If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
Hello,
Today Apple "released" the iPhone 4.0... I believe it has something like a 1.0 GHz processor... I find that quite impressive, 1.0 GHz in such a small package and non-overheating ??? Maybe to good to be true ? I wonder what the future will bring ?... Will we see the rise of "low power/low heat/low noise desktop computers" being powered by ARM-based processors ? Is this the end of Windows because it doesn't work on ARM processors ? Can intel atom processors compete with ARM processors ? What's AMD's answer to atom and arm ? Can an AMD/Intel single 1.0 to 2.0 GHz core be compared to ARM 1.0 to 2.0 Ghz Core ? Would they both be about as fast... or would one win over the other ? To me 1.0 to 2.0 GHz seems to be the magical number/milestone/border/hurdle towards a good to great desktop experience. For 99.9% of my daily PC activity 1.0 to 2.0 GHz would be enough... this almost includes video processing at modest resolutions 640x480 or so... maybe 800x600, maybe even 1024x768... further enhancements/optimizations might enable very large resolutions too but don't count on it For 1920x1200... 4.0 GHz is probably needed to run smooth and cool (strangely enough)... Or a really cool 2.0 GHz processor Only gaming does need stronger graphics cards and stronger cpu's to do more... However software/technology does advance so maybe I could be wrong an maybe people will need more processing power... but I don't think so... Therefore assuming all people need more processing power is a bit dangerous... A good secondary strategy is to focus on low power/low heat/low noise/weaker processors to accomadate non-gaming related and non-high-performance tasks/crowd I do want a low heat, low noise, low power computer, but I also want a strong, high performant computer which can do heavy tasks. I would love to have a computer which can be totally quite thanks to for example a ARM processor or maybe even an ATOM processor. I would also love it if the fans only go on when it's really needed like gaming or maybe huge video's. Thus I guess a system which can do both would be ideal for me. My current PC is already able to do this a little bit: AMD Dual Core Processor and NVIDIA 7900 GTX graphics card. But these two technologies do not take it far enough. The processor still needs a fan to spin. The graphics card still needs a fan to spin. The desktop still needs fans to be constantly on... - This is the biggest problem probably. Therefore what is needed is: 1. A motherboard which can control the desktop fans and even shut them down. 2. Processors/Graphics cards which can do the same. 3. Special software which can regulate this or special hardware. 4. Debuggers to make sure no evil "shut fans down during heat" is in there to kill hardware 5. Temperature meters everywhere for safety... 6. Emergency shutdown in case of emergency/accidental overheat. 7. Fan spin up failure detection. 8. Maybe even blocked air flow detection. 9. Maybe even unacceptable noise detection and throttling of hardware to reduce noise in return lower performance. 10. This would require microphones which might be too privacy-paranoya So not a good idea. 11. Maybe even build in temperature displays in/on the desktop case to show constant temperature of hardware at different locations in the case to feel "safe" Ultimately HEAT is bad though... even for the high performance situation. HEAT is unpleasant for the human beings... it can become to hot in summer. Assuming HEAT can be expelled from CASE and not be a problem could be wrong thing to do. HEAT also leads to bigger fans on buildings which is bad too. However... In the winter HEAT can working as heating device... and the problem is less big... it can actually be nice. Therefore producing more HEAT in winter is more acceptable... unless melting the polar caps is a bad idea ! And yup it could be bad... many countries facing floodings ! So maybe ultimately HEAT = BAD = EVIL. Try to use materials and designs which give great processing power but no to little heat New inventions are done all the time.... Are intel/amd/ati/nvidia up to the task ? Or will ARM take the cookie and the cake ?! (Just some random thoughts of me on the 1.0 GHz in a tiny package There is even talk of 1.5 GHz in iphone 5.0 wow ! ) Please feel free to comment within the lines and fill in the blanks, misconceptions, pipe-dreams, yes/no etc Bye, Skybuck =D |
#2
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
One problem which I see people mention is:
x86 software does not work on ARM... A solution for this problem is the following (Not my idea, but some crazy noob ?): An x86 compiler which compiles x86 to ARM code. It's a bit a crazy idea perhaps... But x86 is a instruction set/asssembly language after all as well... And languages can be ported/translated right ? Then for example Microsoft or the Users themselfes could do it. Microsoft's Windows on ARM could detect that the executable being installed or being tried to run is actually an x86 executable... Windows then starts the x86-to-ARM compiler and compiles the x86 binary to ARM binary... saves it and then runs it. With arm it is possible to add additional co-processors so maybe co-processors could handle some x86 specific tasks... for compatibility sakes or so... Maybe even an ARM/x86 hybrid ! LOL Or how about the ultimate crazy ****: PowerPC/ARM/x86/Motorola/ATI/Nvidia hybrid ?! LOL. Take the best of all or so Bye, Skybuck =D |
#3
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
"Skybuck Flying" wrote in message
b.home.nl... I find that quite impressive, 1.0 GHz in such a small package and non-overheating ??? That's transistor scaling for you. Will we see the rise of "low power/low heat/low noise desktop computers" being powered by ARM-based processors ? To some extent, yes. It's already happening -- look at the success of Intel's Atom CPUs. Is this the end of Windows because it doesn't work on ARM processors ? No. Windows (NT and beyond) has always had a "hardware abstraction layer" -- you'll recall that NT and XP used to come in versions for MIPS and Alpha CPU architectures. Writing an ARM HAL is no big deal at all for Microsoft -- there's rumors they might have already started looking that direction, e.g.: http://www.engadget.com/2009/05/01/a...rm-processors/ Can intel atom processors compete with ARM processors ? Sure. "Compete" means so much more than just MIPS per dollar or MIPS per watt today -- pricing is very important as well. Look at how "successful" Microsoft was in almost completely killing Linux's inroads into netbooks after they started offered XP Home to netbook manufacturers for $10/copy. What's AMD's answer to atom and arm ? They'd tell you Neo, although to date it doesn't compete all that well. For 1920x1200... 4.0 GHz is probably needed to run smooth and cool (strangely enough)... Or a really cool 2.0 GHz processor Well, if you force the CPU to do all the rendering, perhaps so. But plenty of systems ran at those resolutions back in the days of hundreds-of-MHz CPUs with the help of graphics processor ICs to do the "heavy lifting" for the display. Intel has, for years now, been fighting a battle with the GPU manufacturers over whether rendering is best done in their CPUs of the GPUs. (Notice how Intel doesn't have any GPUs to compete with those from AMD or nVidia...) However software/technology does advance so maybe I could be wrong an maybe people will need more processing power... but I don't think so... Most of technology today is driven much more by what people "want" than what they "need." Twenty years ago who would have thought that for ~$50 then (~$99 now) people would be getting CLOCK RADIOS that consist of a 454MHz CPU, 64MB of RAM, and 54Mbps wireless networking? But here we are today... and here's Chumby: http://www.costco.com/Browse/Product...rodid=11529965 I would love to have a computer which can be totally quite thanks to for example a ARM processor or maybe even an ATOM processor. There are entire web sites devoted to building siltent PCs. If you're willing to live with largish heatsinks, it's actually not particularly difficult to do -- fans are ubiquitous because they're small and cheap and effective compared to the brute-force alternative. I would also love it if the fans only go on when it's really needed like gaming or maybe huge video's. Pretty much all laptops and some desktop machines do this. Therefore what is needed is: 1. A motherboard which can control the desktop fans and even shut them down. 2. Processors/Graphics cards which can do the same. 3. Special software which can regulate this or special hardware. Already done. 4. Debuggers to make sure no evil "shut fans down during heat" is in there to kill hardware Mmm... ok... 5. Temperature meters everywhere for safety... Already done. 6. Emergency shutdown in case of emergency/accidental overheat. Already done (by BIOS). 7. Fan spin up failure detection. Already done (usually by BIOS). 8. Maybe even blocked air flow detection. Generally not worth the cost -- the system has enough thermal inertia that the overheat detectors will indirectly detect this before anything fries. 9. Maybe even unacceptable noise detection and throttling of hardware to reduce noise in return lower performance. Somewhat done. E.g., most hard drives have "best acoustic performance" and "best performance" modes. 10. This would require microphones which might be too privacy-paranoya So not a good idea. I suppose so. 11. Maybe even build in temperature displays in/on the desktop case to show constant temperature of hardware at different locations in the case to feel "safe" Already available. Therefore producing more HEAT in winter is more acceptable... unless melting the polar caps is a bad idea ! To a pretty good approximation, your computer produces no more heat in winter than in summer... it's just easier to move around when when there's a large temperature difference. So maybe ultimately HEAT = BAD = EVIL. No, "lack of heat gradients" might be bad or evil, but heat itself is neither. All the heat that will ever be is already here in the universe -- and the universe is a Good Thing -- it's just that it's slowly all mixing together, and once it's the same everywhere... we're cooked. :-) Keeping as much heat locked up in, say, chunks of coal rather than just heating up the air is rather useful. Are intel/amd/ati/nvidia up to the task ? Yes. Or will ARM take the cookie and the cake ?! No. ---Joel |
#4
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
"Skybuck Flying" wrote in message
b.home.nl... The motherboard has only one temperature sensor as far as I know ? For a motherboard in 2006, yeah, that's pretty likely. I also like to be able to see which parts of the motherboard are becoming the hottest. If you save up your allowace you could purchase a thermal imager? :-) Also my hardware from 2006 doesn't fully shutdown the fans the spindle just slowly... Yet you say it has already be done... I doubt it... but if I am wrong...: I've seen many a laptop that completely stops its fans when they're not needed... but personally not a desktop. I don't know why... Also my AMD X2 3800+ Dual Core CPU definetly does not detect CPU spin-up failure ! Blame your BIOS. Also I rather prefer not clunky big heatsinks... it's just heavy... risk of breaking motherboard... and it don't look so nice... it might also obstruct the airflow if it needs to scale up.... Big Clunky Heatsinks are definetly a NO-NO for me =D - They are windscreens... windscreens are evil inside a pc I need all the wind I can get in my PC to cool it down... unless I am in the desert or so which I am not (yet) lol That's kinda the problem with PCs... since it's "any motherboard, any CPU cooler, any case, hopefully it'll be cool enough?" you tend to end up with a lot of brute-force solutions like big fans rather than more elegant designs where hear flow is more precisely engineered to go certain places. (Even the old IBM PS/2 line had some very nice ducting in it...) Well you have made some claims that some to even all if this has already be done... I highly doubt that... but please do provide links to prove me wrong Check out http://www.silentpcreview.com/ -- those guys are serious about quiet computing. Lastly it's amazing to see how fast Apple has launched new products.... like 4 iphones in just 3 years ? Plus an iPad and maybe some PC like thingies... Doesn't sound like much... but I think it is... it requires all of this enginering of hardware and software... quite impressive ?!? People working at Apple all sign a contract stipulating that Steve Jobs gets to bowhunt you and your family on his private island if your performance review is deemed unsatisfactory. But seriously, yes, Apple's execution has been impressive -- and while I don't think that much of the man personally, one has to give credit that a large part of it is directly linked to Jobs. I do wonder what happened to Steve Jobs though... he so thin ?!? Did all that WIFI give him cancer or so ?!? No, but he had a liver transplant last year. Takes the wind out of most everyone for awhile... ---Joel |
#5
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + Highperformance ;))
On Jun 8, 6:30 pm, "Skybuck Flying" wrote:
One problem which I see people mention is: x86 software does not work on ARM... A solution for this problem is the following (Not my idea, but some crazy noob ?): An x86 compiler which compiles x86 to ARM code. Such programs already exist. It is a clever trick that is used to make fast simulations of the ARM on a PC. Doing it the other way also can be done. It wouldn't be super fast but if you weren't trying to run a complete Windows OS, it could be fast enough to be used. Since the ARM can be had as a part of a FPGA, you could add extra stuff to the standard ARM to make the process go a little faster. |
#6
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + Highperformance ;))
On Tue, 08 Jun 2010 19:25:33 -0700, MooseFET wrote:
On Jun 8, 6:30 pm, "Skybuck Flying" wrote: One problem which I see people mention is: x86 software does not work on ARM... A solution for this problem is the following (Not my idea, but some crazy noob ?): An x86 compiler which compiles x86 to ARM code. Such programs already exist. It is a clever trick that is used to make fast simulations of the ARM on a PC. Doing it the other way also can be done. It wouldn't be super fast but if you weren't trying to run a complete Windows OS, it could be fast enough to be used. Back in '87 or so I had an Acorn RISC "PC", which had an ARM-2, and a "PC emulator". It simulated an 8088 and the PC's basic hardware well enough that I was able to use it to run a "scientific" word processor to write my undergraduate thesis. The "feel" was about as fast as an original 4.77MHz PC, but I didn't run any benchmarks. I'm fairly sure that it would have been a straight interpreter: the machine didn't really have enough RAM to be mucking about with JIT compilation. This on a chip with no cache, no 16-bit memory operations, and which ran the processor clock at 4MHz or 8MHz depending on whether the DRAM-fetch in progress at the time was in-page or doing a row access... I thought it was quite a spectacular achievement. Cheers, -- Andrew |
#7
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
Check out http://www.silentpcreview.com/ -- those guys are serious about
quiet computing. Hmm... that's mosterd after the meal... Computer hardware needs to be designed from the start for low heat/low noise and so forth... But seriously, yes, Apple's execution has been impressive -- and while I don't think that much of the man personally, one has to give credit that a large part of it is directly linked to Jobs. He has gained some respect from me... he seems a more honest guy than I had expected him to be... at least in his presentations. However if the world turns into one big cancer infected place because of all the mobile phones and wifi's and gsm's and so forth than nope May he rott in hell then forever as well I do wonder what happened to Steve Jobs though... he so thin ?!? Did all that WIFI give him cancer or so ?!? No, but he had a liver transplant last year. Takes the wind out of most everyone for awhile... What was wrong with his ex-liver ? Cancer from the wifi ? What did he do with his ex-liver ? Bottle it for memories ? :P*** Ain't he afraid of getting cancer from all that wifi ? Bye, Skybuck. |
#8
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
Andrew Reilly writes:
Back in '87 or so I had an Acorn RISC "PC", which had an ARM-2, and a "PC emulator". That must have been the Archimedes A310. The RISC PC did not come out until some time in the 90's, and this used an ARM610. It simulated an 8088 and the PC's basic hardware well enough that I was able to use it to run a "scientific" word processor to write my undergraduate thesis. The "feel" was about as fast as an original 4.77MHz PC, but I didn't run any benchmarks. I'm fairly sure that it would have been a straight interpreter: the machine didn't really have enough RAM to be mucking about with JIT compilation. It was, indeed, an interpreter, but of the 80186 instruction set. File transfers etc. were a lot faster than on a 4.77MHz PC, but a few things were a bit slower. The overall speed was fine for running the occasional DOS application (I used it mostly for games), but for serious work, you would use native applications. This on a chip with no cache, no 16-bit memory operations, and which ran the processor clock at 4MHz or 8MHz depending on whether the DRAM-fetch in progress at the time was in-page or doing a row access... I thought it was quite a spectacular achievement. Indeed it was. Nowadays, you would use a JIT (similar to Digital's fx!32), so the speed would be better. ARM uses arithmetic flags similar to x86, so it is easier for ARM to emulate x86 efficiently than it is for, say, MIPS to do so. Torben |
#9
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + High performance ;))
"Skybuck Flying" wrote in message
b.home.nl... However if the world turns into one big cancer infected place because of all the mobile phones and wifi's and gsm's and so forth than nope I guarantee you that, whatever the potential health hazards posed by WiFi, GSM, etc. may be, there are orders of magnitudes more lives saved by wireless technology than lost due to it. |
#10
|
|||
|
|||
ARM-based desktop computer ? (Hybrid computers ?: Low + Highperformance ;))
On Jun 9, 12:51*pm, "Joel Koltner"
wrote: "Skybuck Flying" wrote in message b.home.nl... However if the world turns into one big cancer infected place because of all the mobile phones and wifi's and gsm's and so forth than nope I guarantee you that, whatever the potential health hazards posed by WiFi, GSM, etc. may be, there are orders of magnitudes more lives saved by wireless technology than lost due to it. I don't know. How often do you drive around people who drive will using a wireless gadget? I think I'd want to do some research before making any guarantees. Robert. |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Studio Hybrid Desktop | darkrats | Dell Computers | 0 | July 29th 08 05:20 PM |
IBM to build Opteron-Cell hybrid supercomputer of 1 PetaFlop performance | [email protected] | General | 27 | September 15th 06 01:22 PM |
IBM to build Opteron-Cell hybrid supercomputer of 1 PetaFlop performance | AirRaid | General | 1 | September 8th 06 09:48 PM |
IBM to build Opteron-Cell hybrid supercomputer of 1 PetaFlop performance | [email protected] | General | 0 | September 6th 06 02:00 AM |