DenverThere must be some inconsistency in the methodology, two identical CPU models (7840u and 7840HS) at the silicon level, are presenting different IPC.

I'd say < 3% difference is well within margin of error.

ImoutoI'd say < 3% difference is well within margin of error.

The 7840U is likely to be using LPDDR5 which has a higher latency than DDR5. I expect the 7840HS to be using DDR5. This would be enough to explain the higher IPC for the 7840HS.

Well LPDDR5(x) probably has worse latency than regular DDR5 or SODIMM modules used in laptops. At least I saw this mentioned somewhere else.

But there's not enough results out there to draw a serious conclusion that IPC has definitely gone down!

TigerfoxAMD did the same with Intel back in the times of Pentium D, because those were just two prescott cores slapped together in one DIE and unable to cummunicate with each other. It was the same for Core 2 Quad AFAIR.

Let's not play the blaming game here. My argument was about something recent, specially adding the "snake oil" PR joke.
You are just remorseful for something that happened fifteen years ago.

It seems new bios revisions have improved things a bit from release/reviewers.

wccftech.com/intel-core-ultra-meteor-lake-cpus-already-seeing-big-performance-efficiency-improvements-with-updated-bios/

The performance shown in the table TechPowerUp provides here shows performance divided by boost clock speed, which is 5.0 GHz for the 13700H (a 45W processor *1) and 4.8 GHz for the 155H (a 28W processor *2). The processors do have similar cache and single-core tests can usually be run well inside the processor's power limits, so in those respects this appears to be a good test.

But turbo clock speed can be all over the place, as this graph of the 155H frequency in a benchmark from Phoronix shows. *3


The 13700H may be able to maintain its rated boost speed more consistently than the 155H, since it's a higher-power processor. Or there could be a big disparity in the cooling the tested laptops provide. Ultimately I don't think this is a good testing methodology for determining IPC. Perhaps a better method would be to monitor the boost frequency while the benchmark runs, determine the average frequency while it was running, and then divide the score by that average. However, the frequency changes so quickly that it would be hard to get an accurate average. And even if that worked perfectly, it wouldn't address disparities in memory. Testing IPC accurately is very difficult especially with laptop processors. Since several variables can't be controlled, multiple different benchmarks need to be run to get a handle on how each variable affects the score, and only then can you begin to take an educated guess about IPC. Any test that fails to find a way to address every variable (other than IPC) is more conjecture than science.

Attribution

*1 Intel's specs for the 13700H:
www.intel.com/content/www/us/en/products/sku/232128/intel-core-i713700h-processor-24m-cache-up-to-5-00-ghz/specifications.html

*2 Intel's specs for the 155H:
ark.intel.com/content/www/us/en/ark/products/236847/intel-core-ultra-7-processor-155h-24m-cache-up-to-4-80-ghz.html

*3 "Meteor Lake EPP Tuning For Greater Performance Or Power Efficiency With Intel Ultra Core 7" by Phoronix:
www.phoronix.com/review/intel-meteorlake-epp/2

What the fuck, lmao.
My day has been stressful so far, but man does this cheer me up, not because I am happy that a company is releasing a bad product, but because it is just plain funny.

For everyone trying to explain away the results, all info and results thus far from Intel and third parties basically imply a lower IPC. This test just confirms it. Intel added so much silicon outside the P-cores that something had to give. We now know that Intel had to compromise on the P-cores to beef up the e-cores, GPU, IO and NPU.

Meteor lake and IMHO Arrow Lake will not be about absolute CPU performance but about advanced packaging, advanced fab nodes and all around performance. If you want max CPU performance going forward, only Zen 5 looks promising on that front.

No one is denying that single-core performance is lower. We're denying that this test proves it's because of IPC and not just clock speed.

DavenFor everyone trying to explain away the results, all info and results thus far from Intel and third parties basically imply a lower IPC. This test just confirms it. Intel added so much silicon outside the P-cores that something had to give. We now know that Intel had to compromise on the P-cores to beef up the e-cores, GPU, IO and NPU.

Meteor lake and IMHO Arrow Lake will not be about absolute CPU performance but about advanced packaging, advanced fab nodes and all around performance. If you want max CPU performance going forward, only Zen 5 looks promising on that front.

Since Arrow Lake will be desktop, I don't see how they can release a new DT CPU that is slower than last 13/14th gen CPU.

I think its more this is focused on low-wattage and multimedia-type workloads.

A big dollop of speculation going on here, mebbe's we should just wait till the actual desktop variety comes along.

Ed_1Since Arrow Lake will be desktop, I don't see how they can release a new DT CPU that is slower than last 13/14th gen CPU.

I think its more this is focused on low-wattage and multimedia-type workloads.

Its Intel. That’s how.

I don't think Meteor Lake is coming to desktop for a reason. This looks like a prototype release to make a little money and prepare for the real chip launch. Then again, Intel's done this before and other idiotic things, so... not really a shock.

I'm hoping they get their act together and bring a meaningful competitor to both the DDR5-based Ryzen and handheld APU lines, but looks like this ain't that gen.

DavenIts Intel. That’s how.

I get that the hate boner for Intel is strong, but that would be a really odd move in the current market. Intel would rather push the power limits beyond reason than suffering such a debilitating loss. At this point you might as well just die shrink RPL tbh

KEK DOUBLE U

dyonoctisI get that the hate boner for Intel is strong, but that would be a really odd move in the current market. Intel would rather push the power limits beyond reason than suffering such a debilitating loss. At this point you might as well just die shrink RPL tbh

Hi,
No hate just facts
Intel knows people will buy anything they put out
That's why they refresh over and over.

Pretty meaningless comparison when you consider that ML is a 28W processor, while the 13700H is a 45W processor.

This means your frequencies will be all over the place.

Most of the actual hands-on reviews of ML laptops show some decent gains. Not enough for me to switch to a laptop again yet, but pretty good. And the AI stuff, after years of waiting, is finally being used effectively in some apps.

I think what people mean to say is, they don't have a use for a 25W laptop. So they compare ML to 45W large / gaming laptops. At least realize that you're comparing apples to oranges here, because the ML line doesn't have an 45W part.

Within its league it's pretty compelling in the x86 space.

RandallFlaggPretty meaningless comparison when you consider that ML is a 28W processor, while the 13700H is a 45W processor.

That review you linked shows the 165H (5.0 GHz boost) barely losing to the 1365U (5.2 GHz boost) in single-threaded benchmarks, which by the methodology being discussed here implies Meteor Lake has slightly higher IPC, since it loses by 1.6% but has a 3.8% lower clock speed. But I would argue again here that this discrepancy is best explained by differences in sustaining boost clock speed.

btarunrDavid Huang in his blog testing contemporary mobile processors found out, through a series of single-threaded benchmarks.

What? This was never kept a secret. Even in Intels own slides they only compare it to 12th gen. 13th gen was wildly inefficient with power. Everything faster if you dump watts into it and turn the clocks up.

Phoronix covers 8th to 12th gen in regards to performance vs power.

www.phoronix.com/review/intel-whiskeylake-meteorlake

john_Chipsets? Apple's chips are SOCs, they don't really need chipsets. So, even if this was a serious idea and Apple wasn't having problems with it, it wouldn't work anyway.

That's exactly why it might work.

IPC is instructions per clock cycle. It's a blanket term that can refer to any performance modifier other than frequency, but most often is used when discussing microarchitecture improvements. Intel claims that Redwood Cove, the P core microarchitecture in Meteor Lake, is pretty similar to Raptor Cove, the P core microarchitecture in Raptor Lake. So they should have pretty much the same IPC. Most reviews do show a small performance regression in single-threaded workloads—smaller than the IPC decrease David Huang is claiming—which could be explained by the fact that the 155H and 165H in reviews have lower boost clock speeds that the Raptor Lake processors they replace.

SquaredIPC is instructions per clock cycle. It's a blanket term that can refer to any performance modifier other than frequency, but most often is used when discussing microarchitecture improvements. Intel claims that Redwood Cove, the P core microarchitecture in Meteor Lake, is pretty similar to Raptor Cove, the P core microarchitecture in Raptor Lake. So they should have pretty much the same IPC. Most reviews do show a small performance regression in single-threaded workloads—smaller than the IPC decrease David Huang is claiming—which could be explained by the fact that the 155H and 165H in reviews have lower boost clock speeds that the Raptor Lake processors they replace.

He also used SPECint 2017 in his testing ... and this exact thing does not exist, apparently. So it must have been either SPECspeed 2017 Integer or SPECrate 2017 Integer. Both are suites of tests that cover some real-world workloads, designed to put most stress on integer execution pipeline, but it's unclear how much the results depend on branch prediction, the speed of the bus, last level cache, memory etc. Sure, those are valid methods of measuring "IPC", but so are many others (which may depend more, or less, on components other than the cores).

ncrsWell, Intel only claimed IPC gains for E-cores in Meteor Lake while P-cores were supposed to increase efficiency ;)

Go back about 2 years and Meteor Lake was being hyped with double digit P core IPC uplifts as well. All through 2022 MLisD was also saying such things and was saying Arrow Lake will further improve IPC by 20%+ over that. I'm already tempering expectations for Arrow Lake being a leap other than again for iGPU if it gets Alchemist+ cores.

SquaredThat review you linked shows the 165H (5.0 GHz boost) barely losing to the 1365U (5.2 GHz boost) in single-threaded benchmarks, which by the methodology being discussed here implies Meteor Lake has slightly higher IPC, since it loses by 1.6% but has a 3.8% lower clock speed. But I would argue again here that this discrepancy is best explained by differences in sustaining boost clock speed.

It can also be from the tiled architecture, as well as the extra cores vs the 1365U. 1365U is a 15W part.

I read up on the tiled arch some time ago, and in general its potential is actually not as efficient nor as powerful as monolithic. It is more for what I call 'industrial engineering', specifically it is about increasing yield on increasingly complex dies along with the ability to do some mix-and-match.

But I basically agree these look like Raptor Lake cores adapted to tiled architecture on ML. The 1-2% losses in single thread are likely due to that tiled architecture when you consider the 1365U is 15W. We're talking about rounding-error percentages here though.

Regardless of all that, ML is currently the best general purpose compute sub-45W x86 chip around when you consider that +60-70% multithread advantage vs comparable RPL SKUs. So within its sphere, it's no laggard.

Minus InfinityGo back about 2 years and Meteor Lake was being hyped with double digit P core IPC uplifts as well. All through 2022 MLisD was also saying such things and was saying Arrow Lake will further improve IPC by 20%+ over that. I'm already tempering expectations for Arrow Lake being a leap other than again for iGPU if it gets Alchemist+ cores.

Originally there wasn't supposed to be a Raptor Lake, nor a RPL refresh, nor a Rocket Lake for that matter. They were going to go AL -> Meteor Lake.

In that context, it would be double-digit IPC improvement.

RandallFlaggSo within its sphere, it's no laggard.

Lets not forget the addition to ARC instead or Iris XE. Compared to 13th gen and below (all the way to HD).

www.phoronix.com/review/intel-gen9-mtl-graphics