Orders of magnitude
-
What impresses me more is the range of complexity we deal with. It is the highest of any profession or trade ever.
And are we paid commencurately for it?
------------------------------------ "I'm going to walk around a field dangling my keys on a bit of string until I hear whistling noises. " Steve Harris 2009
-
Dalek Dave wrote:
And what have we done with all this extra power?
Well, quite a lot really. We've made PCs more useful ( and easier to use ) because they now do more for us. Displays are the obvious example: 640x480x16 vs 1920x1600x4G. But we also process huge amounts of data in real time and most users just take it for granted. Not just games - LOB apps too. A recent app I worked on held data about training for 105 people in minuscule detail and yet was able to present summaries in real time. Nick
---------------------------------- Be excellent to each other :)
Nick Butler wrote:
But we also process huge amounts of data in real time and most users just take it for granted.
There is a mainframe beauro down the road whose existence I cannot figure. They do monthly batch runs (salaries etc.) that I could easily process in real time even with JavaScript.
-
What impresses me more is the range of complexity we deal with. It is the highest of any profession or trade ever.
Brady Kelly wrote:
What impresses me more is the range of complexity we deal with.
Quite so - that's what makes it fun :) It's also constantly changing with the big companies pushing out new frameworks that soak up take advantage of the performance gains. Nick
---------------------------------- Be excellent to each other :)
-
Since it's slow this morning, I thought I'd share something I've been pondering. My first PC had about a 107 byte disk and a clock cycle was about 10-7 seconds. So we as developers had a range of about 14 orders of magnitude to balance - which was pretty impressive even then. Current PC's have 1012 disks and 10-9 processors - a range of 21 orders of magnitude. So in 20 years, the range has increased by 7 orders of magnitude. That's some rate of change! You can also compare the size / Hz ratio. 20 years ago this was about parity, but now we have 103 more data than speed. Since size keeps increasing but clock speeds have topped out, do we need 1000-core processors? I know it's not scientific! I just thought it might be interesting :) Nick
---------------------------------- Be excellent to each other :)
More interesting than just the clock frequency are the advancements in processor architecture. For example: Intel 80286: 1.8 MIPS at 12 MHz -> 0.15 MIPS/MHz Intel Core i7 Extreme 965EE: 76,383 MIPS at 3.2 GHz -> 23.860 MIPS/MHz
-
Nick Butler wrote:
But we also process huge amounts of data in real time and most users just take it for granted.
There is a mainframe beauro down the road whose existence I cannot figure. They do monthly batch runs (salaries etc.) that I could easily process in real time even with JavaScript.
This goes back to the "If it ain't broke, don't fix it." mentality, coupled with lethargy from their customers.
Henry Minute Do not read medical books! You could die of a misprint. - Mark Twain Girl: (staring) "Why do you need an icy cucumber?" “I want to report a fraud. The government is lying to us all.”
-
More interesting than just the clock frequency are the advancements in processor architecture. For example: Intel 80286: 1.8 MIPS at 12 MHz -> 0.15 MIPS/MHz Intel Core i7 Extreme 965EE: 76,383 MIPS at 3.2 GHz -> 23.860 MIPS/MHz
-
More interesting than just the clock frequency are the advancements in processor architecture. For example: Intel 80286: 1.8 MIPS at 12 MHz -> 0.15 MIPS/MHz Intel Core i7 Extreme 965EE: 76,383 MIPS at 3.2 GHz -> 23.860 MIPS/MHz
Good point - that's 2 orders of magnitude accounted for :) But look at transistor counts: 80286: 134,000 i7: 731,000,000 So we could have a single chip containing 5,455 286s! Nick
---------------------------------- Be excellent to each other :)
-
Since it's slow this morning, I thought I'd share something I've been pondering. My first PC had about a 107 byte disk and a clock cycle was about 10-7 seconds. So we as developers had a range of about 14 orders of magnitude to balance - which was pretty impressive even then. Current PC's have 1012 disks and 10-9 processors - a range of 21 orders of magnitude. So in 20 years, the range has increased by 7 orders of magnitude. That's some rate of change! You can also compare the size / Hz ratio. 20 years ago this was about parity, but now we have 103 more data than speed. Since size keeps increasing but clock speeds have topped out, do we need 1000-core processors? I know it's not scientific! I just thought it might be interesting :) Nick
---------------------------------- Be excellent to each other :)
-
What impresses me more is the range of complexity we deal with. It is the highest of any profession or trade ever.
I think you'll find that engineers in all fields are dealing with a ranges of complexity that didn't exist 20 years ago, not just software engineers. :)
Chris Meech I am Canadian. [heard in a local bar] In theory there is no difference between theory and practice. In practice there is. [Yogi Berra]
-
I think you'll find that engineers in all fields are dealing with a ranges of complexity that didn't exist 20 years ago, not just software engineers. :)
Chris Meech I am Canadian. [heard in a local bar] In theory there is no difference between theory and practice. In practice there is. [Yogi Berra]
Yes, but we hold the record, across all fields. :)
-
This goes back to the "If it ain't broke, don't fix it." mentality, coupled with lethargy from their customers.
Henry Minute Do not read medical books! You could die of a misprint. - Mark Twain Girl: (staring) "Why do you need an icy cucumber?" “I want to report a fraud. The government is lying to us all.”
Maybe they should hear my pitch? :cool:
Last modified: 17mins after originally posted --
-
Good point - that's 2 orders of magnitude accounted for :) But look at transistor counts: 80286: 134,000 i7: 731,000,000 So we could have a single chip containing 5,455 286s! Nick
---------------------------------- Be excellent to each other :)
I remember many years ago seeing on TV a graphics house put out a little tower with 256 gfx chips, one per pixel, haha.
-
harold aptroot wrote:
They're like 3 years behind on Intel.
Get back to the Valley! Like!
------------------------------------ "I'm going to walk around a field dangling my keys on a bit of string until I hear whistling noises. " Steve Harris 2009
-
We already have 1600 core chips. Raedeon 5870[^]
The latest nation. Procrastination.
Radeon HD 5870 X2 - 2x2154 = 4308000000 transistors. That number won't even fit in a DWORD. The mind boggles...
-
More interesting than just the clock frequency are the advancements in processor architecture. For example: Intel 80286: 1.8 MIPS at 12 MHz -> 0.15 MIPS/MHz Intel Core i7 Extreme 965EE: 76,383 MIPS at 3.2 GHz -> 23.860 MIPS/MHz
ZX Spectrum Z80: 0.12 mips at 4 MHz -> 0.03 MIPS/MHz (13 000 transistors) Storage: 1365bits per seconds on audio tape (a whopping 900k on a C90 tape)
Help me! I'm turning into a grapefruit! Buzzwords!
-
Good point - that's 2 orders of magnitude accounted for :) But look at transistor counts: 80286: 134,000 i7: 731,000,000 So we could have a single chip containing 5,455 286s! Nick
---------------------------------- Be excellent to each other :)
or 56230 Z80s
Help me! I'm turning into a grapefruit! Buzzwords!
-
Since it's slow this morning, I thought I'd share something I've been pondering. My first PC had about a 107 byte disk and a clock cycle was about 10-7 seconds. So we as developers had a range of about 14 orders of magnitude to balance - which was pretty impressive even then. Current PC's have 1012 disks and 10-9 processors - a range of 21 orders of magnitude. So in 20 years, the range has increased by 7 orders of magnitude. That's some rate of change! You can also compare the size / Hz ratio. 20 years ago this was about parity, but now we have 103 more data than speed. Since size keeps increasing but clock speeds have topped out, do we need 1000-core processors? I know it's not scientific! I just thought it might be interesting :) Nick
---------------------------------- Be excellent to each other :)
Nick Butler wrote:
and 10-9 processors
Dang. What's left? The hair on the chiny-chin-chin of the squirrel? Marc
I'm not overthinking the problem, I just felt like I needed a small, unimportant, uninteresting rant! - Martin Hart Turner
-
Yes, but we hold the record, across all fields. :)
Hmm... I wonder if the engineers in other fields would agree. Mechanical engineers have to deal with new materials, reliability concerns, and safety issues all the time. Chemical engineering is ever-more complicated by environmental concerns. Electrical engineers are dealing with parts with sub-nanosecond switching times and effects caused by the behavior of countable numbers of electrons.
Software Zen:
delete this; -
Nick Butler wrote:
and 10-9 processors
Dang. What's left? The hair on the chiny-chin-chin of the squirrel? Marc
I'm not overthinking the problem, I just felt like I needed a small, unimportant, uninteresting rant! - Martin Hart Turner
Marc Clifton wrote:
Dang. What's left? The hair on the chiny-chin-chin of the squirrel?
Not far off - about 4" at the speed of light :-D Nick
---------------------------------- Be excellent to each other :)