Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Will Intel share PPC's fate?
- Thread starter RedCroissant
- Start date
- Sort by reaction score
And therein lies the problem: They have shown that they have zero desire to do so. A few "old Mac-heads" complaining won't magically convince them to come out with a desktop or portable-computer-appropriate CPU again.
IBM gets a lot of money from Main Frames for corporations. It is not lucrative to start focusing on making a laptop/desktop Power 8 processor.
It's possible, but it isn't going to happen in the near future. The Star Trek project actually had some fruit to it (this was back in the days of OS 7) when it came to creating an Intel port. The same happened with Rhapsody. Then, we found out about Marklar...
Personally, it wouldn't surprise me if Apple has secret ports of the current Mac OS for ARM or even true PowerPC, just in case either architecture becomes more capable than Intel. BUT...going back to PPC or swapping to ARM would carry issues of its own, like the compatibility that Intel Macs offer (i.e. WINE).
Personally, it wouldn't surprise me if Apple has secret ports of the current Mac OS for ARM or even true PowerPC, just in case either architecture becomes more capable than Intel. BUT...going back to PPC or swapping to ARM would carry issues of its own, like the compatibility that Intel Macs offer (i.e. WINE).
ARM is PowerPC-like (i.e. Apple A4). True PowerPC is what we've been seeing in Macs until the Intel switch.
ARM is not at all "PowerPC-like" any more than MIPS or SPARC are PowerPC-like. They are all "RISC" processors, yes. But that's the only similarity.
"PowerPC" has a specific meaning: A processor that uses the PowerPC Instruction Set Architecture. ARM does not use the PowerPC ISA. It uses the ARM ISA.
And even the latest POWER/PowerPC processors (where "latest" goes back to the G3,) use the same basic idea as the x86 line has used since Pentium Pro - the "core instruction set" is only used at the highest level - the actual low-level execution engines use so-called micro-ops that break the higher-level instruction set down in to smaller pieces to process.
I don't believe 64-bit ARM chips themselves will be what threatens Intel, but rather the shifting consumer market. As mobile devices like smartphones and tablets become more powerful and versatile, more and more people will use them in place of conventional computers. Eventually the market for powerful x86 machines will dwindle to certain niches like software development and 3D rendering
Bingo.
It's not that we'll see ARM on the desktop, it's that we'll see devices that are currently ARM become powerful enough that they REPLACE desktops.
Heck, I had a "high-power" desktop plus a "medium power" laptop for a long time.
Then I got an iPad. It completely replaced my laptop, and for enough uses replaced my desktop that I sold my high-power desktop and went back to using an older/slower one. (It also helped that I could never hack OS X on to my high-power desktop, and I really don't like using Windows at home.)
I still have a "workstation" system for when I really need the extra power, but I rarely even turn it on any more.
Wow... You honestly don't have a clue. Do you?
ARM != POWERPC/POWER or anything else.
POWERPC/POWER != ARM or anything else.
ARM = ARM
POWERPC/POWER = POWERPC/POWER
Last edited:
Same here. The ibook just feels for me like an iPad with keyboard lol at least iPad 1 and I rarely turn my windows computer on
Did you see the benchmarks for the A7 processor?
http://anandtech.com/show/7335/the-iphone-5s-review
Impressive, especially since the 1.3ghz A7 only uses ~2w or less of power.
If next year's A8 doubles performance again, Apple's processor will be within striking distance of the Intel haswell CPUs in performance used in the MacBook Air and using less than 1/5 of the power.
No, they'll be "within striking distance" of the lowest-end Haswells, which even Apple doesn't use. They'll be roughly at par with Intel's Celeron and Pentium branded chips.
And you're also then assuming that Intel won't have performance increases as well.
There will be a point at which they can no longer double. There will be a point at which the generation-over-generation performance increases will be roughly equal to Intel's. Yes, they may very well be able to develop an ARM-based chip that performs on par with Intel chips, but they would likely draw the same power as well.
Not true that if doubling performance won't reach Haswell numbers at least according to Geekbench3 results.
The 1.3ghz A7 scores about 2550 in multicore result (64 bit)
The base 1.3ghz Haswell in the MacBook Air is at 5000 (64bit)
The 1.7ghz i7 is at about 6200.
Yeah it is a big if that Apple can double performance again next year...
As an FYI, the Apple A7 already outperforms the 2010 era macbook airs with Core 2 duo chips..that's quite amazing considering that we are talking about a 4" smartphone.
Intel's performance increases have been very mediocre ever since Sandy Bridge, so I don't expect a big increase with Broadwell.
I highly doubt that A7 approaches Core2duo performance. You would have to go further back. It's still impressive, but what makes it impressive is what you can do from a handheld device today. It might be closer in productivity to Core2duo hardware even if its specs do not add up.
If we are to believe Geekbench3 results, performance per MHz wise, it kind of did, actually.
iPhone6,2 vs MacBook Pro (13-inch Early 2010)
Edit: woops, sorry- wrong link , just fixed it
Last edited:
If we are to believe Geekbench3 results, performance per MHz wise, it kind of did, actually.
If we, the people in the PowerPC section, know anything, it's that benchmarks are not a good measurement of real world power. If you were to transcode a vob file to a h264 file without using a hardware encorder, the Core2Duo would beat the A7 by a few hours.
Geekbench results agree with that, since that specific Core 2 delivers quite a bit more punch than the A7...but running close to 2x the clock speed.
And Anand agrees too:The iPhone 5s Review
If we are to believe Geekbench3 results, performance per MHz wise, it kind of did, actually.
And Anand agrees too:The iPhone 5s Review
The A7 actually scored higher running at just over half the frequency.
Amazing performance given the A7 uses ~2W. I think that particular Core 2 Duo was a 25W chip.
Graphics performance is neck to neck as well according to gfxbench:
http://gfxbench.com/compare.jsp?D1=...MacBook+Pro+2010+(NVidia+GeForce+320M)&cols=2
Last edited:
Interesting. If that is true, may be (MAY BE) we would see a Macbook Air powered by an A7 in the future?
I think they're waiting until they get to the A10 and then they'll switch to roman numerals. OS X on Apple's AX. Synergy, people. .....synergy........[whispered]
I'd doubt they'd put it in their entire Mac line. Remember what happened with all the software and games last time? I'd hate that to happen again.
You are looking at the overall score, which includes AES and SHA encryption- ARM V8 includes special instructions for that, and is naturally much faster. I read somewhere that Primatelabs are considering to separate them from the final score in the future.
Here is some better comparison point:
MacBook Air (11-inch Late 2010) vs iPhone6,1
1.3 vs 1.6GHz, no turbo-crap.Both run 64-bit operating systems, both Darwin-based. Yeah, I know it's not that simple, but it's as close as possible.
If we look at the specific tests, Cyclone is still faster in most of them. In Twofish (integer) and Mandelbrot (FP), C2Duo still has the lead.
Absolutely- the idea that a phone have performance comparable to a 2010 notebook is amazing, especially considering that one of the main feature/drawback, (depending on how you look at it) of the IPhone is that it's kind of small-ish. But let's be more realistic here- it is slightly outperforming 10-Watt, 45nm cpu.
More interesting is how it stands against the latest Atom:
iPhone6,1 vs Dell Inc. Venue 11 Pro 5130
Single-core performance is what is interesting here- in integer tests it ranges from equal to better for Cyclone. In floating point, Cyclone has much bigger advantage, and in N-Body and Ray Trace it's ahead even in absolute performance.
A couple of notes- different operating systems, Windows is 32-bit. Quad core Bay Trail, quite possible that during the whole test it was turbo-ing close to the max possible 2.4 GHz. Should I say that you won't find that in a phone? Nope, most likely won't go in a phablet too.
No. I don't see the point in that. It's a phone chip. A-something-else, however- Why not?