to be the current King of Cores doing actual work on this board (as Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters), and yet even I can't arsed to get off an i7 4790 as my main as 4 cores are all I need for anything except the final crunch. The fact is desktops are losing market share to dual-core laptops and phones as those are apparently all the majority of users need. The idea that the difference between having 99 and 100 cores on a workstation CPU is somehow a life-or-death issue to the general populace's computing needs is quite farcical. Nearly no personal computing workloads are embarrassingly parallel and adding more cores past 4 does nothing for the average user but waste electricity on silicon that they will never use. You AMD fanbois may be desperate to recommend the 2990WX to every grandmother whose Pentium M laptop just died, but the fact is their user experience simply isn't going to be any better for the difference between that and a quad. The performance simply isn't there for their use and so any comparison chart seeks to condense the information for them which sits there and ranks by embarrassingly parallel performance is nothing but BS.
I am using my i7 4790 over my dual X5670 workstations because four fast cores beats 12 slower ones for general use. Userbenchmark has a multicore score which would clearly show the dual X5670's have better multicore performance -- something which anybody who runs these types of tasks can easily scroll down to see. Their overall ranking, however, mirrors my own, and so I take no issue with it whatsoever.
You can pick up four 10-core Xeon E7-4870's for the price of a new quad and get a 4P motherboard for cheaper, too. That doesn't make them a better choice for the majority of people, and a recommendation chart that is aimed at that majority rather than the tiny minority who can fill out those 40 cores should probably reflect this.

I've heard of poking the bear, but you seem intent on poking the whole animal kingdom. Good luck.

 

[rbk123]

The amount of estrogen in this thread is overwhelming.

I seem to be the current King of Cores doing actual work on this board (as Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters), and yet even I can't arsed to get off an i7 4790 as my main as 4 cores are all I need for anything except the final crunch. The fact is desktops are losing market share to dual-core laptops and phones as those are apparently all the majority of users need. The idea that the difference between having 99 and 100 cores on a workstation CPU is somehow a life-or-death issue to the general populace's computing needs is quite farcical. Nearly no personal computing workloads are embarrassingly parallel and adding more cores past 4 does nothing for the average user but waste electricity on silicon that they will never use. You AMD fanbois may be desperate to recommend the 2990WX to every grandmother whose Pentium M laptop just died, but the fact is their user experience simply isn't going to be any better for the difference between that and a quad. The performance simply isn't there for their use and so any comparison chart seeks to condense the information for them which sits there and ranks by embarrassingly parallel performance is nothing but BS.
I am using my i7 4790 over my dual X5670 workstations because four fast cores beats 12 slower ones for general use. Userbenchmark has a multicore score which would clearly show the dual X5670's have better multicore performance -- something which anybody who runs these types of tasks can easily scroll down to see. Their overall ranking, however, mirrors my own, and so I take no issue with it whatsoever.
You can pick up four 10-core Xeon E7-4870's for the price of a new quad and get a 4P motherboard for cheaper, too. That doesn't make them a better choice for the majority of people, and a recommendation chart that is aimed at that majority rather than the tiny minority who can fill out those 40 cores should probably reflect this.

Now *this* is a truly awesome rant. Magnificentlyincoherent, devoid of logic, and so completely off the mark you can't help but admire the plane of reality he swims in.

Please post more often, my good sir.

 

[Markfw]

The amount of estrogen in this thread is overwhelming.

I seem to be the current King of Cores doing actual work on this board (as Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters), and yet even I can't arsed to get off an i7 4790 as my main as 4 cores are all I need for anything except the final crunch. The fact is desktops are losing market share to dual-core laptops and phones as those are apparently all the majority of users need. The idea that the difference between having 99 and 100 cores on a workstation CPU is somehow a life-or-death issue to the general populace's computing needs is quite farcical. Nearly no personal computing workloads are embarrassingly parallel and adding more cores past 4 does nothing for the average user but waste electricity on silicon that they will never use. You AMD fanbois may be desperate to recommend the 2990WX to every grandmother whose Pentium M laptop just died, but the fact is their user experience simply isn't going to be any better for the difference between that and a quad. The performance simply isn't there for their use and so any comparison chart seeks to condense the information for them which sits there and ranks by embarrassingly parallel performance is nothing but BS.
I am using my i7 4790 over my dual X5670 workstations because four fast cores beats 12 slower ones for general use. Userbenchmark has a multicore score which would clearly show the dual X5670's have better multicore performance -- something which anybody who runs these types of tasks can easily scroll down to see. Their overall ranking, however, mirrors my own, and so I take no issue with it whatsoever.
You can pick up four 10-core Xeon E7-4870's for the price of a new quad and get a 4P motherboard for cheaper, too. That doesn't make them a better choice for the majority of people, and a recommendation chart that is aimed at that majority rather than the tiny minority who can fill out those 40 cores should probably reflect this.



Trolling and calling AMD enthusiasts, fanboys, is not allowed.


esquared
Anandtech Forum Director

I know you can't reply for a few days, but do you realize that I an trying to assist 2 of the most prominent universities, Berkley and Stanford, in trying to cure cancer ? Not to mention that I have cancer, and lost to it, I have no bladder now, or prostate, and the cancer drugs they gave me to try and cure the cancer, only made me deaf and dizzy. I am legally deaf now.

Not to mention all the other insane statements you have made that have no bearing in reality. I just sold all my 14 core Xeons, why ? becuase they are slower than the 8 core Ryzen 3000 series, and they use more power.

Sorry all for the OT rant, but I think its deserved.

Lets get back on topic.

 

[Space Tyrant]

I seem to be the current King of Cores doing actual work on this board (as Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters),

Yet somehow Mark doesn't seem to be the one e-peening and virtue signalling here...

 

[DrMrLordX]

The amount of estrogen in this thread is overwhelming.

And people wonder why there aren't more women in tech.

Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters)

Already addressed. But hey, don't let that stop you.

You AMD fanbois may be desperate to recommend the 2990WX to every grandmother whose Pentium M laptop just died

. . . what?

Their overall ranking, however, mirrors my own, and so I take no issue with it whatsoever.

So you think the i3-9350k is faster than the 9980XE? In anything?

Do you think the i3-9350k is comparable to the 9900k? In anything?

a recommendation chart that is aimed at that majority rather than the tiny minority who can fill out those 40 cores should probably reflect this.

Lots of people are buying budget 6-core and midrange 8-core CPUs, and modern software - even games! - can use them all. Don't you think the recommendation chart should reflect that instead of trying to steer everyone to quads?

 

[PotatoWithEarsOnSide]

I've never seen the letters "ore" substituting for "unt" in a more appropriate post before.
Such a stretched attempt to justify an unjustifiable change in a terrible benchmark.

 

[TheGiant]

The amount of estrogen in this thread is overwhelming.

I seem to be the current King of Cores doing actual work on this board (as Markfw looks to be doing nothing with his but destroying the planet trying to pointlessly increment some online counters), and yet even I can't arsed to get off an i7 4790 as my main as 4 cores are all I need for anything except the final crunch. The fact is desktops are losing market share to dual-core laptops and phones as those are apparently all the majority of users need. The idea that the difference between having 99 and 100 cores on a workstation CPU is somehow a life-or-death issue to the general populace's computing needs is quite farcical. Nearly no personal computing workloads are embarrassingly parallel and adding more cores past 4 does nothing for the average user but waste electricity on silicon that they will never use. You AMD fanbois may be desperate to recommend the 2990WX to every grandmother whose Pentium M laptop just died, but the fact is their user experience simply isn't going to be any better for the difference between that and a quad. The performance simply isn't there for their use and so any comparison chart seeks to condense the information for them which sits there and ranks by embarrassingly parallel performance is nothing but BS.
I am using my i7 4790 over my dual X5670 workstations because four fast cores beats 12 slower ones for general use. Userbenchmark has a multicore score which would clearly show the dual X5670's have better multicore performance -- something which anybody who runs these types of tasks can easily scroll down to see. Their overall ranking, however, mirrors my own, and so I take no issue with it whatsoever.
You can pick up four 10-core Xeon E7-4870's for the price of a new quad and get a 4P motherboard for cheaper, too. That doesn't make them a better choice for the majority of people, and a recommendation chart that is aimed at that majority rather than the tiny minority who can fill out those 40 cores should probably reflect this.



Trolling and calling AMD enthusiasts, fanboys, is not allowed.


esquared
Anandtech Forum Director

you have a point for joe default user, enthusiasts here have higher expectations
but you should have said it in better way

dont dismiss what other ppl run on their computers, even if they loop cinebench and that makes them happy