REGISTER | LOGIN
Breaking News
Comments
Newest First | Oldest First | Threaded View
Page 1 / 5   >   >>
mtx
User Rank
Rookie
My PC history
mtx   4/22/2017 4:40:30 AM
NO RATINGS
Age today: 35

 

xmas 1991. Inherited my brothers Commodore 64. Programmed Basic and played games.

xmas 1993. We got the first family PC, an IBM PS/1 486SX 25MHz, 4MB RAM (later upgraded to 8MB and with CD-ROM). Did basic and assembly programming

Spring 1998: Bought my first PC. P-2 300MHz, 64 MB RAM. It was hugely expensive, $2500. Later upgraded to 128 MB.

Spring 2002: Athlon XP 1700+ don't recall how much RAM but I guess what would have been typical at the time (512 MB?)

Spring 2005: Athlon 64 3000+ 2 GB RAM

Spring 2008: Intel Core 2 Duo E8400 with 4 GB RAM (dual core, 3 GHz)

Spring 2011: Intel i5 2500, 8 GB RAM (quad core, 3.3 GHz)

Summer 2015: Intel 6700K with 16 GB RAM (quad core, 4 GHz)

 The last upgrade seems a bit wasteful. Despite of being one of the longest times between upgrades the benefit was not huge. In some things it is an improvement, in others not so. And application requirements are stagnant so it is only for compiling code etc. I can appreciate the difference. But I had the upgrade itch and could afford it ;) In the 2002-2015 era I kept a lot of the HW from machine to machine so it was mostly the motherboard, CPU, RAM and occationally GFX I upgraded.

The next worthwhile upgrades seems to be for higher core counts like AMD Ryzen. Unfortunately I am still early into the 6700K, so will sit out on that one. So I expect to upgrade to either 6 or 8 core in 1-2 years depending on what offerings there is at the time. Could be AMD's next-gen Ryzen og maybe Intel's IceLake. 

 

mtx
User Rank
Rookie
Re: Vocabulary
mtx   4/22/2017 4:29:50 AM
NO RATINGS
Win 3.1 / 3.11 were multitasking but it was cooperative multitasking, meaning the program had to yield control of the CPU to Windows. Windows would never context switch by itself (it would handle interrupts but would not switch between programs based on e.g. timer interrupt). Yielding come mostly naturally in the Windows programming model without much programmer intervention, since control is automatically yielded in the message loop (when the program calls GetMessage()). But if a program starts a huge calculation when processing a message it will lock up the system and even the clock wouldn't update. This required the programmer to manually split up the calculation into small chunks and set up a timer message to trigger the calculation of the individuals chunk. Well-behaving programs could easily multitask. It was actually Win9x (i.e. pre NT) that introduced pre-emptive multitasking and could switch to other programs even if a program didn't yield. It seems the scheduler was not too aggressive/sophisticated about it (probably due to the limited resources available) so in practice a program not yielding could significantly affect the reponsiveness but it usually wouldn't crash the system. WinNT and successors refined the scheduling and priority algorithms further and further and combined with much greater HW resources things are much more  smooth today. But it was Win95/98 that introduced true pre-emptive multitasking. 

raslaje
User Rank
Author
Re: Punch Cards
raslaje   3/6/2017 4:27:02 PM
NO RATINGS
"He also predates the abacus, but don't tell anyone."

But he looks too young in his photo. Why's that?

Maybe he took it with a Kodak?

TonyTib
User Rank
Author
Re: Personal history
TonyTib   3/6/2017 2:15:40 PM
NO RATINGS
Some family notes:

My father used the IBM 1620 in college.  Somewhere, I still have a 1620 manual (I'm not getting rid of it).

My brother designed and built an 8080 system in the 1980's, hand assembled assembly language, and then entered it into the EPROM using a DIP switches (he made his own EPROM programmer) - and only made a few mistakes!

My first computer was the Atari 520ST.  I still have it lying around  - I need to fire it up and show my kids; I think they'd enjoy some of the old games (I had a lot of fun playing Rampage -- need to find a couple joysticks for it).

MeasurementBlues
User Rank
Author
Re: Punch Cards
MeasurementBlues   3/6/2017 10:33:26 AM
NO RATINGS
Max wrote "I predate EDA as we know it"

He also predates the abacus, but don't tell anyone.

traneus
User Rank
Author
Personal history
traneus   3/5/2017 7:35:50 PM
NO RATINGS
1966 Fortran II punched cards on IBM 1620 at university.

1974 Homebrew DRAM-based text dumb terminal via 300-baud modem to Xerox Sigma-6 timesharing system.

1977 Added homebrew 8080 processor to dumb terminal, using high-speed cassette-tape storage.

1987 Heathkit clone of IBM 8088 PC using MS-DOS. I still use this machine for text editing.

1991 Started using Unix workstations in graduate school.

1994 Upgraded to Pentium 90 running MS-DOS for transient analysis of analog/digital circuits.

1995 Upgraded to Slackware Linux.

2001 Upgraded internet access to cable modem.

2005 After hardware failures, upgraded to AMD Athlon 64-bit processor. I still use this machine for most work, now running 64-bit Slackware using 1 GB of DRAM.

traneus
User Rank
Author
Pre-NT and pre-OSX multitasking
traneus   3/3/2017 5:03:48 PM
NO RATINGS
I hardly used pre-NT Windows, and I never used pre-OSX Macintoshes, so I thank those who corrected my recollections. Both NT and OSX (which is Unix at its core) are true pre-emptive mutitasking, and complete rewrites from their predecessors.

perl_geek
User Rank
Author
Practical queueing theory
perl_geek   3/3/2017 3:07:33 PM
NO RATINGS
@Elizabeth "We had a slightly different system..."

Did the administration use the system as a case study in the behaviour of systems with multiple queues? It's actually surprising how many things can be modelled that way, especially in OSs.

(And as a footnote, did you notice that "queueing" has five vowels in succession. There can't  be many words with that count or higher.)

Karnaugh maps and Boolean algebra still have a role, if you're trying to get to the essentials of logic. Even when circuits are cheap and microscopic, hence not worth optimising in themselves, it's worthwhile to boil a problem down to its essentials.

 

Max The Magnificent
User Rank
Author
Re: Punch Cards
Max The Magnificent   3/3/2017 12:24:26 PM
NO RATINGS
@Elizabeth: ...you remember using Karnaugh Maps and Boolean Algebra to get to the minimum possible gate count....

You betcha -- did you see my column on Karnaugh Maps?  And Reed-Muller Logic? And Gray Codes? And Positive vs Negative Logic? And Assertion-Level Logic? And LFSRs?

Phew -- I'm tired just listing them all LOL

elizabethsimon
User Rank
Author
Re: Punch Cards
elizabethsimon   3/3/2017 11:27:45 AM
NO RATINGS
@ raslaje

That was pretty efficient compared to my school.

Yes it was pretty efficient for that day and got even more efficient when I discoverd that the wait times were significanlty reduced if you came in at 7 or 8 AM instead of PM

Page 1 / 5   >   >>


Like Us on Facebook