I am not sure what all is required to use the cores effectively. I don't usually have any problems with processes taking too long except in a few circumstances. For example when I am rendering a movie in FCP, why can't it use 100% of all 8 cores? It doesn't even come close maybe 50% if I'm lucky. Secondly Safari goes into beach ball mode far too often but is not using any significant amount of CPU % when it does.
Lots of issues with parallelizing apps. An application has to be suitable to making parallel, or multi-threaded. Frankly, a lot of apps aren't suitable. If the problem at hand is applicable, it then has to be to be coded properly to take advantage of all the cores. In many cases, all threads are not designed to max out a core and only one thread really needs a lot of CPU: so one main thread of computation with a bunch of helper/listener threads. I think this is the situation for the vast majority of consumer applications with the exception of video/audio (if one can call that "consumer"). In embarrassing parallel applications, yeah, you can max out all the cores.
In the case of FCP7 and prior, the basic architecture of the code is 10+ years old, and Quicktime was even older. Both were designed in the era when uni-processors were dominant in PCs. Obviously, there were many a multi-socket, multi-processor workstation around at the time, including PPC 604 Macs, but I imagine Ubillos wanted the biggest target Mac market, and this likey resulted in a software architecture that really only utilized one core. That's why you see such poor utilization in FCP7 on multi-core Macs.
With FCPX, Apple (with Ubillos still there I think) basically rearchitected the software for a 64-bit, multi-core environment among other things. A do over. As a result, FCPX is really an immature v1.0 product that doesn't have all the features of FCP7, but on the other hand, it can use all of the cores and the GPU. If you ran FCPX, you should see all the cores utilized. In may be possible that a prospective FCPX for iOS with a quad-core A9/A15 could outperform FCP7 running on an 8-core Mac in rendering due to this.
For Safari 5.0.x and prior, it's not that great in terms of threading/processes. Safari 5.1 however, which uses Webkit 2.0 and its multi-process model, it should beachball less, especially in regards to switching between tabs and windows. This is definitely my impression between Safari 5.0 and 5.1 on 10.6.
I am not sure what all is required to use the cores effectively. I don't usually have any problems with processes taking too long except in a few circumstances. For example when I am rendering a movie in FCP, why can't it use 100% of all 8 cores? It doesn't even come close maybe 50% if I'm lucky. Secondly Safari goes into beach ball mode far too often but is not using any significant amount of CPU % when it does.
Have you tried FCPX -- it uses all cores and renders in the background.
One of the amazing things is that you can begin editing -- even while ingesting, transcoding and analyzing in the background. Further, as clips are finished transcoding, they are substituted on-the-fly.
IMO, FCPX is the poster-boy app for exploiting the hardware and OS -- and it is built with some frameworks first developed for iOs.
As to Safari -- I think there are memory leaks and RAM issues. My habit is to have a main window with about 15 tabs open -- then other windows for articles I'm reading, surfing, references, tutorials, etc.
Usually about 10-15, single-tab windows. After a while, performance degrades, I get a message that It needs to reload all the web pages. I normally just quit Safari, restart it, then reopen all windows from last session -- then all is fine. It's like Safari experienced an enema.
This has only happened in the last few months -- on SNL as well as Lion.
Have you tried FCPX -- it uses all cores and renders in the background.
You know I haven't yet, mainly because we use a plugin for our subtitles which is not available for X. All our educational and government work needs to have English subtitles at minimum but lately we are doing a lot of Spanish subtitles as well. Since we rely heavily on various versions of older projects and mix and match different scenes switching to FCPX is not possible.
You know I haven't yet, mainly because we use a plugin for our subtitles which is not available for X. All our educational and government work needs to have English subtitles at minimum but lately we are doing a lot of Spanish subtitles as well. Since we rely heavily on various versions of older projects and mix and match different scenes switching to FCPX is not possible.
Well, like @shrike said above, it is a 1.0 product -- with lots of things missing. I expect that we'll see improvements from Apple and [especially] 3rd-party developers -- as they realize that porting their existing code gives them 21st-century performance, and a whole new source of revenue.
It is a shame that Apple did not make a "best-effort" to import FCP7 projects ala Premiere.
FCPX is flexible enough that I think you can approximate an FCP 7 sequence. I tried it, manually, with a simple project. I was able to stack clips, and put them anyplace I wanted. When I came across something (like a title effect plugin) that didn't have an FCPX equivalent -- I just imported the FCP 7 render file.
AFAICT, Apple has made an SDK available to some developers that will allow creation of a FCPX project file (an SQLite db rather than an XML file) and all the event files (XML). So I hope before too long, we'll have a way to "open" FCP 7 sequences.
Finally, FCPX was supposed to establish a new baseline and standard to code against, for developing follow-on enhancements. If true, then we should expect enhancements this year?
I don't go that far back with FCP 7, but I've heard that it took about 3 years before FCP was a fully-useable product.
With FCPX, Apple (with Ubillos still there I think) basically rearchitected the software for a 64-bit, multi-core environment among other things. A do over. As a result, FCPX is really an immature v1.0 product that doesn't have all the features of FCP7, but on the other hand, it can use all of the cores and the GPU. If you ran FCPX, you should see all the cores utilized. In may be possible that a prospective FCPX for iOS with a quad-core A9/A15 could outperform FCP7 running on an 8-core Mac in rendering due to this.
Good post. Yep, Ubillos is still there. Almost everything in FCPX is focused on making the editing experience fast and efficient -- while letting the computer do the heavy-lifting, in the background, when it will-not degrade the editing experience.
Whenever you touch a key or move the mouse for editing, background tasks will be paused -- if their resources are need in the editing UX. You can even set the time the foreground must be idle before the background resumes -- down to .1 second.
Finally, you just know, that somewhere within Apple FCPX is running on an ARM/iPad.
Oddly, FCPX has a lot of very small icons/buttons -- that wouldn't work very well in multitouch -- but those can be addressed. Then, there are some things that just scream out for a multitouch UI.
Well, you know, all those non-business degree guys seem to be doing pretty well with Apple, well, very well, oh, wait - they're one of the absolute top companies in the world. Dang. Think what would have happened if they'd all gotten their degrees!
I am not sure what all is required to use the cores effectively. I don't usually have any problems with processes taking too long except in a few circumstances. For example when I am rendering a movie in FCP, why can't it use 100% of all 8 cores? It doesn't even come close maybe 50% if I'm lucky. Secondly Safari goes into beach ball mode far too often but is not using any significant amount of CPU % when it does.
The beachball only happens when the main thread is busy, The other threads can be maxed to 100% and the beachball will not appear.
GCD and blocks make threading fairly easy and intuitive. Most iOS apps will use the cores.
Finally, you just know, that somewhere within Apple FCPX is running on an ARM/iPad.
For iOS or Mac OS X?
99% odds that they have Mac OS X running on ARM with a few test applications here and there. But my interest would lie with a FCPX version running on iOS with a touch UI, and it would be too much of a fantasy there.
There may actually be a market for a high end $1299 14"/15" landscape tablet with something like 1536x1024, 2056x1080 or all the way to 4k at 4096x2048. Quad core ARM with a PowerVR 554MP or 6-series GPU.
It could make FCPX pretty interesting to use. One could run 2 iPad apps side by side. Multitasking with two or more visible apps could be a reality.
Heavy though. 3 lbs? And it would flex a bit. And oh the glass would seem fragile at that size.
OS X and iOS have technology to spread work among the cores. 4 cores mean 4 instructions executing simultaneously.
Operating systems are complex beasts, especially the mobile variants. They have a baseband running the cellular networking, they are open for notification processing, playing audio, running apps (not just third party apps). So much to be done that you can be sure 4 cores will be used.
Are the Cortez-A15s going to be ready for early 2012?
Well Yes, Cortex A15 Dual Core Could land as early as first half 2012 for Apple. But if it really is Quad Core then it will more likely be a A9 Quad Core .
The Sad thing is Cortex A15 Dual Core is actually faster then Quad Core A9, And it seems Apple will use Power VR 543MP4, if that is the case then it is basically the same chip used in upcoming PlayStation Portable.
Now Sony Ericisson is suppose to sample their Dual Core A15 + PowerVR 6 soon..........
Well Yes, Cortex A15 Dual Core Could land as early as first half 2012 for Apple. But if it really is Quad Core then it will more likely be a A9 Quad Core .
The Sad thing is Cortex A15 Dual Core is actually faster then Quad Core A9, And it seems Apple will use Power VR 543MP4, if that is the case then it is basically the same chip used in upcoming PlayStation Portable.
Now Sony Ericisson is suppose to sample their Dual Core A15 + PowerVR 6 soon..........
Yup, and Sony/Ericsson is doing real well I hear... </sarcasm>
99% odds that they have Mac OS X running on ARM with a few test applications here and there. But my interest would lie with a FCPX version running on iOS with a touch UI, and it would be too much of a fantasy there.
There may actually be a market for a high end $1299 14"/15" landscape tablet with something like 1536x1024, 2056x1080 or all the way to 4k at 4096x2048. Quad core ARM with a PowerVR 554MP or 6-series GPU.
It could make FCPX pretty interesting to use. One could run 2 iPad apps side by side. Multitasking with two or more visible apps could be a reality.
Heavy though. 3 lbs? And it would flex a bit. And oh the glass would seem fragile at that size.
I was thinking FCPX running on iOS on ARM.
In addition to higher resolution display, I would like ti see a higher resolution (more granular) touch sensor .
This could support, but not require a stylus and crosshair hockey puck. The technology exists (Wacom) for a BT pressure sensitive stylus).
A design that appeals to me is the OTPC (was OLPC):
The tablet appears to be encased in plastic, including the (matt, outdoor) display -- and is targeted at $100 retail. I don't believe that anyone but Apple could make that happen in this decade.
But the plastic would help address the weight and flexibility problems of larger tablets.
I hope TSMC doesn't screw up like with Nvidia's transition to 40nm, but they're running into issues yet again with Nvidia's "Kepler". Y'all don't be so sure it'll be peachy just dumping Samsung's fabs overnight.
Foundry Delays to Push Back NVIDIA Kepler Launch to 2012
It looks like foundry issues are back to slow down the launch of NVIDIA's next generation high-end GPU, codenamed Kepler. The delay may push Kepler's launch to 2012, it was expected to launch by late 2011. The 28 nanometer silicon fabrication process at TSMC, a principal foundry partner of NVIDIA, is producing unsatisfactory yields. Add to that, Kepler's performance is lower than expected.
TSMC's 28 nm process at Fab 15 facility has already seen delays, which have even shaped AMD's designs in a big way. AMD had originally planned to design high-end VLIW4 chips on the 32 nm process at TSMC, but later decided to wait for the 28 nm process, leading to plans of 32 nm GPUs being scrapped by both GPU designers. TSMC was supposed to be in a position take orders of high-end 28 nm GPUs by Q4 2011, and was set to start pilot production for its 20nm process technology in the third quarter of 2012.
TSMC's major client apart from GPU vendors, that relies on new bulk manufacturing process, Qualcomm, hasn't yet adjusted its 28nm process schedule and is set to launch three new products, 8960. 8270 and 8260A using dual-core Krait architecture in the fourth quarter of 2011.
AMD's first 28 nm GPUs, under the "Southern Islands" family are on course for 1H 2012. Around the same time, AMD will also release APUs based on the 28 nm process, codenamed "Krishna" and "Wichita", that replace Ontraio and Zacate.
Comments
I am not sure what all is required to use the cores effectively. I don't usually have any problems with processes taking too long except in a few circumstances. For example when I am rendering a movie in FCP, why can't it use 100% of all 8 cores? It doesn't even come close maybe 50% if I'm lucky. Secondly Safari goes into beach ball mode far too often but is not using any significant amount of CPU % when it does.
Lots of issues with parallelizing apps. An application has to be suitable to making parallel, or multi-threaded. Frankly, a lot of apps aren't suitable. If the problem at hand is applicable, it then has to be to be coded properly to take advantage of all the cores. In many cases, all threads are not designed to max out a core and only one thread really needs a lot of CPU: so one main thread of computation with a bunch of helper/listener threads. I think this is the situation for the vast majority of consumer applications with the exception of video/audio (if one can call that "consumer"). In embarrassing parallel applications, yeah, you can max out all the cores.
In the case of FCP7 and prior, the basic architecture of the code is 10+ years old, and Quicktime was even older. Both were designed in the era when uni-processors were dominant in PCs. Obviously, there were many a multi-socket, multi-processor workstation around at the time, including PPC 604 Macs, but I imagine Ubillos wanted the biggest target Mac market, and this likey resulted in a software architecture that really only utilized one core. That's why you see such poor utilization in FCP7 on multi-core Macs.
With FCPX, Apple (with Ubillos still there I think) basically rearchitected the software for a 64-bit, multi-core environment among other things. A do over. As a result, FCPX is really an immature v1.0 product that doesn't have all the features of FCP7, but on the other hand, it can use all of the cores and the GPU. If you ran FCPX, you should see all the cores utilized. In may be possible that a prospective FCPX for iOS with a quad-core A9/A15 could outperform FCP7 running on an 8-core Mac in rendering due to this.
For Safari 5.0.x and prior, it's not that great in terms of threading/processes. Safari 5.1 however, which uses Webkit 2.0 and its multi-process model, it should beachball less, especially in regards to switching between tabs and windows. This is definitely my impression between Safari 5.0 and 5.1 on 10.6.
I am not sure what all is required to use the cores effectively. I don't usually have any problems with processes taking too long except in a few circumstances. For example when I am rendering a movie in FCP, why can't it use 100% of all 8 cores? It doesn't even come close maybe 50% if I'm lucky. Secondly Safari goes into beach ball mode far too often but is not using any significant amount of CPU % when it does.
Have you tried FCPX -- it uses all cores and renders in the background.
One of the amazing things is that you can begin editing -- even while ingesting, transcoding and analyzing in the background. Further, as clips are finished transcoding, they are substituted on-the-fly.
IMO, FCPX is the poster-boy app for exploiting the hardware and OS -- and it is built with some frameworks first developed for iOs.
As to Safari -- I think there are memory leaks and RAM issues. My habit is to have a main window with about 15 tabs open -- then other windows for articles I'm reading, surfing, references, tutorials, etc.
Usually about 10-15, single-tab windows. After a while, performance degrades, I get a message that It needs to reload all the web pages. I normally just quit Safari, restart it, then reopen all windows from last session -- then all is fine. It's like Safari experienced an enema.
This has only happened in the last few months -- on SNL as well as Lion.
Have you tried FCPX -- it uses all cores and renders in the background.
You know I haven't yet, mainly because we use a plugin for our subtitles which is not available for X. All our educational and government work needs to have English subtitles at minimum but lately we are doing a lot of Spanish subtitles as well. Since we rely heavily on various versions of older projects and mix and match different scenes switching to FCPX is not possible.
You know I haven't yet, mainly because we use a plugin for our subtitles which is not available for X. All our educational and government work needs to have English subtitles at minimum but lately we are doing a lot of Spanish subtitles as well. Since we rely heavily on various versions of older projects and mix and match different scenes switching to FCPX is not possible.
Well, like @shrike said above, it is a 1.0 product -- with lots of things missing. I expect that we'll see improvements from Apple and [especially] 3rd-party developers -- as they realize that porting their existing code gives them 21st-century performance, and a whole new source of revenue.
It is a shame that Apple did not make a "best-effort" to import FCP7 projects ala Premiere.
FCPX is flexible enough that I think you can approximate an FCP 7 sequence. I tried it, manually, with a simple project. I was able to stack clips, and put them anyplace I wanted. When I came across something (like a title effect plugin) that didn't have an FCPX equivalent -- I just imported the FCP 7 render file.
AFAICT, Apple has made an SDK available to some developers that will allow creation of a FCPX project file (an SQLite db rather than an XML file) and all the event files (XML). So I hope before too long, we'll have a way to "open" FCP 7 sequences.
Finally, FCPX was supposed to establish a new baseline and standard to code against, for developing follow-on enhancements. If true, then we should expect enhancements this year?
I don't go that far back with FCP 7, but I've heard that it took about 3 years before FCP was a fully-useable product.
With FCPX, Apple (with Ubillos still there I think) basically rearchitected the software for a 64-bit, multi-core environment among other things. A do over. As a result, FCPX is really an immature v1.0 product that doesn't have all the features of FCP7, but on the other hand, it can use all of the cores and the GPU. If you ran FCPX, you should see all the cores utilized. In may be possible that a prospective FCPX for iOS with a quad-core A9/A15 could outperform FCP7 running on an 8-core Mac in rendering due to this.
Good post. Yep, Ubillos is still there. Almost everything in FCPX is focused on making the editing experience fast and efficient -- while letting the computer do the heavy-lifting, in the background, when it will-not degrade the editing experience.
Whenever you touch a key or move the mouse for editing, background tasks will be paused -- if their resources are need in the editing UX. You can even set the time the foreground must be idle before the background resumes -- down to .1 second.
Finally, you just know, that somewhere within Apple FCPX is running on an ARM/iPad.
Oddly, FCPX has a lot of very small icons/buttons -- that wouldn't work very well in multitouch -- but those can be addressed. Then, there are some things that just scream out for a multitouch UI.
Well, you know, all those non-business degree guys seem to be doing pretty well with Apple, well, very well, oh, wait - they're one of the absolute top companies in the world. Dang. Think what would have happened if they'd all gotten their degrees!
You mean like Tim Cook?
You mean like Tim Cook?
At least you got the sarcasm.
I am not sure what all is required to use the cores effectively. I don't usually have any problems with processes taking too long except in a few circumstances. For example when I am rendering a movie in FCP, why can't it use 100% of all 8 cores? It doesn't even come close maybe 50% if I'm lucky. Secondly Safari goes into beach ball mode far too often but is not using any significant amount of CPU % when it does.
The beachball only happens when the main thread is busy, The other threads can be maxed to 100% and the beachball will not appear.
GCD and blocks make threading fairly easy and intuitive. Most iOS apps will use the cores.
Finally, you just know, that somewhere within Apple FCPX is running on an ARM/iPad.
For iOS or Mac OS X?
99% odds that they have Mac OS X running on ARM with a few test applications here and there. But my interest would lie with a FCPX version running on iOS with a touch UI, and it would be too much of a fantasy there.
There may actually be a market for a high end $1299 14"/15" landscape tablet with something like 1536x1024, 2056x1080 or all the way to 4k at 4096x2048. Quad core ARM with a PowerVR 554MP or 6-series GPU.
It could make FCPX pretty interesting to use. One could run 2 iPad apps side by side. Multitasking with two or more visible apps could be a reality.
Heavy though. 3 lbs? And it would flex a bit. And oh the glass would seem fragile at that size.
Operating systems are complex beasts, especially the mobile variants. They have a baseband running the cellular networking, they are open for notification processing, playing audio, running apps (not just third party apps). So much to be done that you can be sure 4 cores will be used.
Are the Cortez-A15s going to be ready for early 2012?
Well Yes, Cortex A15 Dual Core Could land as early as first half 2012 for Apple. But if it really is Quad Core then it will more likely be a A9 Quad Core .
The Sad thing is Cortex A15 Dual Core is actually faster then Quad Core A9, And it seems Apple will use Power VR 543MP4, if that is the case then it is basically the same chip used in upcoming PlayStation Portable.
Now Sony Ericisson is suppose to sample their Dual Core A15 + PowerVR 6 soon..........
Well Yes, Cortex A15 Dual Core Could land as early as first half 2012 for Apple. But if it really is Quad Core then it will more likely be a A9 Quad Core .
The Sad thing is Cortex A15 Dual Core is actually faster then Quad Core A9, And it seems Apple will use Power VR 543MP4, if that is the case then it is basically the same chip used in upcoming PlayStation Portable.
Now Sony Ericisson is suppose to sample their Dual Core A15 + PowerVR 6 soon..........
Yup, and Sony/Ericsson is doing real well I hear... </sarcasm>
For iOS or Mac OS X?
99% odds that they have Mac OS X running on ARM with a few test applications here and there. But my interest would lie with a FCPX version running on iOS with a touch UI, and it would be too much of a fantasy there.
There may actually be a market for a high end $1299 14"/15" landscape tablet with something like 1536x1024, 2056x1080 or all the way to 4k at 4096x2048. Quad core ARM with a PowerVR 554MP or 6-series GPU.
It could make FCPX pretty interesting to use. One could run 2 iPad apps side by side. Multitasking with two or more visible apps could be a reality.
Heavy though. 3 lbs? And it would flex a bit. And oh the glass would seem fragile at that size.
I was thinking FCPX running on iOS on ARM.
In addition to higher resolution display, I would like ti see a higher resolution (more granular) touch sensor .
This could support, but not require a stylus and crosshair hockey puck. The technology exists (Wacom) for a BT pressure sensitive stylus).
A design that appeals to me is the OTPC (was OLPC):
The tablet appears to be encased in plastic, including the (matt, outdoor) display -- and is targeted at $100 retail. I don't believe that anyone but Apple could make that happen in this decade.
But the plastic would help address the weight and flexibility problems of larger tablets.
2009
http://www.brightsideofnews.com/news...are-gone!.aspx
2010
http://semiaccurate.com/2010/02/27/w...ing-40nm-woes/
2011
http://www.techpowerup.com/148722/Fo...h-to-2012.html
Foundry Delays to Push Back NVIDIA Kepler Launch to 2012
It looks like foundry issues are back to slow down the launch of NVIDIA's next generation high-end GPU, codenamed Kepler. The delay may push Kepler's launch to 2012, it was expected to launch by late 2011. The 28 nanometer silicon fabrication process at TSMC, a principal foundry partner of NVIDIA, is producing unsatisfactory yields. Add to that, Kepler's performance is lower than expected.
TSMC's 28 nm process at Fab 15 facility has already seen delays, which have even shaped AMD's designs in a big way. AMD had originally planned to design high-end VLIW4 chips on the 32 nm process at TSMC, but later decided to wait for the 28 nm process, leading to plans of 32 nm GPUs being scrapped by both GPU designers. TSMC was supposed to be in a position take orders of high-end 28 nm GPUs by Q4 2011, and was set to start pilot production for its 20nm process technology in the third quarter of 2012.
TSMC's major client apart from GPU vendors, that relies on new bulk manufacturing process, Qualcomm, hasn't yet adjusted its 28nm process schedule and is set to launch three new products, 8960. 8270 and 8260A using dual-core Krait architecture in the fourth quarter of 2011.
AMD's first 28 nm GPUs, under the "Southern Islands" family are on course for 1H 2012. Around the same time, AMD will also release APUs based on the 28 nm process, codenamed "Krishna" and "Wichita", that replace Ontraio and Zacate.