Itanium machines have been shipping (as servers, not for the normal people) for almost a year now. Their performance has been very bad and thus you never heard about them much, and they may have been halted by now, waiting for improvements in design and speed.
Coding for a chip requires that the design for the chip is so called "taped out", ie frozen,mand won't be changed until the next revision. as soon as a chip is taped out, there is also physical material available, at least in sample quantities, most certainly running into production right away.
Also, you can't make a compiler for a Virtual chip, at least a working prototype is always needed. Just loojk at BitBoys OY, the company that claimed it was going to revolutionize the graphics card market. They never hard any silicon to do with, they used a complete PC in emulation mode to "emulate" the graphics card. Things have been silent about them since they went belly up about 1.5 years ago.
The Apollo is coming, that has been officially confirmed more than once, and it's logical. However that does not mean it's going to be used in PowerMacs, or any Macs at all. Some of you probably remember IBM's 620 chip back in the days of the 603 and 604. That chip as NEVER used in any retailed Mac, although it was by far the fastest chip of the whole 6xx family. IBM used it for their servers.
The Book E architecture has been there for a long time already, way enough time to develop the G5 to a level that may well have reached mass production by now. the transition to the 8xxx naming doesn'tmean anything either.
G1/G2 also had a 6xx naming, G3/G4 had a 7xx/x naming, so why not have the G5/G6 have a 8xxx naming scheme? Also who tells you Apple hasn't already started developing their version of the G5 themselves. Just because we never heard about an agreement between Apple and motorola, doesn't mean there never was one.
Concluding, all arguments speaking AGAINST a near release of the G5 (within 6-10 months), have failed to convince, an are basically based on a stubborn thought of some people here, while arguments speaking towards the release have shaped up more and more and are starting to go beyond the level of rumors, more to the level of facts.
I don't be getting specific on any of the claims made, I'll just wait and see. Watch for my LOL @ suckfuldotcom, as soon as the G5 has arrived.
[quote]Some of you probably remember IBM's 620 chip back in the days of the 603 and 604. That chip as NEVER used in any retailed Mac, although it was by far the fastest chip of the whole 6xx family. IBM used it for their servers.<hr></blockquote>
[quote]The same people collecting the "evidence" will bitch the loudest when the G5 is not making it´s way to the next PowerMac update.
<hr></blockquote>
I've been collecting "evidence" on the G5 since November. I've got a list of references in my bookmarks. And you will notice that I didn't bitch at all when the Powermacs were not updated at MWSF.
I can't stand the way some people here feel the need to bitch and complain about how some OTHER people here supposedly "whine". In reality, most whining is constructive criticism of Apple's products or strategy. When you love a computer as much as we all love Macs, you want to see Apple do more than just stay alive. We want to see Apple thrive, to grow and innovate.
I'd also think that Apple has a couple of these in their labs -- it would have made a good test machine for a 64-bit OS well before an G5 showed up.
As for the comment that you can build a "virtual machine" to test compilers against... this is true for basic functionality, but Absoft has had a PowerPC compiler for ages so they already have the basic ISA functionality. For any real performance work you need real hardware to test on because an emulated machine's performance characteristics never match the real hardware's. I don't remember Absoft beating a real machine out the door, either.
That said, the Absoft ad isn't particuarly compelling... it doesn't give us any indication of timeframe, just that 3rd party developers have prototype hardware which only the most skeptical didn't already believe.
yeah...when i said "never seen the light of the day" i rather meant the chip was never commercially sold. sure there were protos...whatever...who gives a shit nowadays... :cool:
however, you may have a point here:
[quote]No, read the page, he was at the time employed at the Somerset facility, which is the PowerPC design centre.<hr></blockquote>
"you may have" and not flat out "you have" because ibm was designing fish and chips for themselves too at somerset. smells fishy..
<strong>I can't stand the way some people here feel the need to bitch and complain about how some OTHER people here supposedly "whine".</strong><hr></blockquote>
that is not quite my point.
I say if you cut yourself in the arm don´t blame the knife, that´s all.
I lumped you in with some others here and I shouldn't have. I guess I was just fed up with a particular person's post style, and your's reminded me of it and I went off on you when I shouldn't have. My apologies. You were making a good point.
Itanium machines have been shipping (as servers, not for the normal people) for almost a year now.
</strong><hr></blockquote>
OK, got me, sorry 'bout that.
[quote]<strong>
the transition to the 8xxx naming doesn'tmean anything either.
</strong><hr></blockquote>
Why then rename it?
[quote]<strong>
G1/G2 also had a 6xx naming, G3/G4 had a 7xx/x naming, so why not have the G5/G6 have a 8xxx naming scheme?
</strong><hr></blockquote>
From Moto's website:
PowerPC ISA -> Products:
"MPC7XXX, MPC7XX, and MPC6XX Host Processors"
"MPC8XXX Integrated Host Processors"
[quote]<strong>
Also who tells you Apple hasn't already started developing their version of the G5 themselves. Just because we never heard about an agreement between Apple and motorola, doesn't mean there never was one.
</strong><hr></blockquote>
This might be the case, but I know of no serious hints leading to that conclusion.
In other words, there is no difference between coding for G4 or G5.
Ergo, there is no difference between a G4 coder and a G5 coder.
Ergo, the alleged 'need' for G5 coders because of the imminent release of the G5 is nonsensical.
Ergo, we have to state that there continues to be no good evidence of an imminent release of the G5, whereas the evidence behind the Apollo continues to mount.
Ergo, despite all of the Tinkerbell Effect (if you wish hard enough....) in the world, the G5 will continue to be seriously in-the-future tech.
You know what?
I'd like to be wrong about the G5. I really would.
But I'm not.
Say hello to Apollo, kids! 'Cuz here it comes!
SdC</strong><hr></blockquote>
I have done machine level programming for MOS, Zilog and Motorola processors. In my eyes your statement proves that you don't have the slightest idea about processor design.
There are even noticable differences between the 7400 and the 7450, due to the longer pipeline for example, the L2 cache and more. A G4/7400 compiler will never generate optimum code for a 7450 unless you tell your compiler how to create better code for it.
Application code compiled for the G4 will run without changements on the G5 (in 32 bit mode, as far as reported), but as soon as you wish to generate code that's really optimized you will have to use (or write) a compiler that knows about the G5's pipeline, cache,... Maybe you also wish to add 64 bit support to your compiler.
Compiling an OS is even more special, as most processors offer instructions that can only be used in a special mode reserved for the OS. Usually there are changements in this area if you pass from one chip generation to another.
Did anyone see that news about Linux being the first OS ported to that new Intel Itanium (sp?) CPU. Itanium wasn't even out to developers yet but because they had a good emulator they were able to get a first pass Linux compiled for it. That was months ago and still the Itanium a while off.
AbSoft maybe to working on a compiler or an emulator to aid people getting ready for the G5. We all know the G5 is coming but I'm not sure this news helps us to know when.
<strong>Did anyone see that news about Linux being the first OS ported to that new Intel Itanium (sp?) CPU.</strong><hr></blockquote>
This is correct. Intel had a quite good cpu simulator that was reliable enough to run an entire OS. I worked with some 8 and 16 bit cpu simulators a long time ago, but unfortunately not enough to be able to provide useful information here...
Comments
Itanium machines have been shipping (as servers, not for the normal people) for almost a year now. Their performance has been very bad and thus you never heard about them much, and they may have been halted by now, waiting for improvements in design and speed.
Coding for a chip requires that the design for the chip is so called "taped out", ie frozen,mand won't be changed until the next revision. as soon as a chip is taped out, there is also physical material available, at least in sample quantities, most certainly running into production right away.
Also, you can't make a compiler for a Virtual chip, at least a working prototype is always needed. Just loojk at BitBoys OY, the company that claimed it was going to revolutionize the graphics card market. They never hard any silicon to do with, they used a complete PC in emulation mode to "emulate" the graphics card. Things have been silent about them since they went belly up about 1.5 years ago.
The Apollo is coming, that has been officially confirmed more than once, and it's logical. However that does not mean it's going to be used in PowerMacs, or any Macs at all. Some of you probably remember IBM's 620 chip back in the days of the 603 and 604. That chip as NEVER used in any retailed Mac, although it was by far the fastest chip of the whole 6xx family. IBM used it for their servers.
The Book E architecture has been there for a long time already, way enough time to develop the G5 to a level that may well have reached mass production by now. the transition to the 8xxx naming doesn'tmean anything either.
G1/G2 also had a 6xx naming, G3/G4 had a 7xx/x naming, so why not have the G5/G6 have a 8xxx naming scheme? Also who tells you Apple hasn't already started developing their version of the G5 themselves. Just because we never heard about an agreement between Apple and motorola, doesn't mean there never was one.
Concluding, all arguments speaking AGAINST a near release of the G5 (within 6-10 months), have failed to convince, an are basically based on a stubborn thought of some people here, while arguments speaking towards the release have shaped up more and more and are starting to go beyond the level of rumors, more to the level of facts.
I don't be getting specific on any of the claims made, I'll just wait and see. Watch for my LOL @ suckfuldotcom, as soon as the G5 has arrived.
G-News
<img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" />
as far as my dudeness knows, the 620 never saw the light of the day. there were no computers at all with this chip.
gimme some prove, g-news. until then, i stick to my story.
Also, you can't make a compiler for a Virtual chip, at least a working prototype is always needed. [/QB]<hr></blockquote>
Well, I guess yoou are not a computer scientist...
Anyway, it could be that they are going to make a 32 bit compiler ( for the e500 core ).
<strong>a near release of the G5 (within 6-10 months)</strong><hr></blockquote>
´cept for: that is not a near release, at least not in the sense of most of the optimist here (who vote for a Feb./March release).
6-10 month is quite realistic.
bye.
<strong>
you all do know that ibm has a g4/g5/g6 line of computers/processors which is totally unrelated to anything from apple/motorola?
so, if our friend amir talks about his career and the g5, he certainly isn't talking about our g5.
[ 01-12-2002: Message edited by: Strangelove ]</strong><hr></blockquote>
No, read the page, he was at the time employed at the Somerset facility, which is the PowerPC design centre.
Michael
<hr></blockquote>
I've been collecting "evidence" on the G5 since November. I've got a list of references in my bookmarks. And you will notice that I didn't bitch at all when the Powermacs were not updated at MWSF.
I can't stand the way some people here feel the need to bitch and complain about how some OTHER people here supposedly "whine". In reality, most whining is constructive criticism of Apple's products or strategy. When you love a computer as much as we all love Macs, you want to see Apple do more than just stay alive. We want to see Apple thrive, to grow and innovate.
<a href="http://www.zdnet.com/eweek/news/0408/08ppc62.html"; target="_blank">http://www.zdnet.com/eweek/news/0408/08ppc62.html</a>;
I'd also think that Apple has a couple of these in their labs -- it would have made a good test machine for a 64-bit OS well before an G5 showed up.
As for the comment that you can build a "virtual machine" to test compilers against... this is true for basic functionality, but Absoft has had a PowerPC compiler for ages so they already have the basic ISA functionality. For any real performance work you need real hardware to test on because an emulated machine's performance characteristics never match the real hardware's. I don't remember Absoft beating a real machine out the door, either.
That said, the Absoft ad isn't particuarly compelling... it doesn't give us any indication of timeframe, just that 3rd party developers have prototype hardware which only the most skeptical didn't already believe.
[ 01-12-2002: Message edited by: Programmer ]</p>
<a href="http://www.zdnet.com/eweek/news/0408/08ppc62.html"; target="_blank">http://www.zdnet.com/eweek/news/0408/08ppc62.html</a><hr></blockquote>;
yeah...when i said "never seen the light of the day" i rather meant the chip was never commercially sold. sure there were protos...whatever...who gives a shit nowadays... :cool:
however, you may have a point here:
[quote]No, read the page, he was at the time employed at the Somerset facility, which is the PowerPC design centre.<hr></blockquote>
"you may have" and not flat out "you have" because ibm was designing fish and chips for themselves too at somerset. smells fishy..
[ 01-12-2002: Message edited by: Strangelove ]</p>
<strong>I can't stand the way some people here feel the need to bitch and complain about how some OTHER people here supposedly "whine".</strong><hr></blockquote>
that is not quite my point.
I say if you cut yourself in the arm don´t blame the knife, that´s all.
bye.
I lumped you in with some others here and I shouldn't have. I guess I was just fed up with a particular person's post style, and your's reminded me of it and I went off on you when I shouldn't have. My apologies. You were making a good point.
After all very few people know for real what the deal with the G5 is. So speculating and looking for infos is all that is left for us.
bye.
<strong>Ignorance is bliss...
Itanium machines have been shipping (as servers, not for the normal people) for almost a year now.
</strong><hr></blockquote>
OK, got me, sorry 'bout that.
[quote]<strong>
the transition to the 8xxx naming doesn'tmean anything either.
</strong><hr></blockquote>
Why then rename it?
[quote]<strong>
G1/G2 also had a 6xx naming, G3/G4 had a 7xx/x naming, so why not have the G5/G6 have a 8xxx naming scheme?
</strong><hr></blockquote>
From Moto's website:
PowerPC ISA -> Products:
"MPC7XXX, MPC7XX, and MPC6XX Host Processors"
"MPC8XXX Integrated Host Processors"
[quote]<strong>
Also who tells you Apple hasn't already started developing their version of the G5 themselves. Just because we never heard about an agreement between Apple and motorola, doesn't mean there never was one.
</strong><hr></blockquote>
This might be the case, but I know of no serious hints leading to that conclusion.
Bye,
RazzFazz
<strong>
This is from hmurchison in a different thread.
In other words, there is no difference between coding for G4 or G5.
Ergo, there is no difference between a G4 coder and a G5 coder.
Ergo, the alleged 'need' for G5 coders because of the imminent release of the G5 is nonsensical.
Ergo, we have to state that there continues to be no good evidence of an imminent release of the G5, whereas the evidence behind the Apollo continues to mount.
Ergo, despite all of the Tinkerbell Effect (if you wish hard enough....) in the world, the G5 will continue to be seriously in-the-future tech.
You know what?
I'd like to be wrong about the G5. I really would.
But I'm not.
Say hello to Apollo, kids! 'Cuz here it comes!
SdC</strong><hr></blockquote>
I have done machine level programming for MOS, Zilog and Motorola processors. In my eyes your statement proves that you don't have the slightest idea about processor design.
There are even noticable differences between the 7400 and the 7450, due to the longer pipeline for example, the L2 cache and more. A G4/7400 compiler will never generate optimum code for a 7450 unless you tell your compiler how to create better code for it.
Application code compiled for the G4 will run without changements on the G5 (in 32 bit mode, as far as reported), but as soon as you wish to generate code that's really optimized you will have to use (or write) a compiler that knows about the G5's pipeline, cache,... Maybe you also wish to add 64 bit support to your compiler.
Compiling an OS is even more special, as most processors offer instructions that can only be used in a special mode reserved for the OS. Usually there are changements in this area if you pass from one chip generation to another.
AbSoft maybe to working on a compiler or an emulator to aid people getting ready for the G5. We all know the G5 is coming but I'm not sure this news helps us to know when.
... Apple didn't use the PowerPC 620 because of:
1. it's high price compared to the 604
2. fpu performance was about 50% higher compared to 604, but integer performance was not much higher.
3. There was no 64 bit support in the OS
... only 3400 Itanium systems were sold in 2001 (all other machines were prototypes or demo units).
<strong>Did anyone see that news about Linux being the first OS ported to that new Intel Itanium (sp?) CPU.</strong><hr></blockquote>
This is correct. Intel had a quite good cpu simulator that was reliable enough to run an entire OS. I worked with some 8 and 16 bit cpu simulators a long time ago, but unfortunately not enough to be able to provide useful information here...
Check out the Linux for 64 bit PowerPC page:
<a href="http://www.linuxppc64.org"; target="_blank">www.linuxppc64.org</a>
<strong>
You did realize the "for Intel's IA64 (Itanium)" part, right?
Seen any shipping or near-shipping IA64 systems lately?
Bye,
RazzFazz</strong><hr></blockquote>
IBM Intellistation Z Pro
<a href="http://www.sgi.com/workstations/750/"; target="_blank">SGI 750</a>
<a href="http://www.hp.com/products1/itanium/servers_workstations/index.html"; target="_blank">HP Itanium workstations and servers</a>
<strong>
Seen any shipping or near-shipping IA64 systems lately?
</strong><hr></blockquote>
Yes, and they were benchmarking them over in the Ars Battlefront forum using ArsBench.
To put it bluntly, their performance was shockingly bad. IIRC, I think my Powerbook G3 250 narrowly beat out the Itanium they were testing.
Looks to me like EPIC is a bust. Depending on the compiler for all of your prediction and instruction scheduling never made to much sense to me.