Core M v.s. A8X in Geekbench 3

[Arachnotronic]

http://browser.primatelabs.com/geekbench3/compare/922000?baseline=1061742

Pretty epic results from A8X.

 

[monstercameron]

what are the clock freq ranges on these parts?

 

[Arachnotronic]

what are the clock freq ranges on these parts?

Intel part goes to 2GHz, Apple goes to 1.5GHz.

 

[monstercameron]

Intel part goes to 2GHz, Apple goes to 1.5GHz.

thanks, and how comparable is geekbench cross os and cross isa?

 

[Arachnotronic]

thanks, and how comparable is geekbench cross os and cross isa?

I've seen Intel parts do about the same on Windows as they do on Android.

Both test results linked are 64-bit versions of GB3, AFAIK.

 

[witeken]

http://browser.primatelabs.com/geekbench3/compare/922000?baseline=1061742

Pretty epic results from A8X.

What is an IPad 5.4? Why don't we have Sunspider or Kraken results? I haven't heard much positive about Geekbench recently. I'd rather like to see Cinebench scores.

 

[Khato]

I've seen Intel parts do about the same on Windows as they do on Android.

And there's a reason for that - http://www.primatelabs.com/blog/2013/12/developing-a-cross-platform-benchmark/

The problem with that approach being that it only applies to x86. (Not to mention, this appears to state that in some cases they're re-coding algorithms to avoid compiler optimizations while in others they're working around compiler shortcomings.) What happens for all of the ARM crowd though? Because there basically aren't any alternate OS options there to use for comparison - would be interesting to know if they even attempt to do the same level of leveling the playing field with respect to compiler optimizations or not.

 

[monstercameron]

What is an IPad 5.4? Why don't we have Sunspider or Kraken results? I haven't heard much positive about Geekbench recently. I'd rather like to see Cinebench scores.

hmmm how would you see cinebench scores on an ipad?

 

[Arachnotronic]

And there's a reason for that - http://www.primatelabs.com/blog/2013/12/developing-a-cross-platform-benchmark/

The problem with that approach being that it only applies to x86. (Not to mention, they're basically admitting to re-coding their algorithms to avoid compiler optimizations on x86.) What happens for all of the ARM crowd though? Because there basically aren't any alternate OS options there to use for comparison - would be interesting to know if they even attempt to do the same level of leveling the playing field with respect to compiler optimizations or not.

jfpoole typically answers questions on this forum. I hope he'll chime in!

 

[asendra]

I hadn't noticed until now how close this are getting to my current main machine, a Macbook Air 2012. Wonder how much Apple could push realistically the A8X in a mb air like device, today.

PD. I seriously need a broadwell rMPB right about now.

 

[Eug]

What is an IPad 5.4? Why don't we have Sunspider or Kraken results? I haven't heard much positive about Geekbench recently. I'd rather like to see Cinebench scores.

iPad5,4 is the cellular version of the iPad Air 2. And I think you're a little bit too stuck in Intel's world. You're not going to be running Cinebench on an iPad any time soon.

 

[kimmel]

Wake me when we have cross platform benchmarking that isn't terrible.

 

[jpiniero]

What is an IPad 5.4? Why don't we have Sunspider or Kraken results? I haven't heard much positive about Geekbench recently. I'd rather like to see Cinebench scores.

Sunspider would probably be about 300, maybe a little less.

 

[Sweepr]

Wake me when we have cross platform benchmarking that isn't terrible.

And Core M-5Y70 (up to 2.6GHz) should be quite a bit faster than the Core M-5Y10 (up to 2GHz) tested.

 

[Roland00Address]

Once again arm kills in sha encryption in geekbench.

 

[ShintaiDK]

SHA and memory speed seems to be the A8X main benefits.

But again, the cross platform issue.

Is the Asus one even using dual channel?

 

[Phynaz]

Any idea why the A8X is listed as 3 cores?

 

[Eug]

Any idea why the A8X is listed as 3 cores?

Because it has 3 cores.

 

[ShintaiDK]

Any idea why the A8X is listed as 3 cores?

Because it got 3 cores?

 

[ams23]

http://browser.primatelabs.com/geekbench3/compare/922000?baseline=1061742

Pretty epic results from A8X.

Enhanced Cyclone in A8X is quite good but probably not quite a match for the performance of Core M in most scenarios. The Core M variant tested above exceeds A8X in single-core performance and nearly matches A8X in multi-core performance while having one fewer CPU core (and most mobile apps will probably not make good use of anything more than two CPU cores in the first place).

What would be even more interesting to see would be performance and power consumed by each respective CPU with a given CPU-intensive application.

 

[Steelbom]

Enhanced Cyclone in A8X is quite good but probably not quite a match for the performance of Core M in most scenarios. The Core M variant tested above exceeds A8X in single-core performance and nearly matches A8X in multi-core performance while having one fewer CPU core (and most mobile apps will probably not make good use of anything more than two CPU cores in the first place).

What would be even more interesting to see would be performance and power consumed by each respective CPU with a given CPU-intensive application.

I suspect that this third core is for the rumoured split screen multitasking on iPad. Running two apps at the same time would definitely need another core.

 

[ams23]

Does anyone know if there is a 64-bit version of Geekbench 3 for Windows x86 machines, and if so, what type of performance improvements are expected vs. the 32-bit version?

The Core M results from Asus TF300A are using the 32-bit "Geekbench 3.1.2 Tryout for Windows x86 (32-bit)": http://browser.primatelabs.com/geekbench3/922000

The A8X results from iPad Air 2 are using the 64-bit "Geekbench 3.2.2 for iOS AArch64": http://browser.primatelabs.com/geekbench3/1061742

 

[Arachnotronic]

Does anyone know if there is a 64-bit version of Geekbench 3 for Windows x86 machines, and if so, what type of performance improvements are expected vs. the 32-bit version?

The Core M results from Asus TF300A are using the 32-bit "Geekbench 3.1.2 Tryout for Windows x86 (32-bit)": http://browser.primatelabs.com/geekbench3/922000

The A8X results from iPad Air 2 are using the 64-bit "Geekbench 3.2.2 for iOS AArch64": http://browser.primatelabs.com/geekbench3/1061742

Ahh! Good catch on the 32-bit tryout mode!

 

[jpiniero]

Enhanced Cyclone in A8X is quite good but probably not quite a match for the performance of Core M in most scenarios. The Core M variant tested above exceeds A8X in single-core performance and nearly matches A8X in multi-core performance while having one fewer CPU core (and most mobile apps will probably not make good use of anything more than two CPU cores in the first place).

I tend to agree that performance of Core M looks like it's going to be all over the map; although at a size that's comparable to the Air 2 it would be quite worse than the T300FA, other models with 6.5 W TDP and a fan will perform much better, etc.

 

[III-V]

http://browser.primatelabs.com/geekbench3/compare/922000?baseline=1061742

Pretty epic results from A8X.

Cross OS comparisons are pretty much useless for evaluating hardware, unfortunately. I say this in spite of what you've stated earlier in this thread. Both processors are very impressive, though.

I like how Apple got around the whole "4 cores = bad" Chinese cultural thing by choosing three. Perhaps that was not the major motivator, as their cores are quite beefy and therefore expensive, but that should definitely help their sales over there.

Not to mention, this appears to state that in some cases they're re-coding algorithms to avoid compiler optimizations while in others they're working around compiler shortcomings

Well that's silly.