Also, those DSP cards (cause in effect that's what it is) are designed to a very specific set of instructions/tasks. My bet is if Apple had wanted to use it they would have had to license filter code from Pinnacle. And then come the Endians...
But seriously, if they had such a kick ass system that would have outdone FCP, don't you think they would have ported it to the Mac to take the market? Nobody that heads such a company wakes up one morning and says, "nope, don't feel like doubling my market share today. Think I'll just hand it to my competitors."
The "Endians" can be solved in a matter of hours. No biggy there. In our advanced assembly x86 class we actually had to port a ppc (big endian) to x86 (little endian) and make a windows gui app out of it. It was pretty fun. I wish I could find more things to do assembly in. I actually do some inline assembly but I have to write it genericly to stay compatible with everything.
<rant>Keeping it off the internet is about the only way. Seriously, do you realize that there have been over 23,000 (yep, twenty-three thousand!) spyware/adware applications being installed unknowingly in the past year-and-a-half on Winblows boxes. The ONLY way to keep them running safely is off the internet. </rant>
So true. Which is why I stated there is not going to be a direct connection to the internet. There is no way. I've played with a PC enough to know how to maintain it well enough, but probably not well enough to stop the vast abuse to the OS from outside infiltrators. New viri, and other tactical abusive definitions pop up daily, and sooner or later (most likely sooner) 20 - 2000 abusive packages would get through to what I thought was a clean (but crap) OS before they were defined by the powers that be.
In any case... Does anyone care to speculate on the next processor, and motherboard design that will go into the next PowerMac upgrade, or will it just be an update?
The "Endians" can be solved in a matter of hours. No biggy there. In our advanced assembly x86 class we actually had to port a ppc (big endian) to x86 (little endian) and make a windows gui app out of it. It was pretty fun. I wish I could find more things to do assembly in. I actually do some inline assembly but I have to write it genericly to stay compatible with everything.
You can do anything in a matter of hours, given enough hours. Try converting a million line project which makes extensive use of binary file formats and/or high performance code. Then put it through a full testing cycle. Its not the biggest conversion problem, but it can't be ignored in many pieces of software either.
You can do anything in a matter of hours, given enough hours. Try converting a million line project which makes extensive use of binary file formats and/or high performance code. Then put it through a full testing cycle. Its not the biggest conversion problem, but it can't be ignored in many pieces of software either.
Agreed, but how many projects are million lines of code of assembly. Most of it is written in C and the very cpu intense stuff is inline or stand-alone. It doesn't have to be ignored but it also isn't the top on the list of problems.
The "Endians" can be solved in a matter of hours. No biggy there. In our advanced assembly x86 class we actually had to port a ppc (big endian) to x86 (little endian) and make a windows gui app out of it. It was pretty fun. I wish I could find more things to do assembly in. I actually do some inline assembly but I have to write it genericly to stay compatible with everything.
Disclaimer: I haven't done any coding for close to a decade at this point.
However, the endian issue with such a card is not in the software, although that's where the fix would have to be, it's in the DSP's very complex instructions, so nearly every instruction would need to be rewritten.
Disclaimer: I haven't done any coding for close to a decade at this point.
However, the endian issue with such a card is not in the software, although that's where the fix would have to be, it's in the DSP's very complex instructions, so nearly every instruction would need to be rewritten.
I'm not a driver programmer, but I do know that intel and most chipset companies use C now as opposed to raw assembly.
I fail to see how this would have to be re-written for a compiler would take care of the problem for you... no?
Agreed, but how many projects are million lines of code of assembly. Most of it is written in C and the very cpu intense stuff is inline or stand-alone. It doesn't have to be ignored but it also isn't the top on the list of problems.
The use of C/C++ doesn't mitigate the need to deal with endian issues. If the hardware on the card expects little-endian format then the driver source code will need to be modified to byte swap the values before writing them to the hardware, unless they were written with that in mind in the first place -- generally an unlikely occurance.
Xserve could be the answer, and I actually thought it would be when it came out, but there is still the Pro graphics card lineup missing. I like the Xserve idea for 3D. I'd use a Graphics slated Xserve w/ 2 cluster nodes if one existed, but I'd still rather have a one PowerMac that had alienwares BTO highend features available.
I think I'd be the only one buying an Xserve for that though. PowerMacs would sell better.
Some of this stuff is waaay over my head, but you've made the crucial point that Apple SHOULD offer at least as much expandability in BTO
configurations as possible.
Expandabilty and user upgrades are what keep me away from Mac mini,
PowerBooks and iMacs.
I've also looked at the Alienware workstations, but of course to me
Windows is the big turn off. Otherwise they offer exceptional BTO options. I had trouble keeping the configurations lower than $5000.
I still do not understand why BTO options for Apple gear are so limited.
10,000 RPM HD's, 512 MB GPU's, internal pro sound cards etc.
The modular concept does seem to have wings separating the desktop
controller and monitor from storage devices.
I guess that's why RAID clusters are available.
I can see where an expandable rack mount storage system for heavy demands would be ideal.
If nothing else, the modular concept might help clean up your work space environment.
Hopefully the one thing that we can STILL count on from Apple's next
I'd be amazed to see a G5 upgrade in a G4, and I'd be further amazed if it performed at more than 50% of a native G5 system. As far as G5 system upgradability, I don't think it'd even be reasonable to expect until the memory controller is on the CPU, since the CPU doesn't seem to support enough bus multiplier options.
In the end upgradability sells machines, not upgrades, as the lacklucter performance of the upgrade vendors over the last 10 years or so shows.
See those two little silver boxes inside each G5? What if you could buy those seperately? I'm talking about upgrading Powermac G5s to be faster, when faster processors come out. Say a corporation bought tons of Xserves . They could swap in new, faster G5 processors, taking out the original G5 processors. I'm not talking about G4>G5. That's just stupid.
See those two little silver boxes inside each G5? What if you could buy those seperately? I'm talking about upgrading Powermac G5s to be faster, when faster processors come out. Say a corporation bought tons of Xserves . They could swap in new, faster G5 processors, taking out the original G5 processors. I'm not talking about G4>G5. That's just stupid.
I've been wondering about upgrades too if the new dual core modules
will plug into existing G5 towers.
I guess the only place to ask might be xlr8yourmac.com
Hasn't Apple been fighting the notion of "Drop in G5 Upgrades" with policies like Thermal Calibration and such. Of course they don't like processor upgrades and with the advent of the Mini that market is looked less and less profitable.
Take this news of a Dual G4 for the Cube. It was the thing of legends for a long time. Reps from PowerLogix and others would say, "Were still working on it." over and over and over. Finally we have seen pictures of it and it would of course cost more then a Mini.
Equating a Mini to a Dual G4 isn't really comparable to newer G5 chips in Power Macs. But if Apple continues making life hard for 3rd party processor companies and fewer and fewer people purchase these upgrades where will the R&D money come from for G5 upgrades?
Another thing that s ruining the PowerMac is the speed war/race. It's been over for a while, (G4) but with IBM on board Apple appeared like it was making progress to overtake the likes of intel, and AMD, but that dream seems to be dwindling as well.
I remember a few of us back then were waiting for the Dual 3GHz G5 PowerMac at WWDC 04 before we would be buying one, but that didn't happen as planned, and I'm not sure that is ever going to happen.
I had expected a few things to happen when that machine came out.
Graphics card manufacturers like PNY (example) could take notice, and possibly offer mac drivers for some of their highend 3D cards.
We could have heard of more VFX houses taking a closer look into the viability of using more Mac's in their workflow, at levels possible with Shake, or as 3D workstations.
This could have cause more highend switchers, and speed hungry users to consider using Mac's.
We would probably have all already played DOOM 3 on a Mac.
The what if's are endless, but I think the speed war/race an exciting time that turned the economy into what was a seemingly endless up swing.
I think the low cost computer is an effective machine to that will generate sales, but your highend machine is what spawns user, and customer confidence in your capabilities.
It's also something that needs to be updated, and upgraded soon!
Another thing that s ruining the PowerMac is the speed war/race. It's been over for a while, (G4) but with IBM on board Apple appeared like it was making progress to overtake the likes of intel, and AMD, but that dream seems to be dwindling as well.
I remember a few of us back then were waiting for the Dual 3GHz G5 PowerMac at WWDC 04 before we would be buying one, but that didn't happen as planned, and I'm not sure that is ever going to happen.
I had expected a few things to happen when that machine came out.
I don't know how many times this has to be repeated before it starts to sink into people's heads. Especially you onlooker, I know you read these boards daily.
The speed of the G5 cpu itself isn't a problem. If you compare CPU benches in software like DNETC (first written on x86) and Cinebench etc... the wars are a dead even race right now. The cpu is fast... and fast enough to compete (and many times beat) what the wintel world offers now.
Then you ask yourself... if the g5 is so fast how come benches are down in programs like... Adobe Premier. This software has been ported from the Wintel side, like most other software that is slow on macs... and it is NOT optimized.
Compare a 1.8ghz g5 iMac to a 2.8ghz p4... pretty damn close eh? Now compare a 2.5ghz g5 to a 3.2ghz p4... pretty damn close eh? Now compare a dual 2.5 g5 to a 3.8ghz p4 (are they officially out yet?)... Point is the cpu isn't the problem, its doing fine.
The good news is a new generation of g5's will be out soon. The thing I keep crossing my fingers for is on chip memory controller. The main bottleneck in the dual proc g5's is the shared memory controller. Once that is taken care of I can see a dual g5 machine taking on a dual opteron machine easily.
Though I completely agree with Onlooker about the graphics cards. I believe that is the main turn off for professional graphics workshops right now. Though I don't see a huge difference between a 6800ultra and FireGL or Quadro in maya right now. This isn't my field so I can't say any more than that. But I know what would make Onlooker happy is the newest Quadro, or maybe even any quadro, card released for the mac. Is this true with all professionals in this field?
I don't know how many times this has to be repeated before it starts to sink into people's heads. Especially you onlooker, I know you read these boards daily.
The speed of the G5 cpu itself isn't a problem.
That wasn't the point I was getting at. I should have wrote more on it, but I was trying to keep it short.
The point was -> that the speed race/war was exciting, and spawned a frenzy of computer buying in the early 90's. If Apple, and IBM were to have broken loose at WWDC 04, and inch noticeably ahead it could have spawned another frenzy in the Mac's favor.
It could have Also triggered PC, Mac discussion, and AMD, Intel, PPC. discussion, and options. What to buy, what was fast, and what isn't. Why to buy, and why not. This time it could have been in Apples favor.
Now that you have started on this path look at it this way.
Now that they are all dead still in the water. All the highend options lay in the hands of the X86 community. People still look at Mac's and still think pay more get less. Even with the new Mac Mini - if a PC user wanted to switch over, and try a Mac he could easily talk himself out of it looking at the Highend system, and what it can, and more so what it can't do, or can't be equipped with.
Can't is a bad word, but it comes up a lot when you talk about upgrading a PowerMac to anything beyond a Power-Desktop.
The good news is a new generation of g5's will be out soon. The thing...
Says who?
This "news" is the reason I started browsing these boards. I haven't seen any rumors to suggest that anything is in the works for improvements to the G5. Or by "soon" do you mean WWDC?
Comments
But seriously, if they had such a kick ass system that would have outdone FCP, don't you think they would have ported it to the Mac to take the market? Nobody that heads such a company wakes up one morning and says, "nope, don't feel like doubling my market share today. Think I'll just hand it to my competitors."
Originally posted by Rhumgod
<rant>Keeping it off the internet is about the only way. Seriously, do you realize that there have been over 23,000 (yep, twenty-three thousand!) spyware/adware applications being installed unknowingly in the past year-and-a-half on Winblows boxes. The ONLY way to keep them running safely is off the internet. </rant>
So true. Which is why I stated there is not going to be a direct connection to the internet. There is no way. I've played with a PC enough to know how to maintain it well enough, but probably not well enough to stop the vast abuse to the OS from outside infiltrators. New viri, and other tactical abusive definitions pop up daily, and sooner or later (most likely sooner) 20 - 2000 abusive packages would get through to what I thought was a clean (but crap) OS before they were defined by the powers that be.
In any case... Does anyone care to speculate on the next processor, and motherboard design that will go into the next PowerMac upgrade, or will it just be an update?
Originally posted by emig647
The "Endians" can be solved in a matter of hours. No biggy there. In our advanced assembly x86 class we actually had to port a ppc (big endian) to x86 (little endian) and make a windows gui app out of it. It was pretty fun. I wish I could find more things to do assembly in. I actually do some inline assembly but I have to write it genericly to stay compatible with everything.
You can do anything in a matter of hours, given enough hours. Try converting a million line project which makes extensive use of binary file formats and/or high performance code. Then put it through a full testing cycle. Its not the biggest conversion problem, but it can't be ignored in many pieces of software either.
Originally posted by Programmer
You can do anything in a matter of hours, given enough hours. Try converting a million line project which makes extensive use of binary file formats and/or high performance code. Then put it through a full testing cycle. Its not the biggest conversion problem, but it can't be ignored in many pieces of software either.
Agreed, but how many projects are million lines of code of assembly. Most of it is written in C and the very cpu intense stuff is inline or stand-alone. It doesn't have to be ignored but it also isn't the top on the list of problems.
Originally posted by emig647
The "Endians" can be solved in a matter of hours. No biggy there. In our advanced assembly x86 class we actually had to port a ppc (big endian) to x86 (little endian) and make a windows gui app out of it. It was pretty fun. I wish I could find more things to do assembly in. I actually do some inline assembly but I have to write it genericly to stay compatible with everything.
Disclaimer: I haven't done any coding for close to a decade at this point.
However, the endian issue with such a card is not in the software, although that's where the fix would have to be, it's in the DSP's very complex instructions, so nearly every instruction would need to be rewritten.
Originally posted by ChevalierMalFet
Disclaimer: I haven't done any coding for close to a decade at this point.
However, the endian issue with such a card is not in the software, although that's where the fix would have to be, it's in the DSP's very complex instructions, so nearly every instruction would need to be rewritten.
I'm not a driver programmer, but I do know that intel and most chipset companies use C now as opposed to raw assembly.
I fail to see how this would have to be re-written for a compiler would take care of the problem for you... no?
Originally posted by emig647
Agreed, but how many projects are million lines of code of assembly. Most of it is written in C and the very cpu intense stuff is inline or stand-alone. It doesn't have to be ignored but it also isn't the top on the list of problems.
The use of C/C++ doesn't mitigate the need to deal with endian issues. If the hardware on the card expects little-endian format then the driver source code will need to be modified to byte swap the values before writing them to the hardware, unless they were written with that in mind in the first place -- generally an unlikely occurance.
Originally posted by onlooker
Xserve could be the answer, and I actually thought it would be when it came out, but there is still the Pro graphics card lineup missing. I like the Xserve idea for 3D. I'd use a Graphics slated Xserve w/ 2 cluster nodes if one existed, but I'd still rather have a one PowerMac that had alienwares BTO highend features available.
I think I'd be the only one buying an Xserve for that though. PowerMacs would sell better.
Some of this stuff is waaay over my head, but you've made the crucial point that Apple SHOULD offer at least as much expandability in BTO
configurations as possible.
Expandabilty and user upgrades are what keep me away from Mac mini,
PowerBooks and iMacs.
I've also looked at the Alienware workstations, but of course to me
Windows is the big turn off. Otherwise they offer exceptional BTO options. I had trouble keeping the configurations lower than $5000.
I still do not understand why BTO options for Apple gear are so limited.
10,000 RPM HD's, 512 MB GPU's, internal pro sound cards etc.
The modular concept does seem to have wings separating the desktop
controller and monitor from storage devices.
I guess that's why RAID clusters are available.
I can see where an expandable rack mount storage system for heavy demands would be ideal.
If nothing else, the modular concept might help clean up your work space environment.
Hopefully the one thing that we can STILL count on from Apple's next
lineup is innovation.
Bring on the QuadraMacs! :-)
Originally posted by ChevalierMalFet
I'd be amazed to see a G5 upgrade in a G4, and I'd be further amazed if it performed at more than 50% of a native G5 system. As far as G5 system upgradability, I don't think it'd even be reasonable to expect until the memory controller is on the CPU, since the CPU doesn't seem to support enough bus multiplier options.
In the end upgradability sells machines, not upgrades, as the lacklucter performance of the upgrade vendors over the last 10 years or so shows.
See those two little silver boxes inside each G5? What if you could buy those seperately? I'm talking about upgrading Powermac G5s to be faster, when faster processors come out. Say a corporation bought tons of Xserves . They could swap in new, faster G5 processors, taking out the original G5 processors. I'm not talking about G4>G5. That's just stupid.
Originally posted by Placebo
See those two little silver boxes inside each G5? What if you could buy those seperately? I'm talking about upgrading Powermac G5s to be faster, when faster processors come out. Say a corporation bought tons of Xserves . They could swap in new, faster G5 processors, taking out the original G5 processors. I'm not talking about G4>G5. That's just stupid.
I've been wondering about upgrades too if the new dual core modules
will plug into existing G5 towers.
I guess the only place to ask might be xlr8yourmac.com
Take this news of a Dual G4 for the Cube. It was the thing of legends for a long time. Reps from PowerLogix and others would say, "Were still working on it." over and over and over. Finally we have seen pictures of it and it would of course cost more then a Mini.
Equating a Mini to a Dual G4 isn't really comparable to newer G5 chips in Power Macs. But if Apple continues making life hard for 3rd party processor companies and fewer and fewer people purchase these upgrades where will the R&D money come from for G5 upgrades?
I remember a few of us back then were waiting for the Dual 3GHz G5 PowerMac at WWDC 04 before we would be buying one, but that didn't happen as planned, and I'm not sure that is ever going to happen.
I had expected a few things to happen when that machine came out.
Originally posted by onlooker
Another thing that s ruining the PowerMac is the speed war/race. It's been over for a while, (G4) but with IBM on board Apple appeared like it was making progress to overtake the likes of intel, and AMD, but that dream seems to be dwindling as well.
I remember a few of us back then were waiting for the Dual 3GHz G5 PowerMac at WWDC 04 before we would be buying one, but that didn't happen as planned, and I'm not sure that is ever going to happen.
I had expected a few things to happen when that machine came out.
I don't know how many times this has to be repeated before it starts to sink into people's heads. Especially you onlooker, I know you read these boards daily.
The speed of the G5 cpu itself isn't a problem. If you compare CPU benches in software like DNETC (first written on x86) and Cinebench etc... the wars are a dead even race right now. The cpu is fast... and fast enough to compete (and many times beat) what the wintel world offers now.
Then you ask yourself... if the g5 is so fast how come benches are down in programs like... Adobe Premier. This software has been ported from the Wintel side, like most other software that is slow on macs... and it is NOT optimized.
Compare a 1.8ghz g5 iMac to a 2.8ghz p4... pretty damn close eh? Now compare a 2.5ghz g5 to a 3.2ghz p4... pretty damn close eh? Now compare a dual 2.5 g5 to a 3.8ghz p4 (are they officially out yet?)... Point is the cpu isn't the problem, its doing fine.
The good news is a new generation of g5's will be out soon. The thing I keep crossing my fingers for is on chip memory controller. The main bottleneck in the dual proc g5's is the shared memory controller. Once that is taken care of I can see a dual g5 machine taking on a dual opteron machine easily.
Though I completely agree with Onlooker about the graphics cards. I believe that is the main turn off for professional graphics workshops right now. Though I don't see a huge difference between a 6800ultra and FireGL or Quadro in maya right now. This isn't my field so I can't say any more than that. But I know what would make Onlooker happy is the newest Quadro, or maybe even any quadro, card released for the mac. Is this true with all professionals in this field?
Originally posted by emig647
I don't know how many times this has to be repeated before it starts to sink into people's heads. Especially you onlooker, I know you read these boards daily.
The speed of the G5 cpu itself isn't a problem.
That wasn't the point I was getting at. I should have wrote more on it, but I was trying to keep it short.
The point was -> that the speed race/war was exciting, and spawned a frenzy of computer buying in the early 90's. If Apple, and IBM were to have broken loose at WWDC 04, and inch noticeably ahead it could have spawned another frenzy in the Mac's favor.
It could have Also triggered PC, Mac discussion, and AMD, Intel, PPC. discussion, and options. What to buy, what was fast, and what isn't. Why to buy, and why not. This time it could have been in Apples favor.
Now that you have started on this path look at it this way.
Now that they are all dead still in the water. All the highend options lay in the hands of the X86 community. People still look at Mac's and still think pay more get less. Even with the new Mac Mini - if a PC user wanted to switch over, and try a Mac he could easily talk himself out of it looking at the Highend system, and what it can, and more so what it can't do, or can't be equipped with.
Can't is a bad word, but it comes up a lot when you talk about upgrading a PowerMac to anything beyond a Power-Desktop.
Originally posted by FallenFromTheTree
Yes it's an Illuminati conspiracy to control how much we can do on our Macs <">
You said it.
Originally posted by emig647
The good news is a new generation of g5's will be out soon. The thing...
Says who?
This "news" is the reason I started browsing these boards. I haven't seen any rumors to suggest that anything is in the works for improvements to the G5. Or by "soon" do you mean WWDC?
yours,
one track beagle