AMD Athlon 5350 / 5150 & Sempron 3850 / 2650

[phoronix]

Phoronix: AMD Athlon 5350 / 5150 & Sempron 3850 / 2650

It's been a busy past few days since AMD launched their "AM1" Socketed Kabini APUs. After the initial Athlon 5350 Linux review on launch-day, I did some tests involving a faster kernel and newer Mesa code along with some reference DDR3 memory scaling benchmarks for these APUs with Jaguar processor cores. Since then the Athlon 5150 and Sempron 3850/2650 APUs arrived. After a busy weekend of benchmarking, here's the initial Ubuntu Linux benchmarks of all four AMD AM1 APUs that are available at this time: the Sempron 2650, Sempron 3850, Athlon 5150, and Athlon 5350. With these four new AMD APUs are also a number of thermal and power consumption tests.

http://www.phoronix.com/vr.php?view=20168

[curaga]

Can you mention the PSU?

[Ferdinand]

I thought these were low power parts? 38Watts doesn't sound like very power efficient to me? These parts are 2x faster than Krait cores right? But Krait cores use about 3Watt peak?

[Spittie]

I thought these were low power parts? 38Watts doesn't sound like very power efficient to me? These parts are 2x faster than Krait cores right? But Krait cores use about 3Watt peak?

Those parts have a 25W TDP, which mean that the APU is going to get more than this on full load. It's also total system power, sadly the CPU is not the only thing using power in a computer.
I'm not sure where you're picking up the 2x faster than Krait, I've never seen any benchmarks with them. But If I have to take a wild guess, I'd say that those are quite more faster than only 2 times.

[Michael]

Can you mention the PSU?

Some Cooler Master ~600W PSU.

[DanL]

38Watts doesn't sound like very power efficient to me?

That's for the whole system, and it's also 38W AC, so if the PSU is 80% efficient, it's 30W DC (so the TDP of 25W for the APU sounds correct).

[Spittie]

That's for the whole system, and it's also 38W AC, so if the PSU is 80% efficient, it's 30W DC (so the TDP of 25W for the APU sounds correct).

It's probably even worse than 80%, since the load on the PSU was very low.

[gradinaruvasile]

How was the temperature monitored?
I have an A8-6500 and the k10temp driver reports 6C degrees right now (ambient temp is ~22C). It was the same for my previous A8-5500 (and the Athlon II 250 i had before IIRC).
The ITE monitoring chip from the mobo reports 18C, again off.
Both sources report different temperatures under heavy load ( 4 threads compile) where they become accurate, the k10temp one is ~10C lower than the monitoring chip's (which is probably used in the BIOS/UEFI too, probably with some correction for lower temps).

Another thing is that temperatures can vary very fast with CPUs so rebooting may skey data, unless you enter the BIOS and let it idle for 1-2 minutes then read the temps, because there you dont have any frequency optimizations and the cores are always on full power.

[AJSB]

OC and dGPU tests , YEEESSSSS

I hope that includes the GTX750(Ti)...and that are OC+dGPU tests besides OC and dGPU alone

[Michael]

and that are OC+dGPU tests besides OC and dGPU alone

Unlikely, too time consuming, etc, unless there were requests from many Phoronix Premium readers for that... besides the overclocking is limited and just getting about an extra 100MHz if lucky.