Worst chipset of all time

NeghVar

2[H]4U
Joined
May 1, 2003
Messages
2,671
For me, it was the Intel i820

Intel's deal with Rambus. RDRAM was ridiculously expensive. Then Intel came out with the memory translator hub due to the outrage from consumers over RDRAM prices. This created even more problems. Due to the overhead of resources required for the translation, it cause a performance deficiency.
dimmriser.jpg
 
The i820 may have indeed been Intel's worst chipset, but it was hardly the worst chipset ever produced. The VIA MVP3 was pretty fucking terrible as I recall. Like many Super-7 chipsets at the time, it had AGP graphics card compatibility issues, PCI problems and the usual driver issues. Beyond all that, VIA and SIS chipsets usually didn't perform very well.
 
When I first became aware of "chipset", I had the 430FX. I have never had an SiS chipset. So I cannot say by experience. I did experience the i820 though. Where I worked at the time did a office wide upgrade of our PCs. They chose Systems which had the i820 and used the MTH's. Constant slow performance and instability.
 
When I first became aware of "chipset", I had the 430FX. I have never had an SiS chipset. So I cannot say by experience. I did experience the i820 though. Where I worked at the time did a office wide upgrade of our PCs. They chose Systems which had the i820 and used the MTH's. Constant slow performance and instability.

Yep. We had about a hundred i820 equipped HP workstations where I worked at the time. Because of the recall, we talked to HP and HP sent me the RAM and new daughter boards for all of the workstations we had. I replaced them all myself over the course of a couple months.
 
For me, it was the Intel i820

Intel's deal with Rambus. RDRAM was ridiculously expensive. Then Intel came out with the memory translator hub due to the outrage from consumers over RDRAM prices. This created even more problems. Due to the overhead of resources required for the translation, it cause a performance deficiency.
dimmriser.jpg

Wasn't that long ago I shitcanned a few machines in the attic that were rdram. Feels gross given what those fucking things cost.
 
820/RDRAM failed mainly because of the DRAM cartel that got some DRAM execs jailed and a ~1B$ fines. And that's a real shame. Memory is the last part we got on a parallel bus today unless you go HMC with SerDes.

Worst chipset in my mind was the nforce. Specially those that had the LAN integrated with that active armor junk.
 
Last edited:
Yep. We had about a hundred i820 equipped HP workstations where I worked at the time. Because of the recall, we talked to HP and HP sent me the RAM and new daughter boards for all of the workstations we had. I replaced them all myself over the course of a couple months.
The SiS 655FX was pretty good IIRC, long time ago though.
 
Worst chipset in my mind was the nforce. Specially those that had the LAN integrated with that active armor junk.

Nforce was at least a fast chipset.

Wasn't there a bug with the 810 chipset? Think it was memory corruption? Or Disk?

I remember my dad buying an emachines with a Celery 600, 810 chipset, and WinME pre-release.

The computer crushed itself doing a disk defrag. Something with defrag literally fragged parts of Windows, and then System Restore also had a bug and failed to succesfully restore ANYTHING.
 
IIRC the 810 was quite good, used it with a 633MHz coppermine. No corruption in my case.
 
Nforce was at least a fast chipset.

Wasn't there a bug with the 810 chipset? Think it was memory corruption? Or Disk?

I remember my dad buying an emachines with a Celery 600, 810 chipset, and WinME pre-release.

The computer crushed itself doing a disk defrag. Something with defrag literally fragged parts of Windows, and then System Restore also had a bug and failed to succesfully restore ANYTHING.

WinME was awful in itself. WinME prerelease and wonder why it corrupted? :p
 
WinME was awful in itself. WinME prerelease and wonder why it corrupted? :p

It's true. Maybe it wasn't the 810, but I do remember reading about a certain intel chipset used during that time that could have memory corruption.......
 
It's true. Maybe it wasn't the 810, but I do remember reading about a certain intel chipset used during that time that could have memory corruption.......


I don't remember any Intel chipset having data corruption issues, but I know the VIA 686B southbridge did.
 
i've had a number of nforce chipsets that were incredibly hot to the touch.

mostly sli ones. 780a, 750a, 790i stuff.
 
i burnt out two nf4 chipsets on asus a8n sli deluxe mobos, they were rated for a high working temp, if I recall somewhere in the 80-110oc range but still they got burnt out.

The heat coming from them was nuts and the tiny 20000 rpm fan on the heatsink that would die after a couple of days did 0 to help with cooling.
 
  • Like
Reactions: N4CR
like this
Let's not forget how bad the AMD SB600 South bridge was. It was so shitty manufacturers started using SIS South bridges in place of them.
 
All of the Nvidia Intel chipsets were absolutely miserable. IIRC, the first NForce chipset for AMD was also a PITA, but NF 2, 3,4 were pretty epic.
 
  • Like
Reactions: N4CR
like this
I don't remember any Intel chipset having data corruption issues, but I know the VIA 686B southbridge did.

That's the one people think of when say bad chipsets. It's practically the picture in the dictionary definition.
 
Any SiS chipset.
Don't remind me about that PC-Chips M598 mobo and its SiS 530 chipset that my father built me a computer around back in the day. Just don't. There were so many things wrong with that build, from the chipset to the atrocious board layout (who puts port headers between expansion slots!?) to the use of an AT case and PSU well into the age of ATX (the board itself actually supports both AT and ATX).

I still remember that he bought an AGP card once, then had to take it back because *there was no actual AGP slot; the bus was electrically tied to some garbage integrated graphics that lost DirectDraw acceleration if you updated DirectX to anything later than 7.0 or so.* He came back with an ATI Xpert98 PCI, which did the job, but I wonder why he never bought a 3dfx card in their prime.

Also, I didn't learn until much later that it wasn't a 366 MHz AMD K6-2 he installed, but a 350 MHz one set to run at a slightly higher speed because going from 66 MHz to 100 MHz made the system unstable for some reason. Couldn't have been the RAM, it was a single stick of 128 MB PC-133.

It was a cheap piece of crap, but hey, I cut my teeth gaming on that thing regardless, even if I had no choice but to borrow the later Athlon XP systems from time to time once games got too demanding for it.
 
I'm backing the SIS group. They were awful. Just awful. At least VIA had a few good Athlon chipsets mixed in.

...Or maybe the worst was the Media GX SOC from Cyrix. That thing was simultaneously ahead of it's time, and 3 generations behind.
 
Had to work on a few HP/Compaqs that had socket 478 P4s with ATI chipsets in them and they were absolute garbage.
 
All of the Nvidia Intel chipsets were absolutely miserable. IIRC, the first NForce chipset for AMD was also a PITA, but NF 2, 3,4 were pretty epic.

NF3 was all good to me. Bit unstable the NF3-UD bios - sometimes would lose settings. Took weeks to get them dialed in for air WR level OCs lol.
My vote is for early VIA. I can't remember how bad the SIS chipsets were at the time, so probably avoided them like the plague.

Whatever the sandybirdge rev1 boards ran on was a buggy shitheap too.
 
nforce if i recall was a nighmare, followed by anything SiS
 
nforce if i recall was a nighmare, followed by anything SiS

The nForce chipsets weren't perfect, but on the AMD side they were among the best ever. nForce 2 was excellent as was the nForce 4. Some of AMD's Athlon XP era chipsets were good as well. Anything that originated on the ATI side of the house later on was crap until the 800 series.
 
Back
Top