ARM vs x86

File: 10.png (88KB, 1637x920px) Image search: [Google]

88KB, 1637x920px

>>55798831
The Exyno 8890 is on average substantially behind that i3 in both integer and floating point performance. Remove AES and SHA1 results and there is no advantage shown for the ARM part.

ARMs own A73 isn't much of an improvement over the A72 in raw performance. Its 5%-10% increased performance at 20% reduction in power per op. They aren't even trying to create a chip to compete against X86.

>>55799037
Are you blind m8? The Exyno 8890 scored 65% faster in multi-core integer performance and 3% faster in multi-core floating point performance compared to a 2.4GHz i3-4000M processor.

>>55799090
See all those long blue bars? Thats the i3 offering significantly more performance. The Exynos 8890 pulls ahead by a significant margin only in AES and SHA hashing, because ARM had dedicated accelerators for it.
There are only a select few metrics where the ARM part pulls slightly ahead.

Are you honestly too stupid to read a few graphs?

>>55799133
The biggest factor in how shit or not shit a processor is, is integer and floating point performance though which is what you were talking about.

>>55799090
I think I need to point something out, notice how it says "Multi-core performance"

Does that remind you of anyone on the x86 market, like say AMD?

>>55799163
Here we go again. Not all programs use a single cpu thread m8, it's not 2005 anymore.

>>55799159
The AES and SHA benches contribute to the over all integer score. The "integer" "floating point" and "memory" metrics are aggregate scores of all the individual metrics listed below them.

Its pathetic that this needs to be pointed out constantly. Doesn't matter if its an Apple SoC, the latest Qualcomm chip, or whatever Samsung is using. These ARM chips are not competing against X86 on performance. AES and SHA hashing skews the aggregate score.

>>55799210
No, but processors with better single core performance always win on x86 market.

Besides, it's not like Intel can't roll out multi-core processors if they want to. They even have 24 core processors available if I recall correctly.

>>55799279
>No, but processors with better single core performance always win on x86 market.
Then why are laptops with AMD APUs so god damn popular despite shit single-core performance then? Seriously, every time I walk into a best buy or other electronic store APU laptops are either sold out or there is just one or two left to be sold.

>>55799312
Probably because they are dirt cheap and you're talking about a store that sells computer products to grandma?