Nvidia needs to upgrade Tegra. Although the GPU and video abilities are great, it uses a dual 600mhz ARM11, whereas all the new smartphones are using next-gen ARM Cortex-A8 CPUs and soon dual-core Cortex-A9s that scale to 1.5Ghz each. Not to mention PowerVR SGX GPUS can scale to meet Tegra and dedicated 1080P video abilities are available from all the major ARM manufacturers.
Yeah, I'd say that my biggest worry about Tegra is that outdated (outdating?) processor. Correct me if I'm wrong here, but isn't the processor the key factor in how fast most of the phone's features feel (like apps and web page rendering)? Also, couldn't Nvidia simply make a tegra that uses an ARM Cortex-A8, and maybe a A9 in the future?
Yes, they could, but that requires money and resources and I think they are in a state of mind of "good enough".
Although Tegra uses the ARM dual-core technology, The two processor cores are ARM11 based and can only run around 600mhz. Combined, this should give it a raw capability of ~1500 DMIPS.
The newer ARM Cortex-A8 core (iPhone 3GS, Pre, OMAP3) and Qualcomm's custom, Cortex-compatible Snapdragon core are both significantly faster than the ARM11 on a clock-for-clock basis, with a 600Mhz core coming in at ~1200 DMIPs. More significantly, the Cortex-A8 can easily scale to 1.0Ghz or higher on a 45nm process, and the even faster and more efficient Snapdragon core can go to 1.3Ghz. Additionally, the Cortex generation finally introduces a powerful parallel SIMD engine called NEON, which can be used to speed up intensive multimedia calculations. Finally, having an older dual-core ARM introduces more software complexity than a new single-core ARM of similar power, and as always, you NEVER get perfect core scaling in software, and only certain applications will be able to take advantage of it.
Even a 65nm-produced 800Mhz ARM Cortex-A8 or Snapdragon core would be far faster in real-world use than the Tegra's ~650 mhz dual ARM11.
Additionally, the even newer multi-core capable ARM Cortex-A9 architecture has been available for licensing for a few years now and are just starting to be sampled from major OEMs. No doubt Nvidia could integrate two or more Cortex-A9 cores into a new chip. T.I.'s OMAP4 uses a dual-core Cortex-A9 clocked at 1000mhz+ and an SGX540+ GPU, and is capable of full 1080P encoding decoding. Similarly, Qualcomm has a monster dual-core, 1.5Ghz Snapdragon.
Overall, I hope Nvidia has an updated Tegra platform in the works. Otherwise the GPU and video capabilities are going to far outmatch general processing on all platforms.
Most importantly, having the greatest GPU in the world will do NOTHING to improve load times and rendering times of web pages, nor general application performance.
"I'm looking to grab a new phone using a prepaid service. What's the best prepaid phone service in terms of overall price, phone selection and other bells / whistles. Thanks!"
The most commented posts on Engadget over the past 24 hours.
Now that we've thrown 'em off the trail, use the form below to get in touch with the people at Engadget. Please fill in all of the required fields because they're required.
Nvidia needs to upgrade Tegra. Although the GPU and video abilities are great, it uses a dual 600mhz ARM11, whereas all the new smartphones are using next-gen ARM Cortex-A8 CPUs and soon dual-core Cortex-A9s that scale to 1.5Ghz each. Not to mention PowerVR SGX GPUS can scale to meet Tegra and dedicated 1080P video abilities are available from all the major ARM manufacturers.
Yeah, I'd say that my biggest worry about Tegra is that outdated (outdating?) processor. Correct me if I'm wrong here, but isn't the processor the key factor in how fast most of the phone's features feel (like apps and web page rendering)? Also, couldn't Nvidia simply make a tegra that uses an ARM Cortex-A8, and maybe a A9 in the future?
Yes, they could, but that requires money and resources and I think they are in a state of mind of "good enough".
Although Tegra uses the ARM dual-core technology, The two processor cores are ARM11 based and can only run around 600mhz. Combined, this should give it a raw capability of ~1500 DMIPS.
The newer ARM Cortex-A8 core (iPhone 3GS, Pre, OMAP3) and Qualcomm's custom, Cortex-compatible Snapdragon core are both significantly faster than the ARM11 on a clock-for-clock basis, with a 600Mhz core coming in at ~1200 DMIPs. More significantly, the Cortex-A8 can easily scale to 1.0Ghz or higher on a 45nm process, and the even faster and more efficient Snapdragon core can go to 1.3Ghz. Additionally, the Cortex generation finally introduces a powerful parallel SIMD engine called NEON, which can be used to speed up intensive multimedia calculations. Finally, having an older dual-core ARM introduces more software complexity than a new single-core ARM of similar power, and as always, you NEVER get perfect core scaling in software, and only certain applications will be able to take advantage of it.
Even a 65nm-produced 800Mhz ARM Cortex-A8 or Snapdragon core would be far faster in real-world use than the Tegra's ~650 mhz dual ARM11.
Additionally, the even newer multi-core capable ARM Cortex-A9 architecture has been available for licensing for a few years now and are just starting to be sampled from major OEMs. No doubt Nvidia could integrate two or more Cortex-A9 cores into a new chip.
T.I.'s OMAP4 uses a dual-core Cortex-A9 clocked at 1000mhz+ and an SGX540+ GPU, and is capable of full 1080P encoding decoding. Similarly, Qualcomm has a monster dual-core, 1.5Ghz Snapdragon.
Overall, I hope Nvidia has an updated Tegra platform in the works. Otherwise the GPU and video capabilities are going to far outmatch general processing on all platforms.
Most importantly, having the greatest GPU in the world will do NOTHING to improve load times and rendering times of web pages, nor general application performance.