Apple To Discuss HyperTransport For Future Macs 86
macrealist writes "CNET is reporting that Apple will discuss the use of HyperTransport in Macs at the Developer's conference. The interesting thing is that the article claims that Apple is not likely to use hypertransport to link the CPU to the memory, but instead to link chipsets together because IBM would have to 'to adapt it to the Power architecture.' But according to arstechnica, the 970 does have a frontside bus that operates at similar speeds to Hypertransport."
Re:What about upgrades? (Score:2)
Re:What about upgrades? (Score:5, Informative)
I don't know where these rumors get started.
The combination of fat caches, low latency, and predictive fetching basically negates the memory bus bottleneck in the current-generation (MaxBus-based) Power Macs. Even in SIMD instances, the processor generally doesn't have to wait on data that much. (This is especially true in SIMD instances, because these are almost always sequential-read applications, which makes those fat caches and predictive fetching work up a sweat.) Consider Apple's AltiVec-optimized BLAST, for instance. It's 10X faster than BLAST on a Pentium 4. It's not memory-bound. It's compute-bound.
What's that famous Seymour Cray quote? "A supercomputer is a device for turning compute-bound tasks into I/O-bound tasks."
So if anybody produces PowerPC 970 upgrades with MaxBus interfaces, they're almost certain to be good buys. Unless they cost thousands of dollars, of course. A dual-1.8 GHz (pulled that number out of my ass, guys) Power Mac G5 (pulled that out of my ass, too) will be faster than a dual-1.8 GHz upgrade in a MaxBus G4, but it'll still be considerably faster than the G4 was originally.
Re:What about upgrades? (Score:2)
Re:What about upgrades? (Score:5, Interesting)
Hypertransport as I understand it... (Score:5, Interesting)
Re:Hypertransport as I understand it... (Score:5, Funny)
"We loves it we do! my precioussss"
(twelve powermacs and counting)
Re:Hypertransport as I understand it... (Score:4, Informative)
"Six GDA IP cores are available through IBM Blue Logic IP Collaboration Program including HyperTransport Cave, Tunnel, Host and Bridge, 10 Gigabit Ethernet MAC, and SPI4.2 link controller. Information on these IP cores is available on IBM and GDA web sites."
The above is from this news release from the Hypertransport Consortium http://www.hypertransport.org/pr_050503b.htm
Shadow
Mac Clusters? (Score:1)
Imagine a beow... of MACs?
Re:Mac Clusters? (Score:2)
Re:Mac Clusters? (Score:5, Informative)
<pedantically> I think that Apple has already developed a tried and true solution for external, non-ethernet-based, high-speed data transfer. It is called FireWire800 [apple.com].
Of course, an IP substack can be built on top of the FW, to have additional networking options. (Check out [google.com])</pedantically>
0.02â
Re:Mac Clusters? (Score:4, Interesting)
Re:Mac Clusters? (Score:4, Interesting)
Consider, for a moment, what Firewire is. It's a bus to transfer data from a chain of devices. This is why it supports speeds up to 800Mbps. Individual drives cannot utilize all of this speed by themselves. Therefore, unless you have multiple drives on the same bus (daisy chained), the speed is never fully utilized. That said, why would Apple WANT to use anything but IDE internally? IDE controllers are cheaper, and the IDE interface is plenty fast (100 Mbps) for any drive you can throw at it. In reality, Firewire drives are simply IDE drives with a new interface slapped on. It's cheaper for Apple to ship computers without that extra interface.
Plus, Apple would get a lot of flak for shipping computers with their proprietary standard. And to be honest, I would be one of those people dishing out the flak.
Re:Mac Clusters? (Score:2)
answer - none. All of them use IDE-Firewire bridges. So why add extra firewire inside the machine, and extra firewire on all the drives, when all that's really happening is they're using IDE anyway?
-- james
Re:Mac Clusters? (Score:3, Informative)
I'm not sure about the "sticking out of drive bays" bit, but Macs (and PCs) with FireWire support can generally mount internal firewire drives as well as external, which is what that's for. Since (as you go on to point out) many firewire drives are just IDE drives anyway, there's relatively little point to this.
In reality, Firewire drives are simply IDE drives with a new interface slapped on.
Well the cheap
Re:Mac Clusters? (Score:3, Informative)
Re:Mac Clusters? (Score:1)
Re:Mac Clusters? (Score:1)
Re:Mac Clusters? (Score:1)
Hypertransport could probably be used for inter-processor communication, if you had processors that didn't share their memory but were in the same case.
The key is that it is very fast, switchable, but can't cover great distances or deal with lots of addresses.
It would be important for things like very fast NICs that the old PCI bus just can't keep up with.
So if you want a Beowolf cluster communicating with 10 GigE NICs then hypertransport will help. But to really work well at those rates the appli
Re:Mac Clusters? (Score:1)
Done a long time ago (although not with hypertransport :)
daugerresearch [daugerresearch.com]
Apple feels like ... (Score:4, Interesting)
My next computer will be another powerbook, that's for sure... please continue to rock, Apple.
Re:Apple feels like ... (Score:5, Interesting)
Seriously.
Hear me out.
The Origin family of computers is basically the coolest thing out there. It's getting a little long in the tooth--the node-to-node interconnect is only 3.6 GB/s which isn't awe-inspiring any more--but it's got serious coolness.
IRIX is a great operating system, but it's also long in the tooth, and it's tied to MIPS processors.
The Origin design transcends the CPU, however; behold the Itanium Origin, which SGI calls the Altix. (Yes, it's hideously ugly. I'm getting to that.)
Basically Apple could buy SGI for a song right now. In the transaction, they'd get all of SGI's IP and sales contracts. They could continue operating SGI as an independent company to serve existing customers and customer bases--NASA, NOA, that sort of thing--but shut down everything that's not profitable. Like the nationwide sales channel, for example. What a cluster-fuck that is. Get existing big MIPS customers migrated over to Darwin on PowerPC or Darwin (Linux compatible) on Itanium (I see no good reason to keep going with Linux development) over the next five years or so.
The Origin 350 gets PowerPC processors and Mac OS X Server and becomes the next-generation XServe. Four processors (or maybe even eight; PowerPC's are small and cool) in two rack units, scalable with external 3.6 GB/s high-speed interconnect up to 32 (or maybe 64) processors. No clustering required; that's a single system image for those who need or want it. Ideal price point for an entry-level four-processor system: ~$5,000. Maybe get really cool and do a two-processor system for ~$3,000. Get in bed with Oracle and Sybase (moreso, I mean) and get their stuff running better on PowerPC than it does on any other platform. (Are databases candiates for vector optimizations?) If it makes sense, get big applications like Oracle and Sybase and Web Objects and BLAST and so on running on Itanium 2 versions of the Origin 350-based Xserve. If it makes sense. Keep the PowerPC available throughout the product line, however, for compatibility with existing software.
The Power Mac-based Xserve would remain as the lower-cost and non-scalable option.
If the demand ever exists for it, which it might someday, Apple can put PowerPC's in the Origin 3900 family (the SN-2 family) and sell supercomputers that scale from 4 to 1,024 processors and up. If the demand exists.
As for SGI's visualization systems, IR4 and IP, if it makes business sense, keep 'em around. If it doesn't, kill 'em.
Fuel? Kill it unless it can pay for itself. Same with Octane2. Same with Chimera, the upcoming Origin 300-inspired workstation product.
But the good news is that Apple would be able to take the truly cutting-edge work SGI has been basically pissing away, like GSN with ST for multi-gigabyte-per-second communication over TCP/IP and XFS/CXFS for high-performance direct-attach and fabric-attach storage and FailSafe for high availability and (1) continue developing them in a sort of skunkworks R&D environment, and (2) move them down to the servers and desktops as it becomes practical to do so.
SGI has seriously got it goin on, but they've lost all momentum in the marketplace and need a hero. Apple has momentum in the marketplace and would benefit from a leg up on some cool advanced technologies. Also, Apple has assloads of cache and valuation to trade with.
Please please please please please...
Why wouldn't Apple do it all themselves? (Score:5, Interesting)
Why would Apple buy SGI instead of doing it all themselves? Like you say, in the long term, the OS and the current hardware and the sales organization would be punted. With the 970, Apple looks to be be developing the guts of a strong workstation/server technology on their own. Buying the customers and transitioning over might be possible, but would the (checks NASDAQ.com) $241M be worth it? Wait, $241M? That's all for all of SGI? Well then!
A few things I could see Apple wanting out of SGI:
Maya. Buying that and making it Mac only would be in keeping with all of Apple's purchases lately. Make a free rendering client for Xserve. It'd be neat
The sales organization. Given what SGI is facing in the market place, that they're still around and showing some revenue suggestions SOMEONE is rising to the challenge there.
Existing customer base. Buy the accounts. Make an IRIX compatibility layer for MacOS X.
Engineers. Presumably they've still got some good folks there. Apple could certainly use all the talent they can get in UNIX code, hardware design, etcetera.
I don't see much long term value in SGI's existing products if Apple bought them though, and Apple is certainly willing to give up market share on other platforms in order to make a package Mac-only.
Still, given that the whole company is only $241M, it seems like there might be something worth cherry-picking there.
Re:Why wouldn't Apple do it all themselves? (Score:4, Interesting)
Because the stuff I talked about is a shitload of incredibly complicated hardware design IP. Apple would need years and a whole new staff of scientists and engineers to build it themselves from scratch.
With the 970, Apple looks to be be developing the guts of a strong workstation/server technology on their own.
Chips do not a server make. How you gonna connect 1,024 of those 970's together? Hypertransport? Pff.
Maya.
Maybe, but AW is a wholly owned subsidiary. They might get spun off.
The sales organization.
A thousand times no. SGI's sales organization sucks. Their sales PEOPLE are okay, sometimes, but the organization sucks.
Existing customer base.
Yes. Although IRIX compatability for Darwin ain't gonna happen. Better to migrate those customers to PowerPC or Itanium.
Engineers.
Most of the truly great SGI folks fled a long time ago, with some notable exceptions. But yes, there are some good folks in Mt. View.
I don't see much long term value in SGI's existing products
The Origin architecture is the key to a mid-range scalable server. You've got your 2-p 1-U units, and you've got your 128-p racks. Not much in between but SGI's Origin 350. Sun? Yeah, if you wanna pay through the nose and end up with a system that can't scale. The great thing about the Origin 350 is that you buy one (4 processors) and start using it. You decide you need more oomph, so you buy another one (4 processors) and hook 'em together, and you've got 8 processors. No new software, no migration, just plug it in and go. (Reboot required.)
Apple NEEDS this scalable technology to give them a compelling story in the midrange server market. Combine it with the power of UNIX (TOG's bitching notwithstanding) and the wonderment of Mac OS X Server and you've got yourself a product line.
Apple is certainly willing to give up market share on other platforms in order to make a package Mac-only.
Apple is INCREASING market share (or trying to, and succeeding in large part) by taking top-of-the-line UNIX products like Shake and making them Mac OS X-only. It's a good strategy.
Re:Why wouldn't Apple do it all themselves? (Score:2)
Re:Apple feels like ... (Score:4, Funny)
And then put it on iTMS for $0.99?
Re:Apple feels like ... (Score:1, Interesting)
Apple has instead been hiring the top SGI Enginersm and Marketing people from SGI, in fact the product manager for Apples XServe was the top guy at SGI before comming to Apple, and I expect to see some Government sales go to Apple now.
Bishop do
Re:Apple feels like ... (Score:2)
So that means that Apple could buy SGI more than 200 000 times [apple.com]
Okay. Worst pun ever. Mod me down!
Technically Cool Hardware, or Fast and Cheap? (Score:5, Insightful)
Re:Technically Cool Hardware, or Fast and Cheap? (Score:2, Insightful)
No, the most important performance factor for me is GUI responsiveness -- and that's where Apple really isn't cutting it these days. Neither are the user-friendly Linux desktop apps.
Re:Technically Cool Hardware, or Fast and Cheap? (Score:2, Interesting)
I have a dual-processor 1 GHz Power Mac with a Radeon 9000 card. I run lots of stuff, from basic Internet tasks with Safari and Mail.app to Project Builder/Interface Builder to InDesign and Photoshop to Virtual PC to Microsoft Office (when I have to). I find the UI to be quite responsive. I never have to wait on it, except when I'm running VPC, but that's to be expected.
If you're running Cocoa/Java applications, expect the GUI to be slugg
Re:Technically Cool Hardware, or Fast and Cheap? (Score:1)
I'm talking about extreme sluggishness of Office on a 550 MHz TiBook. Java apps are, as you say, awful but that has to be expected.
Re:Technically Cool Hardware, or Fast and Cheap? (Score:3, Insightful)
Using the GUI in the open/close is slow simply because of design and not pausing or slowdowns. It takes to long to get to where you want to get. Default Folder helps a bit
Re:Technically Cool Hardware, or Fast and Cheap? (Score:3, Insightful)
The alternatives to the Finder really aren't much better. I've used all of
Re:Technically Cool Hardware, or Fast and Cheap? (Score:2)
The advantage to context menus is that you don't have to mouse all the way to the top of the screen to access a menu to do something. Thus there is less movement. Further there is a visual clue since it is an object you are clicking on. Right clicking or Ctrl-clicking provides a list of actions you can do with that object. Since there is a menu, this clearly is visual.
While you may not like context menus clearly many people do, just as they like scroll
Re:Technically Cool Hardware, or Fast and Cheap? (Score:1)
The only thing that pisses me off about my Macs is the damned one button mouse. On my iMac its not a problem, I just plug in a new one and move along, but on my 12" Powerbook, its a pain.
Quartz Extreme? (Score:2)
Switch? (Score:5, Interesting)
I like the idea that Mac develops the hardware and software together under one roof. I think following the process from all angles like that would make for a better product. It's a better philosophy than the Windows/PC mish-mash way of thinking, primarily because no person sees all ends of the production for PC, and you can bet that there are quality issues with computability under PC that just aren't there with Apple (or at least that is what one would expect). So looking at Hyper Transport, at this stage, I'm a tad leery of it because it didn't come from Apple. I'm worried that it might have some kind of negative impact on the technology.
The necessary question is; is this going to be the next evolutionary step for Apple, or is it just an added hardware feature that is relatively minor?
Re:Switch? (Score:5, Insightful)
Apple contracts with dozens of commodity hardware manufactures to build components. Rebranding other-manufacturer items with the Apple logo doesn't make them "come from Apple."
Is this issue really the show-stopper preventing you from using Macs? Seems kind of odd.Re:Switch? (Score:1)
The only thing stopping me currently is a lack of funds, and an uneasy feeling that Apple might not be right for me. I've tried it at college a few times and the graphics generated by Photo Shop for Mac are better than PC for some reason (I have no idea why since it's the same product, different platform).
After cash is concerned, this is going to sound so shallow, but, Mac has typically been a chick thing. Girls ar
Re:Switch? (Score:5, Informative)
Yes, the entire computer is skinnable, user apps included. Now this doesn't include classic apps (which you won't be using much of), unix apps (which don't use NIBS) and monolithic code not in a bundle (like RealBasic). For the rest of the 90% of Mac apps, you can really mod to your heart's delight. Most people don't do this because they *like* the way Apple makes everything work with everything else. But if that's what floats your boat...
Re:Switch? (Score:2, Informative)
Yes, you can. Basically any USB mouse should work. I use an Intellimouse myself - it needs MS drivers to recognize the 4th and 5th button, but the three standard buttons and the scrollwheel work even without those.
Re:Switch? (Score:2, Interesting)
Try the factory-supplied mouse for a month, or even two. You'll be a little frustrated at first because you'll be used to having to do that right-click thing, but learn to us
Re:Switch? (Score:2, Insightful)
I worked in the graphic arts for ten years. Here's the honest-to-god truth: people who use Macs for art applications like Photoshop and Illustrator are generally more talented than people who use PC's.
Yes, I'm painting with a broad brush. But when you pick a random piece of work generated on a Mac, you're going to find that it looks better than a random piece of work generated on a PC.
Cause and effect? No idea. I'm just sayin
Re:Switch? (Score:3, Informative)
This page [mac.com] has a nice selection of themes for Mac OS X. The Rhapsodized and QNX themes are the best of them, IMHO.
bacchusrx.
Re:Switch? (Score:2, Interesting)
The first computer I ever owned was a C64 (yes, I'm that old).
When the first MacIntosh was out, I wanted one. Being a teenager and being broke, I ended up with an Atari ST instead.
Being a student, I was equally broke and happy to replace the ST with old PC boxes I did inherit from their previous users (usually my dad), going 386, 486, P133, P200. Each machine I got was at least 2 years
Re:Switch? (Score:2)
Seriously, the game situation for Macs isn't nearly as bad as it seems. In fact, I was a little surprised the other night going through my CD envelopes and seeing that nearly more than half of my Mac software was games.
The nice thing about gaming on Macs is that if you do find a game, it's usually a tried and true title. Wel
Re:Switch? (Score:1)
You dont' know how true that is. All the good PC games are being ported to Xbox, since it is so easy from a developers standpoint, due to the HDD and Nvidia GPU. Plus UT 2003 is out or coming out for Mac, and I heard Doom 3 will come out on Mac.
Re:Switch? (Score:5, Informative)
Apple is a member of the HyperTransport Consortium. [hypertransport.org] They have a hand in the development of the technology.
Re:Switch? (Score:3, Informative)
Re:Switch? (Score:4, Interesting)
Apple has a long history of choosing technology that it thought was better. Who had SCSI on their full PC line in the mid-80s (which was not invented by Apple) besides Apple? Eventually IDE got good enough that SCSI didn't fit the definition of "better, if more expensive" for their user base and just got to be more expensive so they switched.
A fast system bus is *the* major issue with Apple hardware. The G4 isn't that bad even today (when it's 2x clocked by Intel processors) when it's not starved for internal bandwidth but Apple's current MB designs *do* starve it.
With a faster chip (clock speed) running at 64bits (very good for complex processing) that doesn't have nearly the speed penalty of Intel's 64 bit solution running 32 bit code Apple's going to be doing a lot better than a minor hardware upgrade.
Whether they announce at WWDC or hold off is another question.
Re:Switch? (Score:1)
I guess speed could be in the mind's eye of the beholden?
Re:Switch? (Score:2)
Re:Switch? (Score:2)
Hyper Transport seems to have been collaboratively developed by Apple and others, just as happened five years or so ago with Open Firmware [firmworks.com]. Just as Apple, AMD, and Transmeta all seem to be going in on Hyper Transport as an interconnect between hardware compoinents, Apple [apple.com], Sun [sun.com], and possibly others have been using OF as the "no
CHRP anyone? (Score:5, Interesting)
And yes i will be selling both my macs to get a ppc970 the day they come out.
Re:CHRP anyone? (Score:3, Funny)
I restate: "Asshats."
Re:CHRP anyone? (Score:4, Funny)
A 2Ghz, DDR-compatible CPU appeared on one of their roadmaps about a week ago. Since it just appeared, that means it will be at least 2 years. 2005 for 2 Ghz! If Apple is still using Motorola CPUs by then, I'll personally drive up to Cupertino and put the last piles of earth on top of their grave.
Arglll (Score:5, Informative)
First of all: A "frontside bus that operates at similar speeds to Hypertransport" most likely isn't Hypertransport - just like a car with performance similar to a Porsche isn't a Porsche. So you can't just hook up a 970 (or POWER/PowerPC) to a Hypertransport link.
Furthermore, linking a CPU to main memory via Hypertransport (a point-to-point link) means you can't share the memory with other CPUs (unless you have dual-ported RAM - uhh, yeah, good luck with that plan).
Re:Arglll (Score:2, Interesting)
Re:Arglll (Score:2)
Errm, yes. That is part of the chipset, which is linked together by Hypertransport - just like the article says.
Re:Arglll (Score:2, Informative)
No. At least it is not part of any announced chipsets. What you need is a Hypertransport to ElasticIO bridge. The latter is IBM's CPU-to-CPU bus that has very similiar bandwidth but is not HT. It is likely that this "bridge" will just be the "Northgate" chip for the 970 which also interfaces to memory. You probably don't want to insert the latency of having to switch from ElasticIO , to HT, to DDR on your path to main memory. The parts which hook to the stardard PC I/O (A
Re:Arglll (Score:2)
Oh no, wait, all this talk about Hypertransport is proof that Apple will not use the 970 but the Opteron after all.
Least surprising news ever. (Score:2)
This is great!! (Score:2, Funny)
What I read... (Score:3, Informative)
The new city has main roads that are rings (one or two), rather than grids where the government is focused in one area....industial parks in another.,...and families and fun parks all bunched up in yet another sequestered section. These ring roads serve to generally define city structure.
The dispersed control of new, very large cities is only possible by taking advantage of modern communication and thoughtful agreement to locallized authority.
When city government sits on a throne, and nothing happens without strict review and approval, a city can become bound up in red tape and suffer accordingly.
By applying this logic to chip layouts, the goal of rapid and coordinated decision making can become a more rapid and efficient process.
Let go of the frontside bus logic for a moment or two, and you'll perhaps see how this can be a leap forward, as opposed to an operational liability
ok, i am considering an Apple; advice? (Score:1, Interesting)
For example, if I don't buy a SuperDrive-equipped box now, can I add one later? Are there any other things like this I need to be careful for that are "missing" from lower models? What are the architectural differences between the iMac/PowerMac and iBook/PowerBook? Is the rule about Quartz Extreme acceleration as simple as, "G3's don't support it, G4's do"? Wh
Re:ok, i am considering an Apple; advice? (Score:4, Informative)
Yes: internal or external. Careful about support though: go to Apple's search page [apple.com] and look for DVD-RW. These are supported officially, others are not.
Are there any other things like this I need to be careful for that are "missing" from lower models?
The bottom of the PowerMac line is missing, well, nothing. All PowerMacs have an AGP 4X slot, plus 3 (or four?) PCI slots, so you basically can add what you want.
Connectivity: an airport extreme slot, bluetooth-ready, FireWire 800, USB 2 (the OS doesn't support it yet, but apparently there are some hacks that work), Gigabit ethernet. You can add Fiber Channel (2 Gb/sec) too.
Expandability: you can add 3 more hard disks (RAID support - I think), a second optical drive, go up to 2 GB RAM (maybe 4 GB with 1 GB sticks?). Some vendors sell G4 upgrades (some currently manage to get their Cube at 1,42 GhZ), but at a price.
You can change the video card; all Powermacs come with either Radeon 9000 (dual display, one ADC, one VGA/DVI, adaptor included), 9700, or GeForce 4 Titanium.
What is good about Powermacs (in my opinion) is not that you can upgrade like you would on any PC, but that even without upgrading, well, it still works after 15 years (I have relatives that type text on a Mac Classic / Apple printer). When I replace my G4 (in 3/4 years), it still will do a very sweet SSH / web / email server (maybe Darwin or Linux)
What are the architectural differences between the iMac/PowerMac and iBook/PowerBook?
iMac still has no DDR (!), its SDRAM. You can upgrade to 1 GB. It comes with a SOLDERED video card (GeForce 4 MX). Don't ever think of upgrading anything on an iMac (except RAM and hard disk). You can have airport or airport extreme (depending on the models). But my 3-year old iMac (g3 400 MhZ) runs Mac OS X fine, does Word, internet and email jobs for my dad.
iBook is still G3, maxes out at 640 MB of RAM. I consider it to be a cheap laptop, made for students (well, it manages to do serious DV video editing, so I presume it's powerful enough).
I never really used a powerbook, but they look like sweet machines. DDR, Radeon Mobility 9000, VGA and SVideo out.
As for speed: I have a dual 1 GhZ, it's fast enough. I don't know about your business-type apps (if this means word-processor, well, any mac is enough; if it means Oracle database, well, I just don't know
Quartz Extreme: it's not about the processor (only iBook still has G3 anyways.), but about the graphic cards: you need more than 16 MB VRAM to enable Quartz Extreme (so all the current line, including iBooks, support it).
Anyway: WWDC is REALLY close now, you should wait for the event and decide wether you can wait for the 64 bit processors and huge FSB, or pick up the discounted G4's Apple is sure to sell right before introducing the 970's to the market.
Here, hope it helped. Applestore / Knowledge base webpages should help, or you could check www.xlr8yourmac.com
Re:ok, i am considering an Apple; advice? (Score:2)
If you're going to use VPC on a regular basis, I'd say to go for a G4-based system, as VPC is G4/multiproceesor/Altivec-optimized (though I've never really tried VPC on a G3, maybe it runs just fine..
Decisions... (Score:2)
ARRRRRRGHHH!!!!!