Unrealistic Expectations - Windows Updates on Ryzen 9000

[@bodasactra]

How about showing how Ryzen 5000 series is impacted. I know there are only a few tens of millions of us still on them but, we were hoping not to be pretend forgotten.

[@wii166]

I know i bet there's a lot more AM4 builds haha

[@roohansama]

i got like 20 extra points on my 5600x after the update on cinebench without changing anything

[@kosmosyche]

I walked.

[@Raizazel]

@@roohansama 20 points in multi is nothing margin of error. I have an 5800X3D later today I will run some tests. Then again maybe cinenech is one of those not benefitted the best from branch prediction, cyberpunk on the other hand seems to be really kind of that

[@gabber_]

Zen 3 and below will not be affected by this. The branch prediction update was introduced in Zen4.

[@djnes2k7]

Gotta love how it’s always to do clean install. As if I don’t have like 60 programs installed and keys i can’t find and mods on things it took years to do…..

[@Raivo_K]

You could always use program like Ninite to batch install programs and manually move over the AddData and Document folders. It's trickier with registry. But i hear you. Clean install is a pain even if you have everything ready and automated as much as possible. There's always some stupid nagging setting somewhere that got overlooked and needs to be manually adjusted.

[@yensteel]

They have a script to automate them. But it takes time. At least it's not some disk image the flash onto the drive. That would have been terrible and unprofessional.

[@zebrapluto530]

Nobody said you had to do a clean install. That would be ridiculous. As testers, THEY do a clean install to ensure an even playing field for comparisons is all.

[@ole7736]

Great content, Mike! You guys are the first outlet talking about this optional update and the silent disabling of Memory Integrity. Good job!

[@Ccb559]

Great work. Great video. I think we're all thankful seeing this.

[@ChildoftheKoRnf9]

Not the first time Microsoft changes stuff without communicating with its users that causes unforeseen impacts.

[@Ramu-10]

Not the last time either

[@WatcherKoops4677]

Like the windows vista disaster?

[@bodasactra]

I am betting Microsoft gets a kickbacks from manufacturers to obsolete slightly older hardware with this virtual TPM security BS that AMD is now saying test with it disabled. 24H2 has the key Win 11 security features off by default. Why have it in the first place if this is where we ended up? Millions of upgraded CPUs never needed to happen. Why did Microsoft have Win 11 "ready" affiliate links for new hardware on the Windows 11 update page? They forced CPU upgrades while profiting from it on both ends.

[@lief1]

AMD's marketing department needs to be severely restructured. They are somehow turning easy wins into weird mistakes and failing to get good will in a time where their competitor's products are basically faulty. "Hey, minor improvements this generation but at least they arent dying like intel 14th gen" and more realistic pricing would have been a lot better than this.

[@NadeemAhmed-nv2br]

I mean 5% extra uplift + amd said it working to drastically improve latency with a bios update. Same io die, no way latency increased by 2.5x or 250% on a denser node

[@kevinerbs2778]

@@NadeemAhmed-nv2br could be a design but though too because that larger front-end is either latency bound, or bandwidth bound.

[@HardwareCanucks]

Shareholder perception

[@nayan.punekar]

​@@NadeemAhmed-nv2br5% extra uplift is fake and bs

[@willzyx23]

They apparaently made great efficiency gains. Reaching the same perfomance with less power the previous versions. Whey didnt they release a version then with higher TDP and higher clock speed (name it 9650X or something) , since there obviously seems to be room for it. Release that for the price they are now calling for the 9600, and take 50$ less for the 9600 so it can compete with current prices of 7600, being slightly faster and more power efficient.

[@Shadowized]

Good points made about the memory integrity setting. However I think there's maybe one issue with your CPU testing checklist and it has to do with setting the power plan to high performance, I distinctly remember reading a while back about how with the PPM/X3D driver fails to do core parking properly when that plan is used, even JayzTwoCents talked about it in his "AMD Core-Parking problems FIXED once and for all" video at roughly 3:32 and I'm certain this same gamebar/power plan issue effects Zen5 as well since it uses uneven cores. I'd like to see somebody benchmark with Process Lasso and pin games to just the good CCD.

[@HardwareCanucks]

We copied that list from our Intel methodology. That was our mistake

[@zebrapluto530]

Another point about memory integrity is that you can turn it off in 23H2 also for a performance gain, not just in 24H2. I turned mine off long ago.

[@dbzssj4678]

Microsoft needs to abandon their enshitification project remove all their built in malware, ads and look back at why windows 7 was so successful. Familiar and logical UI, lightweight, stable, and consistent.

[@Traumatree]

I'd prefer Windows XP interface to Win 7 fugly translucent interface.

[@paneerbread]

​@@Traumatree you're the first person I've come across ever since win 7 launched who thinks it's ugly

[@Traumatree]

@@paneerbread Taste is subjective. I prefer the Windows 8.1 interface (not the metro stuff) and this was the best & and fastest Windows ever. As for Win 7, it was just a "working Vista" that I didn't particularly like. The bells and whistles thrown at us visually becomes invisible in the long wrong - even in the way sometimes. But all in all, if we want simplicity, Windows XP/Windows 2003 were the best interfaces where you could reach anything via the GUY via 2 or 3 clicks, vs today's Win 11 where you need to click 6-7 times to find the proper applets that gives the true control - or you run the proper applet by typing it's name but that defeats the purpose of running a GUI OS.

[@paneerbread]

@@Traumatree you're the first person I've come across ever since win 8.1 launched who thinks it's pleasant

[@Traumatree]

@@paneerbread You must be quite young then :)

[@Stevenx01]

That will surely make Steve from Hardwareunboxed stand up in the next video...

[@blackbirdpctech]

This is awesome … thank you for looking at this so thoroughly

[@evenAndre]

Thanks for the video. HUB always turns this off manually in win11, and found 10% perf uplift in their 40+ benchmark in 24h2. Will be interesting to see what you find after even more testing. But def. prob. many reviewers who doesn't touch this setting, and get even more gains,so thanks for the heads up.

[@selohcin]

Thank you guys! This is real journalism, not just tech coverage.

[@TheIndulgers]

I tested with a 7600x and had memory integrity off in both scenarios before and after the update. I found anywhere between a 7-15% increase in Cyberpunk and HD2.

[@LordApophis100]

Thanks for the concise and well explained video. And not re-iterating the whole drama again.

[@AndyHenryROGBUILD]

got the 9950x for my workstation and I'm loving it! waiting for the 9950x3d for my gaming build. Loving Zen5!!

[@01ai01]

Good catch on the memory integrity setting!

[@X1erra]

Thanks to you, I’m not going to be chasing tech news about 24H2 until this drama is settled and I can see a clear path forward. Keep up the good work, though!

[@spladam3845]

thanks for all the hard work team.

[@PaintMasterPoE]

I expected bigger difference so not rushing with update from win 10 to win 11 for now. A lot of people would like to see ryzen 5000, with win 10 included.

[@dg-dm1rr]

Wonder if the X3D versions of the chips are seeing as big a boost as their counterparts

[@andrewvirtue5048]

I only watch for this guy. I dont care about the other guy and his _"enthusiast"_ oversized "full" tower cases.

[@eliteneckbeard8035]

I haven't watched this channel since a channel presenter told me that the Viper Mini was the best mouse and didn't disclose the scroll wheel jump flaw that is on 90% of the Razer Viper mice line.

[@XtergoBlue]

Best video on this topic yet, this is the most significant Windows update for AMD

[@j340_official]

AMD’s marketing and release timing (before windows updates were ready) is one of the biggest self owns of the year. Intel’s raptor lake degradation was reaching fever pitch and AMD missed another opportunity to capitalize. Meanwhile, microcode 0x129 is very stable on my system. Very stable.

[@captainlooney4356]

Good catch on the VBS setting

[@Ryhillchibulls]

4 bots in the first four minutes is wild. clicked to say shocking time for a post.

[@kappab8304]

Thank you for this hard work!

[@RObinetus]

Great job guys, keep this going on all new tech release so that the giants don't release bullshit anymore, and be aware that when they launch something it will be tested more than their own test labs team.

[@gab882]

It just makes 5000 series AM4 CPUs look even more impressive

[@gulk46765]

How? Will they run faster on win11 compared to win10?

[@opensourcedev22]

"You wouldn't download more frames!!" ... Like hell I would

[@Eeshank2]

how big of a difference is it on zen 3 cpu's? mainly 5700/5800x3d?

[@HardwareCanucks]

Working on that

[@1967KID]

​@HardwareCanucks my 5800x3d, 3, 5600x , working good 👍.

[@christophermullins7163]

​@@HardwareCanucksthis is the move. Been hunting for this information for days. Be looking for the video.. thanks for your diligence.

[@bluej511]

At 1440p ultrawide with my 7900XT, didn't see much difference in the 3 benchmarks i ran, F1 21, RDR2 and R6S.

[@Eeshank2]

@@HardwareCanucks looking forward to it!

[@DarthV506]

No 23h2 with memint off for the 2 different patch levels? Great work on showing the various results from the different builds of windows.

[@gabber_]

it doesn't matter if it's 3%, 5% or 10%. free performance is free performance. It makes me wonder how much more there is locked away in the bloat that is Windows 11.

[@juts89]

Doesn't really have anything to do with what would be considered bloat.It's adapting the scheduler to newer CPU architectures and finding more efficient ways to schedule tasks.

[@gabber_]

@@juts89 I know, I was just wondering how much more could be gained, from elsewhere.

[@bgtubber]

What about Windows 10? Or is it only Windows 11 that was gimping Ryzen?

[@andrea99887766]

will be interesting to see how compares win 10 vs win 11

[@nolan8226]

So what exactly does the memory integrity do? Is this something we should even be turning off or does it increase system vulnerability enough to not be worth it? AMD was already telling us to use the system admin profile enabled just for a bit of gaming performance at unreasonable risk.

[@Lakefield_]

it kinda double checks if the code is trusted or not. If you use software from known sources, it's not a big of a deal turning it off. Also if you use older software or hardware, with drivers that doesn't support this feature, this can cause issues. So as an experienced user and you know what you are doing, you can turn it off. But as far as i know, this only affects gaming performance and usually you are GPU bound anyway, so you won't see any increases outside of benchmark testing. I'm on a 7800X3D paired with a 4070 Ti Super. So the CPU is a bit bored anyway. :D

[@nolan8226]

@@Lakefield_ Yeah I realized I actually have several drivers that won't let me turn this on and clicking on "learn more" windows doesn't recommend removing drivers just to turn it on. I've got a 7600x + EVGA 3080 and I think the CPU is actually a bit of a bottleneck, sometimes 1% lows appear to drop with PresentMon and I've noticed that the CPU busy time is pretty much on par with GPU busy (GN video recommended CPU waiting on GPU a bit) and small spikes in GPU wait tend to coincide with 1% lows dropping considerably (not often though)

[@Floreum]

I got a 9950X, fresh install and updated to the new 23H2 update that brought back these features from 24H2 and noticed my Memory Integrity is also off by default.

[@HardwareCanucks]

We installed the latest update of Windows 23H2 from the ISO and memory integrity is ON. Are you sure you aren't on a Preview build?

[@Floreum]

@@HardwareCanucks I am certain I am not on a preview build. I did however install my OS using Rufus with some of the options it provides to turn off certain Windows features, so maybe that's why it's already off?

[@Adel-sj4ue]

But arent off give better results as shown in the vedio This is a point i cant understand

[@PowellCat745]

@@Floreum Rufus does disable memory integrity from my experience.

[@frommatorav1]

@@Adel-sj4ue Memory integrity being off does give a lot better results but it is unsafe in many ways. Hardware Canucks as a rule for security, keeps it On.

[@tanmay5570]

very detailed investigation

[@mteng1]

how is the gain with older cpu like the 5800x3d?

[@joeybuddy96]

So what's the fastest for single-core performance post-update? 7800x3D, or something else?

[@coolbvwes1]

What about the new 23h2 update with vbs off vs 24h2 vbs off curious about those results.

[@Mutio86]

So it’s better to have memory integrity off?

[@MikeNovelli]

2:00 likely because the degreaded performance was due to the Spectre/Meltdown mitigation.....

[@johann3029]

Why not turning Mem. Integrity OFF on 23H2 for Testing? just to see ...

[@BadRusMan]

In short, if you want max fps - win 11 24h2 with mem integrity off, otherwise - just update your 23h2

[@Francis-p5i]

Next fan roundup when?

[@HardwareCanucks]

Before end of year hopefully

[@seamon9732]

I wish someone would finally include Win 10 in those tests. 50.16% of PCs on Steam are still on Win 10. Overall Win 10's market share is still 68%. Common guys, more than half of gamers want to know.

[@razpor]

Whats the effect on zen 3 ....5900x specifically??

[@WSS_the_OG]

This reminds me that I have to prepare for a clean Windows install for the release of 24H2. :/ Geesh.

[@optimusprime9213]

So, I'll just turn of memory integrity and enjoy the double digit performance boost. Not gonna hurt I guess + the free insane performance!

[@FrotLopOfficial]

Bro has a forehead crease like _______

[@SergeisAdventures]

😂😂😂😂😂

[@rewind841821]

Very nice video , thank you . Also what base are you using for the Edifiers speakers ?

[@FatetalityXI]

When linux performs way better than windows in same supported software/games then I guess windows has some serious issues under it's belt.

[@nurbson3202]

Finally we are back to windows 10 performance 😅

[@EhNothing]

While there definitely have been hits to AMD's rep from all this, Intel still has self destructing CPUs and handled it so poorly that it would be virtually impossible for AMD to dig its way down to where Intel is. Yeah it sucks, but it could be worse.

[@maroosagaming]

the new optional update did absolutely 0 for my 7800x3d, ill have to wait for the 24h2 release it seems

[@HardwareCanucks]

You won't see anything. I mention this. It's mostly snake oil.

[@coolbvwes1]

I do see a little difference, not big one but some games are 3/5% faster. And im running revi os where vbs was already turned off. I manually updated 23h2 because for me it didn't show in additional update.

[@Raivo_K]

It did improve 0.1% lows on my 5800X3D but i already had VBS and memory integrity off and was running with system admin account.

[@jjlw2378]

The update reduced stuttering in Dragons Dogma 2 by .05%! The 7800x3d is absolute garbage in this game.

[@zomberkay]

Just mod your game ​@@jjlw2378

[@Real_SkyRipper]

is an update like this coming to Windows 10? free fps is always welcome, not changing to Windows 11 tho, to limited, can't even have the task bar at the top which for me is a huge deal breaker.

[@middle_pickup]

New branch prediction you say? I'm sure that won't come back to bite us in the ass.

[@NotNoord]

Wait, what?! Turning off Memory Integrity gives additional 17 fps in Baldur’s Gate 3, same 17 fps in Cyberpunk 2077 and 62 fps in CS2? Gaming on windows is insane!

[@Henrik_Holst]

Whatever the change is in 24H2 it's not some new branch prediction algorithm because that is not how this works at all. For your code (aka the games and applications that you run) to benefit from it they would have to be recompiled with zen5 optimizations enabled. Much more likely is that these is some residue from the Spectre mitigations that affected the branch predictor (possible due to extensive flushing) on context switches.

[@HansElzer]

Very enlightening 👍

[@djayjp]

MS is just returning gains lost going from W10 --> W11....

[@HaveYouTriedGuillotines]

I went from W10 --> Linux.

[@OV3R-9K]

@@HaveYouTriedGuillotines you mean W10 --> I have lots of free time so i will tinker

[@HaveYouTriedGuillotines]

@@OV3R-9K You have to tinker with Windows as much if not more on account of how frequently Microsoft keeps bungling things up, as well as the general Windows ecosystem beginning to fall behind Linux in some very critical ways. You've probably heard about how hard installing drivers on Linux can be. You've probably also heard that arch is hard. The thing no one will tell you is that 90% of the time your drivers are already in the kernel, thus installed by default, and Arch has this thing called the AUR, and it often has any drivers not in the kernel. You can just install a graphical package manager like Bauh or Octopi that can automatically pull and build packages from the AUR, and then installing any drivers hosted on the AUR is literally just a few clicks to install. Truth be told, the last time I had to tinker with something seriously was over 4 months ago, when I had to do some sleuthing to figure out how to get AI image generation working on an AMD card. And this wasn't a problem with Linux insomuch as it was an issue with the technology being cutting edge and in no way standardized or polished.

[@HaveYouTriedGuillotines]

@@OV3R-9K Nope. Installing Linux is literally just a few clicks these days. You have to tinker with Windows as much if not more on account of how frequently Microsoft keeps bungling things up, as well as the general Windows ecosystem beginning to fall behind Linux in some very critical ways. On my system, I didn't even have to install any drivers. Everything just worked out of the box. Even my 8-Bitdo arcade stick. It's all in the kernel.

[@HaveYouTriedGuillotines]

@@OV3R-9K You've probably heard about how hard installing drivers on Linux can be. You've probably also heard that arch is hard. The thing no one will tell you is that 90% of the time your drivers are already in the kernel, thus installed by default, and Arch has this thing called the AUR, and it often has any drivers not in the kernel. You can just install a graphical package manager like Bauh or Octopi that can automatically pull and build packages from the AUR, and then installing any drivers hosted on the AUR is literally just a few clicks and your password.

[@Karti200]

It was not mentioned in this video - what a surprise... - that performance boost is BECAUASE microsoft took down security updates related to meltdown from Windoiws soo same shit as we had in the past with spectre unlock for older intel iCores, difference is for Intel we had to do it manually, for Ryzens windows jsut removed security patches from 24H2

[@dm1i]

But some testers on KZbin are always turning Memory integrity off. So they should be unaffected by this update. But they still found up +30% fps boost in some games anyway.

[@HardwareCanucks]

They would still be affected by the branch prediction uplift as we showed

[@capitoneo]

Can you make a follow up video with comparison to latest Windows 10?

[@Oscee613]

Worth noting that turning memory integrity off is a security risk - not the end of the world but better to keep in mind.

[@HardwareCanucks]

Correct which is why we always leave it on....except...

[@ihadlunch8575]

Should I upgrade my 8700k?

[@profosist]

Memory integrity being odd on the preview 24H2 is likely because it's the release preview early branch.

[@luckrequired5382]

I really can download more memory now???

[@Gattberserk]

Look like AMD internal politics is a lot more serious than anyone would have guessed.

[@pretentiousarrogance3614]

Are you guys gonna test Zen 2?

[@tomwaller6893]

The patch is already out as a standard Windows update and is giving me a 10% boost on R23 on a 7600 X 6 core 12 thread CPU with a score of 15325 using DDR5 6000 CL30 and 15 in undervolt in PBS.

[@HardwareCanucks]

We covered it in the last half of the video. ;)

[@brucelee5247]

Good work!!

[@pyric]

Amd has needed to clean house in their marketing department going back to the early Ryzen days. Almost every single launch they do something stupid to hurt the company's reputation.

[@arp1]

Hardware unboxed test always with VBS off. And they measured sometimes even 20-30% in non GPU-bounded situations.

[@HardwareCanucks]

Different methodologies will yield different results. We don't recommend our viewers turn off VBS and our methodologies reflect that.

[@cowymtber]

At AMD, AI and EPYC are making gaming and desktop into the redheaded step child.

[@ThaexakaMavro]

Good job!

[@jouniosmala9921]

Does The memory integrity toggle have bigger impact on Zen5 compared to Zen4 on games? Zen5 has additional hardware changes to disable Spectre style attacks at the expense of performance. If another thread is run in a core that thread cannot access data another thread in L1 or L2 caches but goes directly to L3 cache. Now if two threads share common data, then that data is duplicated in L1 and L2 caches and synchronization is done through L3 cache. And it shows up in Anandtech communication latency between two virtual SMT cores in same core, when compared to same benchmark in Zen4 review. And Linux benchmarks show large performance increase from Zen5 in benchmarks where said latency is not an issue. For instance 18% Javascript on chrome and 30% fps for a game made with modified quake3 engine. The weakness of Zen5 is tightly coupled threading with many threads on a scheduler that moves those threads around that it got from a security feature. But that weakness encompasses modern game engines which seems to be primary review criteria for desktop processors. The actual fix is probably partially in Windows scheduler and some modification to underlying game engine.

[@kevinerbs2778]

the windows update addresses a specific Spectre/Meltdown patch problem that wasn't compatible with AMD's predictor. Technically Zen 5 has dual predictors on its front-end. Not mention that it also has dual pipelines & dual decoders too along with dual micro-op cache. but normally in gaming it's only executing at 4 way & not using both it just uses a single one for a single thread. They dual work fine for specific code & SMT thread through a core, so it's working fine with SMT. Microsoft might have double up the fix for Zen 5on Spectre/Meltdown. That situation you are talking about is from a pipeline hazard of over lapping of data with two different threads, it's there prevent repeating instrctions that overlap. It also helps that they're run at different cycles one is 1 cycle the other is 2 cycles.

[@jouniosmala9921]

@@kevinerbs2778 What I'm describing, is a different problem, it's about stuff that could be read in all the old processors in L1 or L2 cache will hit L3 cache in games. It's a side effect of preventing other threads unrelated to currently running thread issuing instruction to hit data of current thread in L1 or L2 cache, but it also prevents game threads sharing data in L1 or L2 cache, and its only available in Zen5. I'm curious if that feature is something that could be disabled to actually test if that's a culprit of Zen5 gaming performance issues. And if it is, then windows scheduler needs to reduce amount moving around of threads, and there's a possibility of minor modifications to game engines helping with Zen5.

[@kevinerbs2778]

@@jouniosmala9921 To my understanding, wouldn't that matter if L2 cache was inclusive vs exclusive? pretty sure Zen 5 was changed to be inclusive from exlscuive from the original Zen, because of what you described. Although it depends on how inclusive it is. Since older Intel cpus used to duplicate all L2 instructions to all other cores L2 cache just in case it was needed on repeat loop instructions which many games use.

[@jouniosmala9921]

@@kevinerbs2778 Inclusive vs Exclusive is another matter, this is specific decision to check if data belongs to thread requested and if not then treat it as a cache miss. But slowdown comes, when thread 1 reads shared game object and after working on it gets to sleep and thread 2 gets awakened and reads the same game object that in all the other CPU:s is in L1 cache but with Zen5 only allows it to be read from L3 cache. So any time you switch what thread runs on a core all shared game objects become cache misses for L1 and L2 caches.

[@kevinerbs2778]

@@jouniosmala9921 That Probably explains why Wendal on level1tech's was seeing cache misses on things that are normally not cache misses. I wonder why they designed that kind of limitation on Zen 5? you think it has to do with context sensitivity for threads. I mean it has to wait for the other thread to be empty, even if it's SMT thread to clear to do anything with. This why I don't like shared resources with limits.

[@gucky4717]

It is a culmination of the past uplifts each gen and the fact that AMD took 2 years from Zen4 to Zen5. Intel only takes one year and nobody expects a two-digit uplift from them. And of course the bad marketing from AMD that said 16% IPC uplift...

[@lizardx6504]

It takes Intel 1 year to rename the 13th generation to 14th without changing the chip revision. Are you even aware of this? And Skylake kernels from the 6th to the 10th generation are great progress, you're funny

[@thegeekishgamer]

You guys are the goats 👌🏽

[@thedigitallens]

Can we see performance change for apps like premiere pro & davinci resolve?

[@Kahana82]

Wouldn't it be useful to be able to disable the memory integrity on a per-executable basis, or is that a non-issue ? Similar to what the "Exploit Protection" alows to do through the "App- and Browser control" in the Windows Security panel.

[@xmaverickhunterkx]

This is like Fine Wine for CPUs

[@MrJMHP]

AMD and Intel are both squeezing x64 to the limit... And everyone is already facing that limit. Performance gains are increasingly scarce and at the expense of more watts.

[@ismaelmeurer8446]

The problem is that Windows sucks, especially for AMD processors. Phoronix analysis showed that Linux had almost a 20% increase in performance for AMD 9000 series, overall.

[@tommihommi1]

that was with a looooot of AVX512 tests

[@Raizazel]

Windows does not suck for AMD, AMD might suck for Windows. If your statement was true it would shift the blame to the other direction. Windows is what it is with all its updates and all, you manufacture CPUs knowing that most of them are going to be running windows, if you don't optimize for that it is your fault. Then again, you might see most benefits for AMD 9000 in AVX512 flows, nothing to do with windows.

[@syafri88]

@@Raizazel please, he is just protecting AMD. Can't have AMD being portrayed as bad in any videos

[@Soussa-h4e]

@@syafri88indeed, anything that is from AMD has to be protected by these crazy people. All hail AMD.

[@HardwareCanucks]

You might want to recheck their updated testing. Once Linux specific AVX optimizations are taken out of the equation, it's more like 7%.

[@dez7roy3r]

Excellent job all of you. AMD and Intel are disappointing quite a bit. Nvidia, you're up. Kidding.

[@FlanK3rCZ]

but boost on Intel is not so high as on AMD Ryzen overall (of course, some games who favourites Intel, yes). Intel is trash due many, many issues and crazy power consumption (not because performance). There is still not real official fix, so I agree with Steve.

[@Pachupp85]

We need to wait for hardawre unboxed for zen1 2 and3.

[@robertt9342]

It entirely possible that someone screwed up the testing and AMD has a hard time walking back their marketing.

[@GOD999MODE]

Pr0n bots all over the goddamn comments section.

[@HardwareCanucks]

We're on it.

[@nbc8944]

I wish there was a super simple gaming windows os that ist super light and focused on best performance and low telemetry

[@lohaswinner]

I would like to know the performance of Intel CPU with Windows 23H2 and KB5041587 installed.

[@jrhymescrypto]

This is such a boneheaded test, firstly, you do NOT set the power plan to high on Ryzen chips, you set it to balanced or core parking / scheduling gets screwed, secondly, you need to ensure the chipset drivers are installed, thirdly, you need to in game set SMT to on (or high core count sku’s will get hurt).

[@HardwareCanucks]

Boneheaded comment. With the latest Windows power plan revisions Core Parking remains active regardless of the setting.

[@Globespy]

@@HardwareCanucks - I thought that (at least for AM4) the best Windows power setting is Balanced (not balanced performance/balanced power saving)

[@inphuriated]

what is, and where is, the impact of cl28 vs cl30 on intel and amd systems, super rare to see any any other ram than cl30 in tests/reviews also what was the reason to go with cl28 and not the more common cl30?

[@james...cardinal]

is there a roadmap for processor releases beyond amd threadripper 9000 or intel emerald rapids ?

[@ThaexakaMavro]

SO why memory integrity so important to win11 when it comes disabled in win 10?

[@Derekmeetsworld]

so is it worth it to upgrade to windows 11 from 10 yet?

[@profosist]

Watch out using High Performance power plan. Ryzen and Intel since 10.5gen have recommended Balanced. The you can toggle in the new power menu.

[@HardwareCanucks]

Yeah we've found it makes no difference in CPU testing. It's effect is very much laptop focused.

[@eman6784]

Good Visuals 👍👍

[@ryanng7594]

If there's any possibility that it's not AMD but Microsoft to be blamed😓 Especially considering phoronix's benchmark post🙃