Intel Core Ultra Arrow Lake Preview

[Daven]

It's marketing man they come up with all kinds of BS we will have reviews soon.

The slides for Zen5 were really bad hopefully these are at least better than that lol.

But I don't even know what it means. Three equal height bars with different numbers above them compared to an ON PAR baseline. WTF!

 

[Chrispy_]

I'm stoked for this; it's been a long time since Intel was an appealing option to me - and a focus on efficiency is good. $309 is a lot to ask for 6 P-cores with a mandatory new motherboard on top of that, but the CPUs always find a street price that's relevant to their performance and positioning in the market, and nobody needs a flagship z-series board with an i5.

Cherry-picked results in marketing slides are so rarely accurate indicators of the real performance, as ever I'll be looking forward to independent reviews later this month. I really hope they don't suck because both AMD and Intel do better when there's good competition. Whether you're an AMD fanboy, and Intel fanboy, or someone who treats both companies with a healthy dose of cynicism...

 

[Macro Device]

rip gamers

Excellent gaming performance is available for way below 300 USD already if we talk CPUs. Heck, even a 100ish dollar 12400F grinds through all recent releases with a single digit worth of exceptions at over 90 FPS. Only the most CPU intensive and action packed games might use a tad more but we're talking extremely niche cases. 99+ % gamers, if they get a 245K (let alone 285K), will still be heavily GPU/RAM limited in 99+ % their beloved games. It's really hard to mind having basically same stuff but for much less wattage. Don't see anything wrong in this.

What is terrible is the GPU situation.

I also heavily despise that non-square socket shape. If you want it to run stable you MUST make it square. Multi billion mega corporations... they never learn.

 

[csendesmark]

Wow, rip gamers..... 2024 has been rough for gamers.

Hopefully the 9800X3D can save the day.

What do you mean?
The Ryzen 9900X and 9950X are quite good for gaming too

So they finally admitting that the 14900K is s#@&?

 

[AnotherReader]

I will wait for an actual review before I spout off too much.


The number of admissions by Intel is telling.

Their nodes were fluff, TSMC is still more advanced and all the BS talk was just that, BS.

Power draw...... simply amazing what a better node will get you. No doubt Intel made some awesome discovering in power domain management and it's going to help a lot with this series.

Chiplet based design. The reality is they copied AMD, they used a tile like AMD did to mount GPUs with a silicon interposer, they can call it whatever but like the Barenaked Ladies sang, it's all been done before....

The most exciting parts are the removal of HT and the Ecores on the ring bus. I hope they put their big boy pants in and approached it with a "how can this F us" security mindset, I say this with the best of intentions as I think they have a winner with this design if it doesn't need 5 patches to slow it down and validate security on code.... They may have just placed themselves a generation ahead of AMD on architecture and planning. AMDs only move to counter is a little.BIG core where a LOT of resources can be turned off on a big core until needed or a set of stripped down cores with commonly used hardware. The proof is in the pudding that HT/SMT is is dying a rapid death and the silicon space it used is better suited to cache and or branch prediction.


The best part is the pricing, it tells almost the complete story.

Intel's server parts based on these P cores still use SMT and are expected to continue using it.

 

[Steevo]

But I don't even know what it means. Three equal height bars with different numbers above them compared to an ON PAR baseline. WTF!

Normalized. They set the base point 0 to be the performance of AMDs chip then rate their performance as a deviation of they set point. Its better for them to manipulate numbers that don't mean anything off of.

Like the new Civ game, 13% faster...... at 200FPS, where it doesn't matter.

Intel's server parts based on these P cores still use SMT and are expected to continue using it.

Maybe if Intel were clever and the mask were the same they would let some setting slip to enable HT on these chips for the enthusiasts who like to tweak....

 

[N/A]

But I don't even know what it means. Three equal height bars with different numbers above them compared to an ON PAR baseline. WTF!

It means that you have the ability to choose between 125, 175 and 250 watts in the BIOS. And the performance per watt is on par for all the modes. Meaning 125 watts 125 FPS, 175 watts 175 FPS and 250 watts 250 FPS all performed on the one and only 285K /no AMD here/, which doesn't make sense because that's not how power works. Usually doubling the power means 25% more performance.

 

[phints]

Small performance increase and large efficiency increase. Basically win-win and efficiency is by far what Intel needed to improve the most.

Excited for 265K vs. 9800X3D, but this looks like a good turn-around.

 

[Steevo]

Small performance increase and large efficiency increase. Basically win-win and efficiency is by far what Intel needed to improve the most.

Excited for 265K vs. 9800X3D, but this looks like a good turn-around.

It's not really beating the 7800X3D where it counts.

 

[persondb]

But I don't even know what it means. Three equal height bars with different numbers above them compared to an ON PAR baseline. WTF!

It just means that no matter the PL1 you choose, the gaming performance is the same.

Though I have no idea if it's the same as Raptor Lake at 250W or at the same PL1.

 

[Squared]

Intel more energy efficient than AMD? Honestly hard to believe.

It'll be a little disappointing for Intel if this generation is not more power-efficient. It uses TSMC N3B for CPU, N6 for I/O, and N5 for graphics. Ryzen 9000 uses TSMC N4B for CPU and N6 for everything else.

Also Intel's tiles on a silicon interposer with all CPU cores on a single tile should be more efficient than AMD's solution, and this should be more apparent near idle.

 

[Chrispy_]

Intel's server parts based on these P cores still use SMT and are expected to continue using it.

I was under the impression that SMT is too much of a security risk for client devices exposed to the web. Servers at least (should) sit behind many layers of security and default-closed firewall rules that allow only explicit traffic to/from them.

Personally I do not really care if a CPU has SMT or not. All that matters is performance and if Intel can get that without resorting to SMT, then great! Intel's HT implementation was always pretty weak compared to AMDs, whilst still offering all the downsides of SMT when the Application/OS/Scheduler gets it wrong. If the performance is there without the potential downsides of SMT then everyone wins.

 

[Sarajiel]

Three equal height bars with different numbers above them compared to an ON PAR baseline. WTF!

Read the footnotes: "on par" means the results are within 3% of each other.

Like the new Civ game, 13% faster...... at 200FPS, where it doesn't matter.

It's the old Civ VI, and they probably just used the turn time benchmark, which basically only scales with 1T performance.

 

[evernessince]

Similar gaming performance to current top performers, with an apparent efficiency and MT advantage over anything on the consumer desktop market is nice. Having great IO, along with an NPU is a perk too. I am a bit puzzled to see gen 1 Arc though.

I wouldn't be surprised to see the U5/U7 models have more performance in gaming than their i5/i7 predecessors too, despite 14900K-285K not having much of an uplift. Realistically most people aren't using $700 CPUs and $1600 GPUs in their gaming rigs, so advancing the mainstream is great, considering the competing X3D chips essentially start at the high end unless you're willing to buy an AM4 CPU in 2024.

I'm wouldn't be surprised if gaming performance is similar to the non-X3D parts but I'm not so sure their efficiency averages out to beat Zen when we see actual benchmarks. We'll just have to see. At the very least, one can hope there is in fact a large improvement to efficiency.

The problem I think many people will have is that Intel is comparing against the 9950X, of which was an incredibly underwhelming launch that didn't move the needle at all. Being equal to an unappealing option isn't really an enviable position. If people wanted that level of performance with that level of efficiency, they could have had it years ago with Zen 4. Waiting this long just to get it and having to buy a new motherboard is a worse proposition than AM5 was.

 

[ZoneDymo]

im not even impressed by Intel's OWN FREAKING CLAIMS....

also they are once again pulling shenanigans with the graphs, why is this even legal here is just one graph I adjusted to be more real with the claims made:

 

[AnotherReader]

I was under the impression that SMT is too much of a security risk for client devices exposed to the web. Servers at least (should) sit behind many layers of security and default-closed firewall rules that allow only explicit traffic to/from them.

Personally I do not really care if a CPU has SMT or not. All that matters is performance and if Intel can get that without resorting to SMT, then great! Intel's HT implementation was always pretty weak compared to AMDs, whilst still offering all the downsides of SMT when the Application/OS/Scheduler gets it wrong. If the performance is there without the potential downsides of SMT then everyone wins.

I believe it's an even greater risk for cloud providers though it can be mitigated by ensuring that different VMs don't share the same physical core.

 

[ThomasK]

Interesting how tables have turned. When Ryzen launched, it had better MT performance and worse gaming performance.

Also, no word on platform longetivy. Guess Arrow Lake's customers will be throwing their MB away after 1 or best case 2 generations.

Next.

 

[Darmok N Jalad]

Losing SMT means leaning even more on scheduling for MT tasks. I’m still not convinced that the hardware and Windows are really ready for this. It might be some great engineering, but the user experience is what really matters. I’m a broken record on this, but MT benches load all the cores and gets a magical score not achieved in reality. I run tasks daily that should load all the cores on my mobile Adler Lake, but it doesn’t play out like that at all. The result is a crawl. Windows 11 isn’t working any miracles either.

 

[Kapone33]

Please include Space Marine 2 and City Skylines 2 in your Gaming benches as those Games are super CPU intensive.

 

[Mack4285]

Given the very different architectures and packaging strategies between Ryzen (Zen 5) and Core Ultra (Arrow Lake-S), it's looking like they will be remarkably similar in power usage and performance.

What about idle power consumption, similar there as well? Because AMD with chiplet design can draw 30W doing nothing... I hope that Arrow Lake has low idle power consumption.

 

[kondamin]

Please include Space Marine 2 and City Skylines 2 in your Gaming benches as those Games are super CPU intensive.

If we are doing requests, please add a run where you are running one of the big anti-virus software packages while doing the benchmarks and do this for every subsequent benchmark runs.

 

[Eternit]

Why they have models without GPU but no models without useless npu.

 

[marios15]

On paper looks good, but this is going to be Intel's first "chiplet" IMC....although the interposer will definitely help compared to AMD's approach I expect similar limitations.
Then you also have the mixed cores with missing instruction sets.

This will probably be great on Linux, but Windows just won't give it any chance.

 

[LittleBro]

But I don't even know what it means. Three equal height bars with different numbers above them compared to an ON PAR baseline. WTF!

There is not any AMD CPU mentioned in that slide, so one can only guess that it means same gaming performance at different power draw. Which is hilarious. Winter is coming, so having 250W at same performance as with 125W while keeping yourself warm is indeed handy.

 

[Vayra86]

There are multiple games I'm already cpu limited in with X3D 7000 processors. I agree though most people probably are not.

Which? And what are you targeting?

But I don't even know what it means. Three equal height bars with different numbers above them compared to an ON PAR baseline. WTF!

I think it tries to say that this 285K will run at 125W doing the same performance as it does on 250W.

Which ultimately makes us question Intel's power management once more. Yeah. Great marketing lol, why could the CPU even run at 250W for 'on par' performance to half the draw? Extra heating?

Please include Space Marine 2 and City Skylines 2 in your Gaming benches as those Games are super CPU intensive.

Can you even get a good canned bench for SM2 and CS2? I mean... its all about assets on screen and in game logic. And isn't CS2 still a big pile of doodoo?

It's not really beating the 7800X3D where it counts.

Yeah for all the gaming efficiency they gained, they're still comparing it to a non X3D in their own slides, nuff said, this gen isn't going anywhere unless you're an Intel-only buyer.