UserBenchmark Adjusts Algorthym to Favor Single-Thread Performance, Causing Strange Results

[Franck]

Still useless change for me. When will i have IPC per seconds per core stats directly that i don't have to waste 10 minutes per processor in order to compare them to know which one is actually the fastest for each specific software we make. It's not difficult it's a simple base value which they hide everywhere and you have to rely on wikichips or things like that to figure it out. 

[Bananasplit_00]

Well then, guess user benchmark isn't as useful for ultra quick overviews anymore. Really strange that single core is now somehow more important than before. Are there even relevant games that can't use more than a single core anymore? I don't know of any

[MMKing]

Why is the combined score even a thing? Why not just rank all the CPUs in 3 different categories ''gaming'', ''workstation'' and ''desktop''. Let the end user decide which metric to list the CPUs by.

 

cpu-monkey.com is a better web page all round. No silly combined leader board. You pick the metric you want the CPUs sorted by. You also get much more detail on the CPUs.

[Stroal]

9 hours ago, wanderingfool2 said:

It all depends on workload really.  The way I look at it, I can see the merits of putting more weight on single core performance, because in my opinion the larger majority (and not talking about tech people here), will likely not require as many threads (and single core performance is likely to make things feel snappier).

 

An example being people who may utilize excel...it typically is stuck to one core....or when using a web-browser, while it may utilize more threads it is very unlikely to utilize all of the threads (and in some real world cases I found it more pinning a single core with the others doing smaller tasks).  So I am just saying I can see merit in the fact of weighting single core performance heavily.  (That isn't to say I would prefer a faster 2 core over 4 core vs 8 core...but I think with most people here being more tech oriented there is a tendency to forget that the non-techy people I think outnumber us)

Ok, but an i5 4770k is being shown as better than an i7 4770k (https://cpu.userbenchmark.com/Compare/Intel-Core-i7-4770K-vs-Intel-Core-i5-4670K/1537vs1538). That's insane. I mean no one in their right mind would recommend an i5 4670k right now, but non-tech people use this to make purchasing decisions, and that's a problem. 

 

Add in the fact that next gen consoles are going to be on Zen 2, and not jaguar, multi core + multi threads are going to be the way of the future. 

 

 

[Brooksie359]

11 hours ago, Arika S said:

Userbench should not be used for any serious comparison, it's good to get a general idea, but you should ALWAYS go and find full reviews on both products and compare them yourself

Honestly I dont even think it's useful for a general idea either. It's so off sometimes that it isn't reliable enough to give a general idea. 

[Princess Luna]

Yeah they broke userbenchmark now it's useless for the only thing it was decent at: rough yet simple estimations for newbies.

 

Now it's no better than CPU Boss.

[Neftex]

4 hours ago, Blademaster91 said:

Except if Intel were involved in Userbench being so skewed towards fewer cores, they wouldn't be showing the as already mentioned result of a i3-9350KF being "better" than a i9-9980XE.

I've never found Userbench very useful though, even before they changed the scoring, it was obviously a site someone shouldn't use to make a decision without looking at real benchmarks first.

the type of person whos looking to buy 9980XE doesnt give a damn about userbenchmark. its for people who type "CPU X vs CPU Y" into google to look up whats better in their budget and click first few links. this affects majority of non tech people

[poochyena]

10 hours ago, Arika S said:

Why do you need anything more than a general idea when comparing old stuff?

I want to know how much faster one cpu is over another. I can't do that with reviews since reviews never compare current CPUs with much older ones. A review telling me the ryzen 3700 gets 200 fps in CSGO doesn't tell me anything if idk how much my current cpu gets.

[pas008]

never really liked userbench at all

 

 

but single core is king for me and majority of software out there period atm

 

but times are changing for multicore so this is what they should of had from the get go

 

[poochyena]

12 hours ago, pas008 said:

but single core is king for me and majority of software out there period atm

Its not so much a problem that extra weight is given to single core speed, since it is more important, the big problem is its given way too much weight. Right now the site says a cpu with ~10% faster single core speed is slightly better than a cpu with 800% faster multi-core speed. No one in their right mind would recommend sacrificing 800% multi-core speed for a small 10% single core speed outside of VERY specific use cases. 50% or 100% multi-core speed? Sure, but not 800%.

[Beskamir]

On 7/24/2019 at 8:16 PM, emosun said:

I like how slowly over the last decade cpus are adding more and more cores to solve the speed problem..... yet somehow theres this small very vocal few who think a quad core is the cpu of the future.

I think userbench is catering to the mostly gamer user base as opposed to really measuring hardware speed.

Well multithreaded programming is a fair bit harder than just using a single thread so that's probably partly why it's not catching on as quickly and as much as it should be.

[ZcanKal]

18 hours ago, Bananasplit_00 said:

Well then, guess user benchmark isn't as useful for ultra quick overviews anymore. Really strange that single core is now somehow more important than before. Are there even relevant games that can't use more than a single core anymore? I don't know of any

Minecraft..?

[Mihle]

Can Userbenchmark please go and die?

Now it's totally useless and misleading.

 

I bet non techies and children will continue to use it and be mislead to buying a worse product...

[MMKing]

What do we want? More cores

When do we want it? Every year

 

9350 4 core > 9960 16 core
6% faster single core speed

3% faster quad core speed

400% behind on multi-core speed

= 2% faster than 9960

 

3800x 8 core > 3900x 12 core

3% single core speed

2% quad core speed

42% behind on multi core speed

= 1% faster than 3900x

 

So what is going on here. Is that Userbench is not actually comparing individual CPUs against one another. They seem to be testing system against each other. In essence, a system with a 3600 + 2080Ti vs a 3950X + bargin bin GPU for a display output. Will add the scores against each other, and the system will believe the 3600 is a fantastic gaming CPU while the 3950x is practically useless.

 

It seemingly takes all these metrics, and spits out a % as a relation to the ''best CPU'' the 9900k, which is a baseline 100% in gaming, desktop and workstation. All other CPUs are ranked with the 9900K as the base line. So the % you see on the different metrics on other CPUs, is how well it performs in relation to the 9900k. The 9700k is 98% in gaming and desktop, and 76% in workstation, making it the 4th best CPU. While the 3900x is 94% in gaming and desktop, and 131% in workstation. Making the 3900x the 6th best CPU... obviously.

 

The benchmark leaves allot to be desired as well. The 3D demos would have been impressive had they been released in the mid 90s. All in all, Userbench is worse than first suspected.

[Loote]

I can see a method to this madness, with the increasing number of cores, the multi core score increases nearly linearly , which means in 4/8 CPU the multi core could amount to 50% of total score, and in 12/24 CPU it'd be 75%, they tried to find a compromised and failed miserably, that's all.

[mr moose]

I think people are confusing single core performance with only having a single core.     Needing a 6 core CPU does not mean single core performance is not important. I am a light  to moderate gamer and until I decided to get in to video editing 4 cores was well enough for me.  Maybe sometime in the future the minimum requirement will be 6 cores or 8 cores to run word and edit a photo for facebook, but for now it is still four cores and for majority of users single threaded performance is more important. 

 

Besides, when was the last time anyone actually got useful information out of those aggregate comparison websites?

[Blademaster91]

20 hours ago, Neftex said:

the type of person whos looking to buy 9980XE doesnt give a damn about userbenchmark. its for people who type "CPU X vs CPU Y" into google to look up whats better in their budget and click first few links. this affects majority of non tech people

But anyone building their own gaming pc isn't going to find userbenchmark as useful info,or would also look for actual reviews. As for it being in one of the first few links, there are reputable benchmarks in the first few links also.  And "non tech" people aren't going to care what CPU is in their prebuilt PC.

[pas008]

10 hours ago, poochyena said:

Its not so much a problem that extra weight is given to single core speed, since it is more important, the big problem is its given way too much weight. Right now the site says a cpu with ~10% faster single core speed is slightly better than a cpu with 800% faster multi-core speed. No one in their right mind would recommend sacrificing 800% multi-core speed for a small 10% single core speed outside of VERY specific use cases. 50% or 100% multi-core speed? Sure, but not 800%.

what if you are saving 1 min a task and doing 80 tasks a day you are more productive

 

simple fucking math

 

single core is king for the time being plain and simple majority of software shows this but for the software that more cores matter see below

if i needed multi core and saved the same

same difference hence why x series and tr exists period for those workloads

[poochyena]

58 minutes ago, pas008 said:

what if you are saving 1 min a task and doing 80 tasks a day

then multi-core speed is better. What task are you saving a minute in due to a 10% increase in single core speed and 90% less multi-core speed?

[Kukielka]

This made me chuckle.

[pas008]

1 hour ago, poochyena said:

then multi-core speed is better. What task are you saving a minute in due to a 10% increase in single core speed and 90% less multi-core speed?

Many things too many to list

 

But for my needs cad and gaming

Single core performance is king for my needs in both

I'm liking more and more gaming devs are implementing multi core but when you are playing at higher res it's starting to matter less anyways

 

 

 

 

 

[emosun]

12 hours ago, Beskamir said:

Well multithreaded programming is a fair bit harder than just using a single thread so that's probably partly why it's not catching on as quickly and as much as it should be.

So we just need every programmer from 2003 to work their way out of the industry before it can move on

[S w a t s o n]

Yikes, could they have made it look like they were in Intel's pocket any further than this?

[MeatFeastMan]

Very uh...interesting timing. I mean, there was definitely no launches recently was there...nothing to speak of! I smell a few bribes from a certain blue themed company to ensure those blue themed processors would top the tables.

 

They know the sales are with the 10 year old Fortnite lovers who are only interested in a 2080ti and 9900k because they've been told by their Intel/Nvidia influenced famous streamer friends that AMD heats your house up in the winter. They'll be looking at those userbenchmark tables and thinking the same they've always felt about AMD. Shame on UserBenchmark.

 

But no longer! Whether those battle royale streamers like it or not, the Ryzen 5 3600 WILL overcome those challenges and shut those Intel-kissing youtube and twitch streamers right up. Even the most Intel biased fanboys know that the 3600 is coming for that marketshare.

 

In all seriousness though, it's time people were made more aware that it's okay to buy AMD. The tech youtubers are working hard on it, but it's being cancelled out by rubbish like we've seen with userbenchmark. The bribing must stop.

 

 

 

 

 

[mg7134]

On 7/24/2019 at 10:28 PM, SenioRR said:

I mean the site was already a very rough way of estimating performance, although quiet convenient so a lot of new users would look at UserBenchmark.

I mean just look at this:

https://cpu.userbenchmark.com/Compare/Intel-Core-i3-8350K-vs-AMD-Ryzen-TR-2990WX/3935vsm560423

Or this:

https://cpu.userbenchmark.com/Compare/Intel-Core-i7-4770K-vs-Intel-Core-i5-4670K/1537vs153

Even this!

https://cpu.userbenchmark.com/Compare/Intel-Core-i3-9350KF-vs-Intel-Core-i9-9980XE/m775825vsm652504

 

Either the people behind UserBenchmark have lost all of their brain cells OR there was a sizeable cheque sent their way.

Anyway, I usually go to GamersNexus for benchmarks but an upcoming PC enthusiast might look at UserBenchmark as a reliable source of information on part performance.

I agree because no one care about overall performance