- Joined
- 24 Jan 2002
- Messages
- 12,388
To be clear for those believing Apple proved ARM is a better platform over x86 far from it;
So far they’ve proven opposite.
Apple and ARM deliberately leaves out the actual factor, (I'll add that for ya), M1 performance isn’t because of ARM, it’s because, (thanks to the buying up the first 5nm lithography, nobody else has it).
M1 is the a beneficiary of 60 percent more transistors, (that's 6 billion more) over AMD double transistor count over Intel.
Apple m1, 16 billion transistors, Ryzen mobile 10 billion, Intel 8 billion!
Even with that extra 6 billion transistor with closer proximity for superior communication, the best ARM still can’t beat the best x86 mobile chips.
When AMD debuts with the same transistor count as M1 the same lithography, early samples demonstrate 25% gain in ipc and 40 percent overall performance!
AMD best mobile chip beats ARM best (M1) now, give it the same transistor count as M1 watch it absolutely smoke ARM.
That’s not superior anything but battery life.
Nor is ARM strictly “RISC”, it also sports CISC” nor do AMD/INTEL sport only” x86, they also incorporate RISC.
When AMD//Intel are on the same transistor count as the best ARM, that’s when you can actually discover which is superior architecture, not before.
Sadly for x86 players, by the time they get their first run of 5nm, Apple has already bought the first run of the next generation. (we don’t know if it’s 4 or 3nm)
Transistor count isn’t the only benefit to smaller lithography, closer transistors means lower latency, even if transistor count is the same, close transistor proximity improves performance.
Nobody should be thinking ARM is superior to x86 until they’re both on the same lithography.
Hopefully, the beginning of next year AMD will get 5nm before Apple gets 3nm, that’s when we find fair comparison.
Time tells
So far they’ve proven opposite.
Apple and ARM deliberately leaves out the actual factor, (I'll add that for ya), M1 performance isn’t because of ARM, it’s because, (thanks to the buying up the first 5nm lithography, nobody else has it).
M1 is the a beneficiary of 60 percent more transistors, (that's 6 billion more) over AMD double transistor count over Intel.
Apple m1, 16 billion transistors, Ryzen mobile 10 billion, Intel 8 billion!
Even with that extra 6 billion transistor with closer proximity for superior communication, the best ARM still can’t beat the best x86 mobile chips.
When AMD debuts with the same transistor count as M1 the same lithography, early samples demonstrate 25% gain in ipc and 40 percent overall performance!
AMD best mobile chip beats ARM best (M1) now, give it the same transistor count as M1 watch it absolutely smoke ARM.
That’s not superior anything but battery life.
Nor is ARM strictly “RISC”, it also sports CISC” nor do AMD/INTEL sport only” x86, they also incorporate RISC.
When AMD//Intel are on the same transistor count as the best ARM, that’s when you can actually discover which is superior architecture, not before.
Sadly for x86 players, by the time they get their first run of 5nm, Apple has already bought the first run of the next generation. (we don’t know if it’s 4 or 3nm)
Transistor count isn’t the only benefit to smaller lithography, closer transistors means lower latency, even if transistor count is the same, close transistor proximity improves performance.
Nobody should be thinking ARM is superior to x86 until they’re both on the same lithography.
Hopefully, the beginning of next year AMD will get 5nm before Apple gets 3nm, that’s when we find fair comparison.
Time tells
Last edited: