Well, I can remember when Apple Marketing decided to retroactively double the 040 Megahertz rating for no good technical reason. But I don't think Apple ever shipped a 66Mhz 040. Only 33Mhz and 40Mhz (Q840AV).
This entire post is qualified with an "IIRC"...
There was a 66MHz '040, as well as one that ran at 80MHz. See the thing is, back then PCs and Macs were rated by the speed of their FSB. So a Quadra 840AV had a 40MHz FSB, but internally the CPU ran at 80MHz. When Intel switched the way they marketed their chips (I believe the 486DX2 was the first to use the new naming system), Apple switched the way they marketed their computers.
This all happened a long time ago, and I wouldn't be surprised in the least if I'm wrong.
The XServes still haven't shipped. They were announced on Jan 6th.
Conspiracy theory: remember the Xserve images that briefly appeared on the Apple Store showing 2.3 GHz? Maybe the plan has been all along for the Xserve to be at 2.3, but they have to wait for the next tower revision before they start shipping. If Apple planned to bump the towers in late January or February but ran into delays, that could have forced them to delay the Xserve as well. (If they shipped an Xserve at 2.3, tower sales would come to a screeching halt until the bump).
No, I don't actually think this is the case, but it's at least possible...
Conspiracy theory: remember the Xserve images that briefly appeared on the Apple Store showing 2.3 GHz? Maybe the plan has been all along for the Xserve to be at 2.3, but they have to wait for the next tower revision before they start shipping. If Apple planned to bump the towers in late January or February but ran into delays, that could have forced them to delay the Xserve as well. (If they shipped an Xserve at 2.3, tower sales would come to a screeching halt until the bump).
No, I don't actually think this is the case, but it's at least possible...
Cool theory... but why do you think they didn't just announce the Towers at MAC World too?
There was a 66MHz '040, as well as one that ran at 80MHz. See the thing is, back then PCs and Macs were rated by the speed of their FSB. So a Quadra 840AV had a 40MHz FSB, but internally the CPU ran at 80MHz. When Intel switched the way they marketed their chips (I believe the 486DX2 was the first to use the new naming system), Apple switched the way they marketed their computers.
This all happened a long time ago, and I wouldn't be surprised in the least if I'm wrong.
The 80486 and 68040 were the first mass market desktop processors which clocked the processor higher than its bus so nobody had figured out how to market that yet. Apple just went with the external clock speed, but Intel went with the internal speed... causing Apple to shift its policy later. First they went to a 33/66 kind of notionation, and later just to calling it 66.
The 80486 and 68040 were the first mass market desktop processors which clocked the processor higher than its bus so nobody had figured out how to market that yet. Apple just went with the external clock speed, but Intel went with the internal speed... causing Apple to shift its policy later. First they went to a 33/66 kind of notionation, and later just to calling it 66.
I always thought that was pure hot air as far as the 68040 went - it couldn't do a string of NOPS at 66Million per second, never mind any other interesting instructions. There might have been some section of the 040's guts that was double clocked but not the whole shebang. I think there were some FPU instructions that demonstrated slightly improved latency due to that detail but the basic heartbeat of the chip was 33 million instructions per second (peak) at 33mhz.
The 486 DX2 on the other hand could in fact execute twice as many instructions per second as the 33MHz part in the best case.
Conspiracy theory: remember the Xserve images that briefly appeared on the Apple Store showing 2.3 GHz? Maybe the plan has been all along for the Xserve to be at 2.3, but they have to wait for the next tower revision before they start shipping. If Apple planned to bump the towers in late January or February but ran into delays, that could have forced them to delay the Xserve as well. (If they shipped an Xserve at 2.3, tower sales would come to a screeching halt until the bump).
No, I don't actually think this is the case, but it's at least possible...
The only problem is that nothing would keep them from only shipping a 2.0 Xserve today, unless they never intended to release them at 2.0 and don't have a way of down-clocking them to 2.0. I would not be supprised if Apple releases a 2.3 after the PM is updated but I would be if Apple does not release any Xserves at 2.0.
The 80486 and 68040 were the first mass market desktop processors which clocked the processor higher than its bus so nobody had figured out how to market that yet.
Except that the 040 didn't really run faster than its bus. Here's the first post on Google I found, where an Apple engineer is quoted "The 68040 is not clock doubled ... Instructions are not executed at a 66 MHz rate."
Except that the 040 didn't really run faster than its bus. Here's the first post on Google I found, where an Apple engineer is quoted "The 68040 is not clock doubled ... Instructions are not executed at a 66 MHz rate."
Oh well, no use dredging up 10 year old flamewars, so I'll leave it at that.
Good find, it brings back memories. Strangely we found that the DX2 wasn't particularly faster than the 040 at double the clock rate, which may be why I was thinking it was running internally at double the rate. That was the beginning of Motorola's falling behind, however, especially when Intel went to a 3x clock. The PowerPC was mainly an injection of IBM technology to keep Moto in the game, but eventually even that wasn't enough. Fortunately IBM has picked up Apple's business.
Comments
Originally posted by IntlHarvester
Well, I can remember when Apple Marketing decided to retroactively double the 040 Megahertz rating for no good technical reason. But I don't think Apple ever shipped a 66Mhz 040. Only 33Mhz and 40Mhz (Q840AV).
This entire post is qualified with an "IIRC"...
There was a 66MHz '040, as well as one that ran at 80MHz. See the thing is, back then PCs and Macs were rated by the speed of their FSB. So a Quadra 840AV had a 40MHz FSB, but internally the CPU ran at 80MHz. When Intel switched the way they marketed their chips (I believe the 486DX2 was the first to use the new naming system), Apple switched the way they marketed their computers.
This all happened a long time ago, and I wouldn't be surprised in the least if I'm wrong.
Originally posted by emig647
The XServes still haven't shipped. They were announced on Jan 6th.
Conspiracy theory: remember the Xserve images that briefly appeared on the Apple Store showing 2.3 GHz? Maybe the plan has been all along for the Xserve to be at 2.3, but they have to wait for the next tower revision before they start shipping. If Apple planned to bump the towers in late January or February but ran into delays, that could have forced them to delay the Xserve as well. (If they shipped an Xserve at 2.3, tower sales would come to a screeching halt until the bump).
No, I don't actually think this is the case, but it's at least possible...
Originally posted by 3.1416
Conspiracy theory: remember the Xserve images that briefly appeared on the Apple Store showing 2.3 GHz? Maybe the plan has been all along for the Xserve to be at 2.3, but they have to wait for the next tower revision before they start shipping. If Apple planned to bump the towers in late January or February but ran into delays, that could have forced them to delay the Xserve as well. (If they shipped an Xserve at 2.3, tower sales would come to a screeching halt until the bump).
No, I don't actually think this is the case, but it's at least possible...
Cool theory... but why do you think they didn't just announce the Towers at MAC World too?
Originally posted by Whisper
This entire post is qualified with an "IIRC"...
There was a 66MHz '040, as well as one that ran at 80MHz. See the thing is, back then PCs and Macs were rated by the speed of their FSB. So a Quadra 840AV had a 40MHz FSB, but internally the CPU ran at 80MHz. When Intel switched the way they marketed their chips (I believe the 486DX2 was the first to use the new naming system), Apple switched the way they marketed their computers.
This all happened a long time ago, and I wouldn't be surprised in the least if I'm wrong.
The 80486 and 68040 were the first mass market desktop processors which clocked the processor higher than its bus so nobody had figured out how to market that yet. Apple just went with the external clock speed, but Intel went with the internal speed... causing Apple to shift its policy later. First they went to a 33/66 kind of notionation, and later just to calling it 66.
Originally posted by Programmer
The 80486 and 68040 were the first mass market desktop processors which clocked the processor higher than its bus so nobody had figured out how to market that yet. Apple just went with the external clock speed, but Intel went with the internal speed... causing Apple to shift its policy later. First they went to a 33/66 kind of notionation, and later just to calling it 66.
I always thought that was pure hot air as far as the 68040 went - it couldn't do a string of NOPS at 66Million per second, never mind any other interesting instructions. There might have been some section of the 040's guts that was double clocked but not the whole shebang. I think there were some FPU instructions that demonstrated slightly improved latency due to that detail but the basic heartbeat of the chip was 33 million instructions per second (peak) at 33mhz.
The 486 DX2 on the other hand could in fact execute twice as many instructions per second as the 33MHz part in the best case.
Look what I started... hehe
Originally posted by 3.1416
Conspiracy theory: remember the Xserve images that briefly appeared on the Apple Store showing 2.3 GHz? Maybe the plan has been all along for the Xserve to be at 2.3, but they have to wait for the next tower revision before they start shipping. If Apple planned to bump the towers in late January or February but ran into delays, that could have forced them to delay the Xserve as well. (If they shipped an Xserve at 2.3, tower sales would come to a screeching halt until the bump).
No, I don't actually think this is the case, but it's at least possible...
The only problem is that nothing would keep them from only shipping a 2.0 Xserve today, unless they never intended to release them at 2.0 and don't have a way of down-clocking them to 2.0. I would not be supprised if Apple releases a 2.3 after the PM is updated but I would be if Apple does not release any Xserves at 2.0.
Originally posted by Programmer
The 80486 and 68040 were the first mass market desktop processors which clocked the processor higher than its bus so nobody had figured out how to market that yet.
Except that the 040 didn't really run faster than its bus. Here's the first post on Google I found, where an Apple engineer is quoted "The 68040 is not clock doubled ... Instructions are not executed at a 66 MHz rate."
http://groups.google.com/groups?f&se...fellow.MIT.EDU
Oh well, no use dredging up 10 year old flamewars, so I'll leave it at that.
Originally posted by IntlHarvester
Except that the 040 didn't really run faster than its bus. Here's the first post on Google I found, where an Apple engineer is quoted "The 68040 is not clock doubled ... Instructions are not executed at a 66 MHz rate."
http://groups.google.com/groups?f&se...fellow.MIT.EDU
Oh well, no use dredging up 10 year old flamewars, so I'll leave it at that.
Good find, it brings back memories. Strangely we found that the DX2 wasn't particularly faster than the 040 at double the clock rate, which may be why I was thinking it was running internally at double the rate. That was the beginning of Motorola's falling behind, however, especially when Intel went to a 3x clock. The PowerPC was mainly an injection of IBM technology to keep Moto in the game, but eventually even that wasn't enough. Fortunately IBM has picked up Apple's business.