Old motherboards still doing you proud :)

[Smoblikat]

Originally Posted by Accutech

Still have a Timex Sinclair c/w ZX80 processor 2 Kb ram no colour. Still runs like new.
Also have a commodore vic20 and a commodore 64 both still running like champs. The 64 even has an external (get this) 300 baud modem. It will even get on the internet, Awful slow but still it gets there. I also have a collection of processors starting with Intel 8086 and up to Duo core centrino. Most are in systems that still run, A 386sx with a math co-processor. A few AMD based systems, but I always preferred Intel simply because of the fact that AMD has always had an issue with thief FPU (Floating decimal point) If you take an Intel based system and an AMD based system in a business environment The Intel system is always right exactly to the penny while without fail the AMD system will be off by a penny. So what we say but when it is time to do a monthly report for example, If using an AMD system you could be out by dollars (Then you spend a huge amount of time looking for the error, BUT if using an Intel baswd system you are right on to the penny everytime.
How's that for a bit of trivia.

Is there proof of this?

[VirtualLarry]

Originally Posted by Smoblikat

Is there proof of this?

It's bullshit.

Edit: When I overclocked my Athlon XP 1800+ system, I used the OpenGL "3D Pipes" screensaver. I would let it run for a while, and if the overclock was unstable, then it would start to display rendering errors. This is because on the Athlon XP, the FPU was the "critical path" (well, that's simplifying a bit), and would show errors first, while the integer calculations were fine.

So yes, if the AMD Athlon XP rigs were running out of spec, then there could have been FPU problems. But I've never heard that mentioned as a general rule, for AMD rigs running in-spec.

Edit 2: The Athlon XP, I believe, required you to manually set the vcore for the CPU. I don't know if it could send a signal to the mobo to auto-detect the voltage. Some AMD CPUs of the same family had different vcores. If one didn't set the vcore correctly, when setting up the system, then it could lead to the same sort of instability from running out of spec as overclocking would.

This does not mean that there was any kind of flaw in AMD's CPUs, but rather, that the system integrator didn't do a complete job.

[C1]

I do remember this flap: http://en.wikipedia.org/wiki/Pentium_FDIV_bug

I remember some guy on a radio talk show (Norm LaFave - what ever happened to him?), telling how this wouldnt matter for the average home use. Had to call in and tell him that it could be a big deal. Couldnt convince him though.

Also, when it came to hand calculators, the TI's were a touch more accurate than some of the ealier HPs (eg, HP45).

I actually had a problem in a grad school (QC and Reliability course) because of the HP as some of the very complex chain calculations would be off a bit & my instructor accused me of an incorrect solution (marked the problem wrong & I had to show him the methodology was able to give a correct equivalent answer when using the more accurate TI.

Gotta watch out with those seemingly itsy-bitsy-tiny-weenie inaccuracies as they could easily bite you in the ass building, say, an atomic power plant or even a rocket ship.

Best not to use the C-64 for such project, lol (I still have my SX-64 - taught me to be an expert in character string handling which turned out to be very helpful in my careers; saved a guys ass by writing an octal to decimal conversion macro for Quatro so he could use Quatros graphic/plot features in a job related scientific analytic endevor; he never even thanked me).

[Slammy1]

I believe that's a problem with the data precision. Consider a decimal:
...ax10^2+bx10+c+dx0.1+ex0.01...

When it gets converted to binary it's of the form
...a'x2^2+b'x2+c'+d'x0.5+e'x0.25...

Each manipulation requires that the number be converted to binary, manipulated, then converted back. That's where 8/16/64 bit precision comes into play. Basically, a given binary number will represent a range of decimal values which varies depending on how much data you use to store the number.

EDIT: My P3B-F and P4P800 both died around the same time this year. I'm sure I could resuscitate the P3B, just don't want to invest the time nor the money, the P4P's MoBo shot craps and is dead dead.

[WhoBeDaPlaya]

I have a EPoX 8K3A+, DFI AK76-SN and ECS K7AMA still running strong

[Rhoxed]

ABIT NF7-S V2
ASUS P4P800-E DELUXE
MSI K8N Neo 4 Platinum
MSI K9A2 Platinum

[DLeRium]

Originally Posted by MegaWorks

The legendary DFI LanParty NF-4 Ultra-D

I have two of these boards running and mining bitcoins. They've been fully stressed this past year....

[stargazr]

ASUS P4PE w/P4 Williamette.

[OmegaShadow]

m2a-vm

if only i can get my hands on a phenom ii x4.

[jumpncrash]

I still have a dual Slot 1 board from an HP server that still ruins 24/7 as a router, in fact I have a whole bunch of slot1 boards, also have a pentium 90, but it's not doing anything

[Gs dewd]

I am still running my Abit KT7A with a Axp 2400+ running as a file server. 12 or so years it's been running.

[OVerLoRDI]

I have a buddy who is still running one of the first ASUS AM2 boards. Don't remember the exact model number, but he "upgraded" from a socket 939 system to the first gen AM2 stuff. At the time the difference was basically zero and offered few advantages. We mocked him slightly.

Fast forward to today. He is sitting pretty with a Phenom II x6 in the same motherboard.

[Name User]

This 965P DS3 just won't quit.

6 years, taking an E4300 from 1.8 to 3.0 and a Q9550 from 2.83 to 3.4. It was never even meant for 1600 FSB.

Finally had some hiccups these past few weeks, I figured the MCH is probably cooked at this point, RAM coming up a fraction of its speed, running in weird single and flex modes for no reason at all.

So I underclocked, undervolted, tinkered with things, was hoping to get it stable enough to maybe ride this thing out for another week while my backup laptop arrives. Then out of boredom thought what the hell, lets max out the voltage, relax the timings and see what happens. It's back, stable as a rock, and I think I'm about to reboot and try for 3.6

By the look of it, this thing will hold out until Haswell.

[pcunite]

Anyone remember the FIC SD11? What a sweet board for an Athlon 650mhz to work on. Very stable board. I ran it for years as an IPCop router.

[Dkcode]

Originally Posted by PM650

A7N8X Deluxe - retired it last year, it's now my mom's machine.
Abit KR7A (KT266A) - been my dad's machine since I upgraded to the A7N8X years ago.

Both had bad caps (although neither died as a result), both still working flawlessly to this day after the caps were replaced a year ago

A7N8X Deluxe was the first high performance system I built for myself, I always settled for the cheaper crap before as I wasn't made of money.

Great board, in fact I only looked it up the other day as I forgot what it was called.

[WT]

Wow, blast from the past for me yesterday, as I upgraded a co-workers very old PC. I opened it up to scavenge parts and found an Epox 8RDA+ board that I had forgotten about. Other than being dusty as hell, it still works fine. I had upgraded it with a heatsink fan on the chipset ran than the cheap Epox heatsink, so it will probably go in the freebies thread.

[Rubycon]

The plug in VRMs on my Shuttle HOT-613 Pentium Pro (Socket 8) dual boards were made by Raytheon and actually had the word "SKYNET" silkscreened on them.

[Dkcode]

Originally Posted by Rubycon

The plug in VRMs on my Shuttle HOT-613 Pentium Pro (Socket 8) dual boards were made by Raytheon and actually had the word "SKYNET" silkscreened on them.

Excellent!

[Rubycon]

My memory is about as dusty as the modules it would seem...

Not silkscreened but printed on the label. Nonetheless here you go!

[Insert_Nickname]

I have a 13 year old Epox MVP3-G2 with a K6-2 500MHz and 128MB PC100 SD-RAM that still gets turned on occasionally for retro gaming...

[Archman]

Still running:

Abit KT7 Tbird 1.0 GHz (upgraded to XP 2200+ running at 2.0GHz) with 1.25GB ram for a basic home computer with an ATI Radeon 7200...

Capacitors are showing signs of age, but soon to be upgraded to an MSI RS482M4-IL with a AMD 64 3700+

The ABIT KT7 will be recycled for a NAS build with NASlite-M2 with old EIDE 300GB hard drives

[DublinGunner]

It still works, but unfortunately I (intentionally) destroyed the AMD 486DX4-100 that used to run in it in order to take some processor die shots.

But, I have a MINT condition Shuttle 80486 Deep Green motherboard, that tomorrow I'm giving to Dublin's oldest electronics shop (where I bought it 18~ years ago) to add to their PC museum display.

This was the first motherboard I ever bought, and my first ever PC build, when I was 13 or so. It will make me happy to see it on display each time I go into the store

Incredibly, I still have the full length ISA card that had to be bought separately, ya know, to add such high tech items as Hard disk drives

[VinDSL]

Originally Posted by adairusmc

A relative still has my old computer running an Abit IC7-Max III motherboard and my old P4 EE 3.2, and as far as I know they still use it as a basic browsing machine.[...]

Still using my Intel P4 Extreme Edition 3.4 (Gallatin) daily.

Damn thing's a flame thrower!

• DFI LanParty PRO875B rev B1 mobo
• Crucial Ballistix Tracer PC4000 1GB RAM
• XFX 7600GT 560M AGP (PV-T73A-UDF3)
• Mountain Mods U2-UFO Opti-1203 case
• Corsair HX520W Modular PSU
• Tuniq Tower 120 CPU cooler
• CoolViva Z1 GPU cooler