G5 chip will ship in august
It's difficult to have any information from Motorola concerning their roadmap, because of the culture of secret from Apple.
However, the AMD roadmap is not secret.
There will be a new ATHLON based upon 0,13 micron process in may, and in august AMD will introduce a new ATHLON (codename Barton) on SOI 0,13 micron process.
Considering the fact that AMD and MOT cooperation for FAb process, the release of the G5 based upon 0,13 SOI will appear the same time.
The only problem with this kind of prediction, is that it's mean that the release of the appolo chip will occur in august too. So there will be the G5 for the powermac line and the G4 appolo for TI books and I mac. During this time we can only see a small speedbump of the G4 up to 933 mhz - 1000 mhz.
I expect that i am wrong, and that we will soon be able to see SOI chip soon.
However, the AMD roadmap is not secret.
There will be a new ATHLON based upon 0,13 micron process in may, and in august AMD will introduce a new ATHLON (codename Barton) on SOI 0,13 micron process.
Considering the fact that AMD and MOT cooperation for FAb process, the release of the G5 based upon 0,13 SOI will appear the same time.
The only problem with this kind of prediction, is that it's mean that the release of the appolo chip will occur in august too. So there will be the G5 for the powermac line and the G4 appolo for TI books and I mac. During this time we can only see a small speedbump of the G4 up to 933 mhz - 1000 mhz.
I expect that i am wrong, and that we will soon be able to see SOI chip soon.
Comments
You may be right about August though, based on Fred's speech yesterday.
An 800 MHz G4 is just about on par with a 2 GHz P4. When AMD makes a 2 GHz chip, the Power Mac will be behind.
I sincerely hpe the rumors of the 1.6 GHZ G4 are valid.
[quote]Originally posted by Programmer:
<strong>You may be right about August though, based on Fred's speech yesterday.</strong><hr></blockquote>What about the conference call yesterday makes you think about August?
<strong>I don't think the sharing between Mot and AMD is nearly as widespread as everybody seems to think. Fab sharing was considered once a couple of years ago, but I don't know that there real evidence of cooperation since then.
You may be right about August though, based on Fred's speech yesterday.</strong><hr></blockquote>
He's probably referring to their joint development of the HiPerMOS 7 process, rather than mere sharing of fab facilities (AMD is flat out as i hear
<strong>
He's probably referring to their joint development of the HiPerMOS 7 process, rather than mere sharing of fab facilities (AMD is flat out as i hear
Yes, that's clearly what he was refering to... but what hard info is it based on?
<strong>
Yes, that's clearly what he was refering to... but what hard info is it based on?</strong><hr></blockquote>
Having worked for one of the above mentioned companies I can CONFIRM as it were that AMD and Motorola do in fact share R&D on new process technology including copper, HiP7 (130nm), HiP8 (90nm), and SOI. They do not however share fab space, they produce their own products in their own facilities. This is public knowledge however. Below is an excerpt from the press release:
[quote]Under the terms of the planned seven-year agreement, Motorola will license its current copper interconnect technology and High Performance Logic Process (HiPerMOS) to AMD. The companies will collaborate on the development of future logic process technology platforms featuring copper interconnects. These advanced process technology platforms will be used to build powerful microprocessors: PowerPC? processors for Motorola and K86? Microsoft Windows compatible processors for AMD. Microprocessors featuring clock speeds of one gigahertz and higher will require copper interconnect technology.
In turn, AMD will license its flash-memory technology to Motorola, and the two companies will co-develop process technology for embedded flash. AMD is a leading supplier of low voltage, high density stand-alone flash devices; Motorola has demonstrated leadership in flash memory for embedded applications. Motorola expects the agreement to accelerate its embedded flash leadership and enhance its position as the world's number one supplier of embedded solutions. Motorola also will receive rights to use certain networking technology from AMD. <hr></blockquote>
AMD is not moving to SOI and .13 until the second half of this year because they don't need to move to it any time sooner. Motorola could plausibly be moving to .13 and SOI very soon, but that doesn't mean you'll see a G5 very soon. More likely a .13 SOI version of the G4 first.
[ 01-17-2002: Message edited by: Eskimo ]</p>
<strong>
An 800 MHz G4 is just about on par with a 2 GHz P4. When AMD makes a 2 GHz chip, the Power Mac will be behind.
</strong><hr></blockquote>
Hmmm... Now i hate to rag on apple in any fassion, i love macs and for as long as apple is in business, my desktop and or laptop will be a mac, but to say that they aren't behind the PC world as far as high-end hardware goes is rediculous. The things they keep in are a few Altivec / MP Photoshop & final cut pro tasks. (im sure there are more, im exagerating but you get the point).
Im am truely worried for apple now as a stock holder seeing as the new imac is 10x the deal compared to a powermac anywhere near its price.
I like the exictment of apple's secrecy but cummon, time to get the new powermac out the door.
<strong>
Having worked for one of the above mentioned companies I can CONFIRM as it were that AMD and Motorola do in fact share R&D on new process technology including copper, HiP7 (130nm), HiP8 (90nm), and SOI. They do not however share fab space, they produce their own products in their own facilities. This is public knowledge however. Below is an excerpt from the press release:
</strong><hr></blockquote>
Thank you, Eskimo, I hadn't seen that before. It is interesting to note that AMD is using Moto's process technology -- so clearly the deficiences in the G4's clock rate are due to design issues in the G4. After all, the Athlon is on the same process and running @ 1.6 GHz.
<strong>I think tha it's just fine that every produt except the iBook has a G4 chip. Now if only the PowerMac was a gigahertz Mac. Or faster. If Apple is going to diversify, it has to happen this year or the speedup MUST OCCUR SOON.
An 800 MHz G4 is just about on par with a 2 GHz P4. When AMD makes a 2 GHz chip, the Power Mac will be behind.
I sincerely hpe the rumors of the 1.6 GHZ G4 are valid.</strong><hr></blockquote>
Uh, the classic iMacs still have G3s. It doesn't sound to me like they are planning on stopping its production any time soon. And good thing, too.
Now that OS X takes advantage of altivec, there really is no excuse to sell any hardware using the G3 processor. With all Macs running on Altivec, developer's will have no excuses left not to optimize their products for altivec. This will benefit all G4 and G5 owners, in some cases the benefits will be immense. Not all apps can take advantage of altivec, but there are plenty out there that could benefit but don't due to lack of effort on the part of the developers. And until recently, the developers had ever reason not to bother with altivec, since only a fraction of Apple's user base had the hardware for it.
With G4s across the line, and G5s in powermacs, Apple will put many current users in a position where upgrading to new hardware will result in a significant performance gain. This is just what Apple needs to get a shot in the arm.
I think before the end of April the G5 will appear in Powermacs, at the very least it will be announced, although it may ship as late as august. But still in time for the next school year (yeah powermacs aren't really aimed at the edu market, but universities buy them, and demand will probably increase because of OS X's unix roots. At least, that's what I'm hoping!).
<strong>
Thank you, Eskimo, I hadn't seen that before. It is interesting to note that AMD is using Moto's process technology -- so clearly the deficiences in the G4's clock rate are due to design issues in the G4. After all, the Athlon is on the same process and running @ 1.6 GHz.</strong><hr></blockquote>
It has a longer pipeline. That's why it's at 1.6.
<strong>
It has a longer pipeline. That's why it's at 1.6.</strong><hr></blockquote>
That's part of it, although I suspect there are other issues involved as well. The amount of time between the 7450 and the 7460 (Apollo) could be explained by the time required for adjusting the chip layout to address the original design issues which contributed to the clock rate bottlenecks (as well as doing the SOI migration). I expect that they want to push the 7460 quite a ways in terms of clock rate -- 1.8 GHz as an eventual goal has been mentioned, which is a doubling of performance over the 7450.
<strong>But the G5s pipeline isn't going to be shorter, relative to the pentium 4 and the Athlon, isn't it? At least, that's what I remember reading somewhere. If so then the G5 will probably complete more instructions per clock cycle, and thus it will pound pentium ass!
Longer pipeline != better performance.
Generally a longer pipeline allows clock rate to scale better, but has a negative effect on performance at a given clock rate. There are many other factors in keeping a chip's pipelines all as full as possible, and the G5 will no doubt address those aggressively. Don't be disappointed if the G5's pipelines are "only" 10-14 stages long, that's plenty long enough. In the end what matters is how smokin' fast the chip runs the software.
on <a href="http://www.clg-info.fr"; target="_blank">www.clg-info.fr</a> i read that Apple is going to release new mac based upon appolo G4 ranking between 1 ghz and 1,4 ghz. The mobo will be new and will have 66 Mhz PCI able to transfer 512 MByte/sec. It would have AGP pro bus instead agp 4X (does it means that in the future they will have 3D pro video card ?)
on <a href="http://www.gigapole.fr"; target="_blank">www.gigapole.fr</a> they told that the G5 rev 0,8has been push at 2,6 ghz, and Motorola expect to release 2 ghz chip. Perhaps this chip will came later in order to optimize it and to remove most of the bugs. Apple needs time also in order to have the chipset of high technology necessary to run this monster.
I read also in a PC revue that ATA 100 is limited in capacity approximatively to 160 GB, at the contrary of ATA 133 and serial ATA (first serial ATA in late 2002). HD greater than 200 GB are expected in late 2002 (IBM and seagate are able to produce 40 GB per plate HD). So it's seems logical to have ATA 133 in the future mobo rather ATA 100.
can someone check this out?
jobs.apple.com
<strong>But the G5s pipeline isn't going to be shorter, relative to the pentium 4 and the Athlon, isn't it? </strong><hr></blockquote>The G5 is rumored to have 10 pipelines (but who knows, right?), and the Pentium 4 has 20 (?).
<strong>any one have an account on Apple site to search the job listings? A friend of mine mentioned somthing about Apple looking for "G5 testers" on THIER OWN SITE. I haven't seen them mysefl, as I haven't had time to go snooping.
can someone check this out?
jobs.apple.com</strong><hr></blockquote>
You don't need an account to just search the jobs, afik. I did a bit of snooping and found a couple that are sort of interesting (nothing about G5s, though):
Title: Sr. Software Engineer
Req. ID: 1778173
Location: Santa Clara Valley, California
Country: United States
Software engineer responsible for designing and implementing key features of a Mac-OSX based multimedia application. You will be working on a small team focused on delivering a high performance system using cutting edge development techniques
{TJM: So it looks like Apple is working on a multimedia app - could be Quicktime, or iMovie, or something else?}
Title: Senior PowerPC Validation Engineer
Req. ID: 1783559
Location: Santa Clara Valley, California
Country: United States
The senior engineer filling this position will be doing hands on microprocessor validation for future versions of the Power PC. This person will be designing the hardware needed to exercise the processor under all conditions of frequency, voltage and temperature. He/she will be leading the efforts of other hardware, firmware, and diagnostics people with the common goal of finding all the bugs in a given microprocessor.
{TJM: It looks like Apple is planning on there being future versions of the PowerPC chips, anyway.}
Title: VLSI Physical Design Manager
Req. ID: 1752804
Location: Santa Clara Valley, California
Country: United States
As a Physical Design manager, you will be responsible for all aspects of chip assembly for reasonably fast (~200MHz) multimillion gate Asics. Responsibilities include: leading and managing chip assembly team; establishing physical design methodologies; chip floorplan; power/clock distribution; chip assembly, place and route, timing closure, power analysis; and working with ASIC partners in co-managing the chip assembly efforts on Apple's ASIC designs for Macintosh Computers.
{I don't know enough to know if this is significant - are 200 MHz ASICs impressive or not?}
Also, many of the network jobs wanted experience in Fibre Channel as well as Ethernet. Again, don't know if it's significant or not, but they seem awfully interested in it. Maybe Gigawire will be Apple's implementation of Fibre Channel?
Searching is a pain - it's by keyword only (not text), so it's tough trying to guess what keywords would be associated with interesting things. G5, Hypertransport, and RapidIO all came up empty. Searching for VLSI did not bring up the above ad - "chips" did. Go figure. Also, WebObjects keeps resetting you back to the main jobs page after 10 minutes or so.
Anyway, that's my best shot at the moment.
<strong>...upgrading to new hardware will result in a significant performance gain. </strong><hr></blockquote>
I'm sure this qualifies as "significant performance gain"
current Mac: 266MHz G3 (beige)
hoping for at MWT: 1.4-1.6 G5