Samsung Galaxy S 4 Review - Part 1
by Brian Klug on April 24, 2013 12:01 AM ESTCPU Performance
For our CPU analysis we're left with our usual browser based benchmarks. Again this isn't an ideal list of tests but it's the best we've got for now. Where necessary we'll show results using both stock and Chrome browsers. We did notice a single case of thermal based throttling under SunSpider 0.9.1 (the benchmark alone is ine, but running it after a bunch of others caused throttling), so we're once again presenting results in our standard test environment as well as inside of a freezer to show peak performance. Although the Galaxy S 4 managed to throttle in one of our tests, the device never felt all that warm to the touch. We could be seeing some of the same aggressively set thermal governors that we saw back with the Nexus 4. It's also worth pointing out that we're simply in an era of pushing the limits of just how fast you can go at 28nm LP in many of these smartphones. The mobile SoC vendors also need to do a better job of power management, enabling controlled bursting to these high frequency states vs. sustaining the higher frequencies until there's a serious enough thermal issue that the CPU cores have to throttle themselves significantly.
Under Kraken in particular we see a measurable improvement in performance over the 1.7GHz S600 used in the HTC One. Qualcomm still can't attain the peak performance of ARM's Cortex A15, but once again we're looking at a much lower power profile.
Facebook
Twitter
RSS
335 Comments
View All Comments
blue_urban_sky - Wednesday, April 24, 2013 - link
I am very interested in the sock browser you mention."I find it difficult to hover my finger over the screen without touching it, but in the gallery and sock browser it is admittedly sometimes useful."
wenugopal - Wednesday, April 24, 2013 - link
Just curious, isn't galaxy s4 using LPDDR3 RAM, as opposed to LPDDR2 which is used in HTC One ? Hence, higher speeds ?Brian Klug - Wednesday, April 24, 2013 - link
Absolutely true, there's more to it than that though for things like CPU performance.-Brian
scaramoosh - Wednesday, April 24, 2013 - link
I have big doubts that DDR2 and 3 makes much difference, we never see much of a difference on the PC in day to day use, only benchmarks.B3an - Wednesday, April 24, 2013 - link
DDR3 is significantly faster than DDR2 on PC now. You can definitely notice a difference without benchmarks depending on what you're doing. Stupid comparison though because phones are too different.UpSpin - Wednesday, April 24, 2013 - link
because the RAM gets used for both CPU and GPU I think the upgrade to DDR3 makes a noticable difference, at least for the GPU. (the same way dual channel gives Intel IGP performance a boost compared to single channel)Mithos - Wednesday, April 24, 2013 - link
Man, I DO value removable battery as I'm the type that likes buying an extended battery for business trips. I rely on the GPS, constant mail checking and of course calls.But all that aside, I prefer the construction and UI the HTC One presents, yet it does not have expandable battery.
Ohhhh cruel world and it's first world problems.
Nice review, thank you for the hard work.
Gigaplex - Wednesday, April 24, 2013 - link
What's stopping you from getting a portable battery powered phone charger?Toss3 - Wednesday, April 24, 2013 - link
Removable battery does not just limit you to buying a separate one, but enables you to buy a 4500mAh after market battery that lasts two full days without charging. You cannot do that with the HTC One.UpSpin - Wednesday, April 24, 2013 - link
but this dramatically increases the size of the smartphone. I would rather buy a battery powered smartphone charger, which I can store in a bag or jacket and keep the already large smartphone comfortably sized. It's already thin, no question, but it's wide and long, and if it becomes thicker it becomes harder to use and to store it in a pocket.