Ryan you should be very proud of what you have written thus far. As all the making s of a great set of articles on computer memory. I Eagerly await your next installment ;)
Just nitpicking, but the definitions of Mega, Giga etc are incorrect.
The english definition of a Gigabit, is one thousand million and a terabit is one billion i.e. a million million. I take it that the author uses US English rather than British English...
Originally Posted by Shielder Just nitpicking, but the definitions of Mega, Giga etc are incorrect.
The english definition of a Gigabit, is one thousand million and a terabit is one billion i.e. a million million. I take it that the author uses US English rather than British English...
You're correct in your definition but isn't that exactly what the green table on Page 1 shows anyway? I'm confused
Good article. Although on page 5, all of the diagrams say "MEMROY CONTROLLER"
Quote:
Originally Posted by Shielder The english definition of a Gigabit, is one thousand million and a terabit is one billion i.e. a million million. I take it that the author uses US English rather than British English...
The US definition of a billion (109) is the SI standard for tera-. Does anyone actually use the British definition any more?
Well, I've read all of this.....And, I like it.....You got a good way to explaining thing.....Your analogy to things is also make it really easy to understand.
I am @ a Information System course now, but I really love to learn about the Computer Organizing things....:D.
This most recently update I believe, bring me more into this, rather than learning bout the x86, some motorola legacy procs.., and a out-dated memory hierarchy (SSD wasn't listed^^). And even a DDR is just a scratch....:D.
Now I understand more about dual channel, quad, what's called miss-cached, and etc...
Tera uses the short scale and always has. Anyone still using long scale notation is an old decrepit man.
Plus, you have the oddity of flash and ddr memory systems using 210 binary system (kibibyte) and other media like CDs and harddrives using the 103 decimal system (kilobyte).
Interesting stuff, but I'd hope most of the OCers here know this already.
It is a bit like buying a 160GB Hard Disk and after formatting it, only 149GB is available for your data. This is due to a portion of the drive being occupied by the system level information such as Master Boot Record (MBR) and Partition Tables.
Im sorry, what!?!?
Most of that is lost due to the definition of a byte on a HDD vs the operating system
The percentage that is lost to the system level information is a much much much smaller percentage (that percentage varies alot depending what settings you format with though, and what filesystem you use) - also, that is mainly the partition table as the MBR is 512kb ....
Yes. you are very sharp on that and right about the HDD part.
The original paper was over 250 pages with huge amount of engineering details. We spent months de-constructing and simplifying ideas down for casual readers. I avoided the 1024 vs 1000 argument and explanation.
The above HDD statement was more of an analogy focused on the myth that full MHz was solely used to carry application data. Hence, the word "It is a bit like..."
The objective was to capture a more identifiable situation from perspective of non-initiated computer users, to help explain how all the memory subsystem works.
It always comes down to who do we write for - the engineering community or non-initiated readers. It is a hard balance to find.
We appreciate your thoughts and sharpness on picking that up.
Originally Posted by Shielder Just nitpicking, but the definitions of Mega, Giga etc are incorrect.
The english definition of a Gigabit, is one thousand million and a terabit is one billion i.e. a million million. I take it that the author uses US English rather than British English...
Andy
/nitpicking
Good article though.
I understand the confusion.... blame it on the long and short scaling....
A decision was made to use what general consumers understood, which is in terms of $$$. We didn't want to dive into it as it is purely academic.
The full material has been completed a while ago. It is huge and packed full of diagrams. It is constantly (and painfully) being revised and edited. Big is not always good.
Bindibadgi is right, it is not ready for casual readers. A lot of the current work is repacking the writings into a more concised form for the website.
The final book will be released 100% FREE because the real credit goes to countless scientists, engineers and systems designers. They are the inventers and builders. I just write about what has been done.
Do understand that... It is EXTREMELY difficult in simplifying complex engineering concepts for casual readers because the explanation process tend to loose specificity and prone to mistakes. We want to write for the average users and enthusiasts who does not have a 4-year EE (Electronics Engineering) degree. You may be a precocious 12 year old kid. :)
The latest and the best material is on Bit-Tech.
Things will become very interesting in later parts of the DDR Memory series, where users will have a chance for hands-on experience. There will be extensive list of acronyms that we hope will help users in doing your own study or research as well.
After the series, if wonderful readers like you are still interested, maybe that will inspire you to do something greater.
Tune-in closely at Bit-Tech, you'll get the latest stuff here.
Originally Posted by ryanjleng Yes. you are very sharp on that and right about the HDD part.
The original paper was over 250 pages with huge amount of engineering details. We spent months de-constructing and simplifying ideas down for casual readers. I avoided the 1024 vs 1000 argument and explanation.
The above HDD statement was more of an analogy focused on the myth that full MHz was solely used to carry application data. Hence, the word "It is a bit like..."
The objective was to capture a more identifiable situation from perspective of non-initiated computer users, to help explain how all the memory subsystem works.
It always comes down to who do we write for - the engineering community or non-initiated readers. It is a hard balance to find.
We appreciate your thoughts and sharpness on picking that up.
Salute to the geeks!
I somewhat understand what your saying, however, i think what you said is probably more likely to confuse people because you are sort of saying the HDD space problem is actually a different issue entirely, instead of it really being due to the definition of a bit
Most knowledge'd people will probably understand that most of the space you loose is really due to the definition of a byte, but people who are unsure could easily get the wrong end of the stick
Ryan, I'm really looking forward to the full work. I've only got a physics degree, but I do try and stay abreast of things. Thanks for a great article!
Ryan, let me ask you a pragmatic question for us regular consumers to see if this can apply to everyday decisions. How much of a performance difference, if any, would you realize by using 2 GB of Dual Channel DDR2 SDRAM memory at 800 MHz memory (two sticks) instead of the standard 2 GB of Dual Channel DDR2 SDRAM at 677MHz memory (two sticks), if you were using an Intel E4500 Core 2 processor running Vista and a 256MB video card with a 800MHz controller speed?
Or to put this another way, would you choose the 800MHz memory over the 677mHz memory if you only had to pay $10 more per stick?
Comments 1 to 25 of 32
Why, cos you might actually learn something? :p
More interesting stuff are coming.
The english definition of a Gigabit, is one thousand million and a terabit is one billion i.e. a million million. I take it that the author uses US English rather than British English...
Andy
/nitpicking
Good article though.
Keep them coming!
You're correct in your definition but isn't that exactly what the green table on Page 1 shows anyway? I'm confused
I am @ a Information System course now, but I really love to learn about the Computer Organizing things....:D.
This most recently update I believe, bring me more into this, rather than learning bout the x86, some motorola legacy procs.., and a out-dated memory hierarchy (SSD wasn't listed^^). And even a DDR is just a scratch....:D.
Now I understand more about dual channel, quad, what's called miss-cached, and etc...
Thx Ryan^^.
Plus, you have the oddity of flash and ddr memory systems using 210 binary system (kibibyte) and other media like CDs and harddrives using the 103 decimal system (kilobyte).
Interesting stuff, but I'd hope most of the OCers here know this already.
Most of that is lost due to the definition of a byte on a HDD vs the operating system
The percentage that is lost to the system level information is a much much much smaller percentage (that percentage varies alot depending what settings you format with though, and what filesystem you use) - also, that is mainly the partition table as the MBR is 512kb ....
The original paper was over 250 pages with huge amount of engineering details. We spent months de-constructing and simplifying ideas down for casual readers. I avoided the 1024 vs 1000 argument and explanation.
The above HDD statement was more of an analogy focused on the myth that full MHz was solely used to carry application data. Hence, the word "It is a bit like..."
The objective was to capture a more identifiable situation from perspective of non-initiated computer users, to help explain how all the memory subsystem works.
It always comes down to who do we write for - the engineering community or non-initiated readers. It is a hard balance to find.
We appreciate your thoughts and sharpness on picking that up.
Salute to the geeks!
I understand the confusion.... blame it on the long and short scaling....
A decision was made to use what general consumers understood, which is in terms of $$$. We didn't want to dive into it as it is purely academic.
more info here.
http://en.wikipedia.org/wiki/SI_prefix
http://en.wikipedia.org/wiki/Long_and_short_scales
Kilo, 10^3, 1000, etc is 'k' not 'K'.
Also with that analogy to cars and roads at the bottom of page 1 you're dangerously close to a "series of tubes" comment. :)
I wouldn't have thought 33 was decrepit! ;)
Having reread the article, I found it really interesting. I'm being lazy, but is the full article anywhere to be found (for free)?
Bindibadgi is right, it is not ready for casual readers. A lot of the current work is repacking the writings into a more concised form for the website.
The final book will be released 100% FREE because the real credit goes to countless scientists, engineers and systems designers. They are the inventers and builders. I just write about what has been done.
Do understand that... It is EXTREMELY difficult in simplifying complex engineering concepts for casual readers because the explanation process tend to loose specificity and prone to mistakes. We want to write for the average users and enthusiasts who does not have a 4-year EE (Electronics Engineering) degree. You may be a precocious 12 year old kid. :)
The latest and the best material is on Bit-Tech.
Things will become very interesting in later parts of the DDR Memory series, where users will have a chance for hands-on experience. There will be extensive list of acronyms that we hope will help users in doing your own study or research as well.
After the series, if wonderful readers like you are still interested, maybe that will inspire you to do something greater.
Tune-in closely at Bit-Tech, you'll get the latest stuff here.
Most knowledge'd people will probably understand that most of the space you loose is really due to the definition of a byte, but people who are unsure could easily get the wrong end of the stick
Andy
interesting to learn about ddr ram. look forward for more ddr3 stuff. bring on the tech stuff!!!!
Ryan, let me ask you a pragmatic question for us regular consumers to see if this can apply to everyday decisions. How much of a performance difference, if any, would you realize by using 2 GB of Dual Channel DDR2 SDRAM memory at 800 MHz memory (two sticks) instead of the standard 2 GB of Dual Channel DDR2 SDRAM at 677MHz memory (two sticks), if you were using an Intel E4500 Core 2 processor running Vista and a 256MB video card with a 800MHz controller speed?
Or to put this another way, would you choose the 800MHz memory over the 677mHz memory if you only had to pay $10 more per stick?
Thank you.