User talk:Janagewen
Please do not leave your notes here! This page is only reserved!
.net framework template
| Version | CLR | Release date |
Shipped With Visual Studio |
Preinstalled with Windows | Embraces | |
|---|---|---|---|---|---|---|
| Client | Server | |||||
| 1.0 | 1.0 | February 13, 2002 | .NET 2002 | — | — | — |
| 1.1 | 1.1 | April 24, 2003 | .NET 2003 | — | 2003 | — |
| 2.0 | 2.0 | November 7, 2005 | 2005 | — | 2003 R2 | — |
| 3.0 | 2.0 | November 6, 2006 | — | Vista | 2008 | 2.0 |
| 3.5 | 2.0 | November 19, 2007 | 2008 | 7 | 2008 R2 | 3.0 |
| 4.0 | 4 | April 12, 2010 | 2010 | — | — | — |
| 4.5 | 4 | August 15, 2012 | 2012 | 8 | 2012 | 4.0 |
| 4.5.1 | 4 | October 17, 2013 | 2013 | 8.1 | 2012 R2 | 4.5 |
| 4.5.2 | 4 | May 5, 2014 | — | — | — | 4.5.1 |
| 4.5.3 | 4 | TBA | — | 10 | 2015 | 4.5.2 |
About date formant, because .net framework is a product of Microsoft, a North American international corporation. In order to avoiding the confusion, using American Standard Date format to state its precise release date.
(Tablets) Recreate the picture of Yamhill, EM64T and IA-32e
If I have never argued with professor Jeh I would never have the will to guess how to recreate picture of Yamhill, EM64T and Ia32e. Yamhill is the code name of a project for ever Intel's own 64bit x86 processor, EM64T is a technology to extract the most potential capability of IA-32, making it to address up to 64TiB physical address. IA32e is not a mode of Yamhill processor, but the superset or extension of IA-32e, in other words, when ia32e enabled 32bits and 64bits codes could work seamlessly, because ia32e scale up the original IA-32 by extending the width of registers and execution units, extended segment descriptors and paging mechanism would also introduced with it. IA32e does not make a processor turn into a real 64 bit processor, but just make a 32 bit processor has the capability to do the 64 bit computing from the ground. The benefits of this is that we don't have a second 64bit-lised IA-32, what we need is an architecture completely designed for 64 bit computing from scratch, for example Itanium or something else. --Janagewen on travel — Preceding unsigned comment added by 42.2.159.106 (talk) 00:30, 29 October 2014 (UTC)
Segmentation or Paging is the Essential Way for Physical Addressing
Intel 8086 is the very first processor in the x86 series, it has only one mode, real mode. And this name was later found in 80286. So segmentation had been accompanied with x86 processor since its beginning. Without protection mechanism introduced, segmenataion is the only way to address all possible and potential physical space. Because Segmentation is found in the real mode, so addresses in Segmentation tends to be real, with solid link the physcial address rather than virtual address when in protected mode when paging is enabled. So if it is a real x86 processor, its maximum physical addressing space should be limited by its segmentation mechanism rather than paging.
Protected Mode and Long Mode
Protected Mode was first introduced in Intel 80286, providing mechanism protecting resource and assisting multi-task OS such as UNIX, OS/2 and so forth, and enhanced in Intel 80386 processor by extending computing capability to 32 bits from 16 bits, eventually becomes part of IA-32 architecture. Most x86 OS only running within protectected mode such as Windows NT, Linux, and so many BSD distros; but MS-DOS working under real mode, some extender might switch through and back from protected mode for extending physical memory addressing, likewise, Windows 3.x also transit itself from real mode and protected mode thorugh and ack too. Windows 9x not only depends the real mode as bootstrap, but also sits itself from and to that mode.
Long Mode is a kind of protected mode, not a real mode, but does not belong to the Protected Mode in IA-32, switches itself out of IA-32 dominated Legacy Mode, to a completely self-contained mode, which means even though a x64 processor eliminates its legacy part, it would be fulfunctioned. From the aspect of implementation, or Microarchitecture level, the legacy mode codes could share the same physical execution resources designed for the x64 codes, without needing additional physical execution units for itself, like it found in early Itanium processor. So IA-32 architecture might or might not have relationship with x86-64, x64, AMD64 or Intel64.
IA32e and AMD64
AMD64 is a switched-out architecture from IA-32, providing compatibility mode to support most IA-32 software and representing itself in the Long Mode within AMD64 processors. It is a new architecture, even though most lying on IA-32 architecture.
IA32e is not a mode, but an extention of IA-32, or newer revision of IA-32, introducing 64-bit computing capability for its fundemental architecture. Intel EM64T is the new brought feature to this kind of processors, enabling it for 64-bit computing. Of course, EM64T is not a mode too, but a feature like PAE, could be enabled to extend the 64-bit computing capability. Without EM64T enabled, Yamhill processor would act like a tradtional Pentium 4.
After EM64T enabled, 64-bit registers would be visible to applications. 32-bit traditional OS might need minor fixes to support applications referring 64-bit registers to do the 64-bit computings, but programmers must take care of the computing results without intervented by other processes. But most likely as it in PAE, OS might need a patch to support this feature totally, at least to manage the extended descriptors and extended 64-bit registers and stacks and so forth, but without needing to compile the OS from the scratch like it in AMD64. Virtual Address might be extended up to 52 bits, while physical address up to 46 bits. 16K segments of 4GiB could be spreaded to 64TiB physical space, with or without paging. Because the difference between 32bit and 64bit computing is bit width of register rather than the whole system, so 32-bit and 64-bit codes is also differentiated by the referenced resgistered without needing a submode, such as compatibility mode to support tradtional 32-bit software.
After EM64T disabled, Yamhill processor would act like a Northwood Pentium 4. Janagewen (talk) 10:22, 4 November 2014 (UTC)
Both IA32e and AMD64 are the successors to the then-current 32-bit x86 processors, AMD64 follows the tradtion of x86 processor's mode-transition, first found in Intel 80286, where Protected Mode was introduced alongside with Intel 8086 mode, or Real Mode exactly, and goes much futher. Long Mode is not the third mode in current x86-64 processor alongside with Protected Mode, if Virtual 86 Mode could be counted as one submode of it, but completely a switch-out mode out of the environment of IA-32. Currently, no processor could boot up into this Long Mode directly, but must jump over real mode, protected mode, through compatiblity mode and eventually goes to the pure 64-bit mode. x64 version of UEFI make sure only x64 OS could be installed and boot-up, but without meaning that processor boot into Long Mode directly. Tradtional 32-bit OS would make this kind of processor just jump into the interim stage, Protected Mode, leaving the 64-bit environment and resource never touched. Apple's Mac OS X 10.5 Leopard, 10.6 Snow Leopard and 10.7 Lion could make Intel 64 processor switch to long mode and back when kernel running within Legacy Mode or Protected Mode exactly, where applications are 64-bit. Mode Switch is a long time journey, microarchitecture might also need a system level refresh to do it, so it often tends to be an inefficient algorithm. But to Apple, no matter architectures or processors, everything count as nothing but the whole belief behind them could make everything look the same but different, or else what you need is to customize your own system with your favourate OSes rather than OS X. Mode Transition is a way to succeed the x86 processor into completely a new generation, and if without consideration of back compatiblity, old modes could be further eliminated in the future, but must retain enough essential old instructions to keep the old OS could be run without problems. If what I guess is not too wrong, Intel's initial IA32e is a direct descent of IA-32, without any new mode introduced to the Yamhill processors, but providing the 64-bit extended computing capability. Old OSes could run on this platform without problems, and application programmers could carefully use 64-bit computing resource to accelerate their algorithms in the form of 32 bit execution image. If some a 32-bit OS could be further patched, pure 64-bit application could run under the 32-bit OS without mode transition, but only 4GiB physical space would be address directly simultaneously. If some a 32-bit OS further implement the IA32e feature completely, like PAE, extended segmenting and paging would be supported, then up to 64TiB physical space could be supported, and potential 64-bit virtual address would also be reached. Traditional 32-bit application without needing containers like Wow64 to be executed in, just leaving segmenting and paging up to 4GiB physical and virtual space is enough. Things would be much simpler than what AMD64 did, and 8 additional general registers found in AMD64 might not improve the performance of IA-32 obviously, smarter microarchitecture desgin might play the key role. But Intel never disclose their own IA32e to the public, instead, just reserved the name IA32e and EM64T in their AMD64 compatible processors, Intel 64 processors.
For tablet processors, AMD64 architecture is a monster, and inappropriate. The less number of semiconductors a tablet has, the better power saving would have. The less overhead an OS wastes, the less physical resource would be needed for a configuration. So many China Mainland manufacturers tend to put a desktop computer configuration onto a tablet form only embrace the purpose to support a de-facto desktop OS and make it run fluently. I would never love that taste. If Yamhill does really exist, and IA32e does really ever had been brought out, and that might be the key to conquer the field of today's 64-bit ARM world. Janagewen (talk) 11:29, 6 November 2014 (UTC)
(Game Consoles) Trend of PlayStation 4 and Xbox One
I paid visits to many stores in Hong Kong about PlayStation 4 and Xbox One, but unfortunately, I stood there for minutes but ignored completely, without a welcome to me in any language. Eventually I dropped buying both of them, waiting the next chance paying visits to shops in Japan, when new versions or revisions will come.
For PlayStation 3 and Xbox360, both are using 64-bit PowerPC based processors, all the revisions tend to be the same specification and performance, only different on the semiconductor process, from early 90nms to the latest 40 or 45nms. But for PlayStation 4 and Xbox One, I wouldn't think that would last for this tradition, because they use the common processors from AMD, Jaguar, which would easily be upgraded. So I think the future PlayStation 4 and Xbox One might update their processors not only on the semiconductor process, but also the frequency and even underlying microarchitecture. Janagewen (talk) 23:37, 9 November 2014 (UTC)
Intel i series and AMD A-series processors
I always guess why Intel name their current processors as i3, i5 and i7, and AMD name their mainstream processors as A4, A6, A8 and A10. Or what does "i" and "A" really mean respectively? They both might embrace more meanings, but at least each have two major meanings, for "i", it means Intel and "integrated", integrating processors core, GPGPU and other relative components; while "A" means AMD and "accelerated", with fused GPGPU to accelerate the whole system performance. They both might be the same or similar things for end users, but in essence, they both are completely different things. For Intel i series processors CPU core is the major focus, even though GPGPU also has general computing capability, but just provides for essence needs rather than accelerating the whole system; while AMD put their focus on to the GPGPU or the whole system performance, the fusion exactly. Fusion is different from integration, integration might only put different things together without change their natures, but fusion also use the chemical and physical changes to put different things turn in to the one with their natures changed too. The eventually target for the APU is to merge the different computing resources into one, making different ISAs with the common interface, so merely comparing performance of CPU core of AMD A-series processors seem meaningless. Currently, AMD A-series processors draw the end users a rectangle rather than a line for computing, the AMD Radeon GPGPU lengthen the width (VLIW or RISC SIMD), while the Bulldozer CPU core lengthen the Length (Hyper-Pipelining). They both enlarge this rectangle from different directions, and eventually draw a surface. But Intel i-Series processor might only lengthen a thicker line for performance, does not change the traditional computing way too much.
Of course, Intel and AMD both have successful history for processors, i3 might resemble the most outstanding 80386 processors, while A4 for AMD 486 processors; i5 for 80586/Pentium the very first Intel superscalar x86 processors, while A6 for most successful AMD K6 processors series, A8 for AMD's most self-proud 64 bit hammer processors; i7 for the next generation of x86 succeeding the P6 and its descent, while A10 for AMD's future generation x86 processors.
Future Xbox One and PS4 might Use A-Series Processor
Both Cell and Xenon are two way hyper-pipelining processors, they might seem the Pentium 4 version of Power PC 970. For comparison, AMD Bulldozer have something similar with them, and possess much more flexibilities than ever. Each integer thread is a bit thicker with two symmetric execution units to do actual jobs, and both thread have potential chance to be combined onto a much thicker one; from the other side, each integer thread could be further divided to two much thinner threads, then each module possesses four threads like this, so one cluster might possess eight threads, and two clusters possess 16 threads. These too much thinner threads might seem less important in general computing, but might be much more useful for game console. Today's game console is almost a graphical workstation rather than something like tablets, from this aspect, Bulldozer based A-Series processors much overwhelms Jaguar based traditional simple design, and the advanced features which Bulldozer possesses might enable hybrid OS like it found in PS3/PS4 might make even more room and benefits from each other making an outstanding workstation for game playing specifically. Janagewen (talk) 02:23, 15 November 2014 (UTC)