is AMD Ryzen 7000 actually FASTER than Intel? | (Ryzen Latency Benchmarks)
HTML-код
- Опубликовано: 19 сен 2023
- Is the Ryzen 7950X a snappy CPU? Does it have low input latency? If so is it good enough compared to Intel CPUs vs say Adobe Premiere pro and fast desktop usage? What about opening files, dragging and dropping in windows 10 and 11? Well with today's tests and benchmarks we do all that!
Sponsor Use Coupon (BFTYC) and get 35% off from VIP-SCDKey
Windows10(15$) // bit.ly/BFTYC
Windows11(21$) // bit.ly/BFTYC11
Office2021(49$) // bit.ly/BFTYC2021
✅Shop Aliexpress WorldWide: s.click.aliexpress.com/e/_etUbxJ
✅Shop on Ebay Worldwide: rover.ebay.com/rover/1/711-53...
❤️Become a Tech YES City member and get access to perks
/ @techyescity
⭐Consider Subscribing here bit.ly/3G20vC1
🔐Get a Windows 10/11 Pro Key for CHEAP (Use Coupon BFTYC) - www.bit.ly/BFTYC
💯Merch - www.redbubble.com/shop/techyes...
❤️Support Directly - / techyescity
💻Discord Access - / discord
-------------------------------------------------------------------------------------------------
DISCLOSURES: Generally all links tied to products are either Amazon, AliExpress or Ebay Affilaite links, this means that if you purchase a product we earn a small sales commission, which costs you nothing extra (if you end up purchasing a product). All sponsored content will contain the word "SPONSOR" if directly sponsored or "AD." Any additional revenue stream will be disclosed with similar disclosure.
Music Provided by either: epidemicsound, audio library or royaltyfreeplanet.
#Ryzen #Latency #Benchmark Хобби
For best latency used DOS, Edit with CRT monitor and PS/2 inputs.
Ha-ha. And a 486.
At least you can play Doom in DOS 😂
I think you got this phil
Turn on drive indexing, If you are reusing drives from an older build windows 10/11 will not automatically index the files on that drive. This is why you see that huge delay when looking at MP4 files. Open drive indexing options and add the old drives.
@@techyescity most australian thing I've ever heard
Modern CPU latency is something that not a lot of other people are looking into at the moment, so your reporting on this is very important. Keep it going, man!
It's not really an issue TBH. Windows is the bigger issue with all its new power savings, security updates etc. It's really bogging down.
Go back to Win10 1709 and you'll notice the difference night and day. Even LTSC 2021 is snappy AF.
People have tested Intel vs Ryzen with actual LDAT tools in games and the input latency is the same. Bryan did a desktop test where Windows is trying to save power / has different priorities / affinities etc. It really has to be in a fullscreen application for proper testing.
@@griffin1366ltsc is the way to go
@@griffin1366Wdym proper testing. He specifically said that it's not a gaming benchmark. He only cares about saving time while working on Windows. Did you watch the video at all?
There are few things that can increase responsiveness like disabling E cores, use Performance Power Profile, disabling CPU power saving features in BIOS, etc... Default Windows settings isn't really a great measurement of the CPU's latency.
@@sperrex465 Yes I did. However latency tests on the desktop are largely useless due to how Windows works.
All this snappiness stuff has got me going, wanting more from this series, hope to see more in the future after things settle down for you. It doesn't really matter where you are in the world, quite a few of us enjoy your take, and energy at getting deep into topics. !!Cheers, and keeping working that Tech Yes City Magic.
You may be benchmarking windows explorer bugs when you see latency differences in Win 11 vs Win 10 file explorer. There have been many reports of switching in and out of full screen mode Windows explorer reduces latency. It's a strange bug but it's a real thing. This could be confusing or degrading your testing results.
Facts, I have tons of bugs with windows explorer in w11 on my laptop.
Windows explorer is already a buggy pos in win10. Windows 11’s non existent testing before features are released makes that even worse.
Regarding your recommendation of enabling write caching it's important to point out that you shouldn't enable that setting unless you're using a battery powered device or your PC is connected to a UPS. Otherwise you risk data loss/corruption in an event of a power outage. I know that there is a warning regarding that next to the checkboxes that you showed in the video however some people may miss that if they're not actively watching the video and fail to notice the warning when they enable the setting.
From testing this setting a few years back (pre SSD era), I would never recommend anybody enable it. Seems to almost always result in phantom corruption, even with backup power (system also needs to never crash). On an SSD the performance impact wouldn't be noticable.
@@EricParker I'd be careful about holding on to older results for too long. You are assuming the behavior of the OS, that specific caching feature, and storage devices remains unchanged. I agree about the setting in the past, but I have not seen the same sort of issues in recent years...
@@EricParker That's because there is no setting for enabling cache scrubbing anymore, cache scrubbing resolves that problem.
Let me bring into attention an even worse scenario: ssd with caching disabled and blue screen and reboot. It can signal the OS that write completed but still corrupt data if sudden blue screen or reboot occurs. SSD can have mapping tables in its memory and randomly corrupt the filesystem severely.
I lost trust in ssd for OS drive unless it has capacitors and firmware logic to recover from such events. It's unbelievable how consumer ssd (some very expensive) can ignore this aspect and just gamble with your data. We are talking TB size drives nowadays.
I pressed the bell icon so that I can be snappy every time
These tests are equally as important as benchmarks. Everybody should be doing them.
I remember when we met at an AMD event many years ago, you did bring this matter up during our breakfast. Back then it was a 2000 series Ryzen event and we're talking about 1st gen Ryzen vs Intel, you mentioned that things felt more snappy on Intel. Looks like it's still the same today, though I've been on Ryzen system since 1000 till 7000 now.
And yes the AM4 ones do feel less snappy but I had no issues with them, doesn't bother me. 7000 series is really good though, went from 5900X to 7600 and loving it.
That's a good memory! Do you miss the extra cores going to a 6 core?
@@techyescity don't miss at all, 5900X was for video encoding but since I went for Studio version Davinci Resolve I can enable GPU encoding so the CPU core count is not required anymore.
With the 7600 tuned, I get higher clocks and IPC at far lower power draw. My PSU and AIO fans spin up far less, every fan runs at low RPM.
I really recommend you try 8 core ryzen 5000 & 7000, since they are single CCD
i'd avoid 6 core, since that's a cut die and adds core-to-core latency
oh, and 5800x3d & 7800x3d as well
EDIT: i3 10th gen might have ring bus, being a native 4 core die. and performance is similar to ryzen 3600 (in gaming, usually)
and should be very snappy :) in theory at least
Monodie Ryzen Cezanne (5500, 5600g etc) have the same core-to-core latency as regular 5000 series, but lower core-to-imc. having half the L3 cache, the performance is usually just slightly better than ryzen 3000, but should snap :)
oh, and that FCLK can go up to 2500 :)
Will get onto them. The 7700 felt really fast when I used that. Remember being surprised by it.
Yes, really interested in seeing how a single ccd would perform.
@@techyescity Desktop cpus evolving to unseen server like cpus some years ago is a trend that has some surprising but clear negative return... or deliberately diminissed performance for older characteristics cpu's so they are not a treat/thread and have better clocks, efficincy per core, cache, etc... that could make gaming or single thread task better or equal... But you may be overlooking some major factor in opening files or user input latency! The file system. Are those partitions using NFTS with MFT or GPT? are they recently formatted and installed or they have being used for years with milions of short files being written and erased? MFT can be huge, fragmented, disordered and full of unused old registers.... we are talkin latencies that go form tenths to seconds on HDD and capable of plummet R/W speed from x00/x0 MB/s to x0/x MB/s and so will do the same in SSD's on a faster scale... ms to microseconds... I revived a lot of "old computers" from slow laggy to fast as fresh installed...
@@techyescity R7 5800X/3D has infinity fabric, memory controller and ddr at 1:1:1, R7 7700X/800X3D on the other hand has only memory controller and ddr 1:1 the infinity fabric speed is to slow to keep up.
so in theory the 5000 series could be lower letancy than 7000 series
@@hliasunknown runs at 3x i think, so should be in sync. check buildzoid for better answer
Now you have to test x3d cpus... love your videos!
There it finally is! Thank you so much for making this video!
It's cool that you recognize the part of your work that you enjoy. Hope you and the fam are able to get to Australia in the most optimal situation.
Finally been waiting for this one
Lovin the content man, keep it going!
Great work - Saving us all one follicle at a time! 🙌
Thank you so much for these latency tests, which most reviewers completely neglect/overlook!
Very good quality Content. Was Helpful Learned a Lot About cpu input latency
Thnaks for your works!
Very good video! I would also add that tuning system memory can make a huge difference in system responsiveness. I could feel it when tuning memory from XMP 3600 C18 loose timings to 3800 mhz C14 and tight secondary and tertiary timings on my Ryzen 3800x, 5900x and now on the 5800x3D (not as much in x3D)
Even when Brian's videos do not hit my personal interests.. I almost always watch them. The vibe, voice, accent and flow just levels me out so I throw it on in the background. My setup is primarily a sim racing rig. When he said maybe you just suck at games I died 😂
I have to say I miss your part hunting in Australia, looking forward to more.
yes you confirmed it for me why my INTEL sometimes lag with weird issues you made my next choice clear , Thank You.
Can't wait for the video explaining how to tune windows and the BIOS for better latency. Thx for the video
Excited for this.
These are unique tests that other people aren't doing, so I love to see it!
Excellent vid.
Bryan reminding me that I might just suck at games is the harsh criticism I needed in my life. 10/10 video
Question answered! THANK YOU
Dude, you've got striking silver-blue eyes, a rugged, square jawline, you're tall and handsome. You can relax - even if you go bald. Doesn't matter.
You'll be fine either way. I hope that once you get settled you'll have the time to find that special lady that puts a skip in your step and makes you forget about all of the hardships you've gone through.
She'll be lucky, and you deserve it.
Really looking forward to the BIOS settings video
we love your videos
I really think these tests need to also be performed in linux.
edit: you could also take the series further this way by comparing latency between OS as well as just CPUs. And then between different linux distros as well.
Linux is just not mainstream enough to allocate time benchmarking it let alone multiple distros.
@@techluvin7691 Would be nice to see if these latency issues even exist in Linux or if it's just a windows issue. That's my point. It doesn't matter how many people use Linux or not.
Edit: and I believe that starting a series like that would still attract a lot of views.
@@techluvin7691 and you think testing latency is mainstream?
Curious what w10 tuning has been done because that latency mon result is bad for all 3 cpus. 13700K with highest dpc of 53us after 1 minute and chrome was open. There is a guide by windows titled "How to set up a Device for Real-Time Performance" .
Thank you for latency videos! At work I have a dell with a 12500 w/ 65w adapter and when I tried 13500 model for 2 weeks, it felt slower, even when it had more cores and a 90w ac adapter. I'm thinking maybe cause the base clock was lower even though the boost clock was much higher.
Very interesting video. We are lucky to have Tech Yes City ! Who else would have done that kind of comparison ?
Good job! Finally somebody tested everyday users pains! Thank You! I believed that I'm crazy, but now I know it's true! Fancy incredible benchmark, but higher latency... hmm I think that it's same with smart phones chipset too!!! Thank you again!
You're not going bald Brian and again thank you for the 11th Gen Intel latency report in the other video. I never really cared about the latency only because no one was doing it, however it's always nice to know. But I do care which CPU manufacturers are the most compatible with office apps/browsing (not just gaming) and power efficient. Although I won't be buying any of the 12th gen onwards, until I see this architecture's power efficiency either equal AMD or better them, as well as latency to a minor degree.
I run a 4090/7700 4k combo atm and I'm pretty happy but I look forward to the 8800X3d's release as I'd heard that Zen 5 was going to be a banger for years now.
Running a Ryzen 7 5700x, after using other ryzens in the latest years, and i do fell it has more latency then what i remember in the old Intel cpu's (the last one i used was the 7700k). For what i do it doesn't really bug me that much but i do agree it is a real issue.
Great journalism on this. And don't worry about going bald, brother. It's a powerful look that you'll like if you decide to give it a try.
You've done a good job.
I so want you to test this with the new intel ultra cpus when they come out :D
I remember upgrading from an AMD X2 4200+ to a 4800+ back in the day which had 1MB L2 cache per core vs 512KB for the 4200+, purely for better FPS in Bioshock, but noticed a definite (and unexpected) increase in responsiveness in Windows. I've been interested in UI responsiveness ever since, so this video was fascinating.
Boot up a copy of Windows 7 or 8 on older hardware with an SSD, even a 6th Gen Intel, and turn off Speedstep and the latency is phenomenal. It just feels so much nicer to work on, opening Explorer is instantaneous and everything happens in real time. Once you go Win10 and 11 it's like someone added an extra layer of translation as literally everything is slower.
after the intel 10th gen and new gen intel generations latency comparison I was looking for this video !
Recently? I was waiting on this content since early august. I'm so glad you've done it, I was really waiting! lol
+1 requesting tests on the 11900k if possible (is 11th gen as good as 10th on latency topic)
Most likely not as good considering 10900K clocks ring/ram better without latency penalties
That CPU is the last that used ring bus technology amazing low core to core latency. I've seen some Builzoid videos going down to 30 ns on a i9 10900k, amazing.
They all use ring bus, it's just that it works differently from the 11th gen up.
Interesting series. I can't help wondering if it's just a Windows 10/11 issue or if this is consistent also in Linux (I do most of my multimedia there with gaming in Windows). I just own of the recent Intels a 12700KF and it still seems snappier than the old 9600K, though perhaps the file transfer on older Intels was better, am not sure. Have spare 2nd, 6th, 9th, and am maining on 12th gen still, but it is hard to notice.
These benchmarks are way more important to me than synthetic or gaming fps benchmarks.
If the rumors are true will you do a snap test with the i3 14th gen provided it has the supposed 6 cores and no E cores? Would just be interesting to see how a basic P core intel CPU would do on the latency front.
latest news is that 6p+0e cores was canceled. Beside 14th gen is same as 13th
Japan is a great place to visit but working and living there isn't great. They treat you like an honored guest, but only ever like a guest. Even Japanese people say that it's hard to make new friends and the working culture is harsh. I don't blame you for going back to Australia.
It's so interesting that older CPUs can deliver better performance. Keep exposing this.
Quality content as usual. Thanks for all the effort you put into these.
Intel and AMD both push that their respective CPUs are going to CURE baldness with their sheer insanity of incredibly fast performance. ;) I love this focus on how the CPUs feel in day to day tasks. It also gives a bit of insight into just how much tuning can make a difference. And...for those of us with a lower tier CPU, I suspect the tuning of the OS - as well as hardware - can give noticeable day to day improvement in our overall computing experience. Thanks for taking the time to do these comparisons.
I love you man. You're really neat...
Please do an Windows tuning video shortly, from bios settings to windows itself, with security and without differentiation, and perhaps benchmark the 3 options? Base (all security, no tuning), Tuned w/ security, Tuned w/o security -- that would be interesting
Hey Brian, I'm bald its not that bad hahaha. Shaving your head is a choice vs balding. Great video man! Thanks for posting
great video series. Verry very veeeeeeeeeeeeeeeeeeery interested in the monolithic Ryzens, i.e. 5600G 5700G
Thank you! Thank you! Finally someone is speaking up about the latency issues!
BE CAREFUL THEY'RE GONNA COME AFTER YOU NOW
Always been by fear lol
I am happy for you that you could reconnect with your son, hopefully you can get your family to Australia too.
What would be interesting is to see how lower core count CPUs stack up when it comes to latency. For instance, single chiplet Ryzen CPUs like the 7600X and 7700X could very well have lower latency than a dual chiplet design like the 7950X.
That's the video I wanted
4 mins in to the video, this is real world scenario content that most people can actually relate to.
Thanks for the video, just curious if you have tested or played around the 8core 7700x as it only has one chiplets and the new bios that you can run 8000 DDR 5 on some MB does it do anything you latency or just production work at all. I ask as many people don't use any of the 3D stacked chips for photo editing or video editing. But if the new BIOS update what RAM would be a sweet spot for latency with one chiplet. As the prices are coming down. Thanks again
I'd like to see how fast each system gets the network card info sent.. Id like to see two different systems run CS or Apex on a closed server with two soldiers shooting each other simultaniously. I have a feeling you'd see a big difference and one system beeing considerable faster. I have the oppertunity to play Pubg at different pc's among my friends and how fast the hitreg works varies a lot! Having data on this would open up an entirely new discussion!
Thanks for this Byron. We gamers aren't just using our horrendously overpriced PCs for gaming anymore. My 12600k is struggling with productivity tasks and i need an upgrade :(
you know we need a 7950x3D latency test now LOL
Could you try the 13900k with highly overclocked ram (~7600 CL34 with tweaked subs and tertiaries) as a comparison?
Its a great content/topic, so don't feel weird about it!
Switching from 5600x to 5800x3d made a huge difference in exactly that, latency and overall OS performance. I did the upgrade for gaming ofc, but like most people I do day to day stuff more rather than gaming, so the x3d was fantastic and improved my day to day life a lot.
After that I got a very similar(maybe slightly better) experience from Overclocking 12400F with BCLK to 5,2ghz core and 4,6ghz cache on DDR5. The snappiness of the system was a lot higher after I OCed.
I don't know if any of these CPUs I'm mentioning are going to be good for your video editing scenarios, but I believe 13900K and 7950x are kinda the worst in OS performance and latency compared to Ryzen x3D chip(especially single CCD one) or Intel chips without the performance and efficiency cores. It would be fantastic if you can test those or similar ones as I hope my theories wont let you down in the results areas.
BTW, I did test 5800x3d vs 12400F OCed and the 12400F was consistently beating the 5800x3d in multicore, singlecore, games, benchmarks and pretty much everything even tho its 6 core 12 threads. That might be an interesting topic to REvisit as well especially now when 5800x3d is 300-350$ and 12400F is 100-150$ and the DDR5 parts are not so expensive anymore. Also Ryzen 7500F and 7600x are worse than both of those and you know the prices too.
Also, try the 11900k in this test. That chip is an OC beast and has some of the best latency scores even compared with new tech.
That chip was hot garbage, I can’t believe people even bought it. I hated every second of using that stupid chip
Interesting to test, nice latency, fine overclocker. But nowhere near 10900K/9900K in that regard.
@@TheCompyshop It was stupid from an economics perspective. From a technology perspective it was very interesting, and also quite powerful, especially on performance / core metric (the 11900K was, overall as powerful as 10900K, but with only 8 cores vs 10)
@@Winnetou17 10900K is a better processor straight away. Their ipc claims were kind of a gimmick, they optimized it to get higher synthetic scores (just like Zens usually did) but it lost most of 10900K's extreme ring/ram oc potential and came with core to core latency penalty (what those benchmarks are usually never sensitive to) which translated into barely better, and sometimes worse gaming performance.
Additionally, that cpu started to require insane voltages for core oc with people dayling it at 1.5v or whatever, just a straight silicon waste in my eyes.
Hadn't there been 9900K/10900K that cpu would look fine, but from an evolution standpoint just highly meh.
it's a great conversation topic, i do believe there is far too many variables though - an example is I ran latencymon while watching this video in a browser with 13 other tabs open AND a game running on another monitor.
My highest reported DPC time of literally a 1 and done attempt? 227us
Then I tried to idle on the desktop with nothing open and it was hard to get anything below 300us without retrying a 5+ times
My hunch is the Windows CPU scheduler + power state have more to do with the outcome than the CPUs themselves.
Don't worry about it too much, below 1000 is usually fine.
As a brief owner of 7950x, I did notice slower opening of folders compared to my parallel 13900k system
13900k feels snappier
Not to say 7950x is slow, it is just "slower"
Great work and I am glad you cover this. Being amble to switch programs and change files is really important to me and affects me more than the fps I get in games. Thank you for this kind of content. Keep it up!
Maybe try out the 3d cores - 8 & 16 cores CPUs for snapiness. They might be even better if they have the data already sitting in the huge cache
I have a 7800X3D now, and I have observed that it’s generally much faster at Windows 11 Search for local files, and opening local files, than my 5950X was. However it also periodically chokes (and the 5950X did also) when opening multiple Excel files downloaded from my company’s ERP software. Those files aren’t very big or complicated, yet the PC can take upwards of 15 seconds to open the second or third concurrent file. I wonder if Excel is just terrible on everything, or if Excel loves Intel.
I think it's an Excel problem. It happens when you have > 2-3 Excel files open & you try to open another Excel file. Happens to multiple Intel & AMD laptops I have used at work & home.
In my case, the delay is more than 15 seconds & can take up more than a minute. I sometimes just use Libreoffice to open the next Excel file.
I've been devouring all the news and reviews about the current Ryzen 7000's recently as I'm planning on a new build soon. I'm a gamer and I heard about the weird cache memory allocation between the CCD's in certain models and the most recommendations I heard were for the 7800X3D for gaming purposes. The reason I'm a bit scared to consider the 7800X3D for my upcoming brand new build, despite the promising charts, is its core / thread count & max boost clock vs. future proofing. Saying this because I'm about take a leap from a pretty old 4C/8T 4.5GHz CPU - yes, I like to hold onto things for budget reasons - and I wonder what 8C/16T would look like in terms of game requirements for 4K gaming in a few years? I'm not entirely convinced. Any recommendations?
it's not really the core number that keeps you down but more small cache and slow ram, tuning ram will help a lot to alleviate that and you can still run all games on a 6700, as long as you go with AMD gpus that don't tax it
If you have waited this long, just wait for 8000 series imo.
@@techyescity Thank you kindly Brian, I find it best to sit this one out too then. I wonder if the 8000 series Radeon GPU’s will debut around the same time as well.
Yeah hold onto it for a bit more. Next gen might have 10 core avg on amd and maybe double on intel with their weaksauce cores.
Thank you everyone
We’re all in it for the snappiness
Very useful test for anybody who uses PC for anything else then gaming.
What about Ryzen 5000 series? i'm planning to get a 5800X as upgrade eventually down the line.
Get 5800X3D instead and you’ll be good.
I hope you will test intel 14th gen snappiness when it comes out.
SnapYesCity is back baby!
This video has a very good point to it. Power management drivers are implemented by AMD, process scheduling is implemented by Microsoft/AMD. frequency scaling drivers are implemented by AMD. Chipset drivers in general are made by their respective vendor.
Dude, my Ryzen 7 5800X is snappier than my Ryzen 9 5900X, but that's because it only has ONE infinity fabric vs the TWO in the latter CPU, which imparts latency.
I think you should include a Linux platform in these latency tests. I simply don't trust windows to not be full of spaghetti code contributing to random latency spikes
I just wonder if the quality of a video card has any contribution to the overall snappiness of a system. Like a RTX 3060 vs a RTX 4090.
Everything from Nvidia after 30-series has increased dpc latency introduced by the drivers. We are talking about 250-400 us. AMD is about half of that. I would be much more concerned about tweaking the OS and bios to minimize latency, something this dude has put zero effort into.
I switched from an i5-8400 to the 7600 (full new build). The numbers are higher, loading times are shorter and fps is far higher. But windows "reaction times" aren't any better.
One repair, Razer's programs cause 1 to 2 ms delay in memory latency tests. Open programs like Steam, GOG, EA also cause the same problems. The best results were with the fixed clock and fixed Volt(SMT off). With results 3 ms better than with Best timing use the Ryzer running with pbo and other tricks.
literally one of the only tech channels I follow any more. Thanks for making relevant content
can you also cover the latency difference with Dramless/HMB nvme drives?
Mate, I'm running an R9 7900 and a 7900 XTX, coming from a i58600k and a 1080ti. Best deision ever. Full beast mode
I upgraded from a 6700k to a Ryzen 9 7900 and was disappointed with the snappiness. It's either the same, or worse. I suppose single CCD (7600 and 7700) would be better, maybe this could be the next test?
I really appreciate this testing, in my experience the single die chips are slightly faster than the multi die chips for latency. So it could be interesting for you to test a 7700x.
Ive been having weird latency issues and stutters in games and in windows on my r5 3600( oc 4.15ghz and ram at 3600mhz) and rx 6600 combo. Wondering if its related to security setting in the bios that ypu mentioned, bc ive recently updated my bios and it seems to be worse now.
I always look for the best FAPM on my PC!
My largest problem so far was a multiple years old windows 10 installation. It finally runs nice again lol
I had these weird little fraction second freezes do multiple desktop drag selects. Now it doesn't do that at all.
Have good RAM too... and well tuned.
I have noticed this even since Ryzen 1000 series. Is not like Intel. My friends using Intel say the same. Intel cpus are just snappier. You can tell only in real life usage, cannot be observed in synthetic benchmarks.
Very interesting, good work and well done but for me tho it is not relevant, I'm not power user, I'm happy with my windows 11, happy with my CPU, and those microseconds don't bother me.
I thought this video was going to address AMD's new AGESA which makes 6400MHz RAM the new sweet spot for AM5.
Surely this AGESA update makes the 7950X as good if not better than the 13900K with similar speed RAM? How about making a video on that?
I've actually done some testing with my 7950X and 2x32GB G.Skill 6000MHz CL30 EXPO kit, pushed to 6400MHz CL32 by simply turning on a preset in the new ASUS BIOS.
I've measured substantial gains across the board in AIDA64 Extreme. Here are the results (rounded off to the nearest whole number for better readability):
6400MHz CL32 preset:
Read: 86000 MB/s
Write: 94000 MB/s
Copy: 80000 MB/s
Latency: 61 ns
6000MHz CL30 (EXPO ON):
Read: 76000 MB/s
Write: 82000 MB/s
Copy: 71000 MB/s
Latency: 71 ns
4800MHz CL40 (EXPO OFF):
Read: 66000 MB/s
Write: 69000 MB/s
Copy: 61000 MB/s
Latency: 82 ns
Since Zen 1 I haven't been able to see any difference between Intel and Ryzen.. They are both equally fast. I am still on Zen 3 and very happy.. I probably won't upgrade until Zen 5 is selling for cheap like 4/5 years from now.. I worry more about what GPU I have, more than CPU.. I don't mind being a couple Gens behind as long as I can play the games I want the way I want. For the most part I am always a couple gens behind. I just dial in the settings. I'm sure most of us are.. I do have an XFX 6800XT so I am good right now.. Bought Hogwarts last night and I got 4K Ultra with Zero issues..