Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)

[DisEnchantment]

Speculate at will

 

[blackangus]

From a consumer standpoint, the only thing that warrants being named a delay is when a company fails to meet deadline that they publicly committed to.

So if Nova Lake releases in 2035 I guess it's not delayed.

If a product release doesn't meet its timeline but nobody can tell time is it really delayed?

 

[Josh128]

If a product release doesn't meet its timeline but nobody can tell time is it really delayed?

Does a bear poop in the woods even if nobody's around to see it?

 

[CouncilorIrissa]

Moral of the story: don't continue to employ marketing and public relations people that are terrible at their jobs.

I believe Lisa Su has a certain relative she could ask what good marketing brings to the table.

 

[blackangus]

Does a bear poop in the woods even if nobody's around to see it?

Exactly. Multiple frames of reference.

 

[igor_kavinski]

AMD talks Ryzen 9000 controversy: 'We were as puzzled' as reviewers

AMD's David McAfee joined our Full Nerd podcast to wade deep into the differing benchmark results seen by the company and independent reviewers.

www.pcworld.com

The explanation could be as simple as "AMD hardware guys have a vague idea of how software works and AMD software guys have an even poorer understanding of how their own hardware works".

Not the first time when two different departments in the same company have communication problems.

 

[Abwx]

That’s fine, it just means AMD is purposefully handicapping Intel in their comparison.

No.

Same settings as Computerbase in their latest reviews, set apart for the RAM since they test all CPUs at RAM stock speeds.

AMD also pointed out that in its first-party tests that compare Ryzen 9000 series processor models to Intel's, the company ran the 14th Gen Core chips with the same DDR5-6000 memory speed with the same timings, and used Intel's baseline power profile that uses stock power limits for these processors (125 W base- and 252 W maximum turbo power in case of the Core i9-14900K and the i7-14700K, for example).

AMD Works with Microsoft to Improve Zen 3 thru Zen 5 Performance in Windows 11 24H2

AMD Ryzen 9000 "Granite Ridge" desktop processors have been out for a couple of weeks now, and the "Zen 5" based processors have fallen short of gaming performance expectations, set mainly by some of the numbers AMD put out in its Computex 2024 reveal for the processors. The consensus among the...

www.techpowerup.com

 

[Saylick]

So if Nova Lake releases in 2035 I guess it's not delayed.

Is there an official Intel slide which says 1) that Nova Lake exists, and 2) when it's coming out?

 

[Rigg]

And also ask the reviewers to make broader tests because HWUB FI use only 13 games in their average and most are the same kind, NBC who used a more diverse set of games and in a bigger quantity got results that were close to the ones of AMD.

I can as well by selecting 13 specific games get all possible results that would be at odd the ones with the others, all i would have to do is to either select 13 games were AMD is as weak as posible or another set were they are very good and then get the conlusions i want.

FI Xplane is a game where Zen 5 does very well and none set apart NBC, and eventually THG, used it in their tests, they all rely on first person shooter view games or the likes.

You mean the 9700X review that came out 08/19/2024? NBC put out any 50+ game benchmarks in follow up pieces like HUB does?I'll take HUB's 13 game test suite on launch day over a 20 game test suite (half of which were released 5-9 years ago) in a review that comes out nearly 2 full weeks after launch.

 

[jdubs03]

The explanation could be as simple as "AMD hardware guys have a vague idea of how software works and AMD software guys have an even poorer understanding of how their own hardware works".

Not the first time when two different departments in the same company have communication problems.

The question should then be: why isn't there a team in the middle, with managers from both the hardware and software side that can settle any discrepancies from either perspective? The result would be neither side is then talking past the other.
Seems like that shouldn't be hard to institute.

Is there an official Intel slide which says 1) that Nova Lake exists, and 2) when it's coming out?

Only thing I've seen was the leaked Dell roadmap. Anything else was from my understanding MLID.

 

[Kepler_L2]

Is there an official Intel slide which says 1) that Nova Lake exists, and 2) when it's coming out?

No and no, hence my question. Officially, Nova Lake doesn't even exist. But internally and to OEMs, Intel has shown Nova Lake launching late 2026/early 2027.

So if it launches in 2035, did it get delayed by 8+ years or did everything went according to plan?

 

[Hitman928]

No.

Same settings as Computerbase in their latest reviews, set apart for the RAM since they test all CPUs at RAM stock speeds.

AMD Works with Microsoft to Improve Zen 3 thru Zen 5 Performance in Windows 11 24H2

AMD Ryzen 9000 "Granite Ridge" desktop processors have been out for a couple of weeks now, and the "Zen 5" based processors have fallen short of gaming performance expectations, set mainly by some of the numbers AMD put out in its Computex 2024 reveal for the processors. The consensus among the...

www.techpowerup.com

Your insistence on being wrong astounds me.

From computerbase’s Intel test settings:

A look at the details under “Internal CPU Power Management” then shows the desired values: PL1 = PL2 with 253 watts and the “current limit” at 400 amps.

 

[Abwx]

You mean the 9700X review that came out 08/19/2024? NBC put out any 50+ game benchmarks in follow up pieces like HUB does?I'll take HUB's 13 game test suite on launch day over a 20 game test suite (half of which were released 5-9 years ago) in a review that comes out nearly 2 full weeks after launch.

Still way more representative than HWUB, it s not like a 5 years old game new version is that different from the previous onen, and what about recent games that are still not ironed out..?

This test use several res and settings contrary to HWUB lazy reviewers who are just using what can be automated easily, they better wake up a little earlier and do more actual work.

 

[Saylick]

No and no, hence my question. Officially, Nova Lake doesn't even exist. But internally and to OEMs, Intel has shown Nova Lake launching late 2026/early 2027.

So if it launches in 2035, did it get delayed by 8+ years or did everything went according to plan?

Remember when Nehalem was originally a future architecture based off of Netburst?


When Intel pivoted from Netburst to P6 as being the underlying architecture moving forward, does that mean Nehalem was delayed?

I wouldn't say so. I'd argue that Nehalem, as we know it today, was released on time when Intel wanted it to be released.

 

[Rigg]

Still way more representative than HWUB, it s not like a 5 years old game new version is that different from the previous onen, and what about recent games that are still not ironed out..?

This test use several res and settings contrary to HWUB lazy reviewers who are just using what can be automated easily, they better wake up a little earlier and do more actual work.

Apparently NBC was too lazy to get a review out until 12 days after launch. This is a really bad take, which seems to be par for the course. You do realize HUB goes out of their way to avoid automated game testing and is critical of it right?

 

[CouncilorIrissa]

So if Nova Lake releases in 2035 I guess it's not delayed.

Nova Lake released in 2035 would be hopelessly dated, so you could infer it having been internally delayed from being years behind in performance, unlike GNR X3D released around CES. An absolutely irrelevant example, not to mention that such a product wouldn't ever be released in the first place.

And in the absence of any other supporting evidence, namely roadmaps, the news about delays of unannounced products are about as worth paying attention to as Intel's MLID's alleged cancellation of "Beast Lake", lol.

edit: as I said, from consumer standpoint, the only delay worth talking about is a failure to publicly announced deadline. And if we're talking about internal roadmaps, then the news of said delays better be supported by a shred of evidence.

 

[Abwx]

DP

 

[Abwx]

Your insistence on being wrong astounds me.

From computerbase’s Intel test settings:

So Computerbase are exceeding Intel s own request, surely to not handicapp them in the benchmarks, they are quite forgiving, and i guess that its a definitve prove thet they are not AMD centric the slightest.

Apparently NBC was too lazy to get a review out until 12 days after launch. This is a really bad take, which seems to be par for the course. You do realize HUB goes out of their way to avoid automated game testing and is critical of it right?

They started with the 9600X some time ago, and they had to wait for AMD delivering a 9700X, you are just throwing useless pseudo arguments here.

At least this allowed them to not do some blunders since hey got wind that a fresh W11 installation was required, otherwise they could had used the set up used for the 9600X.

 

[Hitman928]

So Computerbase are exceeding Intel s own request, surely to not handicapp them in the benchmarks

It really is astounding.

 

[Abwx]

It really is astounding.

I edited my post.

As said it s their choice, but AMD cant be put at fault for respecting Intel s official request, that s way more pofessional than comparing your mobile CPU with a dual channel to the competitor one purposely gimped with a single channel, and then when asked why to answer that it was a laptop out of the box

Of course, after you ordered it with the minimal options at Asus.

 

[Rigg]

They started with the 9600X some time ago, and they had to wait for AMD delivering a 9700X, you are just throwing useless pseudo arguments here.

At least this allowed them to not do some blunders since hey got wind that a fresh W11 installation was required, otherwise they could had used the set up used for the 9600X.

Ah yes! I see, My half joking jabs are "pseudo arguments" but your hot takes on HUB's testing are totally grounded in logic and reason. NBC's testing screams automated canned benchmarks. Are you sure that's not what they're doing? Either way It's funny you''d use that as an argument against HUB's methodology. I'm so glad NBC included all of that amazing GTA V data. Most of it is run up against the 188 FPS limit of the engine. Such amazing games testing! So many settings tested! At least people still play that game. Unlike most of the other 5+ year old games they tested.

 

[Abwx]

Ah yes! I see, My half joking jabs are "pseudo arguments" but your hot takes on HUB's testing are totally grounded in logic and reason. NBC's testing screams automated canned benchmarks. Are you sure that's not what they're doing? Either way It's funny you''d use that as an argument against HUB's methodology. I'm so glad NBC included all of that amazing GTA V data. Half of it is run up against the 188 FPS limit of the engine. Such amazing games testing! So many settings tested! At least people still play that game. Unlike most of the other 5+ year old games they tested.

Everyone use automated testings, but i dont think that reviewers use more than a short scene for each game, so if there s not enough games the statistics will be of low value even for the tested games.

Beside if NBC took that long for the 9700X review it s eventually that they made more extensive tests, just their CPU apps perfs tests are much more complete than HWUB that just do a handfull of tests.

Here all their apps tests, and they have also ST for all CB versions, 7 Zip ST, and Wprime with 32m and 1024m and so on.

 

[Hitman928]

I edited my post.

As said it s their choice, but AMD cant be put at fault for respecting Intel s official request, that s way more pofessional than comparing your mobile CPU with a dual channel to the competitor one purposely gimped with a single channel, and then when asked why to answer that it was a laptop out of the box

Of course, after you ordered it with the minimal options at Asus.

You still don’t understand the Intel default settings guidance but I’m not going to pollute this thread anymore with it. AMD isn’t following Intel’s guidance, that’s it.

 

[marees]

The explanation could be as simple as "AMD hardware guys have a vague idea of how software works and AMD software guys have an even poorer understanding of how their own hardware works".

Not the first time when two different departments in the same company have communication problems.

The question should then be: why isn't there a team in the middle, with managers from both the hardware and software side that can settle any discrepancies from either perspective? The result would be neither side is then talking past the other.
Seems like that shouldn't be hard to institute.

Apparently what added to the confusion was MS recalling Recall resulting in delay of a wider rollout of their latest OS

 

[Josh128]

You mean the 9700X review that came out 08/19/2024? NBC put out any 50+ game benchmarks in follow up pieces like HUB does?I'll take HUB's 13 game test suite on launch day over a 20 game test suite (half of which were released 5-9 years ago) in a review that comes out nearly 2 full weeks after launch.

Just an aside, Johnny Mac was 100% justified and correct to freak out in that classic piece. If you watch the shot that caused his outburst, it was called out but was 1000% in, lol.

 

[Rigg]

Everyone use automated testings, but i dont think that reviewers use more than a short scene for each game, so if there s not enough games the statistics will be of low value even for the tested games.

Beside if NBC took that long for the 9700X review it s eventually that they made more extensive tests, just their CPU apps perfs tests are much more complete than HWUB that just do a handfull of tests.

Here all their apps tests, and they have also ST for all CB versions, 7 Zip ST, and Wprime with 32m and 1024m and so on.

You just criticized HUB for doing automated game testing. I point out that they avoid automated game testing. I ask you if that is what NBC is doing, and you reply with - "Everyone use automated testings", and then you shift the goalposts to app testing.

I have nothing against NBC. I honestly don't pay any attention to them. My initial impression is their desktop CPU reviews come off as a discount version of TPU.

It's been pointed out several times in this thread alone, but apparently the point hasn't hit home. HUB's testing methodology is specifically tailored to it's YouTube audience. The test suite is focused on delivering real world, relevant data to the typical desktop PC gamer who constitutes their audience. They do real in game testing, with modern games, and use carefully selected test runs that produce meaningful data. They collect multiple extended test runs to compile their data which they update for all of the hardware in their charts regularity. They don't muddy the waters on launch day reviews with a bunch of low res, low setting, garbage data from old ass games that isn't really relevant to real world usage for most people. They also do follow up testing that paints a more full picture of many of the products they review. This includes 50+ game tests. Are they perfect? No. Do they do a good job providing useful data to their audience? Yes.

If you think anyone is doing a better job than HUB of testing PC hardware (as it relates to gaming) for new product launches, you're completely delusional. There are plenty of others that do good work as well, but HUB is definitely one of the good ones. Your lazy comment is pure bovine excrement.