Ever had faulty CPU launches?

Matmatty65 · 7 Oct 2020 at 20:40

So basically you know like how you can have like faults/issues with GPU launches as an example. Rtx 20 series was artifacting , 5700xt driver issues/green screens. Have you ever seen anything like that on a CPU release? Or is their generally not a lot that can go wrong with CPU's.

Quartz · 7 Oct 2020 at 21:01

Read up on the FDIV bug.

Journey · 7 Oct 2020 at 21:01

Yes, many, many years ago Intel recalled their PIII 1,13GHZ chip as it was factory 'overlcocked' and not fit for purpose, they did it due to AMD's Athlon TBird beating them all over.

Plec · 7 Oct 2020 at 21:26

Journey said:
due to AMD's Athlon TBird beating them all over.

And the might of a graphite pencil - one of the few CPUs i've kept (i killed the temperamental ABIT MB - drove me to my limits).

kitfit1 · 7 Oct 2020 at 23:19

Plec said:
And the might of a graphite pencil - one of the few CPUs i've kept (i killed the temperamental ABIT MB - drove me to my limits).

Lol, did the the same with an Abit......................learnt that lesson and moved on to a DFI.......................the doggies danglies of MB's

CuriousTomCat · 7 Oct 2020 at 23:37

Back in 1994 the Pentium 60 and Pentium 66 had a floating point bug where the CPU would calculate long division incorrectly. These two CPUs were cancelled shortly after release and replaced with a Pentium 75 which didn't have the problem.

I believe the 386 had some bugs but that was before my time.

pc-guy · 7 Oct 2020 at 23:45

Many moons ago when OS and hardwares weren’t as synchronous as nowadays.

I remember the days where I had to manually configure serial and parallel bus allocation in bios and then making sure DOS or whatever windows version at the time used the same resources. It was often problematic when new CPU is launched and then software isn’t updated. Plug and play wasn’t really a thing back then. Everything more or less had to be manually configured and that was the real pain in the backside.

one thing to add is that computer back in those days are a lot simpler than nowadays.

Mr Evil · 8 Oct 2020 at 01:07

All modern CPUs are faulty, it's just a question of how faulty. You can see for yourself by looking for errata; for instance the errata for Ryzen 1 includes some serious bugs. Mostly they can be worked around in software, or require circumstances that don't occur in normal use.

Justintime · 8 Oct 2020 at 02:44

Yep the Pentium bug would probably not be a big deal today but in those early days (i was 11 in 93) many people didn't have pcs and ignorance was the majority so it was blown out of proportion. The 1.13Ghz Slot 1 PIII was a truly faulty cpu and Tom's hardware really went against the grain on that one and probably suffered for it for a bit, at the time i wrote off the P6 Architecture, who knew it would go on to save them from Netburst?

Justintime · 8 Oct 2020 at 02:48

pc-guy said:
one thing to add is that computer back in those days are a lot simpler than nowadays.

Really? i find todays hardware a walk in the park but I actually prefer the old days of jumpers and DIP switches etc.. when you set something it stayed that way without a software layer in between to go funny.

Rroff · 8 Oct 2020 at 03:06

Wow was not aware of the 1.13GHz CPU problem - a friend in the US gave me one for free while I was out there that ran rock solid on my Gateway 2K machine (BX440 board IIRC). I'll see if I still have it.

clowesy · 8 Oct 2020 at 03:10

Plec said:
And the might of a graphite pencil - one of the few CPUs i've kept (i killed the temperamental ABIT MB - drove me to my limits).

I remember the pencil trick too. I managed to unlock the multiplier and gain 100mhz!

Justintime · 8 Oct 2020 at 03:46

Rroff said:
Wow was not aware of the 1.13GHz CPU problem - a friend in the US gave me one for free while I was out there that ran rock solid on my Gateway 2K machine (BX440 board IIRC). I'll see if I still have it.

Sure it was a Slot 1 BX440 and not a newer Tualatin 1.13Ghz in a newer chipset board? I ask because officially the BX440 doesn't do 133Mhz as the AGP runs at 89Mhz so not 66Mhz default evem if it was onboard video it will likely be AGP based) and not every BX440 did 133 (my BH6 did 150Mhz but at 100Mhz AGP!, forgot what gpu might be a Geforce 256 that coped). So a big OEM using a 1.13 on an older 440BX which wouldn't officially support the cpu and fsb sounds suspect

Would've thought they'd use the horrible i815 chipset for that.

Just reread your post, get that you yourself put the cpu

Rroff · 8 Oct 2020 at 03:57

Justintime said:
Sure it was a Slot 1 BX440 and not a newer Tualatin 1.13Ghz in a newer chipset board? I ask because officially the BX440 doesn't do 133Mhz as the AGP runs at 89Mhz so not 66Mhz default evem if it was onboard video it will likely be AGP based) and not every BX440 did 133 (my BH6 did 150Mhz but at 100Mhz AGP!, forgot what gpu might be a Geforce 256 that coped). So a big OEM using a 1.13 on an older 440BX which wouldn't officially support the cpu and fsb sounds suspect

The CPU wasn't provided by the OEM - my Gateway 2K shipped with a 500MHz P3. Don't think Gateway ever sold that model with higher than 800MHz. I'm not 100% what motherboard was in it but I'm 99% sure it was a 440.

I already had a 1.5GHz P4 in my main setup by then which replaced the Gateway system.

Justintime · 8 Oct 2020 at 04:01

Aye reread your post and figured you put it in yourself

its 4am

Rroff · 8 Oct 2020 at 04:08

Justintime said:
Aye reread your post and figured you put it in yourself its 4am

When I was in the US a friend of mine who ran a computer store gave me some stuff from returns that would otherwise be thrown out that I used to upgrade older machines, etc. I'm guessing that is why it was in returns LOL. But it ran in my P3 which was used as a secondary machine by then absolutely fine.

EDIT: I've still got the old Gateway tower so gonna pull it out later see what is in there.

pc-guy · 8 Oct 2020 at 06:35

Justintime said:
Really? i find todays hardware a walk in the park but I actually prefer the old days of jumpers and DIP switches etc.. when you set something it stayed that way without a software layer in between to go funny.

Exactly. In the old days you know what those hardwares are doing. Setting jumpers did x etc. Nowadays there are so many BIOS settings it’s enough to make you want to smash it all when you can’t figure out why certain voltage just won’t stay the same etc.

CuriousTomCat · 8 Oct 2020 at 08:32

Justintime said:
Really? i find todays hardware a walk in the park but I actually prefer the old days of jumpers and DIP switches etc.. when you set something it stayed that way without a software layer in between to go funny.

Yes really. Today's hardware maybe a walk in the park for the consumer, but the actual hardware used to be incredibly simple compared to today.

Early motherboards were homemade. People used to write the schematic, buy the chips, capacitors, resistors etc and solder them onto the board, then plug a TV into it and have fun.

It's physically impossible to build your own Ryzen motherboard today because they're too complicated.

Justintime · 8 Oct 2020 at 11:25

But when last was that true? 40 years ago?

I did a few homebrew projects in the 90s but even then was just historic stuff for fun.

Cyber-Mav · 8 Oct 2020 at 12:13

integer divide by zero. dunno which cpu it was but it was an intel cpu i had and in a dos box do a math sum and divide something by 0 and it would bluescreen/reboot.