Very interesting read. Where are those figures from?
*** The Official Samsung Galaxy S4 Thread ***
- Thread starter Cosimo
- Start date
Very interesting read. Where are those figures from?
That's confirmed it then really!
Soldato
- Joined
- 8 Mar 2007
- Posts
- 10,938
So was the Snapdragon they were comparing it with. It was supposed to be a like for like test on phones from the same gen.
If you disagree and think it was an unfair test then maybe you should vent your frustration at androidauthority.com
It's an almost perfect analogy
IT'S COMPLETELY DIFFERENT!.
I get the analogy, I just don't see what it proves. Intels and AMDs aren't Exynos's and Snapdragons.
Listen, my stance here was why put off buying the S4 because it has the Snapdragon, it is Bloomfield coming in and telling me how wrong that is and how much better the Exynos will be despite few benchmark results.
What i can't see happening is Samsung using two different processors with vastly different performance specs on them. I trust them to ensure that the Snapdragon is on a par with the Exynos.
Wish I'd have just posted that myself now as that what basically what I was saying in my original post that kicked this debate off.
Soldato
- Joined
- 8 Mar 2007
- Posts
- 10,938
Does this mean the Exy is less efficient contrary to popular belief?
Soldato
- Joined
- 29 Jun 2004
- Posts
- 12,957
I love how people are banging on about 4 vs 8 cores etc.
I don't think that there's any question that the Exynos version is faster than the Snapdragon version in overall muscle, probably by around 10-20% depending on the situation, especially in terms of PowerVR GPU used in that version of the phone.
However, will you actually notice any difference in real terms? Highly unlikely.
Agreed there's no way Samsung would give the USA, home of apple and 30/40mil potential sales a nerfed model, they'll both be great, however they have dropped the communications ball this current confusion shouldn't exist.
Soldato
- Joined
- 21 Oct 2009
- Posts
- 2,742
America normally gets the worst model, it's not much difference this time though.
In your original post you said the Snapdragon would be faster with apps using a single core because it was clocked faster than the A15. That's what I said was wrong.
I didn't say the Exynos was going to massively better, I said people want it because it's supposed to be more efficient due to having 4 low power cores instead of having to use the main power hungry cores all the time, and that the Exynos wouldn't be slower when running single threaded apps.
The benchmark you posted is the Snapdragon 600 (that's being used in the s4) against the exynos 4 (used in the s3). It isn't a previous gen CPU.
The Exy will be less efficient at full load, but when doing light tasks it has the quad core a7 for power efficiency.
Anyway, this is my last reply on this topic.
Soldato
- Joined
- 21 Oct 2009
- Posts
- 2,742
There shouldn't be any difference 
.
.
Soldato
- Joined
- 21 Oct 2009
- Posts
- 2,742
By "this topic" I meant what I was discussing with estebanrey, not the whole thread .
.
Soldato
- Joined
- 29 Jun 2004
- Posts
- 12,957
Even if you're doing "light tasks" on a single core 900mhz CPU you shouldn't feel any lag - because the "task is light". I.e. reading messages, phone calls etc.
Even if you have a 4 core, 8 core, 100 core CPU phone you won't in reality feel the difference in the day to day running. The only reason the CPU is different is due to the carrier requirements in each market (US vs UK markets for example).
I really don't know why people are banging on about it. It's a non-issue.
28nm HKMG Cortex A7 (plus any potential overheads for cluster switching) or 28nm LP Krait? That's the real question.. I agree with Bloomfield, there won't be much difference.
That's theoretical peak power consumption, your phone would melt fairly quickly at those power levels.
A phone only ever uses ~1.5W at peak performance (possibly going past 2W with larger phones for a second or two, Samsung are known to push comfort levels to the max
)
These chips (4-6W) are really designed for tablets as they stand a better chance of dispersing all that heat with much larger surface areas, they also have huge batteries. When you go past 7W you almost certainly need a small fan for cooling.
Both Snapdragon and Exynos will consume similar power levels at idle (28nm, power gated etc) and those rare peaks (limited to 1.5-2W by the GS4 chassis/case)
They could use a dual core Krait, calibrate performance to the S4 chassis (a single Krait core will use 1W+ at 1.9Ghz) and you'd get the same performance.
That's theoretical peak power consumption, your phone would melt fairly quickly at those power levels.
A phone only ever uses ~1.5W at peak performance (possibly going past 2W with larger phones for a second or two, Samsung are known to push comfort levels to the max
) These chips (4-6W) are really designed for tablets as they stand a better chance of dispersing all that heat with much larger surface areas, they also have huge batteries.
When you go past 7W you almost certainly need a small fan for cooling.Both Snapdragon and Exynos will consume similar power levels at idle (28nm, power gated etc) and those rare peaks (limited to 1.5-2W by the GS4 chassis/case)
They could use a dual core Krait, calibrate performance to the S4 chassis (a single Krait core will use 1W+ at 1.9Ghz) and you'd get the same performance.
Soldato
- Joined
- 8 Mar 2007
- Posts
- 10,938
I never specifically made that claim because I haven't seen any benchmarks. I was talking in general, you may think I'm splitting hairs but nowhere did I state the Snapdragon 600 would definitely be faster than the Exynos 5.
No it wasn't, did you even click the link? It was the Snapdragon S4 Pro (I think the S4 bit has confused you) which is a previous gen model used in the HTX One X (not to be confused with this year's HTC One or course). They compared the One X against the S3, two phones from the same gen.
http://en.wikipedia.org/wiki/Snapdragon_(system_on_chip)#Snapdragon_S4
Unless it makes a day's difference to the battery life I don't see the massive important (to my life anyway). I have to charge my S2 every night, making it last til 2 am instead of 1 am doesn't get around that.
Last edited:
Soldato
- Joined
- 21 Oct 2009
- Posts
- 2,742
I clicked the link, saw Optimus G then the word pro (Optimus G Pro does have the 600 in it) then closed it again
. The first time I didn't even open the link since in the quote it said Exynos 4412 and you'd just said "Anyway this bechmark test of the latest Snadragon vs latest Exynos".This doesn't count as a reply since I'm not actually arguing anything
.
Associate
- Joined
- 19 Jun 2009
- Posts
- 1,306
Wondering why Samsung has clocked the Exynos at 1.6 ghz and the snapdragon 600 at 1.9ghz. Probably to get parity between the two chips.
My impression is that they will be equal for performance and that the Exynos needs need the A7 quad especially during idle to compete with the efficient snapdragon s4?
My impression is that they will be equal for performance and that the Exynos needs need the A7 quad especially during idle to compete with the efficient snapdragon s4?
By "this topic" I meant what I was discussing with estebanrey, not the whole thread
Just so we're clear the "so" in my post was an auto corrected soz, not an accusatory SO! I expect you know this but clarity is never wrong especially in such a crucially important heated subject like phone preference
Last edited:
Soldato
- Joined
- 21 Oct 2009
- Posts
- 2,742
Just so we're clear the "so" in my post was an auto corrected soz, not an accusatory SO! I expect you know this but clarity is never wrong especially in such a crucially important heated subject like phone preference
Yeah, I guessed that
.