I have a question; As RAM prices continue to plummit, and the Hot Deals section of this forum keeps lighting up with more cheap RAM deals, I'm wondering why is it so cheap? Why isn't other hardware this cheap? Is it because so many companies produce it? Or maybe I'm just a cheap geekjew that's too tired of dropping $1500 a year on computer parts...
Tired of paying thousands for silicon and copper alloys
- Thread starter idea
- Start date
One thing no one mentioned yet: manufacturing efficenses. Making parts on silicon is still a very new industry. The manufacturing technology for integrated circuits has increased dramatically every year since they were first invented in 1974 (I think it '74 anyway. Near there). We have gone from a few hundred transisters per chip to nearing 100 million now. And it is predicted that we will be at 1 billion sometime in 2010. If you wanted 128meg of memory 5 years ago, it would have been made on 1 micron technology on a 4" wafer and you would have spent mucho dough to buy all of those huge chips required to stuff in 128 megs of memory. Only servers and other truely heavy duty applications could justify that kind of expense (or have MBs with enough memory slots to stuff in that much memory!). Now, that same memory will be manufactured on an 8" wafer (and soon 12" will be standard), will use a .15 or .13 micron technology, and they will fit neatly on 8 chips which will be on one memory module. Also, each of these chips is much smaller than the 4 mb chips you would have been buying 4 years ago. It's just much cheaper to make memory today than it was 5 years ago.
CPUs have been following the same larger wafer/smaller manufacturing technology as well, buy there is one major difference: 5 Years ago, it would take some were around 5 transistors for each memory cell. Today, it still takes about 5 transistors per cell (yes, these are ball parts. different memory technology uses different amounts of transistors). A CPU on the other hand, has gone from a few ten's of thousands of transistors in the original Pentium to over 50 million in the PIV. A thousand fold increase in transistors is a huge increase in die size required to hold it. Only the constant shrink on transister size has allow chips like the Athlon or PIV to be produced. Needless to say, it all boils down to the fact that CPU prices have not fallen nearly as fast as memory has and probably never will due to the constantly increasing complexity of CPUs. Of course, we have done a similar thing with memory by going from 8/16 megs of memory standard 5 years ago to 128 megs today (or has it slipped up to 256megs?). But that is only a 16 fold increase in transister density, not a 1000 fold increase.
CPUs have been following the same larger wafer/smaller manufacturing technology as well, buy there is one major difference: 5 Years ago, it would take some were around 5 transistors for each memory cell. Today, it still takes about 5 transistors per cell (yes, these are ball parts. different memory technology uses different amounts of transistors). A CPU on the other hand, has gone from a few ten's of thousands of transistors in the original Pentium to over 50 million in the PIV. A thousand fold increase in transistors is a huge increase in die size required to hold it. Only the constant shrink on transister size has allow chips like the Athlon or PIV to be produced. Needless to say, it all boils down to the fact that CPU prices have not fallen nearly as fast as memory has and probably never will due to the constantly increasing complexity of CPUs. Of course, we have done a similar thing with memory by going from 8/16 megs of memory standard 5 years ago to 128 megs today (or has it slipped up to 256megs?). But that is only a 16 fold increase in transister density, not a 1000 fold increase.
Add to the fact that the memory market is an open market. Prices are based on spot prices.
ie Intel can charge 10X as much for a Xeon chip because they are the only ones that make Xeons where as ram prices are based on the Supply & Demand for them.
Right now there is a major lack of Demand for computer components in general.
For companies like NVidia all they have to do is cut back production to keep there prices constant. Memory manufacturers cant cut production because that would mean higher per unit costs than there competitors. Therefore there competitors can sell the ram at a cheaper rate, thus having better profit margins.
Memory manufcaturers either continuosly lower costs (ie the end price) or go out of business because there memory products are too expensive for the spot market and no one buys them.
FYI for DRAM setting a price low enough will stimulate buying. Take most people on the forums seeing as how low memory prices went they got 0.5Gig+ of ram in there system. A year ago it was considered excessive or just plain Ludacrous.
ie Intel can charge 10X as much for a Xeon chip because they are the only ones that make Xeons where as ram prices are based on the Supply & Demand for them.
Right now there is a major lack of Demand for computer components in general.
For companies like NVidia all they have to do is cut back production to keep there prices constant. Memory manufacturers cant cut production because that would mean higher per unit costs than there competitors. Therefore there competitors can sell the ram at a cheaper rate, thus having better profit margins.
Memory manufcaturers either continuosly lower costs (ie the end price) or go out of business because there memory products are too expensive for the spot market and no one buys them.
FYI for DRAM setting a price low enough will stimulate buying. Take most people on the forums seeing as how low memory prices went they got 0.5Gig+ of ram in there system. A year ago it was considered excessive or just plain Ludacrous.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 20K
-
-
-
Facebook
Twitter