Why is company XXX the best for memory?

[drhowarddrfine]

I designed computers from scratch for large companies as an electronic
engineer. Back then, the only thing I compared was reliability of the
chips made by the manufacturer and the published timing and tolerances
of their chips. Nothing else mattered. If the chip was in spec and
reliable, I bought the chip.

Nowadays, I see people recommending companies, such as Crucial, and
claiming their computers are running better, faster, than some other
brand but, to me, that doesn't make sense. Crucial, iirc, doesn't make
the chips so they have no control over anything else but building the
stick. If the chips are in spec, they should run as fast as Crucial or
anyone else. A car doesn't run faster because it's made by one company
over another if they used the same components.

Am I missing or forgetting something? Or is this just the talk of
12-year olds who don't know any better? (I'm not just picking on
Crucial because I've heard the same of their competitors)

 

[Paul]

I designed computers from scratch for large companies as an electronic
engineer. Back then, the only thing I compared was reliability of the
chips made by the manufacturer and the published timing and tolerances
of their chips. Nothing else mattered. If the chip was in spec and
reliable, I bought the chip.

Nowadays, I see people recommending companies, such as Crucial, and
claiming their computers are running better, faster, than some other
brand but, to me, that doesn't make sense. Crucial, iirc, doesn't make
the chips so they have no control over anything else but building the
stick. If the chips are in spec, they should run as fast as Crucial or
anyone else. A car doesn't run faster because it's made by one company
over another if they used the same components.

Am I missing or forgetting something? Or is this just the talk of
12-year olds who don't know any better? (I'm not just picking on
Crucial because I've heard the same of their competitors)

What exactly is the purpose of this rant ? If a "12 year old" needs
educating, then a professional like yourself is well equipped to
help them out. People measure you by the strength of your arguments,
not by how many degrees you've got.

Paul

 

[drhowarddrfine]

What exactly is the purpose of this rant ? If a "12 year old" needs
educating, then a professional like yourself is well equipped to
help them out. People measure you by the strength of your arguments,
not by how many degrees you've got.

Paul

It's nice to hear from a 12-year old, Paul, but I'm afraid I can't help
you since, like I said, I don't do that work anymore. This isn't a rant
at all but a grown up topic questioning why people focus on a company
name for no known reason. I want to know that reason, if there is one,
because I can't find it anywhere.

Now run along now, Paul. I hear your mother calling.

 

[Rod Speed]

I designed computers from scratch for large companies as an electronic engineer. Back then, the
only thing I compared was reliability of the chips made by the manufacturer and the published
timing and tolerances of their chips. Nothing else mattered. If the chip was in spec and
reliable, I bought the chip.

Nowadays, I see people recommending companies, such as Crucial, and claiming their computers are
running better, faster, than some other brand but, to me, that doesn't make sense. Crucial, iirc,
doesn't make the chips so they have no control over anything else but building the stick.

They obviously get a choice on what chips they buy
and other detail like the contents of the spd etc.

If the chips are in spec, they should run as fast as Crucial or anyone else.

Yes, but they obviously get to not use the worst chips etc.

A car doesn't run faster because it's made by one company over another if they used the same
components.

The reliability is determined by the design of those components etc.

Am I missing or forgetting something?
Yep.

Or is this just the talk of 12-year olds who don't know any better?

That too.

 

[drhowarddrfine]

They obviously get a choice on what chips they buy
and other detail like the contents of the spd etc.


Yes, but they obviously get to not use the worst chips etc.


The reliability is determined by the design of those components etc.


That too.

Yes, Crucial gets their choice of chips to use but so does everyone
else. You say I'm missing/forgetting something but you don't say what
it is. So far, you have only agreed with what I said.

 

[kony]

I designed computers from scratch for large companies as an electronic
engineer. Back then, the only thing I compared was reliability of the
chips made by the manufacturer and the published timing and tolerances
of their chips. Nothing else mattered. If the chip was in spec and
reliable, I bought the chip.

Ok, that is a reasonable thing to do, though there are still
other considerations today with integrated *modules* that
have PCB design differences, and variability in how suitable
a given EPROM SPD programming is compatible with a
particular board chipset, the timings it's bios supports,
and it's fallback/failsafe values if the board cannot run
one or multiple modules at the PROM-spec'd timings.

Nowadays, I see people recommending companies, such as Crucial, ...

Then often do so because people just want a name thrown out
there... "What should I buy?" Rather than a long lecture
about how to discriminate memory and a bunch of technical
terms (that many users will read and come back with a reply
like "I just wanted to know a product to buy." when it
wasn't actually what they asked, often without a context it
is assumed a person asking such a basic question is a
typical end user that just wants to build a new system or
upgrade a single one, the learning curve for discrimination
is a lot of time and research (for a fuller understanding)
with a diminishing return since they're only doing this task
once, or once every few years.

So it's easier to just recommend something one has personal
experience with or even better, that they see sells well and
has fewer user problem reports, suggesting there might be
more stability margin, more conservative specifications for
the chips or whole modules. Then there's also the issue
that major brand memory is more often tested by motherboard
manufacturers, and if certain timings prove problematic for
that board (for whatever reason, it may be a board problem,
not a memory module problem per se) the board manufacturer
may make bios adjustments and even explicitly specify
improved compatiblity with certain modules.

Thus, having some information about something that works, a
known good combination, is greater than merely assuming it's
"supposed" to work based on the specs.

...and
claiming their computers are running better, faster, than some other
brand but, to me, that doesn't make sense.

Well obviously it's untrue, it couldn't run better or faster
given same timings but perhaps they have falsely attributed
benefit to the perceived quality of a module rather than
that their system merely needed more total memory.

Crucial, iirc, doesn't make
the chips so they have no control over anything else but building the
stick.

They certainly do have control over chosing who to buy from,
and what chips they use in a given product. For example,
one manufacturer of low end memory might use chips with very
little margin at 200MHz and 2,2,2,5 timings. Another might
use chips that could do 220MHz at same timings. These specs
have to assume a reasonably ideal system running them,
reasonably good PCB, etc. As an engineer you should be able
to appreciate leaving a little margin in everything because
the more variables there are, the more one is likely to get
into trouble if just scraping by on max specs.

Max specs is key here, a chip isn't speced as 200MHz 2,2,2,5
actually, that's just the maximum it could run stable....
which you probably already know, but it does tend to account
for why some modules with same specs, aren't quite equal in
use in any particular, imperfect system. So there could be
instability, but not the "faster" aspect some falsely
presume (all else being equal).

If the chips are in spec, they should run as fast as Crucial or
anyone else. A car doesn't run faster because it's made by one company
over another if they used the same components.

Absolutely, but it's a bit hard to see what the user meant
and what their understanding was of the technology without a
full context behind their comments.

Am I missing or forgetting something? Or is this just the talk of
12-year olds who don't know any better? (I'm not just picking on
Crucial because I've heard the same of their competitors)

Sometimes it's just easier to tell someone what they want to
hear in the format they want to hear it in... "OCZ is great,
buy that." is a hell of a lot quicker than all the above we
both wrote.

 

[Paul]

Yes, Crucial gets their choice of chips to use but so does everyone
else. You say I'm missing/forgetting something but you don't say what
it is. So far, you have only agreed with what I said.

If you are an insider in the memory industry, then you know what
shortcuts are used. If you are an outsider, then you don't.
None of the company web sites I've looked at, goes into any
detail as to what method they use to produce a quality product.
The difference is in the testing, and testing is the part of
the product cost that can be shaved as needed.

Which means, you can only judge them by results. Have a look at
the Newegg customer reviews, and you can see differences between
brands. I don't have to recommend any brand, and I can let the customer
reviews do the talking. If a particular product is full of reports
of DOA, then there is a good chance the stuff is crap (like some
poorly tested UTT).

As for recommending Crucial, the main advantage is a person who
only knows the name of their computer (or their motherboard), can
find a product to match it. For example, instead of explaining
the difference between high density and low density RAM, for
a 440BX chipset, I can send them to the Crucial site, and they'll
get some RAM that works. Think of it as being closer to a
"full service" vendor.

Paul

 

[Rod Speed]

Rod Speed wrote

Yes, Crucial gets their choice of chips to use but so does everyone else.

Nope, most obviously with the chip manufacturers who
only get to use their own chips on their memory modules.

And some have decided that Crucial doesnt choose to use the worst chips available.

You say I'm missing/forgetting something but you don't say what it is.

Yes I did, that chip choice.

So far, you have only agreed with what I said.

Wrong, as always.

 

[Rod Speed]

drhowarddrfine wrote

If you are an insider in the memory industry, then you know what
shortcuts are used. If you are an outsider, then you don't.
None of the company web sites I've looked at, goes into any
detail as to what method they use to produce a quality product.
The difference is in the testing, and testing is the part of
the product cost that can be shaved as needed.

Its more than JUST the testing, they can obviously choose to
not use the worst of the chips that are commercially available.

Which means, you can only judge them by results. Have a look at the Newegg customer reviews, and
you can see differences between brands.

And some choose to use Crucial ram instead of bothering with that approach.

I don't have to recommend any brand, and I can let the customer reviews do the talking.

Some prefer to rely on the reputation of an operation like
Crucial instead and their guarantee that the ram they
supply will work fine in the system the ram is offered for.

If a particular product is full of reports of DOA, then there is a good chance the stuff is crap
(like some poorly tested UTT).

As for recommending Crucial, the main advantage is a person who
only knows the name of their computer (or their motherboard), can
find a product to match it. For example, instead of explaining
the difference between high density and low density RAM, for
a 440BX chipset, I can send them to the Crucial site, and they'll
get some RAM that works. Think of it as being closer to a
"full service" vendor.

Indeed.

 

[paulmd]

It's nice to hear from a 12-year old, Paul, but I'm afraid I can't help
you since, like I said, I don't do that work anymore. This isn't a rant
at all but a grown up topic questioning why people focus on a company
name for no known reason. I want to know that reason, if there is one,
because I can't find it anywhere.

Have you really never heard of brand loyalty? Do you not already know
that marketing has nothing to do with actual quality?

That said: there are a couple reasons people go with Crucial. As
mentioned, the configurator is handy to get memory that WORKS in a
machine, and customer service does matter. How returns are handled
matters. From what I've seen in the used computer business, crucial
seems a bit more resilient towards mishandling that most others.

Also... the cheap brands have less margin built into their specs. Which
is not helpful to stability. OCZ gears toward overclockers, so they
tend to have some wiggle room. (published specs are at least part
marketing)


There are MANY good companies besides crucial (kingston and kingmax,
among others). And some dogs, too (mushkin seems to have real
compatibility issues, and the here today, gone tomorrow companies are
to be avoided).

 

[drhowarddrfine]

Ok, that is a reasonable thing to do, though there are still
other considerations today with integrated *modules* that
have PCB design differences, and variability in how suitable
a given EPROM SPD programming is compatible with a
particular board chipset, the timings it's bios supports,
and it's fallback/failsafe values if the board cannot run
one or multiple modules at the PROM-spec'd timings.



Then often do so because people just want a name thrown out
there... "What should I buy?" Rather than a long lecture
about how to discriminate memory and a bunch of technical
terms (that many users will read and come back with a reply
like "I just wanted to know a product to buy." when it
wasn't actually what they asked, often without a context it
is assumed a person asking such a basic question is a
typical end user that just wants to build a new system or
upgrade a single one, the learning curve for discrimination
is a lot of time and research (for a fuller understanding)
with a diminishing return since they're only doing this task
once, or once every few years.

So it's easier to just recommend something one has personal
experience with or even better, that they see sells well and
has fewer user problem reports, suggesting there might be
more stability margin, more conservative specifications for
the chips or whole modules. Then there's also the issue
that major brand memory is more often tested by motherboard
manufacturers, and if certain timings prove problematic for
that board (for whatever reason, it may be a board problem,
not a memory module problem per se) the board manufacturer
may make bios adjustments and even explicitly specify
improved compatiblity with certain modules.

Thus, having some information about something that works, a
known good combination, is greater than merely assuming it's
"supposed" to work based on the specs.





Well obviously it's untrue, it couldn't run better or faster
given same timings but perhaps they have falsely attributed
benefit to the perceived quality of a module rather than
that their system merely needed more total memory.





They certainly do have control over chosing who to buy from,
and what chips they use in a given product. For example,
one manufacturer of low end memory might use chips with very
little margin at 200MHz and 2,2,2,5 timings. Another might
use chips that could do 220MHz at same timings. These specs
have to assume a reasonably ideal system running them,
reasonably good PCB, etc. As an engineer you should be able
to appreciate leaving a little margin in everything because
the more variables there are, the more one is likely to get
into trouble if just scraping by on max specs.

Max specs is key here, a chip isn't speced as 200MHz 2,2,2,5
actually, that's just the maximum it could run stable....
which you probably already know, but it does tend to account
for why some modules with same specs, aren't quite equal in
use in any particular, imperfect system. So there could be
instability, but not the "faster" aspect some falsely
presume (all else being equal).

Right, and that is what I'm getting at. Is this really an issue
anywhere? Not asking for specific instances but are these problems
widespread? That is, Crucial and "John's Memory Cards" may use the same
memory chips but John screws up the timing with his interface circuitry.

My point in all this is that I believe that any memory stick designed to
work in my computer will work at full speed no matter who I buy it from,
ignoring any reliability problems due to the manufacturer.

 

[drhowarddrfine]

Have you really never heard of brand loyalty? Do you not already know
that marketing has nothing to do with actual quality?

That said: there are a couple reasons people go with Crucial. As
mentioned, the configurator is handy to get memory that WORKS in a
machine, and customer service does matter. How returns are handled
matters. From what I've seen in the used computer business, crucial
seems a bit more resilient towards mishandling that most others.

Also... the cheap brands have less margin built into their specs. Which
is not helpful to stability. OCZ gears toward overclockers, so they
tend to have some wiggle room. (published specs are at least part
marketing)


There are MANY good companies besides crucial (kingston and kingmax,
among others). And some dogs, too (mushkin seems to have real
compatibility issues, and the here today, gone tomorrow companies are
to be avoided).

Ok but that is my question. Loyalty has nothing to do with the
question. Reliability isn't really part of it either. Some have given
the impression, and even claim, that one manufacturers memory stick will
let your computer run faster than anothers. Unless I'm wrong, I'm
saying it doesn't matter who you buy your stick from, they all run the
same because the timing is the same, excepting reliability of manufacture.

 

[Rod Speed]

Ok but that is my question. Loyalty has nothing to do with the question. Reliability isn't
really part of it either. Some have given the impression, and even claim, that one manufacturers
memory stick will let your computer run faster than anothers. Unless I'm wrong, I'm saying it
doesn't matter who you buy your stick from, they all run the same because the timing is the same,
excepting reliability of manufacture.

Maybe they are just getting confused about the 2, 2.5, 3 ratings etc, and
havent noticed that operations like Crucial to tend to sell the faster ram.

 

[paulmd]

Ok but that is my question. Loyalty has nothing to do with the
question. Reliability isn't really part of it either. Some have given
the impression, and even claim, that one manufacturers memory stick will
let your computer run faster than anothers. Unless I'm wrong, I'm
saying it doesn't matter who you buy your stick from, they all run the
same because the timing is the same, excepting reliability of manufacture.

Well, no. Changing brands of ram rarely improves speed, given identical
specs. Except in the realm of overclocking, where timings are pushed.
Kingmax (or king-something) actually made "PC-150" just for the
overclocker set.

 

[Rod Speed]

Well, no. Changing brands of ram rarely improves speed, given
identical specs. Except in the realm of overclocking, where timings
are pushed. Kingmax (or king-something) actually made "PC-150" just
for the overclocker set.

The original might have meant that stuff like Crucial can be run at the rated
specs fine but that cheap crap may have to be run slower to be reliable too.