Uncategorized

Intel’s Dubious Real-World Performance Claims Continue to Entertain

Ever since AMD launched the Ryzen CPUs, Intel has taken a bizarre approach to tackle Team Red’s resurgence, namely fake it till you make it. That’s quite literally what Intel’s PR teams have been doing over the last few years. Dubbing the few select applications, especially games as “Real World Benchmarks”, we have seen Team Blue make many ridiculous claims in the past. The latest set of slides from Intel compares its latest products against AMD’s 7nm offerings. Let’s have a look, first at the good, then the bad stuff:

These slides are completely justified. It’s true that Intel’s 10th Gen CPUs completely beat their Ryzen counterparts in gaming benchmarks by 10-15% on average (no matter how inconsequential that delta may be). Ergo, if your workloads are strictly limited to gaming, then buying an Intel CPU isn’t a bad idea. It may very well offer a higher price-performance ratio, at least in the desktop market, regardless of how soon the chips become obsolete.

Unsurprisingly, integrated GPU benchmarks of the 10th Gen mobile processors were glaringly absent, no doubt because the Renoir parts utter destroy the Comet Lake parts while also beating the newer 10nm Ice Lake chips by a good 20-30%.

Now, onto the more interesting claims: Intel claims that the AC performance of the Ryzen 7 and Core i7 are roughly on par with each other. Nothing wrong with that. But then goes on to show that on battery, the performance is worse off for the Ryzen parts. That would be okay, but then suddenly, the next slide shows that the battery life of the Ryzen 7 4700U is nearly 100 minutes more than the Core i7-1065G7, despite the fact that the former is an octa-core chip.

The highlight of the slide is that the Ryzen 7 trades some of its performance for a longer battery life which is exactly what the 15W notebook APU should do. Intel seems to thing that that’s a bad thing. Not sure why.

Then we have Intel’s custom RUG performance tests. These are basically a set of scripts prepared by the company (without showing the source code) that emulate mundane tasks like pasting charts from Excel to Word, converting a PPT to PDF and mail merging. Now, I’m not sure why Intel considers these as indicative of “real-world performance” when you can perform all these on the slowest of chips without noticing any major lags.

Digging deeper into the power-frequency scaling, you can see that the Ryzen 7 draws lower power on battery by limiting its boost algorithm while the Core i7-1065G7 consumes slightly more, resulting in the reduced battery life. Not sure why this is supposed to be a bad thing for a mobile processor, but Intel surely thinks that way.

While the 15W U-series Ice Lake CPUs fo have a marginal (imperceivable) lead in single-threaded applications, the 14nm Comet Lake-H CPUs are slower than AMD’s 7nm Renoir-H APUs in Nearly Every application, and as shown by Intel’s own tests have a worse battery life as well

The same behavior is reproduced on the H series chips, where the Ryzen 7 4800H offers a whole extra hour of battery at the cost of slightly reduced performance in a few select synthetic benchmarks. In most applications that actually matter, the 45W Renoir-H APUs lead the Comet Lake-H parts by a good margin.

Of course, Intel is spinning this as a negative aspect, despite the fact that the company is now moving to integrate low-power cores in not only its notebook, but desktop CPUs as well to “save power”. Yet, when the Ryzen 7 trades a bit of performance here for a lofty boost in battery life, it’s a bad thing.

Source

Areej Syed

Processors, PC gaming, and the past. I have been writing about computer hardware for over seven years with more than 5000 published articles. Started off during engineering college and haven't stopped since. Mass Effect, Dragon Age, Divinity, Torment, Baldur's Gate and so much more... Contact: areejs12@hardwaretimes.com.
Back to top button