Performance metrics in real world scenarios.

Grimlaking

2[H]4U
Joined
May 9, 2006
Messages
3,245
I see the reviews of the 9900k vs the Ryzen platform where it is always pulling ahead and that is awesome. I'm still going Ryzen... but what I would love to see that I haven't seen documented is real world.

I load up windows 10 I have a few things that are loading up in the background.
Epic Launcher,
Origins launcher
GOG launcher
Uplay
Blizzards Launcher
Steam
MSI's updater
Corsair CUE
headset software as I'm not using a Corsair headset. (yet)
Gigabytes Video Card Controller
MS's security software,

And that's all from the top of my head.

What I want to know is how do these CPU's with graphics cards work in real world solutions. I love that the Intel is faster on a per thread basis for gaming in this scenario. (Oh lets not forget about patches! Security and otherwise.) But what if you have all of the common launchers running? How about a few chrome tabs on some pages that auto refresh. (Wowhead). Oh and don't forget your Twitch client that will be running, Oh yea and Discord too!

You see my point. Can anyone point me to some real world numbers for modern tests of CPU/GPU performance?
 
Performance is both quantifiable ("measurebators" who measure and analyze data and only go with faster "read newer HW") and subjective nonscientific comparisons (i.e. "feels faster").
For example, my 5820k rig OC clocks lower than my sons 8700K OC but his system is not as snappy as mine (X99 versus 1151).
I buy solely on actual computational needs and price. Hard to compare HW unless everything in your test setups are the same. Impossible to do across AMD and Intel's HW.
I buy based on my preference not some measurbators review. I still read and respect the reviewers expertise and do use it as a baseline when purchasing.
Hard to verify reviewers true expertise and scientific competence,not peer reviewed like my publications are.
I spent 30 yrs as a professional scientist and research director so I know about taking measurements and data. Higher performance is not always better.
I like AMD now over Intel better price/performance.
Brand bias is huge among GPU buyers, if I remember Kyle did a blind free sync to g sync test and most chose the lower performing AMD cards to the Nvidia top tier cards.
 
Last edited:
I see the reviews of the 9900k vs the Ryzen platform where it is always pulling ahead and that is awesome. I'm still going Ryzen... but what I would love to see that I haven't seen documented is real world.

I load up windows 10 I have a few things that are loading up in the background.
Epic Launcher,
Origins launcher
GOG launcher
Uplay
Blizzards Launcher
Steam
MSI's updater
Corsair CUE
headset software as I'm not using a Corsair headset. (yet)
Gigabytes Video Card Controller
MS's security software,

And that's all from the top of my head.

What I want to know is how do these CPU's with graphics cards work in real world solutions. I love that the Intel is faster on a per thread basis for gaming in this scenario. (Oh lets not forget about patches! Security and otherwise.) But what if you have all of the common launchers running? How about a few chrome tabs on some pages that auto refresh. (Wowhead). Oh and don't forget your Twitch client that will be running, Oh yea and Discord too!

You see my point. Can anyone point me to some real world numbers for modern tests of CPU/GPU performance?

You brought this up on the other forum and I've tried to make it clear why reviewers don't do things that way. I'll reiterate here. There are two reasons why reviews don't do that. The way multi-tasking works, you generally aren't going to even notice these things as they do very little to nothing in the background. Check your taskmanager and their CPU time will be so minimal that its hardly worth mentioning. We also do not do reviews with that much software on the test machine because it introduces variables into the testing that we may or cannot always account for or control. It introduces potential for software problems with the OS as well as increased setup time for each test system. Testing for reviews doesn't always go smoothly, and troubleshooting can eat up hours or even days of time you don't often have. When I did the 3900X review, I had to benchmark not just one or two systems, but a slew of them for comparisons as it was our first CPU review on that site.

I didn't have time for all that nonsense. I had the CPU for about 7 days before the review needed to be done. I should have included the other CPU's AMD sent us at the time but couldn't do it because the time just wasn't there to accomplish all of that. Some review sites don't have other jobs they do, but I have a day job. That said, as I said before, I did have allot of that list running on the test systems. All the system configurations I use tend to have most of the game launchers on them. In the past, I wouldn't have even connected the test bench systems to the internet, and now I have to.

As to the second part, of course these CPU's and graphics cards work in real world situations. Why wouldn't they work just from having those things running? Some of the testing scenarios are real world. For example, I test Destiny 2 in our various reviews as its popular right now, looks good, and I actually play the game on what little I have that passes for free time. But, there is no canned benchmark for that. So, I simply play the game for a bit in the most repeatable way possible and show the data I collect. Some of the tests such as Premiere Pro or After Effects are scripted, but they are very much real world examples that you can rely on. Lastly, if these processors and graphics cards didn't work in the real world in real systems, you'd hear about it frequently on various forums and Youtube videos. Destiny 2 being incompatible with the Ryzen 3000 series is a prime example of this. It was known the very day these processors were in the hands of the general public that this was a problem.

I actually knew about the Destiny 2 problem a week ahead of time, but had at that point assumed it was something on my end. I ended up dropping Destiny 2 from that article as I was unable to make it work and ran out of time to troubleshoot it. Thankfully I had a relatively clean system. Had it included extra bullshit software, my troubleshooting task would have been even harder as I wouldn't have known what was specifically to blame for the issue. Some review hardware ends up in our personal machines as well. Thus, this stuff gets used in the same conditions as most people would use it, or perhaps worse. If I found some rampant issue with a CPU in a condition outside the normal testing, I would either include that scenario as a test or I would make another article out of my findings once enough evidence was gathered. Why? Simple. It's extra content which means extra hits.
 
Last edited:
Back
Top