Yes, it is this recent leap up the popularity of the overall performance evaluation software – UserBenchmark
UseBenchmark thinks that CPU multi-core performance over four cores is not worth noting, so its CPU evaluation weight logic is as follows:
Single-core performance: 40% of CPU performance;
Quad-core performance: 58% of CPU performance;
Multicore performance (more than four cores) : 2% of CPU performance.
This APP can run on PC, it is very green, the interface and use design is very human
The current version of the software is also only 5.55mb in size, so you can use it at a click without installing it
Are you living too long in the age of four? With the qing dynasty all gone, don’t you still accept the fact that six and eight nuclear cores have become the mainstream?
Digital pro test, 5 years ago i7-4980hq platform score geometry?
No practice has no right to speak, before the evaluation, or go to the official website to download the test software to use again.
On The UBM website homepage, The Best list is very interesting… . 9400f, GTX1660Super top the list what the heck? Just the rice?
Clicking on the CPU in the previous figure will jump to a link to a CPU test list to see if the list is ranked by performance or shipments
Let’s just “run a run” after downloading the software
This 3D test is as advanced as lu’s
A long story… . ╮ (╯ del ╰) ╭
It is a long story squared ┑ (￣ Д ￣) ┍
It is a long story after (‘ д `)… 彡… 彡
A long story +1… .
Well, after the previous rounds of advanced GPU testing, you can generate the test results by doing disk and memory testing.
There’s also a nice countdown to test completion
Brief evaluation of test results
Digital this video card’s bright machine platform is already four or five years ago the backward “advanced” configuration, but feed GTX1080 class graphics card or easy. UserBenchmark’s results don’t give you an exact score, but instead rank them in order of the results uploaded by the user (the default is automatic), giving you a rough idea of the percentage scale.
So this machine flunked 50% game performance, desktop performance barely passed 69%, and workstation performance flunked 48%
This is not to say, after all, is a few years ago the backward platform ~ look down
CPU and GPU test evaluation
The full 3.8ghz quad-core 8-thread i7 also currently gets a “Very Good” rating of over 74.7% CPU. But this RX570 graphics card is very “Average” (Average), only more than 49.8% of users, more appropriate
Disk performance evaluation (ooh, I plugged in the wrong disk data interface again, both SSDS were plugged into SATA2)
The disk performance tests were all rated “Excellent” and very nicely indicated to the user that you might have inserted the wrong disk data interface (which turned out to be true, I’m going back to SATA3 in a moment). The four 4GB DDR3 dual-channel 16GB got only “Very Good” reviews, given the DDR4 era.
Try to summarize
In terms of ease of use, the 5.5mb software does a pretty good job. But…
The 4-core i3-8350k (83.5) is worse than the 8-core Ryzen7 2700X (81.7), which I can’t understand
Yes, as shown above, even the 8-core, 16-threaded Ryzen7 2700X is a little brother to the i3-8350k in this test software. ┍ ┑ (￣ Д ￣)
Not only that… .
Please Shut the FXXK UP! U “AMD community”
The official response of UserBenchmark to users’ questions about the reduced weight of CPU testing logic and multi-core performance is as follows:
The multi-core performance of the Ryzen3000 series processors, based on the Zen2 architecture, is too high.
Reasonable suspicion that AMD used the test vulnerability to improve test scores;
Since the test software can’t get a “reasonable” score, we’ll have to lower your multicenter performance weight to get a “reasonable” score.
So now slipped Ryzen9 3900 x 2, all threads tear slipped, originally seventh i9-9900 k on title ╮ (╯ del ╰) ╭;
We don’t take money yo, our software is free, only rely on Google advertising revenue, we don’t exactly meal.
Digital gentleman’s comment on the software is:
Software testing programs are too poor to fully capture the performance of mainstream multicore processors, let alone make the leaderboard.
Test logic loopholes, but the Boss to modify the Boss blood slot, but also blame the Boss too strong;
Non-objective evaluation software, can not be called evaluation software;
It is suggested to add a pre-use clause to this software, which can be identified as a CPU with more than 4 cores.
Apologies to AMD users.