I hate synthetic benchmarks and hype-beast clickbait.

You really get a sense that the tech hobbyist scene is dying generation-ally. The commentary on products is niche-ing down to cater to a smaller and smaller group of “tech dudes”.

More people than ever are USING tech, but fewer really seem entertained BY the tech. I digress.

I’ve read through three different hot takes now about Qualcomm’s new X Elite chip, comparing a couple of Geekbench scores against other chips from AMD, Intel, and Apple. They look at a couple number scores, and then draw a conclusion before any of these components have actually shipped in products to consumers. It’s kinda the worst flavor of arm chair tech reviewing.

(Get it? ARM chair? It’s a silicon pun.)

The conversation leading up to a product shipping can be a lot of fun as we speculate on how different core configurations, graphics processing solutions, and power envelopes might impact the user experience.

Just slapping a Geekbench bar graph on a blog post is NOT particularly entertaining. But here we are.

I was YEARS too early in pitching the idea of ARM based laptops while I was hosting and producing at Newegg. I just like to remind folks that I’ve been a LONG time advocate for using ARM chips in consumer PCs.

Our phones have been overkill powerful for years now, I thought we’d see those kinds of performance-per-watt improvements on laptops soon. I was wrong, until recently.

Thankfully Apple was able to rip that band-aid off with the M1 processor. We weren’t excited about MacBooks with ARM chips because they were more powerful than the previous generation of Intel MacBooks. We were excited because they delivered better performance while using a lot less power.

The Windows side of this conversation has struggled. I believe Windows on ARM is near a point today where PRECIOUS few individuals (mostly the techier folks reluctant to try alternatives to the software they currently use) will run into any daily issues using an ARM laptop or tablet.

Often the issue for current Windows on ARM solutions isn’t performance per watt, but performance per dollar. A Surface Pro 9 SQ3 is EXPENSIVE, but my Robo and Kala is hanging in roughly with older Core i3’s, with significantly better battery life, and at a reasonable price for an OLED screened convertible tablet.

But how much power will these Qualcomm laptops use?

Of all the articles I’ve skimmed so far, I will point out that Windows Central did a solid job pointing out the difference between CPU power consumption and system power consumption. It’s a maddening aspect of the conversation, as Apple REALLY likes to highlight the CPU power consumption. It’s almost impossible to get a proper accounting of the total system power consumption.

Conversely, Qualcomm is presenting data based on total system consumption of two test configurations of laptops, but we can’t get a proper accounting of CPU power draw.

We can’t really compare oranges to oranges with this kind of data.

The folks at Notebook Check have detailed previous generations of MacBooks, and I can skim some info from their MacBook Pro review (using the M2 Pro chip), but since we’re only looking at a Geekbench score, that’s not very good data to compare.

I’m sure there are other reviewers on YouTube that have done similar work, and detailed solid comparisons, but searching for ANYTHING Apple mostly just delivers a constant stream of YouTubers making porn faces in their thumbnails. I digress again…

Qualcomm estimates the total system power consumption of their top model around 80W, and they test a configuration of a low power ultra-book style rated for 23W.

We often see TDP listings for the M2 chips floating between 25-35W, we rarely see estimates on the total system. The screen, RAM, storage, fans, GPU, radios, they all draw power too.

Running over to Notebook Check, the new review for the MacBook Pro base model with an M3 (non pro), the authors noted a peak power draw of 67W which leveled off to around 60W under heavy load. The MacBook Pro with an M2 Pro showed a peak power draw of 100W with a sustained heavy use floating just above 70W.

We can clearly see that Qualcomm is trying to battle Apple directly, and they’re putting out parts for next year that should float between the options Apple sells with M2 Pro and M3 Pro configurations.

That’s where the Geekbench scores are a little interesting.

M2 Pro MacBook Pro GB6:
Single-core 2,663
Multi-core 14,568

M3 MacBook Pro GB6:
Single-core 3,099
Multi-core 11,966

80W Snapdragon X Elite Demo GB6:
Single-core 2,940
Multi-core 15,100

23W Snapdragon X Elite Demo GB6:
Single-core 2,741
Multi-core 13,021

In practical situations, I find there’s VERY little difference in real-world performance when we consider up to 200 point differences in Geekbench. It’s crazy, but when you really run a game, a large spreadsheet, or a video project, 200 points can often be margin-of-error territory looking at total GB scores.

Also, we’re kind of taking Qualcomm at their word, with the best possible interpretations of their controlled tests, on demo builds they made in house. We’re not sure what real-world system configurations will look like yet.

For those considerations though, we might be looking at the same kind of generational leap in Windows laptops that we saw when Apple introduced the M1. The 80W configuration slots in nicely between the M2 Pro and the range of scores we’re starting to see on the M3 Pro.

But What About Even LESS Power?

The more exciting device (in my not so humble opinion) though is Qualcomm’s claims for the 23W system. So long as that style of build really can stay under 30W, we’d be looking at REALLY close performance to the Apple M3 MacBook, but at half the power draw.

That would put power draw slightly below my Robo and Kala, but with SIGNIFICANTLY higher GB6 scores.

Intel Core i7 13705H GB6:
Single-core 2,613
Multi-core 14,547

However, putting that chip into a Lenovo, Notebook Check saw idle power consumption floating around 20W with low screen brightness, and under load, the laptop peaked at 150W, which exceeded the capabilities of the 140W included charger. It’s shocking to consider that Qualcomm’s “low end” option might get us close to the performance of Intel’s i7, but at near the same total power draw as the Intel system at idle.

All that’s left for us to do is wait.

Qualcomm is kicking off an exciting conversation in the Windows laptop space, but there’s still a TON we don’t know. Qualcomm can build a test system at 23W, but can HP or Dell do the same, and sell that kind of system at scale, for a price that will attract consumers?

We see Davinci Resolve is moving forward with plans to natively support Windows on ARM, but will Adobe follow? Will we see better compatibility and emulation of all those OLD precious legacy programs techies are so concerned about?

We have to believe there will be a response from AMD and Intel, but hopefully we’ll also see some pressure from Mediatek on lower power Chromebooks.

We’re just getting started. 2024 looks like it’s gonna be a LOT of fun for us PC nerds…

Supporting SomeGadgetGuy!

This post was published early for people supporting my Patreon! Sharing and commenting are always appreciated, but continuing these conversations would not be possible without the generosity of this community. If you would like to receive early access to videos and articles, and see my videos in 4K, please consider joining the rad group of nerds at https://Patreon.com/SomeGadgetGuy