Qualcomm pits its Snapdragon 8cx against Intel's i5 8250U in benchmarks

[midian182]

Recap: Windows on Arm might offer excellent battery life, but there have long been questions over the machines’ performance. Qualcomm promised this would change when it announced the Snapdragon 8cx last December. Now, the firm has shown off some benchmarks for the platform, comparing it to Intel’s Core-i5 8250U.

The largest processor Qualcomm has ever made, the Snapdragon 8cx features eight custom cores in its Kryo 495 CPU. These four high-performance Cortex-A76 cores and four low-power Cortex-A55 cores are all based on the 7nm process and are clocked higher than the Snapdragon 850. In addition to offering twice the performance of that previous chip, the 8cx brings 60 percent more battery life.

Qualcomm ran the 7W Snapdragon 8cx against Intel’s 15W Core-i5 8250U mobile processor using 3D Mark and PCMark. In the application benchmark, the 8cx was ahead in more tests than Intel’s chip, while it beat the i5 8250U soundly in a graphics benchmark and battery life test. It’s worth noting, though, that a 2K screen was used for Intel’s model, while Qualcomm’s design uses an FHD display, helping put the Adreno 680 ahead of the UHD Graphics 620.

Qualcomm also shared a series of videos, which you can see at the bottom of the page, showing how its Snapdragon 8cx outperforms competitors in a series of real-world tasks, such as running programs, multitasking, and browsing the web.

In addition to the benchmarks, Qualcomm announced it is partnering with Lenovo on the first 5G-enabled always-connected laptop, currently known by the codename Project Limitless. It features a second-generation X55 5G modem alongside Cat 22 LTE.

The first devices to feature the Snapdragon 8cx are expected near the end of the year, while the 5G laptop is set to arrive in early 2020.

Permalink to story.

https://www.techspot.com/news/80254-qualcomm-pits-snapdragon-8cx-against-intel-i5-8250u.html

 

[SirDigby]

Vendor produced benchmarks are always biased but using a 2K screen for Intel and FHD for Qualcomm is really on the nose disingenuous. For shame Qualcomm.

 

[VitalyT]

Very impressive results, provided they are not marketing BS. I will believe it when I see benchmarks from independent testers.

 

[TheDevopsGuy]

Vendor produced benchmarks are always biased but using a 2K screen for Intel and FHD for Qualcomm is really on the nose disingenuous. For shame Qualcomm.

Its a pity they're testing two systems on a different OS version 1809 vs 1903 and the display is also biased.

When you're testing two systems all should be on equal terms. Basically all the results you just witnessed are irrelevant for the scope of any analysis.

 

[netman]

Vendor produced benchmarks are always biased but using a 2K screen for Intel and FHD for Qualcomm is really on the nose disingenuous. For shame Qualcomm.

UL aka Futuremark performs the PCMark10 test, not Qualcomm...

 

[TheDevopsGuy]

UL aka Futuremark performs the PCMark10 test, not Qualcomm...

Yes, they're using third party applications to benchmark but apparently these are Qualcomm produced results and presented by them.

 

[yeeeeman]

Call me when you have full compatibility with all x86 and x64 apps.

 

[JakeAndrew77]

UL aka Futuremark performs the PCMark10 test, not Qualcomm...

Yes, they're using third party applications to benchmark but apparently these are Qualcomm produced results and presented by them.

I just cannot believe that you went there... ANYONE using the same platforms under the same configurations will get similar results because they are VERIFIED by Futuremark. To Skew the numbers would be a death sentence to Futuremark.

 

[TheDevopsGuy]

I just cannot believe that you went there... ANYONE using the same platforms under the same configurations will get similar results because they are VERIFIED by Futuremark. To Skew the numbers would be a death sentence to Futuremark.

Theyre not falsifying results. They put a bias in the testing phase to get a better result is what im saying. They did legitimate results and have also been aproved.

This is like comparing a truck vs a car comparison its kinda funny actaully.

 

[Digitalzone]

The OS version should be same, however, I would not expect different results for x86-64 CPU on different Windows versions (unless for HW Intel bugs).
The different resolution is totally biased, shame.
But after all, it seems to me, that Windows on ARM works.

 

[fktech]

Competition, now lower the prices!

 

[Evernessince]

"It’s worth noting, though, that a 2K screen was used for Intel’s model, while Qualcomm’s design uses an FHD display, helping put the Adreno 680 ahead of the UHD Graphics 620."

Taints the results. That's twice the number of pixels. If they are willing to pull those shenanigans I wonder what else they are willing to do.

 

[B5S46M]

"It’s worth noting, though, that a 2K screen was used for Intel’s model, while Qualcomm’s design uses an FHD display, helping put the Adreno 680 ahead of the UHD Graphics 620."

Taints the results. That's twice the number of pixels. If they are willing to pull those shenanigans I wonder what else they are willing to do.

Umm, last time I checked FHD was 1920x1080 while 2K was 2048x1080. That's hardlty 2x the pixels. More like 12.5% more pixels for 2K.

 

[Squid Surprise]

Even if they are valid... what Qualcomm is saying is that their top-of-the-line CPU is equal to one of Intel's slowest mobile CPUs that is already 2 generations old (the 10x series is releasing within a few weeks I believe).

Yes, it's still impressive, but let me know when it performs on par with an i7... from the current generation

 

[Evernessince]

Umm, last time I checked FHD was 1920x1080 while 2K was 2048x1080. That's hardlty 2x the pixels. More like 12.5% more pixels for 2K.

That's Cinema 2K. For monitors 2K resolution is 2560 x 1440. https://www.tomshardware.com/reviews/what-is-fhd-full-hd,5741.html

I'm typing this on a 2k monitor as we speak. So yes, my original point very much stands.

 

[TempleOrion]

Why not just give the actual screen resolution for the avoidance of doubt?

 

[Squid Surprise]

Why not just give the actual screen resolution for the avoidance of doubt?

I suspect that was the reason actually... to keep people doubting

 

[actionjksn]

This isn't something I would probably buy but it would be fine for a lot of people. Some people don't need anything more than MS Office and need high battery life more than anything else, so it would be fine. I think it would be great for long flights, watching movies and stuff.

 

[hover389]

"It’s worth noting, though, that a 2K screen was used for Intel’s model, while Qualcomm’s design uses an FHD display, helping put the Adreno 680 ahead of the UHD Graphics 620."

Taints the results. That's twice the number of pixels. If they are willing to pull those shenanigans I wonder what else they are willing to do.

Umm, last time I checked FHD was 1920x1080 while 2K was 2048x1080. That's hardlty 2x the pixels. More like 12.5% more pixels for 2K.

Yeah, people mistakenly sometimes seem to think 1440p is 2k when its actually 2.5k

 

[captaincranky]

Powering the screen is a huge part of laptop battery life. The Intel laptop has double the resolution, but more than half the battery life. It looks like Intel won to me.

As for the testing procedures, this is the kind of nonsense that went on back in the heyday of home audio, and goes on yet today, in the "vinyl is better than CD audio", power struggle

People who don't understand those spec comparisons, believe whatever makes them feel better about themselves, then sit around listening to clicks, pops, and skips all night, if the kids jump up and down on the floor.

I have mixed feelings about 2K resolution in a laptop from the jump (*). At 27" though which is likely the typical enthusiast's preference, (Let's call it "Full 2K"), the screen accuracy and definition, mops the floor with FHD

(*) I have even more reservations about the the necessity of the resolution with the iPhones, "retina display". I think people may have been sold a phony bill of goods on that one.

 

[Evernessince]

Yeah, people mistakenly sometimes seem to think 1440p is 2k when its actually 2.5k

No, in reality 2k has two different formats

https://www.tomshardware.com/reviews/what-is-fhd-full-hd,5741.html


"2K

2560 x 1440 (typical monitor resolution)
2048 x 1080 (official cinema resolution)"

Calling Cinema resolution 2K IMO is sort of misleading as actual resolution is not really much of a bump up from 1080p at all. If you average the height and width you get 1,564. On the other hand if you average the height and width of 2560 x 1440 you get 2,160. It makes a lot more sense to call 2560 x 1440 2K as the resolution is closer to what 2K implies. Coming from 1080p, anyone looking at 2K would assume about twice as much or at the very least a decent improvement.

They should stop shifting between using the vertical and horizontal dimensions and simply use an average of both. Much more reliable and consistent.

 

[captaincranky]

No, in reality 2k has two different formats

https://www.tomshardware.com/reviews/what-is-fhd-full-hd,5741.html


"2K

2560 x 1440 (typical monitor resolution)
2048 x 1080 (official cinema resolution)"

Calling Cinema resolution 2K IMO is sort of misleading as actual resolution is not really much of a bump up from 1080p at all. If you average the height and width you get 1,564. On the other hand if you average the height and width of the 2560 x 1440 you get 2,160. It makes a lot more sense to call 2560 x 1440 2K as the resolution is closer to what 2K implies. Coming from 1080p, anyone looking at 2K would assume about twice as much or at the very least a decent improvement.

Gee, "official 2K cinema resolution's" aspect ratio is 2:1.

Wait, I better dumb that down so people will understand. Let's call it "16:8".

@Evernessince I do have to say, the screen on the "competitor's laptop", does look like a 2:1 aspect ratio.

 

[Evernessince]

Gee, "official 2K cinema resolution's" aspect ratio is 2:1.

Wait, I better dumb that down so people will understand. Let's call it "16:8".

There is actually a 3rd 2K resolution as well in 2560 x 1080 (16:9) which is used for ultrawide monitors, which is still a significant resolution drop over 2560 x 1440.

 

[captaincranky]

There is actually a 3rd 2K resolution as well in 2560 x 1080 (16:9) which is used for ultrawide monitors, which is still a significant resolution drop over 2560 x 1440.

I think you'll find that's more like 16:7 Actually 16:6.75 (I did the math))

 

[Evernessince]

I think you'll find that's more like 16:7 Actually 16:6.75 (I did the math))

Yep my bad. I believe they officially call it 21:9 even though that's rounded down.