As soon as you can address all your phisical memory with a single register, I'll give the OK; because the opposite, i.e., multiplex registers for addressing more memory is both *slow* and painful.
After a bit of investigation, I found that it is possible to do somewhat like bank switching (AWE), again, a weird thing like back in time the EMMS and XMMS bank switched memory. It could be useful for spare/temporal data, but out of the scope of the compile, i.e., via explicit memory handling.
AFAIK, the 36-bit address it is not intended for a *single* process, but for the OS; i.e., the OS could manage the 36-bit address range, but with 32-bit processes.
P.S. I lived the 16-bit era, and I hate it absolutelly, specially those weird "far calls" (seg:off, as, (seg4 | off)), I hated, and still hate, to program 8086/V20/V30 in a multi-segment approach (the tiny (64K)/small(64/64K) models were ok, but compact(64/XKB) and huge(X/YKB) were terrible to deal , being hard to get a nice code design without ugly tricks (both in C and assembly). I was in joy with the Watcom C 32-bit DOS extensor, also with the DJ Delorie' DJGPP GCC DOS port and its GO32 32-bit DOS extender.
Ok, touched. Then, limit it to people without penal antecedents or with a violent historic.
I do not want to use any demagogic argument, as I'm really sad by the incident at Virginia, but, I want to believe that *many* cases could have been predicted via medical or penal data.
[Disclaimer: I'm against weapons for "personal defense"]
Yes, while these journalists use demagogic arguments to point video games as the "cause", I think that the point is: how the hell could be so easy for anybody to get such arsenal?!
While I respect, but not endorse, the decision of a country/state to allow weapons for personal defense, I think that "serious" psychological analysis should be mandatory before some could bring/buy weapons.
The Spanish Civil War (1936-1939) was started as an insurrection from a fascist movement, intended to finish the Second Spanish Republic
While the fascist movement was backed by both German-Nazi and Italian-fascist movements, the Spanish Republic was backed just by volunteers, known as International Brigades; the external Nazi supplies was key for defeating the democratic goverment by the fascists.
Despite the sad result, I disagree with your arguments; in my opinion, the Nazi/fascist movement was *neglected* until Germany invaded Poland (1939). That is why I can not agree with your argument: the "world" was not "terribly busy" in 1936, at least not Europe (I know that USA was still recovering from its 1929/1930 recession), Europe could have done *a lot more* for Spain at the time. Well, I hope we'll learn from past mistakes, although I'm not that optimistic.
I was skeptic about the Cell chip, but after some homebrewed tests (explicit unrolling, some SPU handmade optimizations, etc.), I have to admit that, despite that both PPE and SPEs have to be programmed with its limitations in mind to avoid in-order CPU with pipeline stalls with a heavy branch penalties -not trivial if you want to get top performance-, it is still a *monster*, a *true* supercomputer on a chip (the PS2's "Emotion Engine" was similar in concept but much limited for interprocess communication). As example, on a given SPU, you can order DMA transfers while working in parallel (working like explicit prefetch or automatic DMA), and that is true for the other SPU's. Such grade of paralelism and throughput is overkill.
P.S. I complain Sony's decission of hiding the graphics chip from the Linux environment: too bad! (Cell + GPGPU == state of the art computing luxury!)
Well, that is one of the sides of the blade. The other implies unique moderation vision. I think that the moderation ban it is too much sensitive, as it has to cut abusive modders, but not to eliminate different points of view. The typical algorithm: "if you're against the 90% of the moderation you're a radical", in my opinion, should not be an argument unless you're that radical in most ways.
I know that diversity preservation it is not easy, as moderation/metamoderation are based in very simple heuristics, which often leave the picture biased (I'm pro-democracy, but I like to see *some* contradiction in moderation).
P.S. yes, I'm not in the moderation "business" for a while;-)
You're right, it also work in Vista (the only intouchable processes are the "protected" ones, like Winlogon). Unfortunately, we are forced to use it, for achieve trivial tasks such grabbing the cursor icon used inside third party application window, and other "accessibility" hooks, as the officially provided by Microsoft are not enough. Personally, I hate to use these tricks, IMO it is the result of a bad OS design, as we are not using it for "evil" applications, just normal ones that make the average Joe's life easier.
I think that something similar is also happening in Barcelona (Spain, Europe): there is stock in every big store (AFAIK). One of the biggest electronic retailer chain (Media Markt) is giving a Blu Ray movie for free with the PS3 (599 euro) since today (no gift was given at the launch day), other stores are offering games with 25% discount (Carrefour, since the first launch day) when bought with the console. FNAC offers 10% discount for games and peripheals if you buy the console there, with an additional 5% (or 2%, please confirm it) discount if you have paid the discount fee (a weird discount subscription).
My intuition tells me that Sony, after the first big fan wave buyers, will have to shrink prices notably, at least in Spain: we'll enter in recession in about 6 to 9 months (European interbank mortgage interest rate, "euribor", is rising fast, and people have serious problems, cutting from superfluous things); a real estate crash is also possible, we are looking very close to the US problems, as we'll probably the next affected.
Re:To advance, correct errors rather than drag it.
on
The Future of the PSP
·
· Score: 1
Stock-zero products, i.e., products that are stored as much a week, can be sell with tiny margins. As example, your ill-fated PS3 is retailed with a 5% margin (at Barcelona, Spain, Europe), with another 5% for the medium distributor, and another 5% for the importer. If you are a huge retailer chain, you can get the whole 15% profit margin (Walmart anyone?) or keep under 10% and smash the little stores. Indeed, you can achieve higher margins, usually related to long time selling articles, such as rare cables/adapters (>70% margin, the less it costs the higher profit margin).
Given a PSP (MIPS R4K-like 333MHz, 3D chip, 32 + 4MB DRAM, etc), with a high quality but small TFT (~4.3"), with just one "all in one big chip" (e.g. 90nm, and 65nm in the future) with few additional ICs. That is suitable for a high integration, cheap, "big" IC; there is nothing exceptional. The only handicap is the screen, but still with a 30% increase, a 5.6" TFT screen is not that expensive. Why chinese manufacturers can do similar things with portable DVDs and Sony with much better fabs couldn't? (!)
Re:To advance, correct errors rather than drag it.
on
The Future of the PSP
·
· Score: 1
Do you? Then, enlightme with your deep world knowledge, please.
Re:To advance, correct errors rather than drag it.
on
The Future of the PSP
·
· Score: 1
ack 1: I don't want to carry UMD disks anymore.
ack 2: Why? TFT screens of 30% more area than current PSP are dirty cheap.
ack 3: Design costs for flow shop multi-million-unit gadgets is below 0.5$/unit, marketing is a lot more expensive.
ack 4: It was about reducing costs plus a bigger screen as a plus, 99$, IMO is an easy price target for a massive integrated device.
ack 5: For a retail price of 99$, including three levels of distribution (import, major distribution, retailer) for 14$, you have 85$ left for the making of the device and keep from 0 to 7% profit margin. Further reading: how to make money selling games.
ack 6: well, was just an idea; I'm glad to ear yours.
To advance, correct errors rather than drag it.
on
The Future of the PSP
·
· Score: 0
1) Remove the UMD disk interface: it is slow, mechanic, and power hungry.
2) Bigger screen.
3) Thinner design, lower weight, lower power requirement, memory-card-only games, allow SD cards for user media.
4) Sell it at 99$ / 99 euro (including european VAT) price point.
5) Profit! (without user torturing, dammit!)
By the way, for the next PSP (e.g. "PSP2"): memory-card-games backward compatible, with better processors, more RAM, and enhanced 3D chip, etc.
You're right; current hypervisors are not really designed for running under themselves nor, in most cases, for running other supervisors, quite primitive everything. I was mostly figuring hypothesis, don't take me seriously:-)
That is because we're assuming one-way-forward interlayer knowledge, i.e., layer 0 emulates layer 1, layer 1 emulates layer 2, but layer 0 knows nothing about layer(s) at level 2. Allowing layer-mixing/inference via abstraction, could increase performance, as soon as higher level I/O bottleneck or CPU stall analysis could allow better instruction mixing or let other layer continue on execution while the conflictive is rescheduled. For me it is likely to how a OS works, but change the basic unit (thread/process) by a virtual machine. Current "hypervisors" do profiling for both CPU and I/O balancing, to maximize throughput, latency, or other exotic profiles.
Why not? Just you get execution speed and available memory progressively downgraded/shrinked (doing it ad infinitum or ad nauseam, until you're out of memory for the next emulation context):~P
Circa 1999 I used to develop under Digital UNIX, with 21164@600MHz CPUs, where longs where 64 bit (Digital C compiler, LP64). Not about elitism, but to broke my many years old 64 bit code snippets (!), which still work flawlessly under most UNIXes/Linux, but are broken for Windows x64. Click here for more info.
Yes, I {agree with | like} your the single/double precision definition, however, it is common to accept single as 32 bit ("float" data type), just like "int" is 32 bit (except on weird 16 bit compilers, where "int" were 16 bits and "long" 32 bits). Sometimes data types are, in my opinion, abused, like the Microsoft LLP64, keeping "long" data type for 32 bits instead the classical UNIX use of the "long" data type related to the microprocessor register width (when the microprocessor processor register >= 32 bits), just for making the application porting easier (usually related to lazy/ignorant programmers). What a wonderful world;-)
PS3 Cells are optimized for single precission operations (32 bits), but future Cells, for other users than PS3s, will support double precission (64 bit) with no performance loss.
I feel a similar sensation in Spain, where I live and work. To be honest, my main preocupation is not related to the CS demand, but the inminent spanish economic recession (we also have a huge risk of real estate bubble bursting).
Despite that "Mac" and "iPod" are wide known trademarks, and the "desktop PC" having just two main uses, office and home, there is a hidden mesh of dependencies that makes you to choice the MS Windows option. Today is much easier to migrate from Windows your e-mail and administrative/office applications, but most people don't want to take care of extra annoyances, that is, IMO, the MS de facto monopoly key point.
While OSX is a good desktop OS, as it is XP or Vista, if you want to live without constraints and breath freedom enjoying a highly customizable desktop, your choice is GNU/Linux. Many fears come to most people: What if my disk fails to boot? Where is the Outlook e-mail? Can I play -put a D3D game here-? (...) Where you reply these questions, most users appear to be discouraged. May be the solution could come from some kind of loadable OS, a la Knoppix, using the disk as CD/DVD image cache (with dynamic patching), fault tolerant (fixable via copying the DVD image again), while the data is in alternate disk partitions... et voilà, you've got a crash-proof OS. The Vista FLASH cache techniques are, IMO, just a imagination deficit: if you have mechanic disks with >50MB/s sequential read, why the hell don't you fill 256 or 512MB of RAM with a previously stored OS image?
Yes, that is because it is nonsense for me: If I have enough rights to install a driver to simulate a Ctrl-Alt-Del, then, why I can not do it from a service in userland with "system" privileges? (as I can do it after all, security through obscurity anyone?!)
I acknowledge that Vista has great architecture enhancements (session 0 isolation, secure processes, process scheduler enhancements, networking throughput, etc.), but IMO, it is nonsense to limit "system" or "local system" userland capabilities when you're able to install a driver to surpass this: it looks like a weird workaround. The security approach itself it is easily bypassable *unfortunatelly*, the big Vista feature for developers is "developer headache".
Slashdot Mirror
As soon as you can address all your phisical memory with a single register, I'll give the OK; because the opposite, i.e., multiplex registers for addressing more memory is both *slow* and painful.
After a bit of investigation, I found that it is possible to do somewhat like bank switching (AWE), again, a weird thing like back in time the EMMS and XMMS bank switched memory. It could be useful for spare/temporal data, but out of the scope of the compile, i.e., via explicit memory handling.
AFAIK, the 36-bit address it is not intended for a *single* process, but for the OS; i.e., the OS could manage the 36-bit address range, but with 32-bit processes.
P.S. I lived the 16-bit era, and I hate it absolutelly, specially those weird "far calls" (seg:off, as, (seg4 | off)), I hated, and still hate, to program 8086/V20/V30 in a multi-segment approach (the tiny (64K)/small(64/64K) models were ok, but compact(64/XKB) and huge(X/YKB) were terrible to deal , being hard to get a nice code design without ugly tricks (both in C and assembly). I was in joy with the Watcom C 32-bit DOS extensor, also with the DJ Delorie' DJGPP GCC DOS port and its GO32 32-bit DOS extender.
Ok, touched. Then, limit it to people without penal antecedents or with a violent historic.
I do not want to use any demagogic argument, as I'm really sad by the incident at Virginia, but, I want to believe that *many* cases could have been predicted via medical or penal data.
In Spain, the margin for console is just a mere 5% (I can confirm it for PS3, but I suppose that figures are similar for the 360 and the Wii).
[Disclaimer: I'm against weapons for "personal defense"]
Yes, while these journalists use demagogic arguments to point video games as the "cause", I think that the point is: how the hell could be so easy for anybody to get such arsenal?!
While I respect, but not endorse, the decision of a country/state to allow weapons for personal defense, I think that "serious" psychological analysis should be mandatory before some could bring/buy weapons.
Please, US people, think about it!
The Spanish Civil War (1936-1939) was started as an insurrection from a fascist movement, intended to finish the Second Spanish Republic While the fascist movement was backed by both German-Nazi and Italian-fascist movements, the Spanish Republic was backed just by volunteers, known as International Brigades; the external Nazi supplies was key for defeating the democratic goverment by the fascists.
Despite the sad result, I disagree with your arguments; in my opinion, the Nazi/fascist movement was *neglected* until Germany invaded Poland (1939). That is why I can not agree with your argument: the "world" was not "terribly busy" in 1936, at least not Europe (I know that USA was still recovering from its 1929/1930 recession), Europe could have done *a lot more* for Spain at the time. Well, I hope we'll learn from past mistakes, although I'm not that optimistic.
I was skeptic about the Cell chip, but after some homebrewed tests (explicit unrolling, some SPU handmade optimizations, etc.), I have to admit that, despite that both PPE and SPEs have to be programmed with its limitations in mind to avoid in-order CPU with pipeline stalls with a heavy branch penalties -not trivial if you want to get top performance-, it is still a *monster*, a *true* supercomputer on a chip (the PS2's "Emotion Engine" was similar in concept but much limited for interprocess communication). As example, on a given SPU, you can order DMA transfers while working in parallel (working like explicit prefetch or automatic DMA), and that is true for the other SPU's. Such grade of paralelism and throughput is overkill.
P.S. I complain Sony's decission of hiding the graphics chip from the Linux environment: too bad! (Cell + GPGPU == state of the art computing luxury!)
Well, that is one of the sides of the blade. The other implies unique moderation vision. I think that the moderation ban it is too much sensitive, as it has to cut abusive modders, but not to eliminate different points of view. The typical algorithm: "if you're against the 90% of the moderation you're a radical", in my opinion, should not be an argument unless you're that radical in most ways.
;-)
I know that diversity preservation it is not easy, as moderation/metamoderation are based in very simple heuristics, which often leave the picture biased (I'm pro-democracy, but I like to see *some* contradiction in moderation).
P.S. yes, I'm not in the moderation "business" for a while
You're right, it also work in Vista (the only intouchable processes are the "protected" ones, like Winlogon). Unfortunately, we are forced to use it, for achieve trivial tasks such grabbing the cursor icon used inside third party application window, and other "accessibility" hooks, as the officially provided by Microsoft are not enough. Personally, I hate to use these tricks, IMO it is the result of a bad OS design, as we are not using it for "evil" applications, just normal ones that make the average Joe's life easier.
I think that something similar is also happening in Barcelona (Spain, Europe): there is stock in every big store (AFAIK). One of the biggest electronic retailer chain (Media Markt) is giving a Blu Ray movie for free with the PS3 (599 euro) since today (no gift was given at the launch day), other stores are offering games with 25% discount (Carrefour, since the first launch day) when bought with the console. FNAC offers 10% discount for games and peripheals if you buy the console there, with an additional 5% (or 2%, please confirm it) discount if you have paid the discount fee (a weird discount subscription).
My intuition tells me that Sony, after the first big fan wave buyers, will have to shrink prices notably, at least in Spain: we'll enter in recession in about 6 to 9 months (European interbank mortgage interest rate, "euribor", is rising fast, and people have serious problems, cutting from superfluous things); a real estate crash is also possible, we are looking very close to the US problems, as we'll probably the next affected.
Stock-zero products, i.e., products that are stored as much a week, can be sell with tiny margins. As example, your ill-fated PS3 is retailed with a 5% margin (at Barcelona, Spain, Europe), with another 5% for the medium distributor, and another 5% for the importer. If you are a huge retailer chain, you can get the whole 15% profit margin (Walmart anyone?) or keep under 10% and smash the little stores. Indeed, you can achieve higher margins, usually related to long time selling articles, such as rare cables/adapters (>70% margin, the less it costs the higher profit margin).
Given a PSP (MIPS R4K-like 333MHz, 3D chip, 32 + 4MB DRAM, etc), with a high quality but small TFT (~4.3"), with just one "all in one big chip" (e.g. 90nm, and 65nm in the future) with few additional ICs. That is suitable for a high integration, cheap, "big" IC; there is nothing exceptional. The only handicap is the screen, but still with a 30% increase, a 5.6" TFT screen is not that expensive. Why chinese manufacturers can do similar things with portable DVDs and Sony with much better fabs couldn't? (!)
Do you? Then, enlightme with your deep world knowledge, please.
ack 1: I don't want to carry UMD disks anymore.
ack 2: Why? TFT screens of 30% more area than current PSP are dirty cheap.
ack 3: Design costs for flow shop multi-million-unit gadgets is below 0.5$/unit, marketing is a lot more expensive.
ack 4: It was about reducing costs plus a bigger screen as a plus, 99$, IMO is an easy price target for a massive integrated device.
ack 5: For a retail price of 99$, including three levels of distribution (import, major distribution, retailer) for 14$, you have 85$ left for the making of the device and keep from 0 to 7% profit margin. Further reading: how to make money selling games.
ack 6: well, was just an idea; I'm glad to ear yours.
1) Remove the UMD disk interface: it is slow, mechanic, and power hungry.
2) Bigger screen.
3) Thinner design, lower weight, lower power requirement, memory-card-only games, allow SD cards for user media.
4) Sell it at 99$ / 99 euro (including european VAT) price point.
5) Profit! (without user torturing, dammit!)
By the way, for the next PSP (e.g. "PSP2"): memory-card-games backward compatible, with better processors, more RAM, and enhanced 3D chip, etc.
You're right; current hypervisors are not really designed for running under themselves nor, in most cases, for running other supervisors, quite primitive everything. I was mostly figuring hypothesis, don't take me seriously :-)
That is because we're assuming one-way-forward interlayer knowledge, i.e., layer 0 emulates layer 1, layer 1 emulates layer 2, but layer 0 knows nothing about layer(s) at level 2. Allowing layer-mixing/inference via abstraction, could increase performance, as soon as higher level I/O bottleneck or CPU stall analysis could allow better instruction mixing or let other layer continue on execution while the conflictive is rescheduled. For me it is likely to how a OS works, but change the basic unit (thread/process) by a virtual machine. Current "hypervisors" do profiling for both CPU and I/O balancing, to maximize throughput, latency, or other exotic profiles.
Why not? Just you get execution speed and available memory progressively downgraded/shrinked (doing it ad infinitum or ad nauseam , until you're out of memory for the next emulation context) :~P
Circa 1999 I used to develop under Digital UNIX, with 21164@600MHz CPUs, where longs where 64 bit (Digital C compiler, LP64). Not about elitism, but to broke my many years old 64 bit code snippets (!), which still work flawlessly under most UNIXes/Linux, but are broken for Windows x64. Click here for more info.
Yes, I {agree with | like} your the single/double precision definition, however, it is common to accept single as 32 bit ("float" data type), just like "int" is 32 bit (except on weird 16 bit compilers, where "int" were 16 bits and "long" 32 bits). Sometimes data types are, in my opinion, abused, like the Microsoft LLP64, keeping "long" data type for 32 bits instead the classical UNIX use of the "long" data type related to the microprocessor register width (when the microprocessor processor register >= 32 bits), just for making the application porting easier (usually related to lazy/ignorant programmers). What a wonderful world ;-)
PS3 Cells are optimized for single precission operations (32 bits), but future Cells, for other users than PS3s, will support double precission (64 bit) with no performance loss.
I feel a similar sensation in Spain, where I live and work. To be honest, my main preocupation is not related to the CS demand, but the inminent spanish economic recession (we also have a huge risk of real estate bubble bursting).
Well, after your exposition, would I feel myself much less human? ;-)
--
The good life is one inspired by love and guided by knowledge.
Despite that "Mac" and "iPod" are wide known trademarks, and the "desktop PC" having just two main uses, office and home, there is a hidden mesh of dependencies that makes you to choice the MS Windows option. Today is much easier to migrate from Windows your e-mail and administrative/office applications, but most people don't want to take care of extra annoyances, that is, IMO, the MS de facto monopoly key point.
While OSX is a good desktop OS, as it is XP or Vista, if you want to live without constraints and breath freedom enjoying a highly customizable desktop, your choice is GNU/Linux. Many fears come to most people: What if my disk fails to boot? Where is the Outlook e-mail? Can I play -put a D3D game here-? (...) Where you reply these questions, most users appear to be discouraged. May be the solution could come from some kind of loadable OS, a la Knoppix, using the disk as CD/DVD image cache (with dynamic patching), fault tolerant (fixable via copying the DVD image again), while the data is in alternate disk partitions... et voilà, you've got a crash-proof OS. The Vista FLASH cache techniques are, IMO, just a imagination deficit: if you have mechanic disks with >50MB/s sequential read, why the hell don't you fill 256 or 512MB of RAM with a previously stored OS image?
Yes, that is because it is nonsense for me: If I have enough rights to install a driver to simulate a Ctrl-Alt-Del, then, why I can not do it from a service in userland with "system" privileges? (as I can do it after all, security through obscurity anyone?!)
I acknowledge that Vista has great architecture enhancements (session 0 isolation, secure processes, process scheduler enhancements, networking throughput, etc.), but IMO, it is nonsense to limit "system" or "local system" userland capabilities when you're able to install a driver to surpass this: it looks like a weird workaround. The security approach itself it is easily bypassable *unfortunatelly*, the big Vista feature for developers is "developer headache".