AN28V72X8SMGAS 1GB SDRAM PC133 144pin SODIMM | eBay
-
Do you mean, why are there ECC SODIMMs when no laptop will take them? The answer is simple, there are some boards that are meant for other applications where ECC is desirable (servers IIRC) that take SODIMMs. I could be wrong about the server part, but those are meant for non laptop boards.
cdoublejj likes this. -
Oh, nostalgia. lol
It's a funny thing, really. Even when ECC ram actually had a purpose, the actual error correction may very well never have actually saved data from being corrupted on any reasonably modern computer. That wouldn't also have been detected by a normal parity check, or rescued by the OS.
In the reverse example, ram with badly set up latency&drive strength settings would still not have been made useable by relying on the ECC. Instead the opposite was actually the case. That the ECC ram would become unstable at lower frequencies compared to non-ECC ram.
..anyway. So this is an 144 pin laptop sodimm intended for some sort of pentium or celeron board. Running at 66Mhz FSB on the laptops, no doubt (but it supports up to 133Mhz. The desktop equivalents belong to the class of the most massively overclocking friendly boards ever made). Probably from around '99 or '00. And since it's ECC supported, this must have been put in one of the Pentium II or III boards on some immensely expensive kit. Possibly one of the IBM Thinkpads?
Yeah, I'm bored. -
I wonder if that ram would work on normal laptops with ecc turned off or the SPD data tweaked. i've seen server stuff work on socket9393 but, the hardware was there to do so. -
Modern ECC SODIMMs are usually for micro servers and stuff.
It's funny that ECC enabled notebook does exist. Some new versions of mobo on the Clevo P570 series supports ECC.cdoublejj likes this. -
google doesn't seem to have any thing what applications would use a PC133 SODIMM.
-
-
It's a weird thing.. it's difficult to figure out exactly when ECC ram would specifically have been useful. Further complicating things is that it's been very common to use parity checks on hdd error checking, with systems that have much higher frequency of hardware errors than ram would ever produce. Factory systems and so on will always also have additional internal error checking algorithms and consistency requirements as well.
Where it makes sense is in level 2 cache, for example. Say, you have an instruction level cache for a large system that over time has a significant amount of recycle operations in a relatively fast piece of memory, that operates on independent logic from the system outside. Then it might makes sense.
And I guess if you thought about the existence of, or designed, software that was completely reliant on every memory operation being consistent without errors (..then you need to basically run manual memory mapping and memory operations..). That you wrote insert hacks into a database that were rarely or never written to object structures and cache outside planned maintenance, etc. Then having a hardware check on an extra chip that is guaranteed to run every time a memory operation is completed might make some sense.
The thing is that then we're getting pretty far back in time, and the probability of overrun errors from everything else than hardware error is massively higher than single memory faults..
..actually, scratch that. I mean, a bunch of old software with c+ hacks still turn up today. I guess there might still be situations where it actually does make the system run with somewhat increased consistency. And I can vividly imagine someone selling that "one error is one error you can't afford!", etc.cdoublejj likes this. -
What the??? Ecc sodimms?
Discussion in 'Hardware Components and Aftermarket Upgrades' started by cdoublejj, Sep 25, 2014.