NForce For The Pentium 4? Oh Yes.
SimGuy
Ottawa, Canada
NForce For The Pentium 4? Oh Yes.
Chairman and CEO won't say no. IBM Microelectronics not mentioned...
Now is not the time to talk about next generation products - Nvidia executive during the conference call.
AN ANALYST at Nvidia's conference call last night tried to get CEO Jen-Hsun Huang to answer a question about when the firm will license Pentium 4 technology.
One little mole tells us that the rumour is that Nforce 5 will arrive sometime in Q1 of next year. Maybe.
Although Jen-Hsun kept wriggling out of the repeated question, it's evident from such wriggling that -- as we've suspected for a while -- such an idea obviously isn't too far from Nvidia's mind.
Jen-Hsun didn't answer the question but instead chose to reply to one that wasn't asked, and said that in the next quarter the firm was committed to align its chipset sales to AMD's K8, or Athlon 64.
As we've revealed recently, Nvidia roadmaps are indeed focused on AMD Athlon 64 processors for quite some while, but that doesn't necessarily mean a thing. Nvidia won't want to rock the AMD boat right this minute.
Us hacks call the Pentium 4 generation of Nvidia chipsets the Nforce 5.
While Nvidia executives went out of their way to praise their partner TSMC, we never heard the word IBM Microelectronics escape their lips, suggesting that it will be a while until we see chips from East Fishkill in any quantity.
Even when Nvidia was asked about IBM Micro, an executive said "it rarely used second sourcing" -- a quite puzzling answer, we think. Again the IBM Micro phrase wasn't mentioned, but TSMC was praised.
Nor is TSMC producing quite as many .13µ chips for Nvidia as the graphics firm would like, it appears. The price for .13µ chips from TSMC is still more expensive than .15µ chips – that's not a yield problem, it's a process problem.
One suit said: "Apart from Intel, no one else but TSMC is shipping as much .13µ technology". Another executive said: "Every die we buy, is a functional die", but .13µ chips come at a cost. Doubt was cast on how many .13µ chips its competitor ATI was achieving.
Why is this important? Well, if Nvidia can get most of its dies out at .13µ, it can likely increase its gross margins, affected by the higher cost of such semiconductors.
Another little moley suggests IBM will ship NV36 this coming Comdex... well, we'll have to wait and see about that too.
Source: The Inquirer
==============================================
If NVidia could pull off a killer system core logic for the Pentium 4 platform, they could seriously give the Canterwood & Springdale core logics a run for their money. Should be interesting to keep in mind...
Chairman and CEO won't say no. IBM Microelectronics not mentioned...
Now is not the time to talk about next generation products - Nvidia executive during the conference call.
AN ANALYST at Nvidia's conference call last night tried to get CEO Jen-Hsun Huang to answer a question about when the firm will license Pentium 4 technology.
One little mole tells us that the rumour is that Nforce 5 will arrive sometime in Q1 of next year. Maybe.
Although Jen-Hsun kept wriggling out of the repeated question, it's evident from such wriggling that -- as we've suspected for a while -- such an idea obviously isn't too far from Nvidia's mind.
Jen-Hsun didn't answer the question but instead chose to reply to one that wasn't asked, and said that in the next quarter the firm was committed to align its chipset sales to AMD's K8, or Athlon 64.
As we've revealed recently, Nvidia roadmaps are indeed focused on AMD Athlon 64 processors for quite some while, but that doesn't necessarily mean a thing. Nvidia won't want to rock the AMD boat right this minute.
Us hacks call the Pentium 4 generation of Nvidia chipsets the Nforce 5.
While Nvidia executives went out of their way to praise their partner TSMC, we never heard the word IBM Microelectronics escape their lips, suggesting that it will be a while until we see chips from East Fishkill in any quantity.
Even when Nvidia was asked about IBM Micro, an executive said "it rarely used second sourcing" -- a quite puzzling answer, we think. Again the IBM Micro phrase wasn't mentioned, but TSMC was praised.
Nor is TSMC producing quite as many .13µ chips for Nvidia as the graphics firm would like, it appears. The price for .13µ chips from TSMC is still more expensive than .15µ chips – that's not a yield problem, it's a process problem.
One suit said: "Apart from Intel, no one else but TSMC is shipping as much .13µ technology". Another executive said: "Every die we buy, is a functional die", but .13µ chips come at a cost. Doubt was cast on how many .13µ chips its competitor ATI was achieving.
Why is this important? Well, if Nvidia can get most of its dies out at .13µ, it can likely increase its gross margins, affected by the higher cost of such semiconductors.
Another little moley suggests IBM will ship NV36 this coming Comdex... well, we'll have to wait and see about that too.
Source: The Inquirer
==============================================
If NVidia could pull off a killer system core logic for the Pentium 4 platform, they could seriously give the Canterwood & Springdale core logics a run for their money. Should be interesting to keep in mind...
0