A 1 micron^2 SRAM cell should be available in a 120nm process. A 0.6 micron^2 cell for 90nm process. Which is about the smallest avaliable now. Thus IBM's SRAM cell must be 0.06 micron^2.
Why can marketing people not put in these basic and important (at least for me) facts?
the IT group did that where I work. Now you never get anything done anymore within 6 to 8 weeks (thats the time needed to get all the signatures on the form). The results is that the users are now annoyed...
yeah right, you are clearly not Dutch. Otherwise you would have known that there are no woman living in Delft. Like 99% of all students are male (technical university), and thats probably the same for people working at TNO.
my home dir:
50Gbyte and increasing; digital photos, no replacement possible.
20G misc data for various projects and CV etc.
300M copies and work files for my WWW pages
18Gbyte is copy of all my CDs, which I can redo if I take some time off.
3G of porn, which I might redowload form a.b.p.sex
Conclusion, some of it I do not need to back up, but most of it I do need a copy! So I have the same data on different computers and some of it backed up on CD and DVDs....
Yeah right, the only real reason I do not make good backups is that there is NO real good solution. How do you expect me to make a backup of 150GB of data? On 32 DVDs? No way, they just get lost and scratched. And I am never able to find anything back on that pile of DVDs.
An external hard disk is probably the only solution but not cheap either. Especially when you want to have an off site backup, you probably need 2 of them.
Actully this is not a very new technique. Back in the golden days of OS/2, (what was it, 10 or 20 years ago?) you could already do an install via NFS. Otherwise you had to install Warp from 30 or 40 floppies....
No, it would be even easier to overclock! Set your CPU in the freezer and it will be faster, no need to change a voltage!!!
Or switch of your heater, that is in winter where I live. Or crank up your AC, if you are living in Florida, thats all.
Please do not forget manufacturing tolerances and that you have to make your self-timed logic 99.99999% certain slower than the slowest asynchronous path. This means that you have to qualify your entire logic library with a specific technology, then guardband it to make sure that when manufacturing shifts due to reasons you cannot explain, your chip still works
Does this not depend HOW you implement your selftimed logic? Do you, by any change, know how Philips did implement this?
Besides that, for most ARM applications it is not the raw power that count most. It is the power consumption, which will be much lower for the same number of MIPS (no, not the other core, but the million instr. per seconds). Which will increase your battery life time on your next phone, PDA, MP3 player etc.
But it is probably a hell of a drive to go skiing... I only have to go for 15 minutes, and thats the reason I live here! 30C in summer, and nice layer of beautifull snow in winter, almost perfect, for me that is.
I am not so sure it will not work. For my heater I need electricity for the small computer that controls the thing and for a small electrical pump (50W) that pumps the warm water through the house.
Even DVD backup is not easy. I have about 20 DVDs lying around with part of my files backed up on them. However, also a few giga byte is still not backed up. How do I determine what is backed up and what not? And how do I find back, the files on all those DVDs? There is not good and easy tool for this (on Widows)....
Of course it is a software problem! Could he not just press the reset button?
But seriously an electronic fail is not totaly impossible. The automotive industry has very high demands for quality for their electronics. They do not accept any technology with more than 10 fails (parts) out of 1 milion chips (10ppm). And they are moving to 1 ppm. For a reference, consumer electronic (eg your cell phone) is more like 300ppm. This means that 1 in 100,000 cars could have some electronic prolems (that is for each chip they have)
If only software could be as reliable as this, imagine a OS that only chrashes on 10 out of 1 milion computers.... I guesstimate that windos has a fail level of 100,000 ppm and I am bing nice here.
Don't know why you had to post that AC.
You would be amazed if you knew how many full spreads where actually shot in jpg (with 6Mpixel camera). Many Pro's do not have the time to fidle with the RAW conversion.
Did you ever compare the RAW data and a high quality JPG? I am sure you will not find many artifacts that are distracting.
16.7Mega pixel means that ther are 16.7Mega photo diodes on the sensor, each with probably 14 bit ADC. Generagting ~250Mega bits (raw uncompressed) thus 4.5 seconds transfer time.
From an article: It has sufficient resolution to produce files which convert to 50MB uncompressed TIFF at 24 bit colour depth The 24 bit color depth means only 8 bits per RGB, while in fact the camera probaly has 14 bits depth per RGBG cel. But each cell (photo diode) is counted as pixel by the marketing guys.
However most photographers, even the pros, will shoot in jpg. A jpg in highest quality for this camera is probably about 6Mega bytes and thus transferd in about 1 second....
maybe that is the case, but that does not make it correct. The same as we say England when we mean GB, everyone does it outside GB but is it correct?
actually in belgium they speak walonic and flemish, which are resp. french and dutch dialects...
Why can marketing people not put in these basic and important (at least for me) facts?
No, S in SDRAM is for synchronous
the S in SRAM does stand for static though...
the IT group did that where I work. Now you never get anything done anymore within 6 to 8 weeks (thats the time needed to get all the signatures on the form). The results is that the users are now annoyed...
do not trust any statistics that you did not falsify your self.
databases yes, but as a single image? Can they print this out, or send as a single tiff file?
And scanning grain is already possible. The newest 4000DPI scanners very often have a grain reduction function...
TNO != University
yeah right, you are clearly not Dutch. Otherwise you would have known that there are no woman living in Delft. Like 99% of all students are male (technical university), and thats probably the same for people working at TNO.
No place left for an other drive in mini PC....
my home dir:
50Gbyte and increasing; digital photos, no replacement possible.
20G misc data for various projects and CV etc.
300M copies and work files for my WWW pages
18Gbyte is copy of all my CDs, which I can redo if I take some time off.
3G of porn, which I might redowload form a.b.p.sex
Conclusion, some of it I do not need to back up, but most of it I do need a copy! So I have the same data on different computers and some of it backed up on CD and DVDs....
Yeah right, the only real reason I do not make good backups is that there is NO real good solution. How do you expect me to make a backup of 150GB of data? On 32 DVDs? No way, they just get lost and scratched. And I am never able to find anything back on that pile of DVDs.
An external hard disk is probably the only solution but not cheap either. Especially when you want to have an off site backup, you probably need 2 of them.
Actully this is not a very new technique. Back in the golden days of OS/2, (what was it, 10 or 20 years ago?) you could already do an install via NFS. Otherwise you had to install Warp from 30 or 40 floppies....
even a better one...
Even better...
No, it would be even easier to overclock! Set your CPU in the freezer and it will be faster, no need to change a voltage!!! Or switch of your heater, that is in winter where I live. Or crank up your AC, if you are living in Florida, thats all.
Does this not depend HOW you implement your selftimed logic? Do you, by any change, know how Philips did implement this?
Besides that, for most ARM applications it is not the raw power that count most. It is the power consumption, which will be much lower for the same number of MIPS (no, not the other core, but the million instr. per seconds). Which will increase your battery life time on your next phone, PDA, MP3 player etc.
very funny, but one question. How do you power those Athlons if there is a power outage?
But it is probably a hell of a drive to go skiing... I only have to go for 15 minutes, and thats the reason I live here! 30C in summer, and nice layer of beautifull snow in winter, almost perfect, for me that is.
I am not so sure it will not work. For my heater I need electricity for the small computer that controls the thing and for a small electrical pump (50W) that pumps the warm water through the house.
Even DVD backup is not easy. I have about 20 DVDs lying around with part of my files backed up on them. However, also a few giga byte is still not backed up. How do I determine what is backed up and what not? And how do I find back, the files on all those DVDs? There is not good and easy tool for this (on Widows)....
Of course it is a software problem! Could he not just press the reset button?
But seriously an electronic fail is not totaly impossible. The automotive industry has very high demands for quality for their electronics. They do not accept any technology with more than 10 fails (parts) out of 1 milion chips (10ppm). And they are moving to 1 ppm. For a reference, consumer electronic (eg your cell phone) is more like 300ppm. This means that 1 in 100,000 cars could have some electronic prolems (that is for each chip they have)
If only software could be as reliable as this, imagine a OS that only chrashes on 10 out of 1 milion computers.... I guesstimate that windos has a fail level of 100,000 ppm and I am bing nice here.
Don't know why you had to post that AC. You would be amazed if you knew how many full spreads where actually shot in jpg (with 6Mpixel camera). Many Pro's do not have the time to fidle with the RAW conversion. Did you ever compare the RAW data and a high quality JPG? I am sure you will not find many artifacts that are distracting.
16.7Mega pixel means that ther are 16.7Mega photo diodes on the sensor, each with probably 14 bit ADC. Generagting ~250Mega bits (raw uncompressed) thus 4.5 seconds transfer time.
From an article: It has sufficient resolution to produce files which convert to 50MB uncompressed TIFF at 24 bit colour depth The 24 bit color depth means only 8 bits per RGB, while in fact the camera probaly has 14 bits depth per RGBG cel. But each cell (photo diode) is counted as pixel by the marketing guys.
However most photographers, even the pros, will shoot in jpg. A jpg in highest quality for this camera is probably about 6Mega bytes and thus transferd in about 1 second....