Others have already pointed out that AMD are the masters of dropping the ball. They had Intel scrambling with Zen 3, on the ropes with the 7800X3D and instead of running away with the title they do a 1995/96 Newcastle. The mind boggles.
AMD keeps messing up on pricing as well, only for them to reduce the price weeks or months later, but the bad PR sticks to how it was on release. AMD needs a big overall on it's marketing and needs to be more aggressive on pricing. The irony is, with Intel messing up, AMD really should be pulling away from them but messing up like this is giving Intel time to get back in the game, and on the gpu front, AMD keeps messing up time after time, not because its products are bad, but because sometimes it messes up on its software early on and because it's pricing its gpu's too close to the Nvidia line up to not make a dent in them, if AMD wants to shake things, up, it needs to be far more aggressive because Nvidia has pulled so far ahead that AMD has a mounting to climb, and again, it's not because the products are bad, but AMD's marketing, software on release and pricing isn't doing AMD any favours and it's remarkable that AMD doesn't see that.
Any your a expert on what a said processor costs to make from design to production to Finnish product? What you think you just pull a number from a hat and put that price on that cpu?
Are we going to get similar updates to Windows 10 as well, or is Microsoft going to try to entice us to switch to Windows 11 by locking this update out of Windows 10?
NGL, I thought "Another one bites the dust" was the music just after the intro before realising it was a sponsor spot 🤣 Thanks for the great testing and reviews you do. Keep up the good work!
Would be amazing if you could find the time to do more tests with 24H2, specially with Zen 3, Intel aka the most used CPUs, like 5800X, 5800X3D, 13600k, 13900k etc.
Just tested this, absolutely no gains compared to a Windows 10 IoT LTSC with ReviOS playbook applied on it. I guess this was a problem for those that use vanilla Windows 11. I already knew it heavily hinders your hardware´s performance, no news here. No one should use that operating system.
@@GeorgeCardoso-mh2ei From what I heard that version of windows 10 got the update, and only that one. Are you sure that you did not get it before you tested?
@@proklet4694 I didn´t because I have all updates services killed on the ReviOs control panel. Windows 11 is a mess. This update might be great for its performance but the OS is still very laggy, awful DPC latency across the board, I´m keeping Win 10 ltsc until it runs out of security updates in 2028
What have we learned? Microsoft has been leaving performance on the table. It must be frustrating even for AMD to know the code is there for insider builds. Microsoft has been working on Qualcom builds of Windows.
Microsoft has only updated its core base for the new Intel big-little CPUs, and left AMD in the dust FOR YEARS performance wise. This could be seen 4 years ago under Linux when the same CPU was able to render stuff 20-25% faster than under Windows.
I don't know. I have my fingers crossed for 9800X3D, but not holding my breath. From what I understand, the 7000 series was already bottlenecked by its memory controller. The 9000 apparently has done little to address this weakness. IF that is the case, I could be that 9000 series does actually have more to offer, but is being held back. IF that turns out to be true, the the extra cache could potentially uncover a pretty big performance uplift, by alleviating the memory bottleneck. We won't know until we see them..
Nice testing Leo - some interesting gains - but its such a poor showing from AMD to leave it all so late before explaining the Windows issues - I bet Intel, even with all their issues will have all this sorted before their launch
nobody is gonna buy Intel. They're about ready to be sued for 13/14th gen. They can't fix the issues, they are still selling the chips with oxidation, and they still haven't recalled plus they're actively refusing or delaying RMAs by a month or more in some cases. DO. NOT. BUY. INTEL.
15:55 This is even more bizzare. If they already knew this and had windows update to work for them better on their laptops, they could've just released everything a month later in september alongside the motherboards and had no issues instead of rushing it out for no reason.
Ryzen 7 5800X 3D was included to give more context as we consider it to be a fine CPU despite its age . The reason I didn't restest it with Windows 24H2 is purely down to time as it would involve working with an AM4 motherboard and DDR4 where all the other CPUs are AM5 on DDR5. I figured it was interesting but not necessarily 'that' interesting. Leo
AFAIK it's the branchprediction that's bringing those improvements, but that's only used on a local admin account on windows, not a regular user account.
Your not wrong. Intel has had to rebrand to Core "Ultra". AMDs marketing and branding is subpar. Consumers are dumb sometimes beautiful boxes sell more.
Yeah that would interest me too. Maybe this complete situation isnt a big problem on win 10? Has this anyone tested so far? @haukionkannel - microsoft gives extended support for win 10 until october 2028, called Extended Security Update-Programm (ESU)
If I understand what the AMD rep in The Full Nerd podcast said, it will be updated for Windows 10 also, and will also improve performance for Zen 3 parts
@@Ben-Rogue Oh, neat! So free performance? I'm using Ryzen 7950x on Windows 10. How big of an increase can I expect and in which workloads, if you don't mind asking?
@@bgtubber Well, if it's anything like the uplift people will see for Windows 11; in CPU bound gaming (1080p type gaming), you'll probably see about 5%-10%, but I'm not sure other workloads will see.
Thanks for setting aside the bling-bling upgrade and digging into this Leo. Great points, I assume it's still "Leo's advice" to "Wait for Zen5 X3D."? FWIW, in my opinion AMD marketing is grasping at straws. Users don't use "Insider previews" of Windows. VBS disabled (i.e. SVM disabled in BIOS) is not how users run their machine. Running in some sort of "super admin" mode or even running a game "as Administrator", users do neither. AMD needs to "own" their marketing (reviewer guide, statements at the Zen5 preview event, etc.) lies.
I agree that AMD's marketing department need to pull their head in regarding performance claims, but I don't read their 21 August blog post as saying typical users _should_ disable VBS, run Windows Insider builds, or run applications in system-admin mode. They do give instructions for users who *are* willing to trade security/stability for more performance, and IMO it's a little irresponsible for them to do so without stressing the potential impact on security &/or stability. However, I'm sure they only intended those instructions for the minority of users who can make a well-informed risk assessment, i.e. they think the overwhelming majority of Windows 11 users are better off just waiting for the mainstream update.
@@wereoctopus They're saying all these options improve performance more on the new ryzens as opposed to the previous gen ones, which supposedly widens the gap to match their original claim of improvements over the previous gen They're heavily grasping at straws however and it's evident by everyone
We all run 'insider previews' eventually. Just more stable versions. What these various testing modes and options indicate is that Zen5 needs software optimizations to realize its full potential. No different than Intel who has been in bed with Microsoft for decades and who undoubtedly got first dibs and preferential treatment when it came to optimizations for its processors. That may be changing, thankfully. I'm not saying AMD should not take some blame, but they were up against the Wintel beast for years.
Damn, what the hell happened with Total War: Pharaoh?? That's a huge performance uplift. Either their code HEAVILY depends on branching or something else is going on.
I wish to see if Zen 2 and Zen 3 also benefit from WIndows 11 24H2. Some of us poors on older processors and are running Windows 10 and we want to see if Windows 11 24H2 improves performance on our old Ryzen cpu's, in other words whether it's worth upgrading to Windows 11 24H2 for the extra possible performance from Windows 10.
This is hands down the most thorough review of the 24H2 out there, Leo really is the GOAT. Appreciate you including the 7950X and 7800X3D in your benchmarks! I'd love to see some productivity comparisons too, like the 7950X vs 9950X and 23H2 vs 24H2.
I hope there will be a huge benchmark video where you test intel 12th gen to the newest arrow lake, zen3,4&5 on windows 10, windows 11 24h2(vbs on & off). If this update optimization also comes to windows 10, I wanna stay on windows 10 as much as possible. It's mind boggling how in some games, the gains are ridiculously high but it could be an error as you mentioned in the video. Clearly AMD should have just delayed a bit more until windows 24h2 was launched. 7800X3D also gaining the same improvement makes it the rare Win for AMD.
Annoyingly, the 9950X utilizes core parking just like the 7950X3D/7900X3D so they will not work properly if the motherboard has already seen a normal chip like a 9700X, and vice-versa. This means a Windows reinstall is required after installing the CPU. You might get away with just reinstalling the chipset drivers though. Perhaps this is the reason Total War: Pharaoh didn't boot with the 9700X?
Nobody cares about testing Intel CPUs right now because they've had 2 generations fail back to back. Intel is basically DOA until they fix their brand name.... somehow.... if that's even possible.
Nothing. As yesterday is 24h2 pre release is very unstable cause the new recovery feature.there is o 24h2 beta. And then dev and canary, no one need test new cpus on dev or canary
@@jonathanwilkerson2859 branch prediction is built-in in the cpu itself and isn't really an os-dependent feature in any way (unless we're talking microcode updates, but these should be independent driver/firmware updates and not tied to the os version). "they optimised FOR the cpu's new type of branch prediction" is probably more correct. it'd still be interesting to learn what exactly they mean by that in more detail however. the way i imagine it, the os, or any other software for that matter, can be coded in ways that work well or not with a certain cpu's branch prediction if some quirks of it are known in advance. for a few probably too simplistic examples: prefer conditions based on carry flag more or less than on overflow flag, zero flag, or sign flag, prefer near jumps to long ones, prefer hardcoded address jumps to indirect ones, prefer conditional jump to conditional call, and such. still, its hard for me to imagine what os component could have such a noticeable performance difference by being optimized in this way, when the majority of branch predictions in all these tests should not be in os code but in game code. was it something generic like the task scheduler? if so should we expect improvements in non-gaming workloads as well? or was it actually gaming related, some part of the rendering pipeline, like a vulkan/directx/opengl library? but then why can't that just get updated separately from the whole os? and should we expect even better performance improvements if game code were optimized as well? would that involve just recompiling it with a compiler version aware of the quirks of the new cpu branch prediction mechanism? have compiler makers been given the information to make those optimizations already? i'd really love if some tech-head can give me a layman overview of what exactly is going on here.
AMD - releasing a help page after the launch is over. I hope someone in the company looks into how badly this was handled when Intel were down and in a perfect position to kick hard
Cope. There are established people in Intel actively sabotaging new additions, effectively killing innovation and any fighting chance they might've had.
Hats off for taking the time to do this, I decided to buy a 9900x and so far I'm pleased as gaming is also a side hobby so that wasn't my main concern plus in single thread with a bit of tweaking I'm ahead of the upcoming Intel Ultra 9 leaked geek bench score in single thread with great temps and wattage so personally I feel Zen 5 is an upgrade, Maybe I got a good chip but for those that like to tweak their systems ( custom water cooling loops etc) I feel buyers will have a fun time with this generation.
Thanks for the information I got my rig parts ready for build i will make sure I got that windows version is there any other setting that needs to be set on windows or some other program?
Pity he didn't have time to test the 5800X3D with 24H2. It would be nice to know if the performance uplift affects older processors as well. I guess we'll wait for Steve from Hardware Unboxed to test, I'm sure he will. The question is, will be standing when he does?
i want to upgrade from a 3400g on a gb x570 board. so i bought the new gb x670 mb planning to buy a 9700x in the next 12month depending on the price along with a pair of new ddr5 modules. im still convinced about the amd concepts hence the last upgrade from an old intel system already 5y ago now practically halfed my energy costs on the 65w tdp.
thank you for doing this. makes no sense to me when single tread performance stayed the same in both windows version while the performance of games go up . hardware and software makers always been corrupting the markets with their influence for more profits . i didn't know they sell BS to .
Wow some weird looking changes at times - that pharoah game runs like pants for me too. They really messed up by not telling the channels about this from the very start- clearly a massive disconnection between AMD marketing and technical teams
the interesting part for me would be the low and midrange PCs (e.g. 5600X to 5800X3D) with games, that do have a CPU bottleneck. Playing a high end game in 180 or 200 fps does not really make a big difference, but even a couple of fps more from 55 to 60 fps does make a difference.
Thats a lot of testing for a few days Leo - thanks - I am not sold on any of these processors right now, waiting to see what Intel do in the next few months
I still love AMD - seems like this was a lot of issues combined into one - but I didnt like the wording on some of their blog post blaming reviewers and how they test. thats nonsense
some of the reviewers made a mess, you compare results and methodology and some shpuld bot dp reviews, others did a great job the company can get upset, but people in the end want more performance, no idea why, but they want it they did not care for the reduced power consumption amd will take note and push more watts in future cpus, so people is happy
Hey brother!Can you help me with a question since you know far more than me?I prioritize NC and comfort and I have these options;(edifier WH950NB, sony wh750b, 1more sonoflow, anker soundcore q30 and space one)what option is better?you can sugest something else if you know 😊
Please report any spam as you see it, thanks! And be sure to let us know what you think of the latest Zen 5 developments...
so AMD is right
extra 4% at least, but can go up to big jump in PERF
@@rodrigorras Right about what?
@@bgtubber watch the video
@@rodrigorras I did watch it and I'm still asking you. What are AMD right about?
Leo wanted some royal bling in his life, but AMD was like "honey, new windows update just dropped"
😂
🤣🤣🤣
😂😂😂
0:24 That PCIE slot is holding on for dear life!
it's dey culcha
Others have already pointed out that AMD are the masters of dropping the ball. They had Intel scrambling with Zen 3, on the ropes with the 7800X3D and instead of running away with the title they do a 1995/96 Newcastle. The mind boggles.
))))))) amazing reference
Amd is always fighting windows and in this case amd focused on winning in workloads.
But Intel can't win the title either, they're still on ropes so it's ball back on AMD ;)
AMD keeps messing up on pricing as well, only for them to reduce the price weeks or months later, but the bad PR sticks to how it was on release.
AMD needs a big overall on it's marketing and needs to be more aggressive on pricing.
The irony is, with Intel messing up, AMD really should be pulling away from them but messing up like this is giving Intel time to get back in the game, and on the gpu front, AMD keeps messing up time after time, not because its products are bad, but because sometimes it messes up on its software early on and because it's pricing its gpu's too close to the Nvidia line up to not make a dent in them, if AMD wants to shake things, up, it needs to be far more aggressive because Nvidia has pulled so far ahead that AMD has a mounting to climb, and again, it's not because the products are bad, but AMD's marketing, software on release and pricing isn't doing AMD any favours and it's remarkable that AMD doesn't see that.
Any your a expert on what a said processor costs to make from design to production to Finnish product? What you think you just pull a number from a hat and put that price on that cpu?
Are we going to get similar updates to Windows 10 as well, or is Microsoft going to try to entice us to switch to Windows 11 by locking this update out of Windows 10?
The amount of detail you need to go to just to get a decent benchmark is crazy.
Some interesting results - doesnt matter at this point, AMD really need a course on how to maximise their market when Intel are struggling.
What's the saying? "AMD doesn't know how to be #1, and Intel doesn't know how to be NOT #1."
@@catsspat Okay!
@@catsspat AMD is #1, straggling to widen the lead, though...
This is mammoth like performance though.
NGL, I thought "Another one bites the dust" was the music just after the intro before realising it was a sponsor spot 🤣 Thanks for the great testing and reviews you do. Keep up the good work!
😁
Curious to see how Zen3 benefits from it. And if they backport it to W10 while it's still officially supported
Curious as well.
Agreed, I have no interest in win 11, or 'upgrading' from my 5700x for quite some time.
hardware unbox already showed that w10 on avg is 5-10% faster than w11 in games for ryzen, w11 is just bad and unoptimized
@@v000000000000v that was not against this new update though.
@@BluD ya it isnt, but w11 was slower than w10 so now it's finally catching up
Thanks Leo for, once again, the no drama version of this update.
very professional Leo
Appreciate the kind words
Would be amazing if you could find the time to do more tests with 24H2, specially with Zen 3, Intel aka the most used CPUs, like 5800X, 5800X3D, 13600k, 13900k etc.
Thank you for including the X3D chips in your testing.
No problem!
What a mad series of events no one ever thought would happen 😮
Leo - old school Journo with old school standards - and no dead chickens in the thumbnails either!
You hurt my feelings excluding zen3 24h2 benchmarks 😭 I feel called out
Thanks Leo, good coverage over this launch you must not have slept much
Hey Leo well explained video, all made sense to me, and often there is so much ranting in videos I cant understand the points
Most of those games I played ages ago, stuff I forgot about.
AMD - what a mess guys, you should all be embarassed this happened. Intel handed you a sales market
Thanks leo, been a good source of info over the last 2-3 weeks appreciate it
Glad to help
@@KitGuruTech Thanks!
@@KitGuruTech Hmm!
Wow. Thanks for checking this!
You bet!
Just tested this, absolutely no gains compared to a Windows 10 IoT LTSC with ReviOS playbook applied on it.
I guess this was a problem for those that use vanilla Windows 11. I already knew it heavily hinders your hardware´s performance, no news here. No one should use that operating system.
@@GeorgeCardoso-mh2ei From what I heard that version of windows 10 got the update, and only that one. Are you sure that you did not get it before you tested?
@@proklet4694 I didn´t because I have all updates services killed on the ReviOs control panel.
Windows 11 is a mess. This update might be great for its performance but the OS is still very laggy, awful DPC latency across the board, I´m keeping Win 10 ltsc until it runs out of security updates in 2028
Cant wait to see how the 3D V-Cache version performs ! Will be interesting for sure.
Thanks Leo , good testing and information, no thanks AMD - too late for me im sticking with my 7000 series
Some good gains ! cant wait for the update to roll out
What have we learned? Microsoft has been leaving performance on the table. It must be frustrating even for AMD to know the code is there for insider builds.
Microsoft has been working on Qualcom builds of Windows.
I have to think that Microsoft has indeed been busy with Windows for Arm
Leo
This is exactly 💯 the case. Sweet Qualcom money couldn't be refused
Microsoft has only updated its core base for the new Intel big-little CPUs, and left AMD in the dust FOR YEARS performance wise. This could be seen 4 years ago under Linux when the same CPU was able to render stuff 20-25% faster than under Windows.
I liked the content here, always very calming and thoughtful without explosions and drama
I don't know. I have my fingers crossed for 9800X3D, but not holding my breath. From what I understand, the 7000 series was already bottlenecked by its memory controller. The 9000 apparently has done little to address this weakness. IF that is the case, I could be that 9000 series does actually have more to offer, but is being held back. IF that turns out to be true, the the extra cache could potentially uncover a pretty big performance uplift, by alleviating the memory bottleneck. We won't know until we see them..
Nice testing Leo - some interesting gains - but its such a poor showing from AMD to leave it all so late before explaining the Windows issues - I bet Intel, even with all their issues will have all this sorted before their launch
nobody is gonna buy Intel. They're about ready to be sued for 13/14th gen. They can't fix the issues, they are still selling the chips with oxidation, and they still haven't recalled plus they're actively refusing or delaying RMAs by a month or more in some cases.
DO. NOT. BUY. INTEL.
Even when Intel shoot themselves in the foot with loads of issues, AMD always find a way to screw themselves over. History has proved it.
It’s more like believing too much hype and not staying in reality more.
@@Typhon888 nah i bet the intel launch wont have these stupid windows issues. im not even talking about the hardware
@@Typhon888 AMD never do good launches when they are 'ahead' its the mentality of an underdog.
Depends on who was to blame on the software part.
Amd have a bad track history of screwing the pooch.
15:55 This is even more bizzare.
If they already knew this and had windows update to work for them better on their laptops, they could've just released everything a month later in september alongside the motherboards and had no issues instead of rushing it out for no reason.
Qualcomm - please step it up, we need a third party to get away from these jokers
Thanks Leo, good viewpoints and honest thoughts
Thanks Leo, appreciate all your work on these - love the channel!
Thanks!
@@KitGuruTech Welcome!
Leo's test setups are always top-notch. Props to kitguru!
I would appreciate if Ryzen 7 5800X3D was tested in Windows 11 24H2 too, I don't know why it is there if it was not retested
Ryzen 7 5800X 3D was included to give more context as we consider it to be a fine CPU despite its age . The reason I didn't restest it with Windows 24H2 is purely down to time as it would involve working with an AM4 motherboard and DDR4 where all the other CPUs are AM5 on DDR5. I figured it was interesting but not necessarily 'that' interesting.
Leo
Cyberpunk - thats quite a decent improvement, what is going on though? windows code optimisations? its not RT so im at a loss
AFAIK it's the branchprediction that's bringing those improvements, but that's only used on a local admin account on windows, not a regular user account.
Since the update will also improve ZEN 4 performance, it's pretty much as you were.
This has all been a great advertisement for Linux, no doubt about it
Certainly hasn’t hurt it !
@@KitGuruTech Ya you are Right!
AMD need to find INTEL who seem to handle their PR and marketing
Your not wrong. Intel has had to rebrand to Core "Ultra". AMDs marketing and branding is subpar. Consumers are dumb sometimes beautiful boxes sell more.
Great job - been an interesting two weeks, thats for sure
Appreciate it thanks
@@KitGuruTech Welcome!
Far Cry 6 is a focus - I dont understand, its an old engine that hasnt been really changed for a very long time
What about Windows 10? Or is this optimization only needed for Windows 11? I don't (and don't plan to) use Windows 11.
Win10… end of line. No need upgrades…
😂
Yeah that would interest me too. Maybe this complete situation isnt a big problem on win 10? Has this anyone tested so far? @haukionkannel - microsoft gives extended support for win 10 until october 2028, called Extended Security Update-Programm (ESU)
If I understand what the AMD rep in The Full Nerd podcast said, it will be updated for Windows 10 also, and will also improve performance for Zen 3 parts
@@Ben-Rogue Oh, neat! So free performance? I'm using Ryzen 7950x on Windows 10. How big of an increase can I expect and in which workloads, if you don't mind asking?
@@bgtubber Well, if it's anything like the uplift people will see for Windows 11; in CPU bound gaming (1080p type gaming), you'll probably see about 5%-10%, but I'm not sure other workloads will see.
Why is Microsoft enabling half of this stuff by default? seems utterly pointless to me
Amd are doing pr for Intel. And Microsoft are doing pr for Linux. Amazing 😅
Thanks for setting aside the bling-bling upgrade and digging into this Leo. Great points, I assume it's still "Leo's advice" to "Wait for Zen5 X3D."?
FWIW, in my opinion AMD marketing is grasping at straws. Users don't use "Insider previews" of Windows. VBS disabled (i.e. SVM disabled in BIOS) is not how users run their machine. Running in some sort of "super admin" mode or even running a game "as Administrator", users do neither. AMD needs to "own" their marketing (reviewer guide, statements at the Zen5 preview event, etc.) lies.
Heck yes, we cannot wait for Zen 5 3D, which is apparently coming in January 2025 i.e. CES. Damnit.
Leo
I agree that AMD's marketing department need to pull their head in regarding performance claims, but I don't read their 21 August blog post as saying typical users _should_ disable VBS, run Windows Insider builds, or run applications in system-admin mode. They do give instructions for users who *are* willing to trade security/stability for more performance, and IMO it's a little irresponsible for them to do so without stressing the potential impact on security &/or stability. However, I'm sure they only intended those instructions for the minority of users who can make a well-informed risk assessment, i.e. they think the overwhelming majority of Windows 11 users are better off just waiting for the mainstream update.
@@wereoctopus They're saying all these options improve performance more on the new ryzens as opposed to the previous gen ones, which supposedly widens the gap to match their original claim of improvements over the previous gen
They're heavily grasping at straws however and it's evident by everyone
@@KitGuruTech agreed. was disappointed to see the latest "leaks" pushing them from "soon" (ala near 800-series chipset launch) out to CES. :(
We all run 'insider previews' eventually. Just more stable versions. What these various testing modes and options indicate is that Zen5 needs software optimizations to realize its full potential. No different than Intel who has been in bed with Microsoft for decades and who undoubtedly got first dibs and preferential treatment when it came to optimizations for its processors. That may be changing, thankfully. I'm not saying AMD should not take some blame, but they were up against the Wintel beast for years.
Damn, what the hell happened with Total War: Pharaoh?? That's a huge performance uplift. Either their code HEAVILY depends on branching or something else is going on.
Microsoft really should have worked with Amd on promoting all this better. But don’t believe anyone should be forced to use an admin mode for this
I am sure they did on some level but no one got any of this info at launch in-depth.
Awesome video Leo! I love your reviews and videos
Would have been interesting to see the scaling for the 7700/7700X, but it will probably be the same scaling as the 7800X3D.
I wish to see if Zen 2 and Zen 3 also benefit from WIndows 11 24H2. Some of us poors on older processors and are running Windows 10 and we want to see if Windows 11 24H2 improves performance on our old Ryzen cpu's, in other words whether it's worth upgrading to Windows 11 24H2 for the extra possible performance from Windows 10.
KitGuruTech, amazing content keep it up
Thanks!
This is hands down the most thorough review of the 24H2 out there, Leo really is the GOAT. Appreciate you including the 7950X and 7800X3D in your benchmarks! I'd love to see some productivity comparisons too, like the 7950X vs 9950X and 23H2 vs 24H2.
This reflects the complexity of high performance PC testing - wow, how to talk down to the press. Intel come back all is forgiven
wow the X3D chips are epicly efficient, it blows my mind
so just imagine what performance we could get if windows will be a bit more quality product from the beginning (i mean everything after xp)
I hope there will be a huge benchmark video where you test intel 12th gen to the newest arrow lake, zen3,4&5 on windows 10, windows 11 24h2(vbs on & off). If this update optimization also comes to windows 10, I wanna stay on windows 10 as much as possible. It's mind boggling how in some games, the gains are ridiculously high but it could be an error as you mentioned in the video. Clearly AMD should have just delayed a bit more until windows 24h2 was launched. 7800X3D also gaining the same improvement makes it the rare Win for AMD.
At least they were open about the issue.
Annoyingly, the 9950X utilizes core parking just like the 7950X3D/7900X3D so they will not work properly if the motherboard has already seen a normal chip like a 9700X, and vice-versa.
This means a Windows reinstall is required after installing the CPU. You might get away with just reinstalling the chipset drivers though.
Perhaps this is the reason Total War: Pharaoh didn't boot with the 9700X?
hope they release same update for win10 too
So what exactly has changed with 24H2? Did AMD also use 24H2 for testing the Intel CPUs?
Nobody cares about testing Intel CPUs right now because they've had 2 generations fail back to back. Intel is basically DOA until they fix their brand name.... somehow.... if that's even possible.
Yes, they mention that in the blog post
Leo
Nothing. As yesterday is 24h2 pre release is very unstable cause the new recovery feature.there is o 24h2 beta. And then dev and canary, no one need test new cpus on dev or canary
They optimized the branch prediction code in 24H2.
@@jonathanwilkerson2859 branch prediction is built-in in the cpu itself and isn't really an os-dependent feature in any way (unless we're talking microcode updates, but these should be independent driver/firmware updates and not tied to the os version). "they optimised FOR the cpu's new type of branch prediction" is probably more correct. it'd still be interesting to learn what exactly they mean by that in more detail however.
the way i imagine it, the os, or any other software for that matter, can be coded in ways that work well or not with a certain cpu's branch prediction if some quirks of it are known in advance. for a few probably too simplistic examples: prefer conditions based on carry flag more or less than on overflow flag, zero flag, or sign flag, prefer near jumps to long ones, prefer hardcoded address jumps to indirect ones, prefer conditional jump to conditional call, and such.
still, its hard for me to imagine what os component could have such a noticeable performance difference by being optimized in this way, when the majority of branch predictions in all these tests should not be in os code but in game code. was it something generic like the task scheduler? if so should we expect improvements in non-gaming workloads as well? or was it actually gaming related, some part of the rendering pipeline, like a vulkan/directx/opengl library? but then why can't that just get updated separately from the whole os?
and should we expect even better performance improvements if game code were optimized as well? would that involve just recompiling it with a compiler version aware of the quirks of the new cpu branch prediction mechanism? have compiler makers been given the information to make those optimizations already?
i'd really love if some tech-head can give me a layman overview of what exactly is going on here.
AMD - releasing a help page after the launch is over. I hope someone in the company looks into how badly this was handled when Intel were down and in a perfect position to kick hard
zen 5 launch and intel selling shit cpus for 2 years is not even close of beeing the same :D
Intel is still down, and falling. It's really bad.
@@PohnnyRico this happens before its just intel is the downside so again its a battle of whos going to give up
Cope. There are established people in Intel actively sabotaging new additions, effectively killing innovation and any fighting chance they might've had.
Great job Leo and team as always.
Glad you enjoyed it
Really interesting to see the results.
the fact that 7800X3D still has performance left on the table is nothing sort of spectacular.
Great Vid, as a 7800X3d owner I just switched to Windows 11 24H2, do I need also to disable VBS? Thx.
Good news to see improvements not only with zen 5 but zen 4 !
Some great performance increases. When does this OS uodate roll out for public ?
AMD made a right shambles of this launch - I liked your videos leo, always just straight to the point. thanks for all the hard work
thanks, glad you like them!
AMAZING LEO
Hats off for taking the time to do this, I decided to buy a 9900x and so far I'm pleased as gaming is also a side hobby so that wasn't my main concern plus in single thread with a bit of tweaking I'm ahead of the upcoming Intel Ultra 9 leaked geek bench score in single thread with great temps and wattage so personally I feel Zen 5 is an upgrade, Maybe I got a good chip but for those that like to tweak their systems ( custom water cooling loops etc) I feel buyers will have a fun time with this generation.
Great work Leo. Loved the coverage from the start. Always objective and calm 🤘🏻
Glad you enjoyed it!
Thanks for the information I got my rig parts ready for build i will make sure I got that windows version is there any other setting that needs to be set on windows or some other program?
Good content, thank you
Thank you for testing it! Thumbs up.
Happy to help!
Will the update improve performance in non gaming tasks too?
Some quick runs in Blender and Cinebench suggest not, presumably because branch prediction plays no part.
Leo
I hope you can also test ryzen 5000 series with the new version
Great job Leo very enjoyable and informative
Glad you enjoyed it
Did you ensure the 9950 was the first CPU installed on the system? Because it is like the X3D models in that regard.
Been a heck of a learning experience for a lot of people this launch in August - been a complete pigears of a launch
KitGuruTech, nice video keep up the amazing work
Thanks glad you like it
Great video and review Leo!
Thanks for adding 5800x3d.
Pity he didn't have time to test the 5800X3D with 24H2. It would be nice to know if the performance uplift affects older processors as well. I guess we'll wait for Steve from Hardware Unboxed to test, I'm sure he will. The question is, will be standing when he does?
Excellent to see thanks Leo
Glad you enjoyed it
Ending my week with a kitguru catch up!! Love this account but amd could definitely do better 😂
i want to upgrade from a 3400g on a gb x570 board. so i bought the new gb x670 mb planning to buy a 9700x in the next 12month depending on the price along with a pair of new ddr5 modules. im still convinced about the amd concepts hence the last upgrade from an old intel system already 5y ago now practically halfed my energy costs on the 65w tdp.
why would you buy the 9700x when you can get the 7800x3d the better chip for 30 pounds cheaper? just looking on amazon rn
9000 series ahead of its time... waiting on everything else to ketchup.. 🤔 Thank you Leo..
I would like to see more games tested. Nice video anyway.
We didn’t have a lot of time.
thank you for doing this. makes no sense to me when single tread performance stayed the same in both windows version while the performance of games go up .
hardware and software makers always been corrupting the markets with their influence for more profits . i didn't know they sell BS to .
Really good to see improvements across the board. 🍺
Wow some weird looking changes at times - that pharoah game runs like pants for me too. They really messed up by not telling the channels about this from the very start- clearly a massive disconnection between AMD marketing and technical teams
The AMD marketing teem has been a bit disconnected from their brains for years.
Amd should have waited a month or two clearly based on the OS changes.
Thank you for the detailed testing.
Our pleasure!
the interesting part for me would be the low and midrange PCs (e.g. 5600X to 5800X3D) with games, that do have a CPU bottleneck. Playing a high end game in 180 or 200 fps does not really make a big difference, but even a couple of fps more from 55 to 60 fps does make a difference.
Thats a lot of testing for a few days Leo - thanks - I am not sold on any of these processors right now, waiting to see what Intel do in the next few months
You and me both!
@@KitGuruTech Surely Mat!
I still love AMD - seems like this was a lot of issues combined into one - but I didnt like the wording on some of their blog post blaming reviewers and how they test. thats nonsense
some of the reviewers made a mess, you compare results and methodology and some shpuld bot dp reviews, others did a great job
the company can get upset, but people in the end want more performance, no idea why, but they want it
they did not care for the reduced power consumption
amd will take note and push more watts in future cpus, so people is happy
Thank you Leo!
Hey brother!Can you help me with a question since you know far more than me?I prioritize NC and comfort and I have these options;(edifier WH950NB, sony wh750b, 1more sonoflow, anker soundcore q30 and space one)what option is better?you can sugest something else if you know 😊
Interesting results 😮
How is it possible that the 7800X3D, in Total War, gains 40% at 1440p but only 10% at 1080p?
Wouldn't the opposite make more sense?
some pretty good gains - interesting theory Leo on INTEL at the end ! who knows. they are very close to Microsoft
Pretty shocked at the gains as well, good news for all Zen owners.
@@jeffreypaul9428 only those who use windows 11.
Zen 3 boost also? Interesting!