Published in Processors

Qualcomm dismisses Tegra K1 benchmarks



Didn’t take into consideration thermal constraints

The first set of Tegra K1 benchmarks emerged yesterday, courtesy of Tom’s Hardware and the numbers were quite encouraging, as we said here

However, Qualcomm is having none of it and it promptly shot off a statement saying that the benchmarks really don’t illustrate real-world performance, as they were carried out on a big 4K all-in-one rather than a mobile device.

“The recently released Nvidia Tegra K1 benchmark scores on an unreleased All-in-One desktop platform are generated without taking into consideration any thermal constraints of a mobile environment,” Qualcomm told us. “For a relevant comparison, we would need to see the Tegra K1 in a mobile use case – assuming it will have some traction in smartphones and tablets – instead of a wired reference design with heat sinks and no need for mobile power management.”

Qualcomm also stressed that it always strives to deliver the best power efficiency and performance possible, hence it designs Snapdragon chips specifically for power envelopes and connectivity requirements in mobile devices, such as smartphones and tablets.

Qualcomm’s goal is to provide the best graphics performance per watt. Oddly enough, the statement reads a lot like AMD/ATI statements of years gone by, especially in the post-HD2900 era, when AMD emphasized performance per watt instead of sheer performance numbers. We guess there is still plenty of ATI DNA in the Adreno team, for better or for worse.

“We look forward to seeing how Tegra K1 commercial devices compare to Snapdragon 800 and Snapdragon 805 processors,” Qualcomm said.

Of course, this may take a while as there are no commercially available products based on either platform. We expect the first batch of Tegra K1 devices to show up in late Q1, with availability in Q2. Nvidia refuses to comment on availability, but the info we have leads us to conclude the first K1 products will go on sale in early Q2.

Rate this item
(0 votes)

Read more about: