Is there any possibility we can reconcile these many disparate view points? Are there any theories that we can discount outright? I'd just like to figure out what's going on here, and this discussion has muddled the issue quite thoroughly.
Claim 1: The current Power Mac's lack of DDR is the bane of our existence. PC133 sucks!
Response: Amorph has pointed out that MaxBus is actually quite good and superior to many DDR implementations. DDR would yield little advantage some allege.
Claim 2: The Power Mac doesn't need DDR. MaxBus is sufficient in its current form.
Response: If that's the case, why does the Xserve perform so well? And please don't explain that away by saying it's just a matter of intense I/O operations. The information still needs to get to the CPU, does it not? Additionally, if that's the case, tell that to BadAndy. He seems to know what's going on.
I don't know which claim offers the highest degree of truth. Yet, one thing I think we all must agree on is that we can't count on fair-weather Mac developers to optimize their code all that well. There is far more incentive to optimize code for the PC, of course. Apple must be proactive here, that's the only solution.
Apple must make sure its compiler technology is superior; it must then ensure its third party developers use those superior compilers. And if such efforts fail to bear fruit, Apple must ensure that its machines are so much faster they'll run the crappy code as fast as the fastest PCs. It doesn't matter if our machines are potentially faster - unless those capabilities are utilized to affect real world gains over Wintel, Wintel will still win.
[[[So I'm still unsure. There's obviously something holding it back, but without an in-depth examination of the code I'm withholding judgment. ]]]
Why don't you ask NewTek just how much parallelism they have in their code. Just be sure to give me the person's name so I can do a follow up on that. ;-)
[[[(For some reason, I thought DDR on a PC was getting up to 1700MB/s... Also, do Athlons and P4s have the same efficiency numbers on DDR? Or is that just for the P4?)\t]]]
*Sigh* If you go back to page 1 of this discussion you will see where I posted snippet from Chris Cox from Adobe. If you have any questions, concerns or disagreements I suggest that you contact *him* for the specifics.
[[[Response: If that's the case, why does the Xserve perform so well?]]]
Maybe it's the code? Doh!
[[[ And please don't explain that away by saying it's just a matter of intense I/O operations. The information still needs to get to the CPU, does it not? ]]]
Um, it's a different aminal ;-)
[[[Additionally, if that's the case, tell that to BadAndy. He seems to know what's going on.\t]]]
Who the hell is BadAndy? Maybe you could bring him here and I can introduce him to *my* friends and colleges and we can all talk.
[quote]Why don't you ask NewTek just how much parallelism they have in their code. Just be sure to give me the person's name so I can do a follow up on that. ;-)<hr></blockquote>
I'm not your errand boy. Ask them yourself. <a href="http://www.newtek.com"; target="_blank">www.newtek.com</a>. Be sure and report back on what you find.
[quote] *Sigh* If you go back to page 1 of this discussion you will see where I posted snippet from Chris Cox from Adobe. If you have any questions, concerns or disagreements I suggest that you contact *him* for the specifics. <hr></blockquote>
Why is Chris Cox the ultimate authority on these matters? I hate to bring up BadAndy's name again , but in the thread linked to below he mentions getting 1.5GB/s bandwidth from a dual athlon machine, so it looks like I was a bit off. Take that as you will.
[quote] [[[Response: If that's the case, why does the Xserve perform so well?]]]
In that thread he talks about altivec programming and mentions bandwidth on the MPX bus along the way.
I really don't see BadAndy making an effort to please you (he's a rather obtuse fellow). But, if you find fault with his reasoning, by all means register at Ars and strike up a conversation. I'd be interested in seeing what course that conversation takes.
<img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" /> This thread is getting lame.
The real reason people use Macs are not for speed but for usability. I myself wouldn't mind seeing faster machines but then I use one of my Macs as an SQL server.
The real reason the PM is going to get a refresh isn't for us in these forums (why? so we can type faster!?) but for Hollywood/video/science/genetics fields. Remember 6-8 months ago when SJ asked Hollywood what it would take to get them to use Macs? I think we'll find out in August, and using common sense, the users in those fields actually stress the system.
[quote] The real reason people use Macs are not for speed but for usability. <hr></blockquote>
Non sequitur. You're the first person to mention usability in this thread.
[quote] The real reason the PM is going to get a refresh isn't for us in these forums (why? so we can type faster!?) but for Hollywood/video/science/genetics fields. <hr></blockquote>
What am I, chopped liver? Do you know what Lightwave is?
"On the IBM side, expect a very different approach in POWER5 for the 2004 - besides the usual clock, cache, memory, CPU execution parallelisation and interconnect improvements, there will be "Fast Path" hardwiring (on-chip hardware acceleration) of some common tasks like TCP/IP processing - later maybe even stuff like high-level database or bioinformatics routines. The 0.13 micron copper POWER5 will have 2-way simultaneous multithreading, just like Xeon, and is expected to run well above 2 GHz. It should be much cheaper, cooler and less power-guzzling than POWER4. Don't be surprised to see Apple Macs or thin blade servers based on POWER5."
[quote] Remember 6-8 months ago when SJ asked Hollywood what it would take to get them to use Macs? <hr></blockquote>
And he obviously got the answer "buy up most of the software we use day in, day out, make sure we can't get it for the platforms we use day in, day out (and here's a clue... none of them are macs and they're not all Windows...) and make us run it on a public-beta os on hardware that's frankly laughably underpowered as we'll never get anything rendered, and also threaten us with buying an unspecified 3d app and doing the same special stop-developing for the platforms we've made a huge investment in, and force us to swap to a platform that's even more comically underpowered for 3d than it is for compositing..."
..though I doubt it.
That's the view from film post-production for you. It's called how to win friends and influence people...
It's interesting to note that when Microsoft was bothering the post-production industry and bought SoftImage, they at least kept developing for all the other platforms it ran on...
If Apple come back with a system that ran at anything like an acceptable speed, and had the bandwidth to throw around HD and 2k film resolution streams (that's a 2048x1536 dpi image 24 times a second for 2k film res, and 1920x1080 25 times a second for HDTV, 235-280 Mbytes/sec per layer and some composites can have dozens of layers...) then I would maybe consider it for production work.
But here's the rub... I can already buy pc and SGI based systems that I know can do this. Why would I want to take a chance with such a flakey company as Apple for mission-critical systems?
And I was a huge Apple fan through the Sculley years and before, if the first time you saw an Apple Computer was your iMac, you might want to take a close look at Apple's history before you start to flame me for this post.
And I was a huge Apple fan through the Sculley years and before, if the first time you saw an Apple Computer was your iMac, you might want to take a close look at Apple's history before you start to flame me for this post.</strong><hr></blockquote>
Well I still own my Apple ][c and after reading dozens of off-topic trolling posts like this one I still want to flame you.
One thing that has failed to be factored in to the speed debate:
PCs (based on AMD or Intel processors) running Windows (to accomodate Lightwave, as well as other programs previously mentioned in this thread) crash regularly. Not very fun to have to reboot inbetween rendering sessions, is it? I work in the computing area of my university and as such, I have to "cure" PCs VASTLY more than I EVER have to "cure" Macs... and a great deal of this healing (of PCs) goes on in our graphics labs and video control rooms; this is not heresay.
Although I realize this is a more hardware-theory discussion (using real-world stats) I find it pertinent to mention that productivity, as a whole, is greatly decreased by working in a Wintel environment. Mac OS simply works better (memory management, stability- I concede to speed in some cases).
In addition, by the way, I would like to compliment those individuals who command a vast repository of information related to the hardware being discussed. You truly have made this thread informative and interesting.
"buy up most of the software we use day in, day out, make sure we can't get it for the platforms we use day in, day out (and here's a clue... none of them are macs and they're not all Windows...... and also threaten us with buying an unspecified 3d app and doing the same special stop-developing for the platforms we've made a huge investment in]</strong><hr></blockquote>
Please correct me if I'm wrong, but didn't Apple say they were continuing devlopment and support for IRIX and UNIX, just not Windows?
As a side question for you, how happy are you about M$ owning the OpenGL patents? How available do you think they'll be for other platforms in the next few years?
Comments
Claim 1: The current Power Mac's lack of DDR is the bane of our existence. PC133 sucks!
Response: Amorph has pointed out that MaxBus is actually quite good and superior to many DDR implementations. DDR would yield little advantage some allege.
Claim 2: The Power Mac doesn't need DDR. MaxBus is sufficient in its current form.
Response: If that's the case, why does the Xserve perform so well? And please don't explain that away by saying it's just a matter of intense I/O operations. The information still needs to get to the CPU, does it not? Additionally, if that's the case, tell that to BadAndy. He seems to know what's going on.
I don't know which claim offers the highest degree of truth. Yet, one thing I think we all must agree on is that we can't count on fair-weather Mac developers to optimize their code all that well. There is far more incentive to optimize code for the PC, of course. Apple must be proactive here, that's the only solution.
Apple must make sure its compiler technology is superior; it must then ensure its third party developers use those superior compilers. And if such efforts fail to bear fruit, Apple must ensure that its machines are so much faster they'll run the crappy code as fast as the fastest PCs. It doesn't matter if our machines are potentially faster - unless those capabilities are utilized to affect real world gains over Wintel, Wintel will still win.
[ 07-11-2002: Message edited by: Big Mac ]</p>
Why don't you ask NewTek just how much parallelism they have in their code. Just be sure to give me the person's name so I can do a follow up on that. ;-)
[[[(For some reason, I thought DDR on a PC was getting up to 1700MB/s... Also, do Athlons and P4s have the same efficiency numbers on DDR? Or is that just for the P4?)\t]]]
*Sigh* If you go back to page 1 of this discussion you will see where I posted snippet from Chris Cox from Adobe. If you have any questions, concerns or disagreements I suggest that you contact *him* for the specifics.
[[[Response: If that's the case, why does the Xserve perform so well?]]]
Maybe it's the code? Doh!
[[[ And please don't explain that away by saying it's just a matter of intense I/O operations. The information still needs to get to the CPU, does it not? ]]]
Um, it's a different aminal ;-)
[[[Additionally, if that's the case, tell that to BadAndy. He seems to know what's going on.\t]]]
Who the hell is BadAndy? Maybe you could bring him here and I can introduce him to *my* friends and colleges and we can all talk.
--
Ed M.
I'm not your errand boy. Ask them yourself. <a href="http://www.newtek.com"; target="_blank">www.newtek.com</a>. Be sure and report back on what you find.
[quote] *Sigh* If you go back to page 1 of this discussion you will see where I posted snippet from Chris Cox from Adobe. If you have any questions, concerns or disagreements I suggest that you contact *him* for the specifics. <hr></blockquote>
Why is Chris Cox the ultimate authority on these matters? I hate to bring up BadAndy's name again
[quote] [[[Response: If that's the case, why does the Xserve perform so well?]]]
Maybe it's the code? Doh!
<hr></blockquote>
Maybe Bigc is talking about <a href="http://www.xinet.com/benchmarks/benchmarks.2002/index.html"; target="_blank">a comparison with the same code run on an Xserve and a G4 Powermac? Doh!</a>
[quote] [[[Additionally, if that's the case, tell that to BadAndy. He seems to know what's going on. ]]]
Who the hell is BadAndy? Maybe you could bring him here and I can introduce him to *my* friends and colleges and we can all talk.
<hr></blockquote>
You've already been given a couple of pointers to him, but here's one again just to get you started:
<a href="http://arstechnica.infopop.net/OpenTopic/page?a=tpc&s=50009562&f=8300945231&m=8790959504"; target="_blank">http://arstechnica.infopop.net/OpenTopic/page?a=tpc&s=50009562&f=8300945231&m=8790959504</a>;
In that thread he talks about altivec programming and mentions bandwidth on the MPX bus along the way.
I really don't see BadAndy making an effort to please you (he's a rather obtuse fellow). But, if you find fault with his reasoning, by all means register at Ars and strike up a conversation. I'd be interested in seeing what course that conversation takes.
The real reason people use Macs are not for speed but for usability. I myself wouldn't mind seeing faster machines but then I use one of my Macs as an SQL server.
The real reason the PM is going to get a refresh isn't for us in these forums (why? so we can type faster!?) but for Hollywood/video/science/genetics fields. Remember 6-8 months ago when SJ asked Hollywood what it would take to get them to use Macs? I think we'll find out in August, and using common sense, the users in those fields actually stress the system.
Non sequitur. You're the first person to mention usability in this thread.
[quote] The real reason the PM is going to get a refresh isn't for us in these forums (why? so we can type faster!?) but for Hollywood/video/science/genetics fields. <hr></blockquote>
What am I, chopped liver? Do you know what Lightwave is?
Lemon Bon Bon
[ 07-13-2002: Message edited by: Lemon Bon Bon ]</p>
<a href="http://www.theinquirer.net/Default.aspx?article=4266"; target="_blank">http://www.theinquirer.net/Default.aspx?article=4266</A>;
"On the IBM side, expect a very different approach in POWER5 for the 2004 - besides the usual clock, cache, memory, CPU execution parallelisation and interconnect improvements, there will be "Fast Path" hardwiring (on-chip hardware acceleration) of some common tasks like TCP/IP processing - later maybe even stuff like high-level database or bioinformatics routines. The 0.13 micron copper POWER5 will have 2-way simultaneous multithreading, just like Xeon, and is expected to run well above 2 GHz. It should be much cheaper, cooler and less power-guzzling than POWER4. Don't be surprised to see Apple Macs or thin blade servers based on POWER5."
And he obviously got the answer "buy up most of the software we use day in, day out, make sure we can't get it for the platforms we use day in, day out (and here's a clue... none of them are macs and they're not all Windows...) and make us run it on a public-beta os on hardware that's frankly laughably underpowered as we'll never get anything rendered, and also threaten us with buying an unspecified 3d app and doing the same special stop-developing for the platforms we've made a huge investment in, and force us to swap to a platform that's even more comically underpowered for 3d than it is for compositing..."
..though I doubt it.
That's the view from film post-production for you. It's called how to win friends and influence people...
It's interesting to note that when Microsoft was bothering the post-production industry and bought SoftImage, they at least kept developing for all the other platforms it ran on...
If Apple come back with a system that ran at anything like an acceptable speed, and had the bandwidth to throw around HD and 2k film resolution streams (that's a 2048x1536 dpi image 24 times a second for 2k film res, and 1920x1080 25 times a second for HDTV, 235-280 Mbytes/sec per layer and some composites can have dozens of layers...) then I would maybe consider it for production work.
But here's the rub... I can already buy pc and SGI based systems that I know can do this. Why would I want to take a chance with such a flakey company as Apple for mission-critical systems?
And I was a huge Apple fan through the Sculley years and before, if the first time you saw an Apple Computer was your iMac, you might want to take a close look at Apple's history before you start to flame me for this post.
[ 07-14-2002: Message edited by: shannyla ]
[ 07-14-2002: Message edited by: shannyla ]</p>
<strong>
And I was a huge Apple fan through the Sculley years and before, if the first time you saw an Apple Computer was your iMac, you might want to take a close look at Apple's history before you start to flame me for this post.</strong><hr></blockquote>
Well I still own my Apple ][c and after reading dozens of off-topic trolling posts like this one I still want to flame you.
PCs (based on AMD or Intel processors) running Windows (to accomodate Lightwave, as well as other programs previously mentioned in this thread) crash regularly. Not very fun to have to reboot inbetween rendering sessions, is it? I work in the computing area of my university and as such, I have to "cure" PCs VASTLY more than I EVER have to "cure" Macs... and a great deal of this healing (of PCs) goes on in our graphics labs and video control rooms; this is not heresay.
Although I realize this is a more hardware-theory discussion (using real-world stats) I find it pertinent to mention that productivity, as a whole, is greatly decreased by working in a Wintel environment. Mac OS simply works better (memory management, stability- I concede to speed in some cases).
In addition, by the way, I would like to compliment those individuals who command a vast repository of information related to the hardware being discussed. You truly have made this thread informative and interesting.
<strong>
"buy up most of the software we use day in, day out, make sure we can't get it for the platforms we use day in, day out (and here's a clue... none of them are macs and they're not all Windows...... and also threaten us with buying an unspecified 3d app and doing the same special stop-developing for the platforms we've made a huge investment in]</strong><hr></blockquote>
Please correct me if I'm wrong, but didn't Apple say they were continuing devlopment and support for IRIX and UNIX, just not Windows?
As a side question for you, how happy are you about M$ owning the OpenGL patents? How available do you think they'll be for other platforms in the next few years?