AMD Execution Thread [2024]

I'm still unclear on where this idea that Zen 5 is great for mobile/laptop is coming from, outside the already debunked idea that Zen 5 is much more efficient?
Strix Point looks to be about 20% faster than Phoenix on the same process, albeit with more cores. While Zen 5 on desktop is 10% faster, but on a better process. So Zen 5 is roughly where we would expect a Zen 4 processor ported to N4P to be, where as Strix Point actually seems to deliver decent generational advancements.
 
But that doesn't make it fine. We should still expect better and it should be reasonable to say so.
It also doesn't make it not fine. There's only so much a bunch of engineers can achieve with a given baseline, time and money budget, and given goals.
Whose to say they didn't do amazing with the constraints and resources available? Or whose to say they didn't do appaling

The competition (intel) is the only thing that put all this into perspective. Fortunatelly we have it in reasonable shape here
 
But that doesn't make it fine. We should still expect better and it should be reasonable to say so.

By your reasoning, we could literally defend any terrible release by anybody ever. If the RTX5090 only improves performance over the 4090 by like 10%, would you say that's all good?
The 9700X isn't a flagship product. A better comparison would be the RTX 4060 only improving performance over the 3060 by 15%, which actually happened. And GPUs have delivered higher generational gains over CPUs for years now.
 
I’m not defending anything. GPU comparisons aren’t relevant since GPUs actually have a history of significant generational uplifts.
I dont think I'm long for this forum reading through so many of these comments. It might honestly be worse than going through the Youtube comment section. I'm gonna go insane, especially when I'm dealing with people who I know damn well should know better.

"I'm not defending anything"

Yes, that's literally what you're doing. You're saying that it's ok that they offer a small level of performance gain because ignorant people with older CPU's wont care or know any better.

Ryzen also has a history of better generational uplifts, but you dont seem to care about that.
 
The 9700X isn't a flagship product. A better comparison would be the RTX 4060 only improving performance over the 3060 by 15%, which actually happened. And GPUs have delivered higher generational gains over CPUs for years now.
The 4060 was widely tarnished by most as a terrible new x60 part. And recognized by some more informed people as the low end part it actually is, being sold as a midrange part.

BUT, it actually did provide a very significant efficiency improvement, unlike Zen 5.

Being a flagship part or not is not the point, and it's very weird to not grasp that. Or maybe you do grasp it, and just are trying to find ANY kind of talking point to push against me, no matter how poor.
 
It also doesn't make it not fine. There's only so much a bunch of engineers can achieve with a given baseline, time and money budget, and given goals.
Whose to say they didn't do amazing with the constraints and resources available? Or whose to say they didn't do appaling

The competition (intel) is the only thing that put all this into perspective. Fortunatelly we have it in reasonable shape here
This is just unbelievable. I mean, I'm near 'tearing my hair out' levels of bewilderment at some of the stuff I'm reading here trying to defend this. Are y'all all just professional contrarians or something?

And no, the competition is not the only thing that puts this into perspective. That's another absolutely terrible take. Even if Intel didn't exist at all, we could still make the same judgements about Zen 5, just in comparison to their own previous efforts. Even without competition, a company offering a new product after two years needs to provide a justification why people should buy it over their last product.
 
Strix Point looks to be about 20% faster than Phoenix on the same process, albeit with more cores. While Zen 5 on desktop is 10% faster, but on a better process. So Zen 5 is roughly where we would expect a Zen 4 processor ported to N4P to be, where as Strix Point actually seems to deliver decent generational advancements.
There's nothing special about Strix Point. It's simply using lower clocked Zen 5c cores, which of course makes a difference. And the 'but it has more cores' part is kinda a big deal, no?

Also you cant just say 20% faster when talking about an APU. 20% faster in what?
 
Of course most Zen 4 buyers shouldn't be looking to upgrade to Zen 5, but the improvements we are seeing with Zen 5 don't bode well for Zen 6 being a properly worthwhile upgrade from Zen 4 and to me that would be a let down. It would be fantastic to have an AM4 CPU one day offer a 40%+ CPU limited gaming improvement from my 7800X3D while staying on the same platform, otherwise the upgrade isn't all that worth while unless heavily discounted at near EOL. If we have to wait for a whole new platform be it Zen7/8 or Intel for this kind of performance upgrade then I'd probably hold off upgrading longer like in the bad old days where I kept my i5 4670k for almost 8 years. Of course there are upsides to such a scenario - but I like to see technology moving forward for the sake of what is made possible in games.
 
I dont think I'm long for this forum reading through so many of these comments. It might honestly be worse than going through the Youtube comment section. I'm gonna go insane, especially when I'm dealing with people who I know damn well should know better.

"I'm not defending anything"

Yes, that's literally what you're doing. You're saying that it's ok that they offer a small level of performance gain because ignorant people with older CPU's wont care or know any better.

Ryzen also has a history of better generational uplifts, but you dont seem to care about that.

I said nothing about ignorant people who don’t know better. I’m talking about regular people living in the real world who aren’t suffering from an elitist forum dweller sense of entitlement. It’s a perfectly fine product for those people who are the majority by the way.
 
There's nothing special about Strix Point. It's simply using lower clocked Zen 5c cores, which of course makes a difference. And the 'but it has more cores' part is kinda a big deal, no?

Also you cant just say 20% faster when talking about an APU. 20% faster in what?
Well for example, Phoronix found across all their benchmarks that the 7840HS was 18% faster than the Ryzen 7 7840U, while consuming 24% less power.

"When taking the geometric mean of the 100+ benchmarks run for this launch day article, the Ryzen AI 0 HX 370 was overall about 10% faster than the Ryzen 7 7840HS and about 18% faster than the Ryzen 7 7840U"

"Where things got really interesting with the Ryzen AI 9 HX 370 was the power efficiency of this Zen 5 laptop SoC. Across the span of all the benchmarks, the Ryzen AI 9 HX 370 SoC was pulling about 20.4 Watts with a peak of 34.2 Watts.... Meanwhile the Ryzen 7 7840HS had an average of 35 Watts and a peak of 60 Watts. The Ryzen 7 7840U had a 27 Watt average and a peak of 51 Watts. The Ryzen AI 9 HX 370 came out faster than those parts while consuming significantly less power."


My claim is that this kind of performance per watt improvement is what we would expect with a new generation of AMD laptop CPU, where my point of comparison would be something like the Zen 2 > Zen 3 transition. If you want to say that delivering what is expected makes Strix Point "nothing special", then fine. That's just a question of semantics. Regarding your other points:

1.) We see a significant improvement in single threaded performance too (16% in Geekbench v6 and Cinebench R23 single thread vs. the 7840U), so it's not just the extra Zen 5c cores that are making a difference.
2.) The Zen 5c cores are part of the design, so I don't see why they shouldn't "count" towards the overall performance improvement. Moreover it may be that Zen 5 does not deliver a larger IPC boost, precisely because AMD was targeting a strict transistor budget that would allow them to fit 12 cores (8 x Zen 5c) into a laptop APU and 16 cores into the "dense" server CCD. We see on the desktop side that the Zen 5 CCD is no bigger than Zen 4's.
 
Of course most Zen 4 buyers shouldn't be looking to upgrade to Zen 5, but the improvements we are seeing with Zen 5 don't bode well for Zen 6 being a properly worthwhile upgrade from Zen 4 and to me that would be a let down. It would be fantastic to have an AM4 CPU ...
But nor should Zen 4 buyers be looking to upgrade to Zen 6.. CPUs have not delivered worthwhile improvements over 2 generations, for gaming.
But it's up to you to decide that , with your purchases of course.

I agree with you it would be fantastic. But 40%+ over 2 generations or in 3-4 years for CPUs, that happens once in a decade if at all.
X3D was a one off thing, and I'm guessing (not having owned one so far) the improvement was noticeable in a few scenarios at least. We should hope for more thechnologies like it, they are reasons for changing CPUs more often indeed.


And no, the competition is not the only thing that puts this into perspective. That's another absolutely terrible take. Even if Intel didn't exist at all, we could still make the same judgements about Zen 5,
People's expectations are fluid and unrealiable. In particular my issue is when they can be arbitrary and ignoring the bussiness engineering context.
I mentioned Intel because their results implictltly consider that context. When we compare a product, i hope we all want more information available for that, not less. Arrow Lake should "inform" our feelings further, be them "spot on" or "misguided.

Of course feelings and expections are a valid way to evaluate any product. I'm dissapointed too about Zen5's performance for desktops. My feelings asside, can't find it to be particularly bad product.
And notebook stuff looks quite positive, otoh.

A company offering a new product after two years needs to provide a justification why people should buy it over their last product.
That's just your expectation.
I think a more useful one for us customers is: Company X launches producty Y. "Is there a justification why I should buy it? No?" Then wait until next price dropts or next generation, and repeat the logic
 
Last edited:
This is just unbelievable. I mean, I'm near 'tearing my hair out' levels of bewilderment at some of the stuff I'm reading here trying to defend this. Are y'all all just professional contrarians or something?
It's a different perspective. Contrariness is inevitable when you get different people together to share their ideas. If people thinking different to you is a problem, there are other corners of the internet full of echo-chambers where everyone can share the same viewpoint unchallenged. ;)
Even without competition, a company offering a new product after two years needs to provide a justification why people should buy it over their last product.
That doesn't make sense to me. Most people didn't buy an AMD processor last year, or the year before, or the year before that. If I go looking for a new processor now, I might consider AMD's options. There's the new CPU and the old one. If the new one isn't noticeably better but isn't more expensive, I'll go with the new one. If the old one is pretty much the same performance but cheaper as it's discounted stock, I'll get the old one...until it's no longer available because it's not being made. Then I get the new one that is no better but also no worse, so long as it's the same price. If it's more expensive, maybe I start look at Intel instead. What were the reasons for considering AMD in the first place?

There are many products that really aren't better on iterations. In fact quite often they are worse. In the space of waterproof cycling gear, companies have tended to keep the same price but cut back on features. The new product is worse than the old one. So why do I buy the new one? Because the old one isn't available any more. As a consumer it sucks, but as a business, that's how it is.

The justification here for this CPU, from scanning this discussion, is that AMD have updated their baseline. That's akin to refactoring code in a game. You put in a year's work, roll out a big update, and it's no different for the end users. But that update enables you to improve on future updates, and it needed to happen.

So what you're suggesting, assuming that's what's happened here, is that AMD should do the work to create a new base, but not release any products until they can expand on that base to something offering better performance than the old base. Skip a generation without any releases. But getting the silicon out there and working is a necessary step towards that progress.

At the end of the day, if this CPU doesn't offer what you want, don't buy it. It's not like the old CPU is being replaced with something worse, so consumers are no worse off. It's only a problem for AMD if their competition is more powerful and rolling out more power every iteration. If not, this is just a refactoring of the current CPU. Think of it like a PS2+. It was basically the same as the PS2 and offered no reason to upgrade, but if you didn't have a PS2, and PS2+ gave you a tiny improvement. It's not a disaster or a failure of AMD. You'd have to see where it leads and if it fits AMD's goals.

I suppose what's really needed to measure if this CPU is a win or not is what exactly were AMD's goals for it? If it met their goals as part of their long-term plan, it's a win.
 

IOActive researchers Nissim and Okupski's Sinkclose technique works by exploiting an obscure feature of AMD chips known as TClose. (The Sinkclose name, in fact, comes from combining that TClose term with Sinkhole, the name of an earlier System Management Mode exploit found in Intel chips in 2015.) In AMD-based machines, a safeguard known as TSeg prevents the computer's operating systems from writing to a protected part of memory meant to be reserved for System Management Mode known as System Management Random Access Memory or SMRAM. AMD's TClose feature, however, is designed to allow computers to remain compatible with older devices that use the same memory addresses as SMRAM, remapping other memory to those SMRAM addresses when it's enabled. Nissim and Okupski found that, with only the operating system's level of privileges, they could use that TClose remapping feature to trick the SMM code into fetching data they've tampered with, in a way that allows them to redirect the processor and cause it to execute their own code at the same highly privileged SMM level.
 
AMD actually makes good products. Their marketing and pricing decisions are killing them. HUB is right about AMD releasing stuff at a stupid price only to drop it a couple months/weeks/days later. It's not worth it, those initial negative reviews stick around forever.

Now our introduction to Zen 5 is to see it getting absolutely dragged by reviewers because AMD decided to change the power profile and wipe out any gains the 9700X might have had. This is another own goal from AMD. Zen 5 isn't that bad.
 
AMD actually makes good products. Their marketing and pricing decisions are killing them. HUB is right about AMD releasing stuff at a stupid price only to drop it a couple months/weeks/days later. It's not worth it, those initial negative reviews stick around forever.

Now our introduction to Zen 5 is to see it getting absolutely dragged by reviewers because AMD decided to change the power profile and wipe out any gains the 9700X might have had. This is another own goal from AMD. Zen 5 isn't that bad.
I mean, Ryzen 9000 series isn't the first Zen 5 on the markets. We got introduced to Zen 5 with Ryzen AI series.
 
I mean, Ryzen 9000 series isn't the first Zen 5 on the markets. We got introduced to Zen 5 with Ryzen AI series.
We know that. Look at the headlines. This is about AMD execution which includes marketing.

Very Efficient Ryzen 7 9700X Held Back by Power Limits!
AMD 9700X and 9600X Benchmarks... OOF
AMD Ryzen 7 9700X Review - Zen 5 Sucks For Gaming!
Wasted Opportunity: AMD Ryzen 7 9700X CPU Review & Benchmarks vs. 7800X3D, 7700X, & More
AMD Ryzen 7 9700X review: YouTube hates this CPU

For the majority of potential customers, this will be their introduction to Zen 5. I know AMD marketing sucks but this is indefensible.
 
Last edited:
Back
Top