Although a novel idea, I dont think you have quite the processor design knowledge to justify your reasoning.
First of all, consider the arbiters and scheduling logic necissary for such a system. Also consider that the majority of the P4 transistors are L2 cache. Do you rememeber why we took cache off the motherboard and started putting it on the processor shortly after the 386?
There are other issues I can think of as well, but for the most part I think that the 386, even in theory, couldn't scale the way you propose.
However, you forget that Nvidia hasn't actually integrated a GPU in their core logic since the nforce2 chipset. ATI, Intel, SIS and sometimes VIA release IGP solutions for every chipset revision. Perhaps Nvidia found that IGP sales hurt their discrete solutions?
I used to have a professor who was an ex-NASA engineer. When the paint thing came up in class one day he gave us this explanation:
The reason why the paint was scrapped was not because it was 600lbs lighter, but beccause NASA came under pressure from environmental agencies to reduce the amount of toxic chemicals in the air as the rocket went into orbit and also when it blissfully burns up in the atmosphere.
Ah, yes, Anand's articles are certainly less fullfilling than thesewonderfulgems of journalistic laurels. After all, why talk about ATI video card availability when you can comment on their booth babes. I feel safe trusting my news to the guys who couldn't find the ATI video cards in the ATI booth.
All notable programmers in the last 20 years that write software you and I use constantly, although you probably don't even know it.
Why DJB isnt on the list is completely beyond me. He wrote 20% of the world's mail servers basically by himself in a little less than a year, and then wrote a BIND replacement soon after.
There are actually a few execeptions to this. The "sounding" of the first syllable is more important than the spelling.
For example, we do not say "an used hairbrush", we say "a used hairbrush". I believe any vowel making the sound of "you" or "won" forces us to use "a".
Conversely, when we are refering to the letter X, we say "an X", because X, when spoken is 'spelled' "ex" - and we place "an" before anything that starts "ex". Thats about the best explanation I can come up with, and I am sure there are more exceptions to the rule.
PCxxxx refers to how many MB/s the memory has for thoroughput. It isn't a BS marketing number, although many feel PC100 and PC133 were named improperly.
Fortunately the nice people at JEDEC have documentation of this for you. Here is a google cache link of an Infineon white paper about the ratings.
The southbridge ICH6-W adder costs $5. Its not "un-enabled" on the Grantsdale (northbridge) chipset. Its an extra chip that plugs into the southbridge that had technical issues when it first came out, and was then suspended from shipments for a while. I am guessing it eventually got scrapped for those issues.
HJ
Slashdot Mirror
Although a novel idea, I dont think you have quite the processor design knowledge to justify your reasoning.
First of all, consider the arbiters and scheduling logic necissary for such a system. Also consider that the majority of the P4 transistors are L2 cache. Do you rememeber why we took cache off the motherboard and started putting it on the processor shortly after the 386?
There are other issues I can think of as well, but for the most part I think that the 386, even in theory, couldn't scale the way you propose.
http://www.atacom.com/program/atacom.cgi?KEYWORDS= VIDR_ATIX_PC_25&USER_ID=www&SEARCH=SEARCH_ALL
So I guess the answer would be the ATI video card.
Instead of integrated graphics, maybe NVIDIA is planning to "bundle" their cheap TurboCache cards with nForce4 motherboards.
I concur. It doesn't sound cool to be but it does sound like progress.
However, you forget that Nvidia hasn't actually integrated a GPU in their core logic since the nforce2 chipset. ATI, Intel, SIS and sometimes VIA release IGP solutions for every chipset revision. Perhaps Nvidia found that IGP sales hurt their discrete solutions?
I used to have a professor who was an ex-NASA engineer. When the paint thing came up in class one day he gave us this explanation:
The reason why the paint was scrapped was not because it was 600lbs lighter, but beccause NASA came under pressure from environmental agencies to reduce the amount of toxic chemicals in the air as the rocket went into orbit and also when it blissfully burns up in the atmosphere.
Ah, yes, Anand's articles are certainly less fullfilling than these wonderful gems of journalistic laurels. After all, why talk about ATI video card availability when you can comment on their booth babes. I feel safe trusting my news to the guys who couldn't find the ATI video cards in the ATI booth.
AnandTech to the rescue, again.i =2317
http://www.anandtech.com/tradeshows/showdoc.aspx?
Click here to download plugin.
?
The 3D1 only works on SLI motherboards, even though it only requires one slot.
Wow it's like CoolTechZone paid to get these posts. Here is an article written by someone who has a clue:
i =2317&p=2 i =2317&p=12 i =2317&p=13
Intel: http://www.anandtech.com/tradeshows/showdoc.aspx?
AMD: http://www.anandtech.com/tradeshows/showdoc.aspx?
Transmeta: http://www.anandtech.com/tradeshows/showdoc.aspx?
Notice the $700 price point on dual core Athlon 64s (socket 939). Start saving up now.
Again, AnandTech saves the day
i =2317&p=15
i =2317&p=6
i =2317&p=5
BTX: http://www.anandtech.com/tradeshows/showdoc.aspx?
Shuttle: http://www.anandtech.com/tradeshows/showdoc.aspx?
More Shuttle: http://www.anandtech.com/tradeshows/showdoc.aspx?
God forbid CTX actually covered the show instead of just rewriting press releases...
A much better review is to be had here:5
http://www.anandtech.com/video/showdoc.aspx?i=231
Have any of you tried the Anandtech price engine? It's like froogle but with less BS it seems.
http://labs.anandtech.com/
Yo Rockway, its Kris. Rock on bro. Kristopher
All notable programmers in the last 20 years that write software you and I use constantly, although you probably don't even know it. Why DJB isnt on the list is completely beyond me. He wrote 20% of the world's mail servers basically by himself in a little less than a year, and then wrote a BIND replacement soon after.
There are actually a few execeptions to this. The "sounding" of the first syllable is more important than the spelling.
For example, we do not say "an used hairbrush", we say "a used hairbrush". I believe any vowel making the sound of "you" or "won" forces us to use "a".
Conversely, when we are refering to the letter X, we say "an X", because X, when spoken is 'spelled' "ex" - and we place "an" before anything that starts "ex". Thats about the best explanation I can come up with, and I am sure there are more exceptions to the rule.
According to the UIC LUG archives, he is giving a presentation on this article on the Nov 20 UIC Linux Users Group.
Details here: http://linux.pharm.uic.edu/
Yes.
You and the grand parent are ridiculous.
: www.infineon.com/cgi/ecrm.dll/ecrm/scripts/public_ download.jsp%3Foid%3D37618%26parent_oid%3D-9475+JE DEC+PC+rating&hl=en&lr=lang_en
PCxxxx refers to how many MB/s the memory has for thoroughput. It isn't a BS marketing number, although many feel PC100 and PC133 were named improperly.
Fortunately the nice people at JEDEC have documentation of this for you. Here is a google cache link of an Infineon white paper about the ratings.
http://64.233.167.104/search?q=cache:pT68a3nJpdkJ
Oh, I dont suppose you are refering to PCWorld Bench?
AnandTech
Yeah, I know. Reading the article is tough, but some of us have to do it. More info here: PCWorld
I actually posted it with "its" - the moderators changed it!
The southbridge ICH6-W adder costs $5. Its not "un-enabled" on the Grantsdale (northbridge) chipset. Its an extra chip that plugs into the southbridge that had technical issues when it first came out, and was then suspended from shipments for a while. I am guessing it eventually got scrapped for those issues. HJ