[H]ardocp (Kyle Bennett) thoughts on ATI/AMD, Intel, and Nvidia.

If you're conjoing a 100M CPU core and a 700M transistor GPU core, who's the daddy?

[In the end this is all just semantics, and politics too (like when two companies "merge" rather than "takeover").]

by swallow i dont just mean pull on die. I mean pull on die with the best performance\mm^2 ratio.
 
You can't parallelize tasks like word processing which are used by all people, so unless you introduce some kind of CPU ISA you won't be able to run a single-threaded program, and the most likely candidate is x86 since it is a legendary ISA. By doing this you go back to my point of the GPPU.

Maybe the chipset would take care of that ;)

I see the CPU becoming a mere periferal like the chipset nowadays, and the "upgrades" more on the side of add-in modules tailored for specific tasks in which they'll be much more capable then a CPU. Not much of a difference compared to nowadays, just on another scale.
 
edit
IIRC, the original "ELSA" was a german company allied with Nvidia in the GPU market (they also made network equipment, i think), but then it went bankrupt and sold the brand name to another company, based in Japan.

ELSA isn't a real nvidia only partner, anyway they never been exclusive before 2000 etc. they just made cards based on anything available, the original Gloria L/XL adapters were based on S3 Virge chips with random amounts of memory.
 
Last edited by a moderator:
ELSA isn't a real nvidia only partner, anyway they never been exclusive before 2000 etc. they just made cards based on anything available, the original Gloria L/XL adapters were based on S3 Virge chips with random amounts of memory.

The S3 Virge (the "De-ccelerator", as it was known back then :D) was around long before the term "GPU" was coined by Nvidia for the original Geforce 256 in late 1999, so i think my statement is still valid. ;)
 
The S3 Virge (the "De-ccelerator", as it was known back then :D) was around long before the term "GPU" was coined by Nvidia for the original Geforce 256 in late 1999, so i think my statement is still valid. ;)

Well, nVidia's GPU term was made up to consists any device that assists the CPU in graphics processing, which definitely includes all those S3 this that had some sort of hardware capability.. just not in dx/glide.
 
GPU came up with the GF256 and it's T&L unit, before that the term was not present.

I think the coining of that marketing term has nothing to do with what we call GPU's today, even nvidia itself changed it's 1999 definition of GPU (T&L +10 million polys) to a more common "any processor to which the CPU can offload graphics tasks"

Part of their nsist campaign since not everyone cares about hardware t&l but like "accelerated photo processing" better...
 
Last edited by a moderator:
What's this ?

Nvidia reportedly to trade technology for cooperation with Intel


Nvidia plans to release some of its graphics processor unit (GPU) technology to Intel in exchange for cooperation with the chip giant, according to sources at motherboard makers.

Nvidia has had discussions with Intel and has already reached several agreements for cooperation, noted the sources, adding that details of the agreements remain unknown.

Nvidia Taiwan declined to comment on the report.

Executives of Nvidia recently met with the company's partners in Taiwan to explain the company's plan in the motherboard market, pointed out the sources. With Intel pushing its GPU development again, Nvidia needs to be careful about trading its GPU technology with Intel in exchange for cooperation, added the sources.

http://www.digitimes.com/mobos/a20070507PD208.html


Humm, "Nehalem" chipset license (that would be strange, considering the object of the exchange -GPU tech-), or (yes, i know this sounds thin, at best) a x86 license ?
 
Trade of technology or are we talking about advanced integration of hardware or IP's?
 
When I read that at Digitimes this morning, fairly or not, my first reaction was that Intel is famous for strong-arming partners into cross-licensing IP as the price for playing on their platform. Dunno that's what's going on here, as the sourcing and details are mighty thin, but it certainly would fit the long-term historical pattern.
 
When I read that at Digitimes this morning, fairly or not, my first reaction was that Intel is famous for strong-arming partners into cross-licensing IP as the price for playing on their platform. Dunno that's what's going on here, as the sourcing and details are mighty thin, but it certainly would fit the long-term historical pattern.

I don't see intel sticking out a hand either, hello nVidia, want to make competing hardware?

Aren't we just talking about SLi plus some low-level communication stuff here
 
I don't see intel sticking out a hand either, hello nVidia, want to make competing hardware?

Aren't we just talking about SLi plus some low-level communication stuff here

Maybe. But if NV foresees the sun setting on a good chunk of their AMD chipset business (and I think it likely they do foresee that), then possibly something a littler broader is afoot.
 
Of course, licensing agreements are secretive, but considering NVIDIA's previous refusal to enter Intel chipset market on account of unfavorable licensing terms, there is reasonable speculation that the current chipset license does not involve royalty payments.

JHH has frequently pointed out that NVIDIA has a lot of IP too. So as mighty as Intel might be in the IP department, NVIDIA is hardly a slouch. And key patents are key patents, no matter how many total patents you have. I don't think it would be too much of stretch to think that NVIDIA has long been building up IP leverage to play against Intel.
 
No surprise really. Nvidia has a choice to remain loyal to AMD and hope they dont get backstabbed from a sinking ship. Or make a deal with the devil with the hope you can beat said devil at its own game.

The Intel market is much bigger and more stable than AMDs.
 
It might be bigger and more stable but Intel doesn't like sharing that market with anyone. BTW, despite JHH rethoric about how they are "the last one standing", "winner by default" and "exactly where they want to be", if Intel makes them a good offer - let's say $42 a share - Nvidia will vanish in a snap of the fingures.
 
It makes me really sad :cry: that the AMD\ nvidia realtionship is no more as good as it was before the ATI acquisition. They made really good partners and were meant for each other.
 
It makes me really sad :cry: that the AMD\ nvidia realtionship is no more as good as it was before the ATI acquisition. They made really good partners and were meant for each other.
Considering they bought nVidia's biggest competitor though, didn't you foresee just a few hiccups in their relationship? :|
 
I think one of the reasons why AMD-Nvidia relationship deteriorated that is often overlooked are the declining fortunes of AMD in general. Intel has always had a bigger slice of the pie, but until recently AMD had a multi-year grasp on the enthusiast community - and was growing across the board. Intel, on the other hand, was knee-deep in Netburst nightmare and was never keen in letting other companies on their turf anyway. The Nvidia-AMD relationship was natural - given a choice between a company with superior product and eagerness to partner and one with inferior CPUs and crappy attitude, who wouldn't go with the former?

Now, let's see what happened since:
1) AMD bought ATI. They said all the right things about keeping things open and blah-blah, but they now they are in direct competition with Nvidia across the board - everything Nvidia sells AMD would rather have you buy from them.
2) Nvidia broke into Intel market. You'd rather have 20% of 80% than 50% of 20%.
3) AMD lost the technological leadership. The same people who were buying Nforce 2s for their Athlons are now shopping for Core2 systems.

3 years ago, both companies needed each other. Today, neither does.
 
3) AMD lost the technological leadership. The same people who were buying Nforce 2s for their Athlons are now shopping for Core2 systems.
And that can change in the blink of an eye - Nvidia is in a sticky position here. There is no guarantee that AMD won't rise again as the darling of the enthusiast community and if they do Nvidia will want to be there with competitive products supporting AMD CPU's.
 
Back
Top