Rumor: Apple's upcoming A9-series chips benchmarked, show 20-30% jump over A8
A graph that purports to show GeekBench 3 scores for Apple's anticipated A9-series application processors paints a picture of significant performance improvements, beating out even Samsung's latest 8-core chips.
Apple's A9 chimes in with a single-core score of 2090 and a multi-core tally of 3569, compared to the respective averages of 1611 and 2892 shown by the A8 found in the iPhone 6 series. The A9X chalks up 2109 and 5101 in the same benchmarks, compared to 1808 and 4526 for the iPad Air 2's A8X.
Samsung's Exynos 7420 --?which powers the flagship S6, S6 Edge, Note 5, and S6 Edge+ --?bests Apple's A9 on multi-core workloads with a score of 5048, but falls behind on single-core tasks by posting a 1486. Huawei's Kirin 950 and LG's Nuclun 2 also beat out Samsung's entry.
The graph --?the authenticity of which cannot be determined -- comes from Weibo via GforGames.
Apple is expected to reveal the A9 alongside a set of new iPhones at a special event early next month. The A9X --?which is likely to debut alongside a new iPad --?might bow at the same event, though it's not clear whether Apple plans to hold a separate iPad introduction at a later date.
Samsung and TSMC are believed to have begun mass production for the A9 in July.
Apple's A9 chimes in with a single-core score of 2090 and a multi-core tally of 3569, compared to the respective averages of 1611 and 2892 shown by the A8 found in the iPhone 6 series. The A9X chalks up 2109 and 5101 in the same benchmarks, compared to 1808 and 4526 for the iPad Air 2's A8X.
Samsung's Exynos 7420 --?which powers the flagship S6, S6 Edge, Note 5, and S6 Edge+ --?bests Apple's A9 on multi-core workloads with a score of 5048, but falls behind on single-core tasks by posting a 1486. Huawei's Kirin 950 and LG's Nuclun 2 also beat out Samsung's entry.
The graph --?the authenticity of which cannot be determined -- comes from Weibo via GforGames.
Apple is expected to reveal the A9 alongside a set of new iPhones at a special event early next month. The A9X --?which is likely to debut alongside a new iPad --?might bow at the same event, though it's not clear whether Apple plans to hold a separate iPad introduction at a later date.
Samsung and TSMC are believed to have begun mass production for the A9 in July.
Comments
Anytime you compare processor performance for a mobile device, you need to also compare battery usage.
Anytime you compare processor performance for a mobile device, you need to also compare battery usage.
The same as all the other tests for the past few years.
Samsung's battery life will suck, as usual.
Apple's battery life will - as usual - be the gold standard.
I'm curious how the Fandroid community will defend this pathetic performance. Apple's A9x with less cores runs faster than that 8-core garbage. Shameful.
But even when this is true, Geekbench doesn't measure GPU performance, which is the most important factor.
Probably another "loaded" result, as is so common on Android devices.
https://mobile.twitter.com/jfpoole/status/632066813380165632
If the A9 and A9X benchmarks are real and accurate, my guess is that they were conducted at the same clock speeds as the shipping versions of the A8 and A8X. Bump the clock up a bit and the A9* will trounce the competition.
doesn’t anyone find it odd that it shows results for 6 unannounced SoC and that all those results came out in the last few days!? Some of those SoC will only launch next year!
How can anyone even take this seriously? It would be one gigantic case of industrial espionage
However, if they are correct, it's disturbing. Yes, I know some people here will hate what I'm going to say, but too bad.
But, what we see in these tests is that Apple's new chips lag several current chips that have been out for months, almost a year, in the case of several.
Previously, Apple's new devices not only led all current devices in performance, but new devices introduced afterwards, several months later, and only near the end of the year had newer devices catching up, and superseding them in some areas.
What we're seeing here is that several older devices are better than Apple's new ones.
Yeah, I know the excuses, we need to consider battery power, thermal throttling, etc. But that still doesn't explain why, this year, Apple seems to be starting at a disadvantage for the first time since they began designing their own SoCs.
So I'm hoping that something is fishy here.
And why would you guess that? If these are real Apple devices, they would run at the speeds the devices are designed to run at.
Excuse me, but I've been seeing tests against current SoCs that show these results.
I've been reading several of these tests, which, I have to remind people, may very well be on unoptimized devices.
However, if they are correct, it's disturbing. Yes, I know some people here will hate what I'm going to say, but too bad.
But, what we see in these tests is that Apple's new chips lag several current chips that have been out for months, almost a year, in the case of several.
Previously, Apple's new devices not only led all current devices in performance, but new devices introduced afterwards, several months later, and only near the end of the year had newer devices catching up, and superseding them in some areas.
What we're seeing here is that several older devices are better than Apple's new ones.
Yeah, I know the excuses, we need to consider battery power, thermal throttling, etc. But that still doesn't explain why, this year, Apple seems to be starting at a disadvantage for the first time since they began designing their own SoCs.
So I'm hoping that something is fishy here.
And this is a problem for Apple why? 30% gain from last year SOC sounds about right to me. Apart from a few games, there is very little in the store that's even using A6X or 64Bit.
I doubt Apple has a stop watch checking what the Android OEMS are doing - if you are into specs, you are on the wrong platform. The OnePlus 2 will map the floor of anything that Apple will ship next month but they will not outsell 6S models.
A graph! I'm convinced! Graphs only ever tell truths!
Yeah, I know the excuses, we need to consider battery power, [...]
My BMW has like 400 HP but it sucks gas like crazy. Balance of processor power and battery life is the key to small portable devices. That is why Apple went with ARM instead of Intel. If it is 30% faster with the same battery usage, I think that is amazing.
Excuse me, but I've been seeing tests against current SoCs that show these results.
Moderator;
Moderate thy self.
The same as all the other tests for the past few years.
Samsung's battery life will suck, as usual.
Apple's battery life will - as usual - be the gold standard.
I'm curious how the Fandroid community will defend this pathetic performance. Apple's A9x with less cores runs faster than that 8-core garbage. Shameful.
Fact check:
Number of cores and clock speed are meaningless when these are all very different types of cores and configurations.
Apple and Samsung's SoCs generally hit the market ~6 months apart from each other. One would expect each SoC to be a performance leader around the time of their respected launches.
Is this the fake benchmark I've read about today?
https://mobile.twitter.com/jfpoole/status/632066813380165632
That's dated August 13th and it appears to be related to a fake benchmark of a "16 core" smartphone.
I've been reading several of these tests, which, I have to remind people, may very well be on unoptimized devices.
However, if they are correct, it's disturbing. Yes, I know some people here will hate what I'm going to say, but too bad.
But, what we see in these tests is that Apple's new chips lag several current chips that have been out for months, almost a year, in the case of several.
Previously, Apple's new devices not only led all current devices in performance, but new devices introduced afterwards, several months later, and only near the end of the year had newer devices catching up, and superseding them in some areas.
What we're seeing here is that several older devices are better than Apple's new ones.
Yeah, I know the excuses, we need to consider battery power, thermal throttling, etc. But that still doesn't explain why, this year, Apple seems to be starting at a disadvantage for the first time since they began designing their own SoCs.
So I'm hoping that something is fishy here.
The competition was not there previously. It was pretty much just Qualcomm and the odd Exynos for select markets.
Going into the end of 2015 and start of 2016, we have Qualcomm, LG, Huawei, nVidia, Samsung, MediaTek and Apple all making high end SoC's. All of them use ARMv8 cores, LPDDR4 memory and they should all be on TSMC's 16 nm FF or Samsung's 14 nm FF.
With the exception of NVIDIA (Denver 2 core) and Qualcomm (Hydra core), all of the other non-Apple companies appear to be using Cortex A72 (ARM's new big core) in their upcoming SoC's.
EDIT: Samsung's Exynos M1 might also be using a custom core.
And why would you guess that? If these are real Apple devices, they would run at the speeds the devices are designed to run at.
As you say, IF these are real, destined-to-ship Apple devices programmed with shipping firmware. Nobody says they are.