Hyperthreading (within single core) is actually closer to SMP than SMT, but the scheduling is different than both. I was wondering if OSX supports single hyperthreading. Maybe this is a non-issue with Yonnah.
If OS X needs to support "Hyperthreading", I'm sure Apple will provide the support for it. It should be quite easy.
Hyperthreading is SMT, simultaneous multithreading. It is simply Intel's marketing term for SMT. When using the term, it only applies to a single processor or a single core. A processor with SMT will present itself to the operating system as a multiprocessor system, as if it were an SMP system.
Once again, or perhaps for the first time:
Thread: operating system term for a stream of instructions. An application is typically a stream of instructions, a single stream of instructions, and hence is single threaded. Applications with more than one stream of instructions are multithreaded.
A non-SMT non-CMP non-SMP CPU can only execute one thread at a time. It cannot have 2 threads mixed in the execution pipeline at the same time. Multitasking operating systems empty the execution pipeline (I think) every time they switch threads. It only appears to be executing at the same time because the switching occurs over microsecond time spans, if not less.
SMP: acronym for symmetric multiprocessor. A machine that has 2 processors like Apple's dual CPU machines. The term came about before the advent of multicore processors came about, and not in use much anymore. These machines can execute 2 threads simultaneously, in parallel, and hence, will provide good performance enhancements for multithreaded applications.
Way back, 6 years ago, people argued the merits of SMP and ASMP (I think), asymmetric multiprocessor. Mac OS 8/9 could only run on one CPU and used the second CPU as a dedicated processor as it were. Mac OS X could run on both processors and is able to run apps/threads on any processor it chooses to, independent, ie, the processors are symmetric.
CMP: acronym for chip multiprocessor. A CPU that has multiple processors on the same die. People now refer it to dual-core, quad-core or multi-core processors. Multiple cores mean multiple threads can be executed at the same time, in parallel. A CMP CPU appears to the operating system as a SMP machine.
SMT: acronym for simultaneous multithreading. A CPU can execute multiple threads at the same time. The instructions from multiple threads could be in the same execution pipeline of the processor. The operating system would typically see a SMT processor as a SMP machine. Hyperthreading is merely Intel's marketing name for it, like Velocity Engine is for AltiVec.
There are some variations in hardware multithreading, such as vertical and horizontal microthreading, which I don't really remember the details of anymore. But the general essence of hardware multithreading is the ability of a single CPU or core to be able to have multiple threads executing, mixed, parallel, in its pipeline.
Just a tiny nuance we should remember about Yonah and Dothan. Yonah is a basically two Dothan "cores" on a die with 2 MB of shared L2, as far as we know. Dothan has 1 core with 2 MB of L2. Text illustration:
/snip/
Whether a dual Yonah system is faster than a quad P4 660 system is all dependent on the clock rate of the CPUs. If Yonah is limited to 2.2 GHz, probably not. Competitive in some things, but on the whole, probably not.
Back to the Yonah vs 970mp comparison. Consider the improvement from the 970fx to the 970mp: 2x the cache per processor. The 970fx has 512k cache, but each core of the 970mp has 1 MB of dedicated L2. Grand total of 2 MB on the die. Twice the cache is the equivalent to about adding 5 to 10% in clock rate (within limitations), so each core of the 970mp could be about as fast a 2.7 GHz 970fx. That's pretty good. If Apple used dual channel DDR2-667 for a dual 2.5 GHz 970mp machine, that would be some good performance improvements for streaming memory apps.
Yonah compared to Dothan on the other hand, is more complicated. Single-threaded performance will be like Dothan, perhaps a little better due to the "digital media boost" improvements Intel is doing. But multithreaded performance, hmm, the cores will be a little bit slower because they have to share the same cache. A Yonah core is almost be like a Dothan with only 1 MB L2 and slightly slower to the real 2 MB L2 Dothan. So, it's not exactly like 2 Dothans on one die.
The multicore speedup factor for Yonah compared to Dothan won't be or shouldn't be as good to the multicore speedup factor for a 970mp compared to the 970fx.
Whether a dual Yonah system is faster than a quad P4 660 system is all dependent on the clock rate of the CPUs. If Yonah is limited to 2.2 GHz, probably not. Competitive in some things, but on the whole, probably not.
From THG: Our Pentium M 770 was rock stable at 2.48 GHz. And: We overclocked the Pentium M 770 to 2.56 GHz by increasing the frequency of the front side bus to 160 MHz, without even having to raise the core voltage.
This page LINK shows that the stock 2.13GHz Dothan is running above the Pentium 660 in this test. Of course the overclocked version is running faster but with little extra heat. So I think that with some improvements a 2.2GHz core will run about the same as the Pentium 660.
As far as the "original statement" ( which was the 7th post in this thread ) goes, which is it "The speed of the Intel dev kits are impressing developers" or some crap about an iBook out running a PowerMac, yea I read that and I thought it was funny, that's it right, a joke. My statements are pointing to something else. The idea of transition times, for the PowerMac. Everyone here appears to believe that it will be very late in '06 or early in '07, I believe that depending on the speeds, and from THG we can see that a Yonah can run faster than a Pentium 660 which is what is in those dev kits. If Apple wanted they could switch sooner than later. One Yonah for the Books and little Macs and two for the PowerMacs, depending on how Intel does at getting the performance up. That is where I am heading, and also note that from what I have read that rosetta is running fine, so once the big major apps have been ported and anything deemed a necessitity then Apple is in the drivers seat and can pick the time to switch of their choosing.
Hyperthreading (within single core) is actually closer to SMP than SMT, but the scheduling is different than both.
I was wondering if OSX supports single hyperthreading. Maybe this is a non-issue with Yonnah.
Hyper-threading within a single core IS SMT. By Intel's definition and the definition of SMT [page 3, top 2 full paragraphs right column]. Hyper-threading is just Intel's registered trademark for it, like Moto's Altivec for the G4 SIMD engine and IBM's VMX for a G5 implementation of that standard. Registered trademarks don't change the underlying technology, just give the marketing department a unique term to tout that the competitors cant use.
Guess THT beat me to it, all of it--even the Altivec stuff! Maybe I should read to the end first?
I would strenuously say it actually is true. What are you looking at right now? That display has been created by many times more code and data read into the processor over the FSB than ever goes out.
Except that most of the operations read the frame buffer, modify it, and write it back.
Quote:
Any type of quality encoding uses many times the amount of input data to the encoded output. Your forte, gaming code, will ship 10's of times more code upstream directly to the graphics card over the bus due to textures.
In the case of 3D graphics, most of the data explosion happens in the GPU not in the CPU. The CPU is usually just copying between buffers, if it touches the data at all.
Quote:
Searches? Tons of data comes in--does it match? No, throw it away unchanged. Yes, keep that one little slice of everything you looked through. I could go on for a long time here.
And I could go through a whole list of algorithms which have both an input and output stream. Sure, if you're spending all your time doing searches then you're right, but most apps aren't doing that.
Quote:
The POWER series and 970 style balanced elastic bus is superior for server type transactions where you expect to ship large amounts of data off the CPU, but not much comparison or computation, and can use the clear one way downstream bus to avoid clobbering the upstream bus when shipping those packets out.
The 970 bus is designed so that the direction of the packets never needs to be "turned around". This saves a considerable amount of overhead compared to a single bi-directional bus. Also, I didn't say input and output are perfectly balanced, I just denied that most bandwidth intensive algorithms are almost completely without either input or output streams. Even if all you are doing is reading, the requests for what to read go down the write bus so it is handling some of the traffic. It also allows cache write back operations to happen in parallel to read operations.
What if the 970 MPs never get released? Do you automatically lose?
Of course not - that's why I said, assuming they are released. If they're not released, the bet would be off, since there is nothing to compare with. Same thing for the Intel Powerbooks...if they're not released, the bet would be off.
If there was a GUI speed benchmark, I think the P4 660 would win hands down. A 2 GHz Dothan should win hands down against a 2, maybe 2.2 GHz 970fx.
Didn't the P4 Macs do surprisingly well in the (emulated) XBench UI test? I forget the numbers that were being thrown around...not all the reports seemed to be consistent with each other.
I suspect the G5 will win in Photoshop. I'm not sure why you limit the Doom3 and Halo to 1280x1024 or above... do you think the Yonah PowerBook would win with lower resolutions? In any case, I think the Yonah will be competitive on the games, even considering laptop GPU suckage. Don't know much about Cinebench or After Effects characteristics, so can't say.
And what about browsing, checking Mail, using iPhoto, copying data over a network, opening/closing/resizing windows, doing searches, or any of a hundred other things the typical user will want to do with their laptop?
I also have no clue what sort of chipset Apple will go with. From history, they could easily hobble their PowerBooks simply to maintain market segmentation. In short, I think the technology will be comparable, but I'd say the odds are worse than 3-to-1 that Apple will screw it up again and squander the opportunity. You also specify the best ATI or nVidia graphics card available... which Apple never ships as a factory configuration anymore.
So I'm not going to take your bet, because 1. I'm not the betting type, and 2. as I maintained earlier, I have no idea whatsoever how these things will benchmark against each other and I don't want to lose money over such a silly argument.
Doom3/Halo: The reason to limit to reasonably high resolutions (1280x1024 and above) is that those are the resolutions that people buying high end machines want to play at.
However, I think that at a low enough resolution (maybe 800x600), the Yonah Powerbook could match or beat the G5 in FPS. Many games are single threaded, so 2 vs 4 cores is irrelevant, and if the game is ported from x86 to PPC, or if it is heavily dependent on L2 cache size (the P4 Extreme has a big L2 cache and is targeted in large part for serious gamers), then I could easily see the Yonah PB producing higher framerates at very low resolutions (where GPU doesn't matter)
Cinebench (and After Effects) rendering scale very well across multiple CPUs, so the 4-core G5 will easily beat the 2-core Yonah. I believe they also make substantial use of FPU power.
Browsing, checking mail, and UI I think will be faster on the Yonah PB, in some cases. However, there are two big caveats here. First, I am typing on a 1.2 Ghz iBook right now, and honestly the mail client, browser, and UI run fast enough that I wouldn't care much if they ran any faster. The limiting factors are generally my own response time (for the UI and Mail), and the bandwidth of my broadband connection (for Safari and Mail attachments). The only time I get serious slowdowns is when I start paging to disk (I only have 768 MB RAM). Obviously, this is a laptop problem that will also affect the Yonah Powerbook...it's just not as easy/cheap to cram a couple GB of RAM into a laptop.
Second, what you are talking about is basically perceived speed - how fast does the system feel doing routine tasks. I have personally found that perceived speed/responsiveness depends a lot on hard drive speed. Obviously, the PowerMac G5 will have a much faster drive than the Yonah Powerbook. So I wouldn't be surprised if the G5 still "feels" faster in a lot of routine tasks, even if Safari benchmarks a bit faster on the Powerbook.
As for iPhoto, it leverages Altivec and the GPU, so it will probably be faster on the quad G5 than the Yonah PB.
Didn't the P4 Macs do surprisingly well in the (emulated) XBench UI test? I forget the numbers that were being thrown around...not all the reports seemed to be consistent with each other.
Wasn't it concluded here that XBench was not a great application for benchmarking?
Wasn't it concluded here that XBench was not a great application for benchmarking?
Never said it was. In fact, I think it's terrible. But that doesn't mean that the information it provides is better than nothing at all, if you know how to interpret it...
Never said it was. In fact, I think it's terrible. But that doesn't mean that the information it provides is better than nothing at all, if you know how to interpret it...
Except that it's inaccurate, unreliable, and never gives the same numbers twice. Interpret it how you will.
Except that most of the operations read the frame buffer, modify it, and write it back.
And most of that is in the GPU, not the CPU for basic GUI related stuff.
Quote:
In the case of 3D graphics, most of the data explosion happens in the GPU not in the CPU. The CPU is usually just copying between buffers, if it touches the data at all.
That's my point.
Quote:
And I could go through a whole list of algorithms which have both an input and output stream. Sure, if you're spending all your time doing searches then you're right, but most apps aren't doing that.
We both can, but 50 years of CS Operating systems research & development pretty much says upstream gets much more traffic in the long haul.
Quote:
The 970 bus is designed so that the direction of the packets never needs to be "turned around". This saves a considerable amount of overhead compared to a single bi-directional bus. Also, I didn't say input and output are perfectly balanced, I just denied that most bandwidth intensive algorithms are almost completely without either input or output streams. Even if all you are doing is reading, the requests for what to read go down the write bus so it is handling some of the traffic. It also allows cache write back operations to happen in parallel to read operations.
Most users will be pleased with the performance of x86 Intel Macs because it is good at what most users do most of the time. Those few who care more about FPU or VPU performance are going to want to hang onto those PPC Macs for a little longer.
No doubt. I wouldn't be surprised if savvy video shops hold onto G5 PowerMacs long after the newer Intel PowerMacs are available. Especially with H.264, HD-DVD etc. on the horizon.
Okay now everyone who actually makes their living with after effects, and cinema 4d raise their hands?
2. Photoshop is much slower on mac according to tomshardware if you try it out on some of his slightly less biased tests.
3. At least with intel chips when Apple announces a new powermac you'll actually be able to buy it instead of waiting 6 months for it to show up only to be outdated.
4. I've never judged the speed of my computer by how long it took the photshop filter to run. I've judged it by the wait time, moving files around, opening windows, copying data, checking email, loading web pages, and navigating the file browser.... Currently this means a beachball for part of your day... Cough cough, on the intel DEV system I have yet to see the beach ball (and yes i have plenty of clean installed super fast g5s), ghz does matter for that "fast snappy" response of the OS. I don't care so much if I wait 2 seconds longer for a filter to render... I care if things are snappy...
Quote:
Originally posted by bigmig
Hmm, that's a good point. Except he is trying to compare a 4 core G5 to a 2 core Powerbook, so what he's really saying is that a single core 2 Ghz Dothan lightweight laptop should be outrunning a two core Dual G5/2.5
And yet strangely, if we assume that a 2 Ghz Dothan is about equivalent to a 3 Ghz P4 (actually, the Dothan is probably a bit slower for the class of applications we're talking about, which benefit from vector, floating point, and Hyperthreading enhancements on the P4, but it won't be a huge difference) we can see that the G5 would be:
2.4 times faster in After Effects rendering
1.9 times faster in Cinema 4D rendering
1.9 times faster in Photoshop MP actions
1.4 times faster in Photoshop SP actions
2.1 times faster in Bryce (actually, I think the Dothan would do somewhat better than the P4 here, but it's not going to change anything in the end)
In other words, the Dual G5 wins in every single benchmark, often by 2x or more.
The Dothan would also lose in games. Obviously there is no Dothan laptop gaming equivalent on Barefeats, but you can compare the Dual G5 FPS numbers to Dothan/Mobility X700 numbers from other sites
Unreal 2K4: G5/X800 gets 60/137 FPS (flyby/botmatch), Dothan/X700 gets 20 FPS (31 FPS with no AA). (this one from PC Mag)
Note that these numbers seriously overstate the Dothan/X700 performance, because the G5 is always tested at 1600x1200 (4x AA where applicable), whereas the Dothans are tested at 1400x1050 for Halo and 1280x1024 for Doom and Unreal.
If we're going to talk about Photoshop, then the truth is that most work does not require the fastest cpu on the block.
If you work with file sizes smaller that about 50MB's, as long as you have enough memory, and a good clean scratch drive, you should be fine. There are few operations that can't be handled quickly enough on a decent G4 system. A 9800 Radeon is fine for this as well.
As you go up in size, it becomes more of a hassle to deal with Gaussian Blurs, and rotations, but until you start to get to 100MB's or so in size, it's not a hardship.
2. Photoshop is much slower on mac according to tomshardware if you try it out on some of his slightly less biased tests.
The Ars Technica Forum PS7 Benchmark compilations show that the G5 machines provide approximately equivalent performance to x86 competitors. For awhile, the dual 2.5 GHz Power Mac G5 was the top performer amongst the forum. The G5 has about the same clock for clock performance with Opteron/Athlon and about +1 GHz to P4 clocks in Photoshop.
Photoshop is definitely an area where the G5 competes very well, and is at bang/buck parity with x86 systems on the high end systems.
Quote:
3. At least with intel chips when Apple announces a new powermac you'll actually be able to buy it instead of waiting 6 months for it to show up only to be outdated.
Don't count your eggs before they hatch. There is only one thing we are relatively sure about right now: Apple will be building some nice computers in some nice form factors with the Intel platform, and that performance will be at approximate parity with Windows/Intel machines.
Other then that there isn't much else. Both AMD and Intel preannounce availability of their high end processors and they are subject to long delays in shipment. It will depend on how plays their cards.
Quote:
4. I've never judged the speed of my computer by how long it took the photshop filter to run. I've judged it by the wait time, moving files around, opening windows, copying data, checking email, loading web pages, and navigating the file browser.... Currently this means a beachball for part of your day... Cough cough, on the intel DEV system I have yet to see the beach ball (and yes i have plenty of clean installed super fast g5s), ghz does matter for that "fast snappy" response of the OS. I don't care so much if I wait 2 seconds longer for a filter to render... I care if things are snappy...
So the UI (launching apps, window server, and filesystem) is faster or appears faster on the dev kits?
I'm fairly confident that's because of better burst memory and integer performance, not really GHz.
Other then that there isn't much else. Both AMD and Intel preannounce availability of their high end processors and they are subject to long delays in shipment. It will depend on how plays their cards.
So the UI (launching apps, window server, and filesystem) is faster or appears faster on the dev kits?
I'm fairly confident that's because of better burst memory and integer performance, not really GHz.
What matters here is that Apple will be subject to the SAME delays. They may not get ahead, but they won't get behind.
Why it's snappier doesn't matter. The fact that OS X has been criticized for being sluggish has been a problem. Hopefully that will go away. Though I thought that some of those reasons, such as the window resizing problem was because of the was the OS accesses the info it needs, e.g.. a programming problem. Supposedly gotten rid of with Quartz extreme, when it works finally.
Comments
Originally posted by skatman
Hyperthreading (within single core) is actually closer to SMP than SMT, but the scheduling is different than both. I was wondering if OSX supports single hyperthreading. Maybe this is a non-issue with Yonnah.
If OS X needs to support "Hyperthreading", I'm sure Apple will provide the support for it. It should be quite easy.
Hyperthreading is SMT, simultaneous multithreading. It is simply Intel's marketing term for SMT. When using the term, it only applies to a single processor or a single core. A processor with SMT will present itself to the operating system as a multiprocessor system, as if it were an SMP system.
Once again, or perhaps for the first time:
Thread: operating system term for a stream of instructions. An application is typically a stream of instructions, a single stream of instructions, and hence is single threaded. Applications with more than one stream of instructions are multithreaded.
A non-SMT non-CMP non-SMP CPU can only execute one thread at a time. It cannot have 2 threads mixed in the execution pipeline at the same time. Multitasking operating systems empty the execution pipeline (I think) every time they switch threads. It only appears to be executing at the same time because the switching occurs over microsecond time spans, if not less.
SMP: acronym for symmetric multiprocessor. A machine that has 2 processors like Apple's dual CPU machines. The term came about before the advent of multicore processors came about, and not in use much anymore. These machines can execute 2 threads simultaneously, in parallel, and hence, will provide good performance enhancements for multithreaded applications.
Way back, 6 years ago, people argued the merits of SMP and ASMP (I think), asymmetric multiprocessor. Mac OS 8/9 could only run on one CPU and used the second CPU as a dedicated processor as it were. Mac OS X could run on both processors and is able to run apps/threads on any processor it chooses to, independent, ie, the processors are symmetric.
CMP: acronym for chip multiprocessor. A CPU that has multiple processors on the same die. People now refer it to dual-core, quad-core or multi-core processors. Multiple cores mean multiple threads can be executed at the same time, in parallel. A CMP CPU appears to the operating system as a SMP machine.
SMT: acronym for simultaneous multithreading. A CPU can execute multiple threads at the same time. The instructions from multiple threads could be in the same execution pipeline of the processor. The operating system would typically see a SMT processor as a SMP machine. Hyperthreading is merely Intel's marketing name for it, like Velocity Engine is for AltiVec.
There are some variations in hardware multithreading, such as vertical and horizontal microthreading, which I don't really remember the details of anymore. But the general essence of hardware multithreading is the ability of a single CPU or core to be able to have multiple threads executing, mixed, parallel, in its pipeline.
Originally posted by THT
If OS X needs to support "Hyperthreading", I'm sure Apple will provide the support for it. It should be quite easy.
It seems that the current x86 version of Mac OS X supports HyperThreading.
Originally posted by THT
Just a tiny nuance we should remember about Yonah and Dothan. Yonah is a basically two Dothan "cores" on a die with 2 MB of shared L2, as far as we know. Dothan has 1 core with 2 MB of L2. Text illustration:
/snip/
Whether a dual Yonah system is faster than a quad P4 660 system is all dependent on the clock rate of the CPUs. If Yonah is limited to 2.2 GHz, probably not. Competitive in some things, but on the whole, probably not.
Back to the Yonah vs 970mp comparison. Consider the improvement from the 970fx to the 970mp: 2x the cache per processor. The 970fx has 512k cache, but each core of the 970mp has 1 MB of dedicated L2. Grand total of 2 MB on the die. Twice the cache is the equivalent to about adding 5 to 10% in clock rate (within limitations), so each core of the 970mp could be about as fast a 2.7 GHz 970fx. That's pretty good. If Apple used dual channel DDR2-667 for a dual 2.5 GHz 970mp machine, that would be some good performance improvements for streaming memory apps.
Yonah compared to Dothan on the other hand, is more complicated. Single-threaded performance will be like Dothan, perhaps a little better due to the "digital media boost" improvements Intel is doing. But multithreaded performance, hmm, the cores will be a little bit slower because they have to share the same cache. A Yonah core is almost be like a Dothan with only 1 MB L2 and slightly slower to the real 2 MB L2 Dothan. So, it's not exactly like 2 Dothans on one die.
The multicore speedup factor for Yonah compared to Dothan won't be or shouldn't be as good to the multicore speedup factor for a 970mp compared to the 970fx.
Whether a dual Yonah system is faster than a quad P4 660 system is all dependent on the clock rate of the CPUs. If Yonah is limited to 2.2 GHz, probably not. Competitive in some things, but on the whole, probably not.
From THG: Our Pentium M 770 was rock stable at 2.48 GHz. And: We overclocked the Pentium M 770 to 2.56 GHz by increasing the frequency of the front side bus to 160 MHz, without even having to raise the core voltage.
This page LINK shows that the stock 2.13GHz Dothan is running above the Pentium 660 in this test. Of course the overclocked version is running faster but with little extra heat. So I think that with some improvements a 2.2GHz core will run about the same as the Pentium 660.
As far as the "original statement" ( which was the 7th post in this thread ) goes, which is it "The speed of the Intel dev kits are impressing developers" or some crap about an iBook out running a PowerMac, yea I read that and I thought it was funny, that's it right, a joke. My statements are pointing to something else. The idea of transition times, for the PowerMac. Everyone here appears to believe that it will be very late in '06 or early in '07, I believe that depending on the speeds, and from THG we can see that a Yonah can run faster than a Pentium 660 which is what is in those dev kits. If Apple wanted they could switch sooner than later. One Yonah for the Books and little Macs and two for the PowerMacs, depending on how Intel does at getting the performance up. That is where I am heading, and also note that from what I have read that rosetta is running fine, so once the big major apps have been ported and anything deemed a necessitity then Apple is in the drivers seat and can pick the time to switch of their choosing.
Originally posted by skatman
Hyperthreading (within single core) is actually closer to SMP than SMT, but the scheduling is different than both.
I was wondering if OSX supports single hyperthreading. Maybe this is a non-issue with Yonnah.
Hyper-threading within a single core IS SMT. By Intel's definition and the definition of SMT [page 3, top 2 full paragraphs right column]. Hyper-threading is just Intel's registered trademark for it, like Moto's Altivec for the G4 SIMD engine and IBM's VMX for a G5 implementation of that standard. Registered trademarks don't change the underlying technology, just give the marketing department a unique term to tout that the competitors cant use.
Guess THT beat me to it, all of it--even the Altivec stuff!
Originally posted by Hiro
I would strenuously say it actually is true. What are you looking at right now? That display has been created by many times more code and data read into the processor over the FSB than ever goes out.
Except that most of the operations read the frame buffer, modify it, and write it back.
Any type of quality encoding uses many times the amount of input data to the encoded output. Your forte, gaming code, will ship 10's of times more code upstream directly to the graphics card over the bus due to textures.
In the case of 3D graphics, most of the data explosion happens in the GPU not in the CPU. The CPU is usually just copying between buffers, if it touches the data at all.
Searches? Tons of data comes in--does it match? No, throw it away unchanged. Yes, keep that one little slice of everything you looked through. I could go on for a long time here.
And I could go through a whole list of algorithms which have both an input and output stream. Sure, if you're spending all your time doing searches then you're right, but most apps aren't doing that.
The POWER series and 970 style balanced elastic bus is superior for server type transactions where you expect to ship large amounts of data off the CPU, but not much comparison or computation, and can use the clear one way downstream bus to avoid clobbering the upstream bus when shipping those packets out.
The 970 bus is designed so that the direction of the packets never needs to be "turned around". This saves a considerable amount of overhead compared to a single bi-directional bus. Also, I didn't say input and output are perfectly balanced, I just denied that most bandwidth intensive algorithms are almost completely without either input or output streams. Even if all you are doing is reading, the requests for what to read go down the write bus so it is handling some of the traffic. It also allows cache write back operations to happen in parallel to read operations.
Originally posted by kim kap sol
What if the 970 MPs never get released? Do you automatically lose?
Of course not - that's why I said, assuming they are released. If they're not released, the bet would be off, since there is nothing to compare with. Same thing for the Intel Powerbooks...if they're not released, the bet would be off.
Originally posted by THT
If there was a GUI speed benchmark, I think the P4 660 would win hands down. A 2 GHz Dothan should win hands down against a 2, maybe 2.2 GHz 970fx.
Didn't the P4 Macs do surprisingly well in the (emulated) XBench UI test? I forget the numbers that were being thrown around...not all the reports seemed to be consistent with each other.
Originally posted by Booga
I suspect the G5 will win in Photoshop. I'm not sure why you limit the Doom3 and Halo to 1280x1024 or above... do you think the Yonah PowerBook would win with lower resolutions? In any case, I think the Yonah will be competitive on the games, even considering laptop GPU suckage. Don't know much about Cinebench or After Effects characteristics, so can't say.
And what about browsing, checking Mail, using iPhoto, copying data over a network, opening/closing/resizing windows, doing searches, or any of a hundred other things the typical user will want to do with their laptop?
I also have no clue what sort of chipset Apple will go with. From history, they could easily hobble their PowerBooks simply to maintain market segmentation. In short, I think the technology will be comparable, but I'd say the odds are worse than 3-to-1 that Apple will screw it up again and squander the opportunity. You also specify the best ATI or nVidia graphics card available... which Apple never ships as a factory configuration anymore.
So I'm not going to take your bet, because 1. I'm not the betting type, and 2. as I maintained earlier, I have no idea whatsoever how these things will benchmark against each other and I don't want to lose money over such a silly argument.
Doom3/Halo: The reason to limit to reasonably high resolutions (1280x1024 and above) is that those are the resolutions that people buying high end machines want to play at.
However, I think that at a low enough resolution (maybe 800x600), the Yonah Powerbook could match or beat the G5 in FPS. Many games are single threaded, so 2 vs 4 cores is irrelevant, and if the game is ported from x86 to PPC, or if it is heavily dependent on L2 cache size (the P4 Extreme has a big L2 cache and is targeted in large part for serious gamers), then I could easily see the Yonah PB producing higher framerates at very low resolutions (where GPU doesn't matter)
Cinebench (and After Effects) rendering scale very well across multiple CPUs, so the 4-core G5 will easily beat the 2-core Yonah. I believe they also make substantial use of FPU power.
Browsing, checking mail, and UI I think will be faster on the Yonah PB, in some cases. However, there are two big caveats here. First, I am typing on a 1.2 Ghz iBook right now, and honestly the mail client, browser, and UI run fast enough that I wouldn't care much if they ran any faster. The limiting factors are generally my own response time (for the UI and Mail), and the bandwidth of my broadband connection (for Safari and Mail attachments). The only time I get serious slowdowns is when I start paging to disk (I only have 768 MB RAM). Obviously, this is a laptop problem that will also affect the Yonah Powerbook...it's just not as easy/cheap to cram a couple GB of RAM into a laptop.
Second, what you are talking about is basically perceived speed - how fast does the system feel doing routine tasks. I have personally found that perceived speed/responsiveness depends a lot on hard drive speed. Obviously, the PowerMac G5 will have a much faster drive than the Yonah Powerbook. So I wouldn't be surprised if the G5 still "feels" faster in a lot of routine tasks, even if Safari benchmarks a bit faster on the Powerbook.
As for iPhoto, it leverages Altivec and the GPU, so it will probably be faster on the quad G5 than the Yonah PB.
Originally posted by bigmig
Didn't the P4 Macs do surprisingly well in the (emulated) XBench UI test? I forget the numbers that were being thrown around...not all the reports seemed to be consistent with each other.
Wasn't it concluded here that XBench was not a great application for benchmarking?
Originally posted by DHagan4755
Wasn't it concluded here that XBench was not a great application for benchmarking?
Damn straight!
Originally posted by DHagan4755
Wasn't it concluded here that XBench was not a great application for benchmarking?
Never said it was. In fact, I think it's terrible.
Originally posted by bigmig
Never said it was. In fact, I think it's terrible.
Except that it's inaccurate, unreliable, and never gives the same numbers twice. Interpret it how you will.
Originally posted by Programmer
Except that most of the operations read the frame buffer, modify it, and write it back.
And most of that is in the GPU, not the CPU for basic GUI related stuff.
Quote:
In the case of 3D graphics, most of the data explosion happens in the GPU not in the CPU. The CPU is usually just copying between buffers, if it touches the data at all.
That's my point.
Quote:
And I could go through a whole list of algorithms which have both an input and output stream. Sure, if you're spending all your time doing searches then you're right, but most apps aren't doing that.
We both can, but 50 years of CS Operating systems research & development pretty much says upstream gets much more traffic in the long haul.
Quote:
The 970 bus is designed so that the direction of the packets never needs to be "turned around". This saves a considerable amount of overhead compared to a single bi-directional bus. Also, I didn't say input and output are perfectly balanced, I just denied that most bandwidth intensive algorithms are almost completely without either input or output streams. Even if all you are doing is reading, the requests for what to read go down the write bus so it is handling some of the traffic. It also allows cache write back operations to happen in parallel to read operations.
We pretty much agree here.
Originally posted by Programmer
Most users will be pleased with the performance of x86 Intel Macs because it is good at what most users do most of the time. Those few who care more about FPU or VPU performance are going to want to hang onto those PPC Macs for a little longer.
No doubt. I wouldn't be surprised if savvy video shops hold onto G5 PowerMacs long after the newer Intel PowerMacs are available. Especially with H.264, HD-DVD etc. on the horizon.
2. Photoshop is much slower on mac according to tomshardware if you try it out on some of his slightly less biased tests.
3. At least with intel chips when Apple announces a new powermac you'll actually be able to buy it instead of waiting 6 months for it to show up only to be outdated.
4. I've never judged the speed of my computer by how long it took the photshop filter to run. I've judged it by the wait time, moving files around, opening windows, copying data, checking email, loading web pages, and navigating the file browser.... Currently this means a beachball for part of your day... Cough cough, on the intel DEV system I have yet to see the beach ball (and yes i have plenty of clean installed super fast g5s), ghz does matter for that "fast snappy" response of the OS. I don't care so much if I wait 2 seconds longer for a filter to render... I care if things are snappy...
Originally posted by bigmig
Hmm, that's a good point. Except he is trying to compare a 4 core G5 to a 2 core Powerbook, so what he's really saying is that a single core 2 Ghz Dothan lightweight laptop should be outrunning a two core Dual G5/2.5
And yet strangely, if we assume that a 2 Ghz Dothan is about equivalent to a 3 Ghz P4 (actually, the Dothan is probably a bit slower for the class of applications we're talking about, which benefit from vector, floating point, and Hyperthreading enhancements on the P4, but it won't be a huge difference) we can see that the G5 would be:
2.4 times faster in After Effects rendering
1.9 times faster in Cinema 4D rendering
1.9 times faster in Photoshop MP actions
1.4 times faster in Photoshop SP actions
2.1 times faster in Bryce (actually, I think the Dothan would do somewhat better than the P4 here, but it's not going to change anything in the end)
Source is http://www.barefeats.com/macvpc.html
In other words, the Dual G5 wins in every single benchmark, often by 2x or more.
The Dothan would also lose in games. Obviously there is no Dothan laptop gaming equivalent on Barefeats, but you can compare the Dual G5 FPS numbers to Dothan/Mobility X700 numbers from other sites
(for example, http://www.hothardware.com/viewartic...eid=637&cid=10 )
Halo: G5/X800 gets 57 FPS, Dothan/X700 gets 30 FPS
Doom3: G5/X800 gets 20 FPS, Dothan/X700 gets 16 FPS.
Unreal 2K4: G5/X800 gets 60/137 FPS (flyby/botmatch), Dothan/X700 gets 20 FPS (31 FPS with no AA). (this one from PC Mag)
Note that these numbers seriously overstate the Dothan/X700 performance, because the G5 is always tested at 1600x1200 (4x AA where applicable), whereas the Dothans are tested at 1400x1050 for Halo and 1280x1024 for Doom and Unreal.
If you work with file sizes smaller that about 50MB's, as long as you have enough memory, and a good clean scratch drive, you should be fine. There are few operations that can't be handled quickly enough on a decent G4 system. A 9800 Radeon is fine for this as well.
As you go up in size, it becomes more of a hassle to deal with Gaussian Blurs, and rotations, but until you start to get to 100MB's or so in size, it's not a hardship.
Originally posted by webmail
2. Photoshop is much slower on mac according to tomshardware if you try it out on some of his slightly less biased tests.
The Ars Technica Forum PS7 Benchmark compilations show that the G5 machines provide approximately equivalent performance to x86 competitors. For awhile, the dual 2.5 GHz Power Mac G5 was the top performer amongst the forum. The G5 has about the same clock for clock performance with Opteron/Athlon and about +1 GHz to P4 clocks in Photoshop.
Photoshop is definitely an area where the G5 competes very well, and is at bang/buck parity with x86 systems on the high end systems.
3. At least with intel chips when Apple announces a new powermac you'll actually be able to buy it instead of waiting 6 months for it to show up only to be outdated.
Don't count your eggs before they hatch. There is only one thing we are relatively sure about right now: Apple will be building some nice computers in some nice form factors with the Intel platform, and that performance will be at approximate parity with Windows/Intel machines.
Other then that there isn't much else. Both AMD and Intel preannounce availability of their high end processors and they are subject to long delays in shipment. It will depend on how plays their cards.
4. I've never judged the speed of my computer by how long it took the photshop filter to run. I've judged it by the wait time, moving files around, opening windows, copying data, checking email, loading web pages, and navigating the file browser.... Currently this means a beachball for part of your day... Cough cough, on the intel DEV system I have yet to see the beach ball (and yes i have plenty of clean installed super fast g5s), ghz does matter for that "fast snappy" response of the OS. I don't care so much if I wait 2 seconds longer for a filter to render... I care if things are snappy...
So the UI (launching apps, window server, and filesystem) is faster or appears faster on the dev kits?
I'm fairly confident that's because of better burst memory and integer performance, not really GHz.
Originally posted by THT
Other then that there isn't much else. Both AMD and Intel preannounce availability of their high end processors and they are subject to long delays in shipment. It will depend on how plays their cards.
So the UI (launching apps, window server, and filesystem) is faster or appears faster on the dev kits?
I'm fairly confident that's because of better burst memory and integer performance, not really GHz.
What matters here is that Apple will be subject to the SAME delays. They may not get ahead, but they won't get behind.
Why it's snappier doesn't matter. The fact that OS X has been criticized for being sluggish has been a problem. Hopefully that will go away. Though I thought that some of those reasons, such as the window resizing problem was because of the was the OS accesses the info it needs, e.g.. a programming problem. Supposedly gotten rid of with Quartz extreme, when it works finally.