Intel's Core Ultra 9 285K Performance Claims Leaked, Doesn't Beat i9-14900K at Gaming

[kondamin]

Man 2024 the year every company decided they dgaf about gamers....

CPUs have only been marginal in boosting gaming performing for many years now, a market segment that’s been guided to getting midrange CPUs and splurging in graphics cards for 2 decades now.

 

[SL2]

Nothing surprising here

 

[efikkan]

TBH yes I agree. I think we are reaching a clock rate max like we did with netburst. We can try and increase IPC even more, but I dont think that will be the primary shift.

Games will only scale with faster CPUs until the CPU is no longer the bottleneck. If you e.g. had a CPU 10x faster per core than current Raptor Lake CPUs, you wouldn't see much of a difference in most games, and most certainly not scale very far with the few games which do. But a few years down the line, you'll probably see a larger difference as game engines become more demanding of CPUs. Hopefully some of this is used for something useful to improve the games, but unfortunately a lot of it will probably be bloat. This is already evident in those games which are very sensitive to large L3 caches, which is a symptom of bloated code.

With coding languages and game engines being more advanced then they were, I actually think we will start to see a shift in gaming whein the parallelism with either the engine, software stack or underlying technology APIs. honestly probably all of them.

For the most part, it's mostly secondary workloads we see more parallelized in games, like loading assets, audio processing or some other async tasks. The two most performance critical elements, the game simulation (which we used to call it "game loop" in the old days) and the main rendering thread will continue to dictate the performance of games. Smaller tasks may be delegated to smaller worker threads, but scaling this greatly increases the risk of delays which results in stutter, or worse glitches like we see in so many games now.

If multithreading is to provide significant gains in gaming performance in the future, there would have to be different kinds of changes than we've seen so far. As latency quickly adds up when trying to synchronize increasing number of threads, efforts to reduce latency or even "guarantee" deadlines would be required. Firstly a much faster OS scheduler, and probably some semi-"RT" like features so threads are undisturbed by other tasks. Secondly graphics drivers etc. would need to behave more like in a RT system, and thirdly possibly HW changes to streamline communication.

But while multithreading often gets the most attention, optimizing for ILP is much more important for performance scaling, whether it's for gaming or user interactive applications. For smaller work chunks which needs to be synchronized, multithreading can only get you so far before overhead or latency bottlenecks it, but modern CPUs are also increasingly superscalar, which means the relative performance gains for writing clean efficient code is larger than ever. And while CPU frontends are increasingly advanced, e.g. Meteor Lake improves branch misprediction recovery, the gains from saturating the pipeline is even greater. The bigger problem here is the software practises which are popular today, especially how OOP, abstraction and generalization are employed. It is remarkable how much having dense logic affects CPU performance. But at some point I would expect compilers and potentially ISAs to evolve in order to scale with wider CPU architectures, hopefully in a better way than Itanium.

So I guess Arrow Lake is going to be like Zen 5, Only slight improvements here and there and the main focus seems to be improving efficiency.

If anything, I'm hoping for more consistent performance. Pushing clock speeds too far leads to very unstable clock speeds, and at least for some of us that may be more annoying than slightly lower but more consistent performance. That is at least my impression from comparing Raptor Lake(i5-13600K) to Comet Lake(i7-10700K) at work, purely anecdotal and subjective impression, even though Raptor Lake has clearly higher peak performance.

Hopefully this will reduce expectations and hype though about Arrow lake performance increases, Zen 5 had such a bad reception at launch due to overhype which led to massive disappointment at launch when people saw Zen 5 did not improve as much as they were expecting.

Overhype servers no one.
Unlike most, I'm not that disappointed with Zen 5, and I'm very curious to see how it performs in upcoming Threadripper models.

 

[kawice]

Weird, arrow lake has a 3 node advantage over raptorlake. It should be doing far better than just use less power.

It's not even that, PL2 is rated at 250W on leaked materials, so that's just 3W less than 14900k.

With HT removal it will be slower in MT workload. ST performance doesn't matter much these days.

The only advantages are native support for DDR5 6400MT memory, raised from 5600MT and I think dedicated PCIe 5.0 NVme port, if someone plans to upgrade now.
Oh and theoretical lack of Vmin Shift hardware bug.

 

[tfp]

It's not even that, PL2 is rated at 250W on leaked materials, so that's just 3W less than 14900k.

With HT removal it will be slower in MT workload. ST performance doesn't matter much these days.

The only advantages are native support for DDR5 6400MT memory, raised from 5600MT and I think dedicated PCIe 5.0 NVme port, if someone plans to upgrade now.
Oh and theoretical lack of Vmin Shift hardware bug.

ST performance does matter and MT workloads only matter up to point. It really depends on how many threads your software really can use. For gaming 6 to 8 cores with or with HT is still enough otherwise AMD 12 and 16 core chips would wipe the floor regardless of some small latency issues like they do in other heavily threaded applications. Really that is what matters, most people don't need more than 8 cores or threads for things to be fast. If someone fits the use cases for a really heavily threaded chip well great buy one but they are not the vast majority of people.

Maybe in the future only MT workloads will only matter and then we will all have slower 100+ core chips, but I doubt it.

 

[ShrimpBrime]

I'm an honest guy. Gonna just tell it like it is. Let me reword the thread title for you.

285K is NOT appealing. Have a great Cinebench score, but games CyberPunk 2077 like shit.

FnA, that's a nail in the coffin pre-release!!! OUCH!

 

[AusWolf]

Hang on... What's that 447 W over there? Are they trying to tell me that this thing compares well to the 7950X3D while eating half a kilowatt?

They must be out of their minds to think that this is acceptable on any level.

 

[SL2]

Hang on... What's that 447 W over there? Are they trying to tell me that this thing compares well to the 7950X3D while eating half a kilowatt?

They must be out of their minds to think that this is acceptable on any level.

You know a 14900K alone doesn't draw 500 W in games, that's insane. I think it's the whole system including a graphics card.

 

[Marcus L]

So AMD Ryzen 9000 series shows 5-12% gains vs 7000 series, epic failure, stagnant, sidegrade, Intel, can't beat last gen in gaming, butttttttttttttttttttt power savings, now 250w vs 350w+ WINNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN

You know a 14900K alone doesn't draw 500 W in games, that's insane. I think it's the whole system including a graphics card. (169.5 W seems to be a copy paste error from average application power draw, it's the same error in the 14700K review.)

So 145 - 80 = 65W in games
View attachment 366869

No not just in gaming but CPU intensive tasks, rendering, editing, CPU intensive tasks, those CPU's literally sup 400w of power and double that if not more than Ryzen in gaming where yes it might not be 500w but still 150-200w vs about 80w

We get the same from the Nvidia apologisers, saying AMD uses 30w at idle with their Intel thermonuclear reactors, it's pretty fucking funny

 

[AusWolf]

You know a 14900K alone doesn't draw 500 W in games, that's insane. I think it's the whole system including a graphics card. (169.5 W seems to be a copy paste error from average application power draw, it's the same error in the 14700K review.)

So 145 - 80 = 65W in games
View attachment 366869

Of course not in games. But I don't want my CPU anywhere near that number in any case. If it's total system power, then I guess that's fine, although with every system being different, such data doesn't say anything to me. I only care about individual component power consumption to figure out cooling needs.

 

[Wasteland]

So AMD Ryzen 9000 series shows 5-12% gains vs 7000 series, epic failure, stagnant, sidegrade, Intel, can't beat last gen in gaming, butttttttttttttttttttt power savings, now 250w vs 350w+ WINNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN

I don't think any sane person criticized Zen 5 solely for its unimpressive performance uplift over Zen 4. The criticisms, rather, were:

- AMD wildly over-promised on performance
- pricing

It has become AMD's habit to set prices too high just long enough to get crucified in initial reviews, then almost immediately drop those prices after the damage is done. It's an unforced error, and it's sad to watch if you're remotely interested in healthy competition. In this case, AMD's went a step further, magnifying their error by arguing with reviewers about their benchmark results. And circumstances magnified the error even more--AMD's flailing over Zen 5 actually took pressure off Intel, which was in the process of immolating its reputation via the ongoing Raptor-Lake-degradation drama.

But sure, I agree; Zen 5 isn't bad. The product itself doesn't deserve much criticism.

 

[SL2]

So AMD Ryzen 9000 series shows 5-12% gains vs 7000 series, epic failure, stagnant, sidegrade, Intel, can't beat last gen in gaming, butttttttttttttttttttt power savings, now 250w vs 350w+

Did you even miss the part where AMD claimed Zen5 being much faster than it is, whereas this whole topic is about Intel saying Arrow lake won't be any faster than Raptor? Do you even understand the difference?

This is the weakest post of the day lol

No not just in gaming but CPU intensive tasks, rendering, editing, CPU intensive tasks, those CPU's literally sup 400w of power and double that if not more than Ryzen in gaming where yes it might not be 500w but still 150-200w vs about 80w

Well look again, It clearly says AVERAGE (in chinese, but no need to know that anyway) FPS (which I bet you can read), yeah that's for games, which means including a high end graphics card, and you know it'll be a 4090 (and not AMD or Arc), and you how much power they draw.

It doesn't say 150-200 W anywhere. I already shown you the comparison with TPU's review. You're just making shit up, but since you mentioned "CPU intensive tasks" in the same sentence TWICE I'd suggest you take a nap before replying.

Of course not in games. But I don't want my CPU anywhere near that number in any case. If it's total system power, then I guess that's fine, although with every system being different, such data doesn't say anything to me. I only care about individual component power consumption to figure out cooling needs.

What we want here is irrelevant. You know it must be the whole system.

 

[AusWolf]

What we want here is irrelevant. You know it must be the whole system.

One can never be sure these days. Anyway, if it's CPU only, then it's insane, if it's total system power, then it's useless info.

 

[AnotherReader]

One can never be sure these days. Anyway, if it's CPU only, then it's insane, if it's total system power, then it's useless info.

It is almost certainly whole system power draw which makes comparisons trickier. Note that the figures below are with a 125 W power limit.

 

[AusWolf]

It is almost certainly whole system power draw which makes comparisons trickier. Note that the figures below are with a 125 W power limit.

View attachment 366876

It makes comparisons near impossible. What game, what resolution, what graphics card, rest of system specs, etc. There's too many unknowns, that's why I'm saying that total system power is useless info.

 

[tfp]

It's not useless if the graphics card is the same and it's just really the CPU/MB that is different and is Intel vs Intel. It shows at a system level there is savings.

The are all rumors and leaks. All it doing is hinting at what we will see in the coming days when real reviews are released. This debate is not going to conclude until we have real numbers from review sites.

 

[SL2]

It makes comparisons near impossible.

You're changing subject, and I don't know why. We're probably just supposed to just compare those two systems in that pic with each other. For the same reason you don't necessarily compare system power draw between different reviewers.

Your initial point was

Hang on... What's that 447 W over there?

I tried to explain that it must be system power, as in there's nothing strange about that power draw.

What game, what resolution, what graphics card, rest of system specs, etc.

TPU doesn't show resolution in power draw tests for CPU's either. It is however included in game efficiency, so I'd guess it's the same resolution in power draw.

Finally, this is LEAKED INFO. There's most likely footnotes about all the settings and specifications, but they're not posted here.

There's too many unknowns, that's why I'm saying that total system power is useless info.

Maybe, but like I said, this is not CPU power draw only.

Again, what we would want from a portion of a leaked presentation under NDA at the time is irrelevant. It wasn't for the public eye to begin with, and it's not complete.

My ugly calculation says it runs at 65 W average in games, right between 12600K and 12700K. We'll see in two weeks how close it is.


I guess we'll get more info in a few hours.

Intel Arrow Lake Leak Confirms October 10 Announcement Date For Core Ultra 200 CPUs

Intel's Arrow Lake Core Ultra 200 series CPUs have appeared in a number of leaks now, and some benchmarks of the new chips have even started appearing on sites like Geekbench, however, there are still some questions surrounding the launch date of the new CPUs. Previous leaks indicated that the...

www.techpowerup.com

________________________________________________________________________________________________________________


Edit: I guess we didn't have to wait that long.

Intel Core Ultra 200S Arrow Lake-S desktop processors announced: Lion Cove, Skymont, Xe-LPG, NPU and LGA-1851 - VideoCardz.com

Arrow Lake-S is here, up to 24 cores, 125W and new socket Intel officially introduces the new series for desktop PCs. The Core Ultra 200 series is an update for desktop systems, focusing on power efficiency. It features newer Performance and Efficient core architectures, new graphics, and...

videocardz.com

Hotspot not in the center, of course.

 

[AusWolf]

You're changing subject, and I don't know why. We're probably just supposed to just compare those two systems in that pic with each other. For the same reason you don't necessarily compare system power draw between different reviewers.

I'm not comparing anything against a 14900K because I don't have one. And since I already look at the 14900K as a power hog, I'd rather not compare anything against it in terms of power.

It's like saying that the new Ford has better fuel economy than the F-150 Raptor. But I don't have a Raptor, so why would I care? I have a Fiesta, so how does it compare to that?

I tried to explain that it must be system power, as in there's nothing strange about that power draw.

TPU doesn't show resolution in power draw tests for CPU's either. It is however included in game efficiency, so I'd guess it's the same resolution in power draw.

TPU shows max power consumption. Here, we don't even know if it's that or something else. It's just a random number thrown onto a presentation slide.

Finally, this is LEAKED INFO. There's most likely footnotes about all the settings and specifications, but they're not posted here.

Maybe, but like I said, this is not CPU power draw only.

Again, what we would want from a portion of a leaked presentation under NDA at the time is irrelevant. It wasn't for the public eye to begin with, and it's not complete.

Let's settle with that. Like all leaked info on any product from any company, this is just as much useless.

 

[SL2]

I'm not comparing anything against a 14900K because I don't have one. And since I already look at the 14900K as a power hog, I'd rather not compare anything against it in terms of power.

It's just as strange as AMD comparing a new 9950X with a 7950X. As they're presenteing a successor, it's the most natural thing to do. Since the last one is famous for actually being a power hog, it makes sense to show that the new one draws less.

This isn't any new "Ford"/Intel, it's their newest desktop CPU compared with the previous desktop CPU.

TPU shows max power consumption. Here, we don't even know if it's that or something else. It's just a random number thrown onto a presentation slide.

We weren't talking about max power consumption to begin with, it was average like I said.

You asked for resolution in power draw tests, and it's not present in TPU reviews either. But again, it's not hard to figure out what it is.

Let's settle with that.

You could have led with that. Cheers

 

[mkppo]

Aren't the reviews supposed to be out today?

 

[SL2]

Aren't the reviews supposed to be out today?

That's on october 24 I think

 

[Bobaganoosh]

That's on october 24 I think

It is such a slimy thing to do to review embargo until launch day. They announced them, let the reviews go.

 

[efikkan]

With HT removal it will be slower in MT workload. ST performance doesn't matter much these days.

The only advantages are native support for DDR5 6400MT memory, raised from 5600MT and I think dedicated PCIe 5.0 NVme port, if someone plans to upgrade now.

This is the common misconception with "single threaded performance". What it actually means is performance per core/thread, and this is in fact the other multiplier for the theoretical limit of multithreaded performance, so "single threaded" performance matters whether you have 2 or 32 cores. Only in applications/workloads with large batch jobs will more slower cores makes up for slower performance per core, and in more user interactive applications or mixed workloads*, having fast enough is the key factor for a good user experience, and this will continue for the foreseeable future.

But as you probably know, over time performance per core has become more and more unpredictable. An i9-14900KS (stock) wouldn't run at 6.2 GHz sustained in all kinds of workloads, and the more load there is on other cores the lower it will boost. This has become so unpredictable that the rated clock speeds are almost useless at this point. It started to get bad with Coffee Lake, but with Alder/Raptor Lake the variance of single core performance has gotten pretty extreme. (And I'm talking about desktop K-SKUs, low TDP SKUs and laptops are even worse) How noticeable this is to the end user depends on the workload and the user. So if Arrow Lake manages to reduce this variance while not advancing the peak performance much further, I would still consider it an improvement. If anything, with current products this might be an overlooked advantage for AMD.

*) By "mixed workloads" I mean typical "prosumer" use running multiple applications at once, typically not "high load" most of the time. The vast majority of benchmarks run one at the time, and only benchmarks peak performance.

I don't think any sane person criticized Zen 5 solely for its unimpressive performance uplift over Zen 4. The criticisms, rather, were:
- AMD wildly over-promised on performance
- pricing

When did AMD over-promise on performance for Zen 5? (I must have missed it)

The big deal-breaker for "prosumers" with Zen 5 is the chipset/motherboard offerings. With too many lanes tied up with USB4, lanes shared between some M.2s and GPU, and only 4 lanes to the chipet, combined with "premium" motherboards which doesn't even maximize the platform IO features, it becomes almost laughable. While offering very affordable and efficient 12 and even 16 cores, with beautiful AVX-512 support, the platform looks very appealing until you start looking at long-term usability. For those who don't replace their machine every 2-3 years, memory bandwidth and PCIe lanes quickly becomes the bottleneck. If they can't offer lanes for a GPU + 3-4 SSDs + 10G NIC + 6-8 SATA devices without significant downgrades in performance, it's really a fail. Intel (mainly W680 motherboards) seem to have an edge here, but even here flexibility for expansion should be the primary focus for picking a motherboard, and it's not easy.

But I have hopes for Threadripper though, to finally unleash the Zen 5 cores.

 

[persondb]

Still, 400W on avg for "gaming" is kind of absurd.

X3D will need not even 1/3rd of that to even produce more or higher FPS.

400W is system load. Most of it is probably being used by the GPU.

I would suspect they used System load because they did not want to admit that the 14900K and co use 150W or more in gaming.

 

[AnotherReader]

400W is system load. Most of it is probably being used by the GPU.

I would suspect they used System load because they did not want to admit that the 14900K and co use 150W or more in gaming.

Unfortunately for the 14900k, system load isn't 400 W. It can be as high as 609 W even after enforcing a 125 W power limit. At default settings, Techspot measured another 100 W over that for over 700 W system power draw.

One big advantage of the 125W 'performance' profile is power consumption. Although still significantly higher than the 7800X3D, it's a noteworthy improvement. For example, in Starfield, the 14900K delivered similar fps performance using either profile, but the 'performance' profile reduced total system usage by a massive 15%, shaving 100 watts off.