News AMD announces Zen 5 Ryzen 9000 processors launches in July — four new Ryzen 9, 7, and 5 processors with a 16% IPC improvement

[jeremyj_83]

So, the first part is suspicious, because 33% seems too close to the 35% PPT ratio to be a coincidence. Whether it was operator error or maybe a bug in the BIOS they were using, it sure seems like they restricted TDP and not PPT.

I assume they are adjusting the TDP and that PPT is not adjustable. That would explain the additional power numbers in y-cruncher that follow the PPT exactly.

 

[TheHerald]

Are you a PROFESSIONAL REVIEWER? No. Of course people aren't going to believe you. You could easily change a setting or two to show results that YOU want. Just like I'm sure I could figure out a way to make my almost 11 year old 4770k faster than a 7800X3D in gaming. People who do this professionally, while they might have bias toward AMD or Intel, have to be as objective as possible. If they aren't then they are out of a job.

Where do you get a degree of professional reviewing? I'm more knowledgeable than a lot of those professionals "unboxed cough cough".

And no, if these people are objective they will be out of a job. Lots of reviewrs have actually said exactly that, that unless their videos praise amd and poop on everything else, they are not getting clicks and views. They have to be as amd biased as humanly possible to get revenues.

 

[TheHerald]

Random forum blog poster vs professional reviewer. I'll trust the professional.

And that's why you are wrong. Cause you are trusting the professional that the 7950x scores 31k at 65 watts. Lol, go buy one and come back to talk about it.

 

[bit_user]

Yeah, that's more like it. Though need to look at it all later. Doubtless it raise as many questions as it answers, especially since it looks optimistic for intel, but better starting point that the Z5 information-wise.

Thanks. 🤙

Protip: I find it's generally quite rewarding to take the time to click through all the slides. They often have details not present in the article text (and sometimes even contradict it).

 

[bit_user]

Random forum blog poster vs professional reviewer. I'll trust the professional.

Yeah, there's a lot we don't know about your (Teh Harold's) setup. Professional reviewers tend to be a lot more careful about keeping their test machines pristine and making a minimal number of controlled changes. Something like which antivirus / anti-spyware program you install, or a quick-launcher utility for some app, could have an impact on your benchmark scores. I once even saw a case where the screen saver a guy had was interfering with his benchmarking.

 

[KnightShadey]

Protip: I find it's generally quite rewarding to take the time to click through all the slides. They often have details not present in the article text (and sometimes even contradict it).

Yeah, I know, been doing that for a long time.

It was in ASUS' slides that I found their glossy product page for the TUF A16 that heralded "With 45 TOPS NPU Next-Gen AMD Ryzen" ( so did the number change from initial info-share with the OEMs, so AMD could beat intel's launch #, or was it a Qulacomm typo?) No mention of TOPs on the A14 slide.
That's also where they had their product-page QR code that was live for a coupla hours, but little info there.

BTW, OLD-PRO Tip: Change URLs provided and sometimes you can find info not intended to be shared (also good way to find FCC/patent type info). Doesn't work as well as it used to because people are getting smarter, and PR/web-managers have been fired,, but still occasionally yields an info goldmine.

Lots intel info (although still some gaps, and question does theory/PR translate to real-world?) , but still a sparsity of AMD info, so looks like reviewers/sites got the intel Deep dive, but not AMD.

I'm sure more will become clear before July, just want deeper dive so we can get insight on ecosystem/limitations for a bulk buy (dozens, not enough to get special access).

Waiting sucks, getting deployment ready for when folks come back from the summer holidays would be nice timing, for which the only question is Strix vs 7945HX , waiting For Halo or intel will be a coupla folks already on Core 7/9 with 4070-90s so they're fine it's the GTX1650/iGP folks I want to get done ASAFP.

 

[TheHerald]

Yeah, there's a lot we don't know about your (Teh Harold's) setup. Professional reviewers tend to be a lot more careful about keeping their test machines pristine and making a minimal number of controlled changes. Something like which antivirus / anti-spyware program you install, or a quick-launcher utility for some app, could have an impact on your benchmark scores. I once even saw a case where the screen saver a guy had was interfering with his benchmarking.

Unlike a lot of "pro" reviewers I know what I'm doing. It's the actual reviewers that are publishing unreliable data. Anandtech does it, hwunboxed does it, techpowerup does it, even gnexus does it. Haven't seen tom's hardware do it yet.

For example

Spoiler

These are your experts with their graphs made on excel. Yes, I get that even reviewers can make mistakes. But not understanding - instantly - just by looking at these numbers that you ***cked up majorly means you have absolutely no clue what the heck you are doing and how cpus work.

Funny thing is, your own logic is failing here. You are trusting "pro" reviewers. Great, which ones? Anandtech says 31k @ 65w, hwunboxed says 32k @ 105 watts. Not all of your pros can be right, but they can definitely all be wrong. But yeah, it's my setup that is fauly, lmao

EG1. You have right there in your signature a video about one of those "pro" reviewers that majorly messed up almost all of his numbers....

 

[jeremyj_83]

Unlike a lot of "pro" reviewers I know what I'm doing. It's the actual reviewers that are publishing unreliable data. Anandtech does it, hwunboxed does it, techpowerup does it, even gnexus does it. Haven't seen tom's hardware do it yet.

For example

Spoiler

These are your experts with their graphs made on excel. Yes, I get that even reviewers can make mistakes. But not understanding - instantly - just by looking at these numbers that you ***cked up majorly means you have absolutely no clue what the heck you are doing and how cpus work.

Funny thing is, your own logic is failing here. You are trusting "pro" reviewers. Great, which ones? Anandtech says 31k @ 65w, hwunboxed says 32k @ 105 watts. Not all of your pros can be right, but they can definitely all be wrong. But yeah, it's my setup that is fauly, lmao

EG1. You have right there in your signature a video about one of those "pro" reviewers that majorly messed up almost all of his numbers....

This sounds like all the climate change deniers. Something like 97% of climate scientists say the climate is changing, however, the deniers know better than the experts. In this instance the professional reviewers are all saying the same thing, but you know better than all of them. Your personal data is the Holy Grail and the only thing that should be believed. Do note that while the professionals won't have the exact same absolute numbers the data they share will tell the same story. In your screenshot from Hardware Unboxed it shows the same story as the Anandtech article but with different numbers. The differences in absolute values at 65W between AMD and Intel are even almost identical between the two reviews as well.

 

[TheHerald]

In this instance the professional reviewers are all saying the same thing

No, they are not. You are either lying or missnformed. One of those so called "experts" has the 7950x scoring 24k at 65w, the other expert has it at 31k. Huge discrepancy.

Do note that while the professionals won't have the exact same absolute numbers the data they share will tell the same story

A 24 to 31k discrepancy isn't trivial. That's a huge difference.

In your screenshot from Hardware Unboxed it shows the same story as the Anandtech article but with different numbers. The differences in absolute values at 65W between AMD and Intel are even almost identical between the two reviews as well.

GREAT, TY, you just proved my point. Steve - hardware unboxed - admitted himself on his twitter that the graph is completely flawed and he **cked up. Your expert said his graph is wrong, and since you are claiming that it's similar to ananndtech's graph, anandtech is also wrong. So all this time you've been missinformed cause you trusted "de expertzz". This is his very own tweet admitting the graph is wrong btw

Spoiler

A 12900k scores higher than a 13900k at ISO power but your expert didn't realize something is wrong for that to be the case, no, someone else had to call him out on that. Great experts you got there bud


Funny thing is, me - a random knownothing forum poster called him out immediately for his faulty graph and that's why he removed it and admitted it was wrong. Go look, you won't find it in his review anymore

Intel Core i9-13900K Review: Hot and Hungry

Hot of the heels of AMD's Zen 4 series launch, here comes Intel with their highly anticipated 13th-gen Core series, codenamed Raptor Lake. Today we're reviewing the...

www.techspot.com

If you are actually interested in knowing the truth instead of made up numbers on excel, at 65w both CPUs score identically in CBR23- around 22 to 23k. At 125w the 13900k scores 31.5k and the 7950x scores around 33-34k. Nowhere near the difference your experts are showing in graphs they themselves admitted are wrong after the fact.

 

[jeremyj_83]

No, they are not. You are either lying or missnformed. One of those so called "experts" has the 7950x scoring 24k at 65w, the other expert has it at 31k. Huge discrepancy.

A 24 to 31k discrepancy isn't trivial. That's a huge difference.

GREAT, TY, you just proved my point. Steve - hardware unboxed - admitted himself on his twitter that the graph is completely flawed and he **cked up. Your expert said his graph is wrong, and since you are claiming that it's similar to ananndtech's graph, anandtech is also wrong. So all this time you've been missinformed cause you trusted "de expertzz". This is his very own tweet admitting the graph is wrong btw

Spoiler

A 12900k scores higher than a 13900k at ISO power but your expert didn't realize something is wrong for that to be the case, no, someone else had to call him out on that. Great experts you got there bud


Funny thing is, me - a random knownothing forum poster called him out immediately for his faulty graph and that's why he removed it and admitted it was wrong. Go look, you won't find it in his review anymore

Intel Core i9-13900K Review: Hot and Hungry

Hot of the heels of AMD's Zen 4 series launch, here comes Intel with their highly anticipated 13th-gen Core series, codenamed Raptor Lake. Today we're reviewing the...

www.techspot.com

If you are actually interested in knowing the truth instead of made up numbers on excel, at 65w both CPUs score identically in CBR23- around 22 to 23k. At 125w the 13900k scores 31.5k and the 7950x scores around 33-34k. Nowhere near the difference your experts are showing in graphs they themselves admitted are wrong after the fact.

Hurray for you do you want a cookie for finding something out online and letting someone know? I have a theory that the 13900k was thermal throttling so you weren't actually right and their numbers were correct. That said I'm done arguing with a brick wall. You have been proven wrong multiple times but just like climate deniers it doesn't matter how much evidence there is to contradict your beliefs you still believe some random soothsayer instead of experts.

 

[TheHerald]

Hurray for you do you want a cookie for finding something out online and letting someone know? I have a theory that the 13900k was thermal throttling so you weren't actually right and their numbers were correct. That said I'm done arguing with a brick wall. You have been proven wrong multiple times but just like climate deniers it doesn't matter how much evidence there is to contradict your beliefs you still believe some random soothsayer instead of experts.

Yeah, the 13900k was thermal throttling at 105 watts even though the reviewer said he **ucked up his settings. Rofl, you are the brick wall bud.

 

[Order 66]

I’m so confused as to why AMD would reduce the base and boost clocks of these new CPUs. I mean, looking back, besides the IPC improvements going from zen 3 to zen 4, the main improvements seemed to be higher base and boost clocks. With zen 5, it doesn’t seem like we’re getting much improvement outside of just IPC improvements. I realize that zen 5 may be more power efficient than zen 4, but I was thinking AMD would keep the power draw roughly the same across the lineup, and increase base and boost clocks.

 

[jeremyj_83]

I’m so confused as to why AMD would reduce the base and boost clocks of these new CPUs. I mean, looking back, besides the IPC improvements going from zen 3 to zen 4, the main improvements seemed to be higher base and boost clocks. With zen 5, it doesn’t seem like we’re getting much improvement outside of just IPC improvements. I realize that zen 5 may be more power efficient than zen 4, but I was thinking AMD would keep the power draw roughly the same across the lineup, and increase base and boost clocks.

9950X (170W TDP) 4.3/5.7 vs 7950X (170W TDP) 4.5/5.7
9900X (120W TDP) 4.4/5.6 vs 7900X (170W TDP) 4.7/5.6
9700X (65W TDP) 3.8/5.5 vs 7700X (105W TDP) 4.5/5.4
9600X (65W TDP) 3.9/5.4 vs 7600X (105W TDP) 4.7/5.3

Generally the boost clocks are the same or slightly higher. However, the base clocks are lower across the board BUT outside of the 9950X the TDPs have been reduced 40% on the 9900X and 60% on the 9700X & 9600X. That generally means that efficiency is much higher and even with the lower base clocks and TDP performance could still be >16% higher. If the CPU is able to stay at higher boost clocks then you get the benefit of higher clocks AND IPC. All while reducing the TDP which makes the system FAR easier to cool.

 

[helper800]

I've shown proof - not just screenshots but videos - numerous times of eg. 12900k beating 7800x 3d in gaming, 14900k capped at 90w in games being faster then the 3d etc.

Where have you shown this? What resolution? What games? What ingame settings? What graphics card? What motherboards? What cooling for the CPUs? Open test bench or are the systems inside a case? What cases used if any? What kits of RAM? What BIOS settings did you use for each CPU? Clean install of Windows? Cleanboot before the benchmarking? Were they controlled benchmarks? How did you log the data from the benchmarks?

 

[TheHerald]

Where have you shown this? What resolution? What games? What ingame settings? What graphics card? What motherboards? What cooling for the CPUs? Open test bench or are the systems inside a case? What cases used if any? What kits of RAM? What BIOS settings did you use for each CPU? Clean install of Windows? Cleanboot before the benchmarking? Were they controlled benchmarks? How did you log the data from the benchmarks?

We were using different keyboards so I think that's what caused the 12900k to be faster.

 

[helper800]

We were using different keyboards so I think that's what caused the 12900k to be faster.

Are you sure it wasn't the the wifi card?

 

[TheHerald]

Are you sure it wasn't the the wifi card?

Well judging from your initial post - yeah, it might have been. Every excuse imaginable will be called upon. What was the wind strength, what was the humidity etc.

We were using same GPU (4090), stock CPUs, tuned ram, I was running passive cooling he was using a 360 AIO, same game settings. Not that the cooler matters but since you mentioned it.

 

[Thunder64]

Where do you get a degree of professional reviewing? I'm more knowledgeable than a lot of those professionals "unboxed cough cough".

And no, if these people are objective they will be out of a job. Lots of reviewrs have actually said exactly that, that unless their videos praise amd and poop on everything else, they are not getting clicks and views. They have to be as amd biased as humanly possible to get revenues.

That's a good one, there's some great pro-AMD conspiracy all over the Internet. You wouldn't be affiliated with Userbenchmark, would you? BTW, practically everyone has been crapping on AMD GPU's, especially RDNA3.

 

[TheHerald]

That's a good one, there's some great pro-AMD conspiracy all over the Internet. You wouldn't be affiliated with Userbenchmark, would you? BTW, practically everyone has been crapping on AMD GPU's, especially RDNA3.

It's not my opinion, they've said it themselves, that unless they praise amd their videos ain't getting viewed, so they have to. There is a guy that literally said that the title of his video was "5800x 3d, best gaming CPU ever" and he did that just to increase his subscriptions. And he did.

Daniel owen said the same, here is the video

Spoiler

View: https://www.youtube.com/watch?v=mCfYxs12Z9s&t=944s

And im quoting him "Especially being negative about nvidia and positive about amd performs better in the youtube algorith - generates more traffic".

Techtesters UK said the same, guru 3d had to remove some graphs cause they were getting death threats from amd fans and the list goes on. Even Steve from hwunboxed, do you know what CPU he uses at home? A 12900kS. Now go check his review of the 12900ks and what his opinion was and how the heck he ended up with one.

And here is a tweet from a reviewer, he said the the whole industry learned to stay quite about amd issues cause the fans don't take it too well.

Spoiler

https://i.ibb.co/92wVx3F/image-2024-06-07-082542445.png

Brian from techyes city - and im quoting "yeah im running the 14900k completely stock cause that's how my viewers are using their pcs, they don't tinker with the bios". Then he went into the bios to fix the 7950x 3d cache priority.....

Spoiler

View: https://www.youtube.com/watch?v=zVJFx1rOr2Q

I can find you hundreds of these. I mean MLID himself is a prime example that spreading misinformation is a good thing as long a you are doing it in favor of amd.

A professional reviewer doesn't mean he is a "pro" (good) at what he is doing, it means he is doing it to make money. And they will create the content that makes them the most money. That's self explanatory I guess. It's like a newspaper, they don't exist to inform you about reality, they exist to sell ads.

 

[helper800]

It's not my opinion, they've said it themselves, that unless they praise amd their videos ain't getting viewed, so they have to. There is a guy that literally said that the title of his video was "5800x 3d, best gaming CPU ever" and he did that just to increase his subscriptions. And he did.

Daniel owen said the same, here is the video

Spoiler

View: https://www.youtube.com/watch?v=mCfYxs12Z9s&t=944s

And im quoting him "Especially being negative about nvidia and positive about amd performs better in the youtube algorith - generates more traffic".

Techtesters UK said the same, guru 3d had to remove some graphs cause they were getting death threats from amd fans and the list goes on. Even Steve from hwunboxed, do you know what CPU he uses at home? A 12900kS. Now go check his review of the 12900ks and what his opinion was and how the heck he ended up with one.

And here is a tweet from a reviewer, he said the the whole industry learned to stay quite about amd issues cause the fans don't take it too well.

Spoiler

https://i.ibb.co/92wVx3F/image-2024-06-07-082542445.png

Brian from techyes city - and im quoting "yeah im running the 14900k completely stock cause that's how my viewers are using their pcs, they don't tinker with the bios". Then he went into the bios to fix the 7950x 3d cache priority.....

Spoiler

View: https://www.youtube.com/watch?v=zVJFx1rOr2Q

I can find you hundreds of these. I mean MLID himself is a prime example that spreading misinformation is a good thing as long a you are doing it in favor of amd.

A professional reviewer doesn't mean he is a "pro" (good) at what he is doing, it means he is doing it to make money. And they will create the content that makes them the most money. That's self explanatory I guess. It's like a newspaper, they don't exist to inform you about reality, they exist to sell ads.

What does any of this prove? The only thing that any of this even suggests is that people want to hear that AMD is doing well, so videos that show such things get more views, this does not prove that the underlying content is fake or incorrect... How can you go from that to AMD is controlling what people say, and punishing those that wrong-think?

 

[Hotrod2go]

I'm really looking forward to real world benchmarks & PBO testing with this new 9000 series. If one is a gamer & on a budget atm, the 7800X3D has some sweet deals now $ wise speaking. However the new 9000 series just around the corner practically speaking will also offer improvement gains even if not X3D for all applications.
July will be an interesting month!

 

[TheHerald]

What does any of this prove? The only thing that any of this even suggests is that people want to hear that AMD is doing well, so videos that show such things get more views, this does not prove that the underlying content is fake or incorrect... How can you go from that to AMD is controlling what people say, and punishing those that wrong-think?

That people are more interested in watching / clicking videos that praise AMD and poop on intel and nvidia. Therefore, according to their own words, creators have to create these types of videos if they want to succeed. I never said AMD is controlling anyone, their fans are. They are the most extreme and vocal members of the pc community.

Have you ever seen a reviewer REMOVE a graph cause the fans of particular CPU that didn't do well were boycotting? Cause that happened. That's called manipulating results to appease a fanbase.

The underlying content doesn't have to be fake or incorrect, but it can be massaged. Take cyberpunk for example, the game has super light areas that the 3d chips are super fast - 20-30% faster than anything Intel. Then there are heavy areas like tom's dinner that the 3ds get crippled. Depending on where you test, the results are wildly different, so you can present the picture you want without lying. Same with factorio, 10k small maps, the 3d chips are literally 3 times as fast as Intel. But since the game in the official servers is locked to 60fps, that lead is useless. Now go to the big 60k maps and Intel takes the lead.

 

[TerryLaze]

and even with the lower base clocks and TDP performance could still be >16% higher. If the CPU is able to stay at higher boost clocks then you get the benefit of higher clocks AND IPC. All while reducing the TDP which makes the system FAR easier to cool.

Please go ahead and explain (to yourself mostly) what companies mean when they state base clocks....

*hint* it's the clocks the CPU will reach when it's mostly fully loaded under whatever load (in this case) AMD considered relevant.

Under load all the new CPUs will be running considerably slower clocks.
You will have to run PBO or otherwise overclock, using much more power than just TDP, to get the same clocks as the previous gen.
The 65W from 105W TDP CPUs will lose ~15% of clocks while gaining ~16% IPC...they will be the same speed as the older gen just with lower power draw.
Basically AMD is going to sell you the T variant as a new gen.

 

[bit_user]

Please go ahead and explain (to yourself mostly) what companies mean when they state base clocks....

Minimum guaranteed clock speed for a realistic worst-case load.

Given that Zen 5 now has a full 512-bit AVX-512 implementation, that would certainly be what they measured. Zen 4 didn't have this, so it's reasonable to consider Zen 5 might use more power at the same base clocks as Zen 4 had, but it doesn't mean efficiency is worse on either those or non- AVX-512 workloads. Since efficiency is measured in perf/W, you can get an increase in efficiency at more power, just by increasing performance more than the power utilization.

Anyway, because they didn't increase the power requirements (but actually reduced them, in most cases), they had to drop the base clocks. Still doesn't mean that efficiency isn't even better than it was in Zen 4. That can potentially lead to higher typical clock speeds.

Under load all the new CPUs will be running considerably slower clocks.

It's not typical load, but a worst case load. In a typical load, like gaming, you don't even have all of the cores utilized. So, that's definitely not a worst case load.

Really, though... why do we have to waste all this time on speculation. Terry, you don't have the data to support what you're claiming. None of us do. All of this is speculation and conjecture. Why can't we just wait until independent, 3rd party benchmarks come in? A month really isn't that long to wait.

...oh right, because you need to spread FUD to fend off that threat from AMD and try to keep people from upgrading before Arrow Lake launches.

 

[ubronan]

My biggest fear is that it is gonna end up as another heavy overpriced renewable toy
I love to see huge lower power consumption so for me that is a bonus
And i think the X3D variants will be so much overpriced that i again will stay on my good old 9900 KF
Ow and BTW who thinks the 14X00K.... mess are any good guess again.
Take a good look how all of a sudden the 13X00 series jumped up in price as they run and perform much better than the 14x00 variants
So claiming that the 14x00 series is any good at all is really weird as it sucks, enough reviewers gave their negative verdict on that series and mention to go for the 13x00 series
Or in many cases say to look at AMD as a better choice especially regarding gaming.
The most funny stuff is that many people i know, buy a slowboat Intel even when they do not do anything else than gaming on their machines.
The software they have does not even make a x486 series wake up at all