Sign in with
Sign up | Sign in
Your question

why dont they just make 1 CPU per model

Last response: in CPUs
Share
February 24, 2002 8:07:32 PM

Why don't they just make 1 CPU per model and allow the mobo manufacturers to allow for setting the clock and FSB multiplier thru the bios? A good temp monitor and thermal protection would provide safeguards for CPU frying. That way, the user could decide their performance/life span ratio. And since all models would be the same, there is no need to worry about unscrupulous OEMs marketing overclocked CPUs.

More about : dont make cpu model

February 24, 2002 9:59:24 PM

That plan makes no sense lol.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
February 24, 2002 10:33:29 PM

lets only sell geforce 4 4600's... and derate them then sell them as Geforce4 MX's... yeah. i can see that working.

why earn $399 when u can get $150?

hmmmmm
yes. a less than genius intellect thread this.
*drool*

and as i depart... a final witticism

"i believe in intellectual osmosis. Therefor, i must stop writing in this thread otherwise i will become a moron"

Overclocked athlon 1200C @ 8.5 x 166FSB + PC2700 = GOOD! :smile:
Related resources
Anonymous
a b à CPUs
a b C Monitor
February 25, 2002 12:45:21 AM

Apparently, the two people who replied before me were morons. The assumed that you would have the Manufacturers sell the chips at thier lowest prices, and let you rev them up. I don't think that's what you were getting at.

Instead of having say, an XP 1600, 1700, ... 2000, why not just release the XP and charge what the XP 2000 costs? Sell only one chip, and charge the highest amount possible, then let you down/up clock it as you wish? Right? Well the problem with that is not everyone wants to spend $400 on a cpu. The marketing plan is to put out the new chip at a high price, and sell a bunch. Then lower the price and bring out a faster chip, and sell a bunch of both - people will buy the lower chip now because it's a deal. Why can't you do this with entire chip technologies instead of clock rates? Cause they can't pump out and produce new technologies fast enough to keep a sustained market place. Sustained market = more cash flow.
February 25, 2002 12:53:03 AM

Some processors will simply not be able to run at the highest speed. Do those get thrown away or sold just like the rest without any indication? Some people like to know what their processor can safely handle. This is why they are all tested. Just because your system will boot up with your processor running at a certain clockspeed, that does not mean it is stable. It could very well end up corrupting your data if it gives incorrect results.

The end-user should not be forced to perform intensive testing on the processor to ensure it is working properly. That should be the job of the manufacturer and is the main technical reason why we have official speed grades. The other reason is of course monetary in nature.

-Raystonn


= The views stated herein are my personal views, and not necessarily the views of my employer. =
February 25, 2002 1:52:59 AM

havnt seen u in a while ray. nice to have u back

"I came, I saw, I overclocked", Julius 'Smokin CPU' Caesar :smile:
a b à CPUs
February 25, 2002 5:20:10 AM

YES! Lets eliminate all NW's except the 1800, if anything won't run 1800, throw it away, otherwize, even one that goes 2200 sells as an 1800. Then charge $300 for it.

As for the AXP, let's eliminate everything except the 1800+, and charge $150 for it.

Then all the performance heads can buy 6 CPU's instead of one, to try to figure out which one works at the highest clock! And screw poor people, they don't need computers anyway, they can go to the library.

Sounds great from a refurbishers standpoint, all systems would cost over $1000, so I could raise the price of my 200MMX units to $300 and not have to compete against cheap new systems!

What's the frequency, Kenneth?
February 25, 2002 10:11:47 AM

Quote:
Instead of having say, an XP 1600, 1700, ... 2000, why not just release the XP and charge what the XP 2000 costs? Sell only one chip, and charge the highest amount possible, then let you down/up clock it as you wish? Right? Well the problem with that is not everyone wants to spend $400 on a cpu. The marketing plan is to put out the new chip at a high price, and sell a bunch. Then lower the price and bring out a faster chip, and sell a bunch of both - people will buy the lower chip now because it's a deal. Why can't you do this with entire chip technologies instead of clock rates? Cause they can't pump out and produce new technologies fast enough to keep a sustained market place. Sustained market = more cash flow.



What if I dont need a 2000+, the idea is retarded IMO.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
February 25, 2002 11:07:34 AM

Yes, I agree, the idea is not very smart marketing, at the very least. What will probably happen is that eventually we'll have unclocked (not unlocked, unclocked!) asynchronous CPUs (those will probably come within 5-10 years). These CPU won't have a clock or clock cycles, but will simply process information as fast as it's able to. Imagine this, rather than having IPC * Hz, you have IPS (instructions per second). I can't really go in too much detail about asynchronous CPUs, but I remember reading a lot about them before. Try to search for it on Google.

AMD technology + Intel technology = Intel/AMD Pentathlon IV; the <b>ULTIMATE</b> PC processor
February 25, 2002 12:38:20 PM

I just noticed Ray said exactly the same thing before me. No point repeating. *Post Deleted*<P ID="edit"><FONT SIZE=-1><EM>Edited by bigrat on 02/25/02 09:42 PM.</EM></FONT></P>
February 25, 2002 10:36:09 PM

Well, I see I have to perform yet another act of intellectual charity. I ought to set up a foundation for tax purposes.

Let me elaborate. For each class of CPU (katmai, coppermine, willamette, northwood, t-bird, xp, t-bred, etc), there are very slight differences. If they focused on just one model, even if it meant the upper range of the CPU speed would be lower, then they could save large amounts on capital costs.

As for marketing, they could sell the CPUs at the weighted average retail price. They could probably get away with selling it at slightly higher (since I am assuming far more people buy at the lower range than at the top speeds for each CPU class). And intel/amd already have cheap versions (duron, celery) of their CPUs, so the argument that they would lose the cheapskate market share is irrelevant.

Also, the amount of R and D money they spent per CPU class would decrease, so as soon as intel/amd puts a successful CPU out to market, they can jump right ahead with researching/marketing the next CPU class. This, in my opinion, would engender a shorter CPU generation span, since intel/amd could spend more money in making quantum improvements rather than in tweaking current technology.

I really don't know of anyone who buys the same CPU at diff speeds for an upgrade (how many people bought a p2 350 then upgraded to a p2 550; or an 800 mhz t-bird then upgraded to a 1.4 t-bird). Hell, most people don't upgrade until at least 2 CPU class generations have passed (and I'm not counting the used CPU market -- intel/amd don't profit from that anyway). So, the lame brained argument about this being bad marketing doesn't wash (much like its authors).
February 25, 2002 10:57:19 PM

Quote:
If they focused on just one model, even if it meant the upper range of the CPU speed would be lower, then they could save large amounts on capital costs.


They dont sit around a design table and design the 1700+ and the 1800+, they ARE all the same chip, just ones with higher yields and better characteristics.

"The Cash Left In My Pocket,The BEST Benchmark"
No Overclock+stock hsf=GOOD!
February 25, 2002 10:58:24 PM

Skim I think most everyone understood your question the first time you posited it.

While I understand where you might feel it makes sense in your mind, economically (for the CPU manufacturers) it might not and here's why I can see it NOT being feasible.

1. As the production process of a new chip ramps up, so does the quality of the chip. While there are some indications (roadmaps) for general purposes to indicate how far a particular chip will go, that can't be guaranteed. So, it makes more sense to produce *faster* versions of a design as the production binsplits improve.

Take for example, the Athlon Palomino core.

Why do you think AMD released the MP and Athlon 4 mobile chips first?

Well, I think the answer is pretty clear. Early binsplits of the Palomino just were not good enough to replace the Thunderbird desktop chip. AMD was getting mostly 1-1.2ghz Palominos. So, they did the logical thing...decided to produce the MP and mobile chips, which were more than adequate in that speed range.

If AMD or Intel decided to sell 1 chip per family, there might be some serious problems. the production process would have to be dead on right.. noticeably improved CPU performance over the previous generation.

And I think this is what folks have been trying to explain in not quite so many words.

Of course, there's another reason why your idea doesn't wash....marketing.

mark-

When all else fails, throw your computer out the window!!!
February 25, 2002 11:10:36 PM

i dont even think of upgrading till cpu's double in speed.

started with a p-166
only upgraded to a p2-300 when the 350 came out.
likewise
a cellery 500 when the 600 came about
then up to my current athlon 1200.


"I came, I saw, I overclocked", Julius 'Smokin CPU' Caesar :smile:
February 25, 2002 11:36:02 PM

This isnt as idiotic of a thought as it sounds, but you guys are right in that itll never happen!

1 more post and I "gain a level"
haha

Jesus saves, but Mario scores!!!
!