<strong> Their next step will be to an on-chip memory controller with per-processor memory. </strong><hr></blockquote>
So that would be an NUMA architecture like the the upcoming AMD Opteron will use?
In AMDs case they had to ahve really fast fat pipe from processor to memory to overcome some of the inefficiencies of that setup. Will there be some kind of speed improvement on the Frontside bus of the new G4s
[quote] Advanced Micro Devices will drop its current Athlon and Duron EV6 bus architecture for its upcoming 64-bit Hammer-series processors to allow for connecting large multiprocessing arrays, the Platform Conference was told Monday.
Bob Mitton, AMD marketing manager for workstations and servers, told the meeting here that the 64-bit processors will use a new NUMA (Non-Uniform Memory Access) bus which can link eight-way or more MPUs for high performance multiprocessing. NUMA uses AMD's projected LDT (Lightning Data Transport) controller to handle both the Northbridge memory and Southbridge I/O buses in an array of processors, he said.
Mitton asserted that NUMA is highly scalable and allows each processor to have full access to the processor bus bandwidth.
By contrast, he claimed Intel Corp.'s new IA-64 architecture for Itanium and the follow-on McKinley processors have a shared processor bus that divides the bandwidth among all the processors.
He conceded that in the NUMA scheme a CPU accessing memory at the far end of the multiprocessor array goes further to fetch data than on a shared bus, but claimed the much-faster LDT offsets any potential delay. <hr></blockquote>
This seems to imply that a significant tweak of the Unix Architecture to adapt to a NUMA architecture (so a chip change like programmer spoke of may well coincide with an OS update)
This study characterizes the performance of a variant of UNIX SVR4 on a large shared-memory multiprocessor and analyzes the effects of possible OS and architectural changes. We use a nonintrusive cache miss monitor to trace the execution of an OS-intensive multiprogrammed workload on the Stanford DASH, a 32-CPU CC-NUMA multiprocessor (CC-NUMA multiprocessors have cache-coherent shared memory that is physically distributed across the machine). We find that our version of UNIX accounts for 24% of the workload's total execution time. A surprisingly large fraction of OS time (79%) is spent on memory system stalls, divided equally between instruction and data cache miss time. In analyzing techniques to reduce instruction cache miss stall time, we find that replication of only 7% of the OS code would allow 80% of instruction cache misses to be serviced locally on a CC-NUMA machine. For data cache misses, we find that a small number of routines account for 96% of OS data cache stall time. We find that most of these misses are coherence (communication) misses, and larger caches will not necessarily help. After presenting detailed performance data, we analyze the benefits of several OS changes and predict the effects of altering the cache configuration, degree of clustering, and cache coherence mechanism of the machine. <hr></blockquote>
Er...well, listen to the wail of the Appleinsider and co post board banshees and draw your own RDF conclusions.
</strong><hr></blockquote>
I'm trying to figure out - are you part of the 15-17 year old "Appleinsider" group (the majority) who's parents buy their Mac's, or part of the 18-20 who should be studying instead of surfing? The wailing comes from kids having too much time on their hands.
Lemon Bon Bon - give it up! We understand you think PowerMacs aren't worth the money! The next time you have the urge to fill up a thread (and you have chosen many threads) with your "Mac's aren't worth the money because blah, blah, blah", take a walk instead.
<strong>I just read something which strongly implies that Motorola will not add DDR support to MPX, and it will not reach higher than 166 MHz (if it even gets there). Their next step will be to an on-chip memory controller with per-processor memory. No time frame was given. While not as factual as a press release from Motorola, this is a source I tend to believe (no I'm not going to publish a link), and yes it does shoot down The Register's 7460 / 7470 rumour.
If we are lucky then the on-chip memory controller will arrive very soon. Otherwise I expect that the next PowerMac will be an XServe in a new case.</strong><hr></blockquote>
Programmer,
Is this the same type of architecture that Dorsal M was talking about in his thread? The "per processor memory" seems to support the boxes Dorsal was working on. And Dorsal said they were "late Beta" machines, and hinted that they could be ready in the MWNY or late summer timeframe. Am I getting false hopes here? I sincerely hope not.
Anyone seen the article on the UK register (www.theregister.co.uk) I know these guys publish every rumour they can get their hands on but still...... Or maybe I'm just dreaming ...again..... <img src="graemlins/oyvey.gif" border="0" alt="[No]" />
<strong>Anyone seen the article on the UK register (www.theregister.co.uk) I know these guys publish every rumour they can get their hands on but still...... Or maybe I'm just dreaming ...again..... <img src="graemlins/oyvey.gif" border="0" alt="[No]" /> </strong><hr></blockquote>
Why are people so down on El Reg? They're one of very few unbiased IT who are happy poke fun at M$, Itanium & Jobs when it's deserved. They've been pointing out 2.4Ghz P4s (or whatever) is just hype for years too.
Remember, all news sites use the same sources. The Reg don't deal in rumours often. And there's nothing in the article that can't happen.
Come July 16, when it turns out to be true everybody will be saying how great The Reg is... <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
My own snippet of info is the same as the Dorsal M architecture... he didn't invent it, if you read the RapidIO documentation it is spelled out in detail.
The Register's latest rumour claims the new G4 will support a DDR MPX FSB. I went back and re-read my other source and it did not explicitly state that DDR wouldn't happen, so I guess The Register's 7470 might be correct. I guess we'll know July 16. Either way, I'm buying a machine, and if you go find my predicitons from a month ago I said DDR @ 1.5 GHz.
I also thought it was perceptive of The Register (or their contributor) to say that Moto will be going wider, rather than faster (albeit 2.4 GHz isn't a slouch). In the long run I think this is the way to go and is more in line with what the PPC philosophy has always been.
I wasnt trying to shoot down the register as i do love the way they stick it to M$ But they're pretty good at wording stuff as if they have a direct channel to Steve Jobs brain.... I just felt that they are a bit too important a news source to go rumour-mongering like some others (who-will-remain-nameless) .....
Anyway its only a rumour and not the end of the world..... <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" /> People do get rather uptight on AI <img src="graemlins/oyvey.gif" border="0" alt="[No]" /> (not you of course!!)
<strong>Either way, I'm buying a machine, and if you go find my predicitons from a month ago I said DDR @ 1.5 GHz.
</strong><hr></blockquote>
In the end, this is what it comes down to. When people are ready to upgrade, they will and do. I too plan to purchase a high end machine after MWNY. In fact, I'll be going to MWNY, anyone else? However, I'm keeping my expectations lower (dual 1.2GHZ G4s, 512k L2 cache, based on Mot's Hip7 process and XServe like DDR implementation. Anything more than that will happily surprise me.
On a side note, the time frame will be very close to nVidia's next generation video card announcement which is scheduled for August. This could be an interesting show.
just an aside, but Tekserve, the big Apple repair and support site in NYC, is moving to a new storefront (previously they were on the 4th Fl of a loft building)during the first week of June. They're having an event each day, catering to the various Apple markets (design, DTP, web, video, etc). Anyone in NY area might want to check it out, as there may be some info to be had. Check their <a href="http://www.tekserve.com"; target="_blank">website</a> for the schedule (Tues., June 4 is pro day).
<strong>Well, for what it's worth, I just read this @ Metafilter. If true, then they should just do it and get the pain and howling over with ASAP.
I think that Apple is preparing to move OSX onto AMD's new 64-bit architecture and that is why these ports are being made. I like Motorola a lot, but if Apple could make the move to the AMD-64 it would leapfrog MS/Intel's problems moving win to a 64-bit architecture and get a 2x-3x clockspeed boost. BSD on 64-bit x86-like hardware has already been done to boot... and the early reports say that AMD is hitting 2+Ghz on thier prototypes with much faster speeds on the way. Only problem is AMD doesn't have anyone to buy these chips from them as they are not Intel compat.
It might be everything we always wanted and more.
posted by n9 at 3:14 PM PST on May 14</strong><hr></blockquote>
This reminds me of an Aerosmith song...
Dream On...
AMD has been trying to woo Apple from Motorola for years. Apple wants to deal with IBM again. Unfortunately, IBM is too large to really do what Apple needs to do to get the job done.
And Apple will NOT use an X86 based system. If they use AMD RISC chips, they will either still be PowerPC designed by Apple, produced by AMD, or they will be a new motherboard design for the AMD RISC chips.
AMD has been trying to woo Apple from Motorola for years. Apple wants to deal with IBM again. Unfortunately, IBM is too large to really do what Apple needs to do to get the job done.
And Apple will NOT use an X86 based system. If they use AMD RISC chips, they will either still be PowerPC designed by Apple, produced by AMD, or they will be a new motherboard design for the AMD RISC chips.
But both are VERY BIG IFFS.</strong><hr></blockquote>
2x to 3x clock boost does not equal to 2x to 3x speed increase, so what would be the point (except for/if 64-bit)
Comments
<strong> Their next step will be to an on-chip memory controller with per-processor memory. </strong><hr></blockquote>
So that would be an NUMA architecture like the the upcoming AMD Opteron will use?
In AMDs case they had to ahve really fast fat pipe from processor to memory to overcome some of the inefficiencies of that setup. Will there be some kind of speed improvement on the Frontside bus of the new G4s
This seems to imply a really fast bus will be needed for a NUMA architecture to work well.
<a href="http://www.ebnews.com/digest/story/OEG20010124S0018"; target="_blank">Article about AMD's use of NUMA</a>
[quote] Advanced Micro Devices will drop its current Athlon and Duron EV6 bus architecture for its upcoming 64-bit Hammer-series processors to allow for connecting large multiprocessing arrays, the Platform Conference was told Monday.
Bob Mitton, AMD marketing manager for workstations and servers, told the meeting here that the 64-bit processors will use a new NUMA (Non-Uniform Memory Access) bus which can link eight-way or more MPUs for high performance multiprocessing. NUMA uses AMD's projected LDT (Lightning Data Transport) controller to handle both the Northbridge memory and Southbridge I/O buses in an array of processors, he said.
Mitton asserted that NUMA is highly scalable and allows each processor to have full access to the processor bus bandwidth.
By contrast, he claimed Intel Corp.'s new IA-64 architecture for Itanium and the follow-on McKinley processors have a shared processor bus that divides the bandwidth among all the processors.
He conceded that in the NUMA scheme a CPU accessing memory at the far end of the multiprocessor array goes further to fetch data than on a shared bus, but claimed the much-faster LDT offsets any potential delay. <hr></blockquote>
<a href="http://www-cad.eecs.berkeley.edu/~jimy/classes/cs294-4/numa_report.html"; target="_blank">An explanation of NUMA</a>
This seems to imply that a significant tweak of the Unix Architecture to adapt to a NUMA architecture (so a chip change like programmer spoke of may well coincide with an OS update)
quoted from: <a href="http://www-flash.stanford.edu/Hive/papers/SIGMETRICS95/abstract.html"; target="_blank">http://www-flash.stanford.edu/Hive/papers/SIGMETRICS95/abstract.html</a>; (google cache)
[quote] Abstract: Memory System Performance of UNIX on CC-NUMA Multiprocessors
------------------------------------------------------------------------
This study characterizes the performance of a variant of UNIX SVR4 on a large shared-memory multiprocessor and analyzes the effects of possible OS and architectural changes. We use a nonintrusive cache miss monitor to trace the execution of an OS-intensive multiprogrammed workload on the Stanford DASH, a 32-CPU CC-NUMA multiprocessor (CC-NUMA multiprocessors have cache-coherent shared memory that is physically distributed across the machine). We find that our version of UNIX accounts for 24% of the workload's total execution time. A surprisingly large fraction of OS time (79%) is spent on memory system stalls, divided equally between instruction and data cache miss time. In analyzing techniques to reduce instruction cache miss stall time, we find that replication of only 7% of the OS code would allow 80% of instruction cache misses to be serviced locally on a CC-NUMA machine. For data cache misses, we find that a small number of routines account for 96% of OS data cache stall time. We find that most of these misses are coherence (communication) misses, and larger caches will not necessarily help. After presenting detailed performance data, we analyze the benefits of several OS changes and predict the effects of altering the cache configuration, degree of clustering, and cache coherence mechanism of the machine. <hr></blockquote>
<strong>
Er...well, listen to the wail of the Appleinsider and co post board banshees and draw your own RDF conclusions.
</strong><hr></blockquote>
I'm trying to figure out - are you part of the 15-17 year old "Appleinsider" group (the majority) who's parents buy their Mac's, or part of the 18-20 who should be studying instead of surfing? The wailing comes from kids having too much time on their hands.
Lemon Bon Bon - give it up! We understand you think PowerMacs aren't worth the money! The next time you have the urge to fill up a thread (and you have chosen many threads) with your "Mac's aren't worth the money because blah, blah, blah", take a walk instead.
<strong>I just read something which strongly implies that Motorola will not add DDR support to MPX, and it will not reach higher than 166 MHz (if it even gets there). Their next step will be to an on-chip memory controller with per-processor memory. No time frame was given. While not as factual as a press release from Motorola, this is a source I tend to believe (no I'm not going to publish a link), and yes it does shoot down The Register's 7460 / 7470 rumour.
If we are lucky then the on-chip memory controller will arrive very soon. Otherwise I expect that the next PowerMac will be an XServe in a new case.</strong><hr></blockquote>
Programmer,
Is this the same type of architecture that Dorsal M was talking about in his thread? The "per processor memory" seems to support the boxes Dorsal was working on. And Dorsal said they were "late Beta" machines, and hinted that they could be ready in the MWNY or late summer timeframe. Am I getting false hopes here? I sincerely hope not.
Thanks for the info!
<strong>Anyone seen the article on the UK register (www.theregister.co.uk) I know these guys publish every rumour they can get their hands on but still...... Or maybe I'm just dreaming ...again..... <img src="graemlins/oyvey.gif" border="0" alt="[No]" /> </strong><hr></blockquote>
Why are people so down on El Reg? They're one of very few unbiased IT who are happy poke fun at M$, Itanium & Jobs when it's deserved. They've been pointing out 2.4Ghz P4s (or whatever) is just hype for years too.
Remember, all news sites use the same sources. The Reg don't deal in rumours often. And there's nothing in the article that can't happen.
Come July 16, when it turns out to be true everybody will be saying how great The Reg is... <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
The Register's latest rumour claims the new G4 will support a DDR MPX FSB. I went back and re-read my other source and it did not explicitly state that DDR wouldn't happen, so I guess The Register's 7470 might be correct. I guess we'll know July 16. Either way, I'm buying a machine, and if you go find my predicitons from a month ago I said DDR @ 1.5 GHz.
I also thought it was perceptive of The Register (or their contributor) to say that Moto will be going wider, rather than faster (albeit 2.4 GHz isn't a slouch). In the long run I think this is the way to go and is more in line with what the PPC philosophy has always been.
[ 05-24-2002: Message edited by: Programmer ]</p>
Anyway its only a rumour and not the end of the world..... <img src="graemlins/bugeye.gif" border="0" alt="[Skeptical]" /> People do get rather uptight on AI <img src="graemlins/oyvey.gif" border="0" alt="[No]" /> (not you of course!!)
<strong>Either way, I'm buying a machine, and if you go find my predicitons from a month ago I said DDR @ 1.5 GHz.
</strong><hr></blockquote>
In the end, this is what it comes down to. When people are ready to upgrade, they will and do. I too plan to purchase a high end machine after MWNY. In fact, I'll be going to MWNY, anyone else? However, I'm keeping my expectations lower (dual 1.2GHZ G4s, 512k L2 cache, based on Mot's Hip7 process and XServe like DDR implementation. Anything more than that will happily surprise me.
On a side note, the time frame will be very close to nVidia's next generation video card announcement which is scheduled for August. This could be an interesting show.
Steve
Thanks for the "sanity check". I'm ready for an upgrade ... hopefully, I'll do it after MWNY.
200-300 Mail-in Rebate
200-300 LCD discount
100 iPod Discount
Rebates on other products with AppleCare
Free Firefly HD if you CTO
There were practically no promotions (except the LCD thing) last January, and nothing that I recall last year at this time.
<strong>Well, for what it's worth, I just read this @ Metafilter. If true, then they should just do it and get the pain and howling over with ASAP.
I think that Apple is preparing to move OSX onto AMD's new 64-bit architecture and that is why these ports are being made. I like Motorola a lot, but if Apple could make the move to the AMD-64 it would leapfrog MS/Intel's problems moving win to a 64-bit architecture and get a 2x-3x clockspeed boost. BSD on 64-bit x86-like hardware has already been done to boot... and the early reports say that AMD is hitting 2+Ghz on thier prototypes with much faster speeds on the way. Only problem is AMD doesn't have anyone to buy these chips from them as they are not Intel compat.
It might be everything we always wanted and more.
posted by n9 at 3:14 PM PST on May 14</strong><hr></blockquote>
This reminds me of an Aerosmith song...
Dream On...
AMD has been trying to woo Apple from Motorola for years. Apple wants to deal with IBM again. Unfortunately, IBM is too large to really do what Apple needs to do to get the job done.
And Apple will NOT use an X86 based system. If they use AMD RISC chips, they will either still be PowerPC designed by Apple, produced by AMD, or they will be a new motherboard design for the AMD RISC chips.
But both are VERY BIG IFFS.
<strong>
This reminds me of an Aerosmith song...
Dream On...
AMD has been trying to woo Apple from Motorola for years. Apple wants to deal with IBM again. Unfortunately, IBM is too large to really do what Apple needs to do to get the job done.
And Apple will NOT use an X86 based system. If they use AMD RISC chips, they will either still be PowerPC designed by Apple, produced by AMD, or they will be a new motherboard design for the AMD RISC chips.
But both are VERY BIG IFFS.</strong><hr></blockquote>
2x to 3x clock boost does not equal to 2x to 3x speed increase, so what would be the point (except for/if 64-bit)
<a href="http://www.dell.com/us/en/gen/default.htm"; target="_blank">http://www.dell.com/us/en/gen/default.htm</A>;
<a href="http://thenew.hp.com/country/us/eng/welcome.html"; target="_blank">http://thenew.hp.com/country/us/eng/welcome.html</A>;
<a href="http://www.gateway.com/index.shtml"; target="_blank">http://www.gateway.com/index.shtml</A>;
<a href="http://www.sonystyle.com/vaio/"; target="_blank">http://www.sonystyle.com/vaio/</A>;
These are just a few Wintel based machines. I am sure you will find something you like here.
Please do it. Your constant pissing, moaning, and whining is getting a little old.
<strong>Just for Lemon Bon Bon:
<a href="http://www.dell.com/us/en/gen/default.htm"; target="_blank">http://www.dell.com/us/en/gen/default.htm</a>;
<a href="http://thenew.hp.com/country/us/eng/welcome.html"; target="_blank">http://thenew.hp.com/country/us/eng/welcome.html</a>;
<a href="http://www.gateway.com/index.shtml"; target="_blank">http://www.gateway.com/index.shtml</a>;
<a href="http://www.sonystyle.com/vaio/"; target="_blank">http://www.sonystyle.com/vaio/</a>;
These are just a few Wintel based machines. I am sure you will find something you like here.
Please do it. Your constant pissing, moaning, and whining is getting a little old.</strong><hr></blockquote>
Yeah and extra bonus is a free copy of WindBlows peepee
"pissing"
Yup.
Lemon Bon Bon <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
Yup.
Lemon Bon Bon
Yup.
Yup."
Thanks...
detah
"quit trying to defend apple's sub-standard hardware offerings, it's really desperate."
Lemon Bon Bon
PS (Cob', quit trying to sell me PCs...one of the bleeders is quite enough... Are you on commission? That's four links...which only highlight the...
...Fact remains Apple's 'power'Macs are sub-standard for the money. Maybe defensive zealots feel otherwise...)