Slashdot Mirror
Speculations Intel's Next Generation
An anonymous reader writes "The Inquirer speculates about the next generation Intel chip. It's low power, 64 bit, multi core (up to 16?) and the real reason for the Apple switch."
← Back to Stories (view on slashdot.org)
Correction: 65 bits. Twice as fast as 64 bits.
Probably will feature an android, a Klingon, and a balding captain.
I'll speculate that Intel is going to create a new 128-bit proc composed entirely of turtles. Does that make me slashdot-worthy?
If they're announcing an archtecture this radical at next week's IDF, what are the chances that it will be available and running well in time for Apple's announced timeline for desktops?
Or is Apple going to sell a lesser version first, in which case why haven't they already switched over to selling it to early adopters already. Yes there really are people who buy systems and wait for the applications to arrive later.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
To me that sounds a lot like Sun's Niagara box. Huge CMT box (8 cores, 4 threads each, 32 way box). With power consumption around 65watts, but faster than 4way Xeon processors and probably more like an 8way depending on application. Intel probably is moving to something similar, maybe not quite that many cores and threads.
This is the same thing Intel did to HP who walked away from PA/RISC, and to SGI who walked away from MIPS, and to Compaq/DEC who walked away from Alpha --- so they turned from the leaders in 64-bit computing to resellers of wintel.
Hey, if it worked last time, let's try it again; and maybe the rest of the 64-bit competitors'll give up.
We'll know more when IDF arrives. Until then its just stuff written to try and hit a bullseye in the dark. Which seems to be everywhere nowadays, Dvorak, The Inq, even my fateful Ars is getting bit by the bug that says every action by anyone in the tech industry must be expounded on in a multipage article worth of /. and the ad revenue it brings..
The Doormat
If you're not outraged, then you're not paying attention.
If a VLIW X86 processor had a "native" mode, one would have to wonder if Apple's Rosetta technology could compile directly to it instead of X86. I mean, it would seem dumb to JIT-compile to X86, which in turn is translated to VLIW.
Any sect, cult, or religion will legislate its creed into law if it acquires the political power to do so.
I suspect Apple's switch wasn't because of any cool chip (it'd be ridiculous to think they are getting intel chips that no PC maker will have access to) but simply because it's one less defensive front - they don't have to worry about getting chips that are competitive anymore, which was getting a problem with PPC as well as the all important Notebook chips - IBM simply wasn't offering anymore competitive PPC solutions.
It's one less thing to defend.
Back when Apple first introduced PPC (1994?), they were hyping it throughout because that was one of the few real tangible differences they could tout - pre-OSX Mac was buggy and unstable single-threaded OS while Microsoft had at least NT technology.
Now OS X pretty much rocks and they still have their excellent hardware integration - they don't need a different chip to differentiate them - OSX is their added value.
...The Farmer's Almanac speculates on the next generation "Beefalo" chip: Running from Longhorns daily into a pasture near you, the new "Beefalo" chip (tm) will multi-thread faster spreading odor and increased fertilization rate. Cores have been increased to 8 semi-solid, virtually discrete units that may be tracked onto the North bridge (If you don't wipe your boot sectors before then). Video processing speed will see a marked increase, although cooling remains a concern for these new chips...
Take the 90-Day Challenge! http://rwmurker.bodybyvi.com/
Who told me? The mold that lives in the back of the fridge in the second snack room on the 7th floor of the 4th building at their 2nd site.
Bwhahahahahahahaha.
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
Apple is not that spectacular in terms of choosing chips for performance, from their past history. M68k: good chip, but it was suffering from old age when they moved to PowerPC. (They could have moved to x86, arm, or other processor at that time.) Now, they announce they are moving to Intel, and suddenly Intel has some super-duper chip up their sleeve? I don't think so.
The article starts from that basis and works up to Intel has some super-killer CPU.
Despite the amount of hype surrounding dual-core, unless you massively change software (likely to happen eventually) to support SMP, things go slower on dual-cores than single core processors, if the dual-cores are clocked lower (Intel's current chips). What the article proposes is to duplicate the mistakes Intel has made with Itanium. (It was announced a decade ago. (If not, near enough to count.))
Itanium 1 stripped out all the branch prediction, and similar things, relying on the compilers to do it. The result was that it got soundly thrashed by other 64-bit archs.
So why does Itanium 2 not suck nearly as bad? HP's engineers mostly went back and put all that stuff back IN, because compilers, and code translators are still (with a very very few exceptions, I can think of 2 (one, FX!32, mentioned in the article)) very slow. Even FX!32's speed wasn't due to the speed of translation, it was due to the huge (at the time) performance of the underlying alphas. Sure, it may have been faster than the fastest x86 hardware implementation, but it was still quite slow compared to the native speed of the chip it was on.
So the article speculates that Intel is indeed going to repeat the mistakes of the past, mistakes that *only* came to market because a) Intel has money b)Intel has pride (oh and c)got others to wipe themselves out... except IBM.) I would think Intel would learn from it's mistakes. Right now they should notice that a)processors can't be fabbed right now to work at ~4GB reliably and they are really hot. b)Going the opposite route of improving IPC almost entirely (IA-64s are not low-powered, nor cheap). Instead they should work on the in-between, which they (again due to Intel having tons of money) have in the form of the Pentium M.
A: A bunch of slashdotters doing the same thing.
Read my blog: HansMast.com
If Apple is interested in the server market, severing ties with IBM is not the smartest move.
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
This article was written by Nicholas Blachford, the same fool who tried to analyze the Cell processor of the PS3 and described it as a supercomputer on a desk while not understanding a single thing about it.
Seriously, it's worth a read for the laugh, but there's nothing worth believing in it, this guy doesn't know what he's talking about.
There's a better explanation of why the Inq article's speculation is bogus here:
i on=detail&PostNum=3655&Thread=3&entryID=55310&room ID=11
http://www.realworldtech.com/forums/index.cfm?act
AGP, by it's design, can only have one per machine.
It was an add-on specifically designed for the pattern of usage that video cards perform - lots of data out, and short requests in.
It was a patch to get us by until the "next PCI" came along - but AGP's great performance was also the reason it's taken so long to get PCI Express going; not a lot of demand for something we don't really need. Old PCI slots still provide ample bandwidth for most other types of expansion cards and on the server side you had 64-bit PCI/PCI-X.
Of course, we still needed PCIe, but it hasn't been a big push. Now, with the dual-video board thing happening, it's definitely helped push the bus into the mainstream.
As far as the changes in CPU slots, well, I agree to a point. While I believe that both Intel and AMD could have done more to keep slot changes to a minimum, a lot of times the chip-set changes along with the CPU requiring a new board anyways. So, why not upgrade the CPU slot to accommodate the new data patterns of the new architecture?
I do feel like I own hardware. Software, on the other hand..
- It's not the Macs I hate. It's Digg users. -
Lame Megafuana died out....they couldn't compete Turtles are still here...the superior solution!
Blar.
Well, OS X comes up real short in most server benchmarks, so I hope it's not that. Anyway, half of what Apple sells is 'coolness', and server farms don't go for that, too much.
"A language that doesn't affect the way you think about programming, is not worth knowing" - Alan Perlis
With this tiny font, I couldn't make out what the word there was, but after reaching for my skeptacles it was all clear. Truly the wealth of alternative spellings on Slashdot never ceases to surprise. I'm not even a native English speaker.
Escher was the first MC and Giger invented the HR department.
The interconnect for intells xeon servers is really poor and at high loads all the processors compete for access to the shared bus and memory. This means it doesnt scale worth a darn. You have diminishing returns for each processor something along the lines of:
1 xeon = 100%
2 xeon = 140%
3 xeon = 160%
4 xeon = 170%
Wheres with the amd opteron with hyperTransport interconnect the processors dont have to fight for resources. And performance scales much better along the lines of:
1 opteron = 100%
2 opteron = 180%
3 opteron = 250%
4 opteron = 310%
We have the best government that money can buy.
In any case, the fact that everyone wants to jump to 64 without testing the waters very carefully first is seriously foolish. I know I'm not the only one who feels this way -- Microsoft's Windows speech recognition subsystem refuses to run on any 64 bit architecture unless all of the OS and applications are strapped to 32 bit mode.
This is possibly worse than five years ago when people were paying absurd premiums to go from 800 MHz to 1.3 Ghz with RAM speeds stagnant. At least then you got something more from algorithms which weren't memory access-bound. From 32 to 64 is a significant step backwards in many cases.