Intel Dominated, AMD Gained in Chip Market in 2011: IDC

[god_43]

AMD saw strong growth in its mobile chip business, but Intel continued to control more than 80 percent of the worldwide microprocessor space in 2011, IDC analysts said.

Advanced Micro Devices got a boost from its Fusion processors in 2011, seeing gains in market share for mobile chips and a slight bump in overall processor share, according to market research firm IDC.

However, Intel only showed a slight decline during a year that saw microprocessor revenues jump 13.2 percent, to more than $41 billion. That, in part, was helped by the fact that systems makers were willing to shell out more for the chips than they paid in 2010, according to Shane Rau, director of PC semiconductor research at IDC.

http://www.eweek.com/c/a/IT-Infrastr...11-IDC-148600/


ouch down to 5.5 percent market share...i hope piledriver can help them gain share back.

[savantu]

If they couldn't get share with BD ws. the 1 and a half year old Westmere, what chances does the 10% faster PD vs. the 30-50% faster SB-E Xeons ? AMD will be driven out of the server market. They will probably be in the noise region , 2-5%, depending on HPC where their main selling point are comparable FLOPs with Xeons but at far lower prices.

[BeepBeep2]

Interlagos was so bad it's clear why they lost market share. The Anandtech review for example...AMD lost pretty much everything and used more power.

[R101]

We need strong AMD products or Intel will rip us off.

But in real life server price is like 25% of the whole thing, software is where we spend our money.

[informal]

The problem with interlagos is not only that design was underwhelming but also GloFo's process sucked badly. They couldn't get to the targeted clock zone,not even close(4.3Ghz for desktop parts as per AT;I suppose ~3.1Ghz for 16C server parts within 140W envelope). The had to incorporate various tricks into PD to get the power draw down and clocks up since they can't rely on GloFo's execution any more. Good news is that PD is supposedly clocking noticeably better and has at least some IPC improvements so they can close the gap to SB-E parts,especially in HPC. But what happens when Haswell arrives bringing 2x256bit FMA capability per core next year? If AMD doesn't retain the present parity flops vise with SR core(implying they need 2x256bit FMAC per core pair/module) then they will be out of HPC market come 2014.

[STaRGaZeR]

Keep doing small crappy cores and that market share will continue to shrink

[Piotrsama]

Keep doing small crappy cores and that market share will continue to shrink

It didn't shrink... only the server side.

Intel held 80.1 percent of the worldwide unit market, a slight drop of 0.6 percent over 2010, according to IDC. AMD’s market share stood at 19.7 percent, a bump of 0.7 percent. Via Technologies’ share was 0.2 percent, a drop of 0.1 percent.

AMD was bolstered, in particular, by its performance in the mobile PC chip space, where its share grew to 16 percent, a 2.7 percent increase.

[Tim]

It's been years for AMD to come up with another smash hit like the Athlon 64s, to be honest I couldn't care less any more. The GPU side is interesting, CPU it's Intel all the way.

[highoctane]

AMD actually has more the desktop market share than I thought at 26% after dropping 1.6%, I'm actually kind of surprised they lost marketshare in the desktop space.

Intel also grew its share of the desktop chip space, seeing a 1.7 percent jump to 73.8 percent. AMD’s 26 percent share was a loss of 1.6 percent.

[STaRGaZeR]

It didn't shrink... only the server side.

It did, read the article again. I was referring to server thou.

[BeepBeep2]

The problem with interlagos is not only that design was underwhelming but also GloFo's process sucked badly. They couldn't get to the targeted clock zone,not even close(4.3Ghz for desktop parts as per AT;I suppose ~3.1Ghz for 16C server parts within 140W envelope). The had to incorporate various tricks into PD to get the power draw down and clocks up since they can't rely on GloFo's execution any more. Good news is that PD is supposedly clocking noticeably better and has at least some IPC improvements so they can close the gap to SB-E parts,especially in HPC. But what happens when Haswell arrives bringing 2x256bit FMA capability per core next year? If AMD doesn't retain the present parity flops vise with SR core(implying they need 2x256bit FMAC per core pair/module) then they will be out of HPC market come 2014.

You don't even know how good or bad GloFo is. You just presume it is the cause because you don't want to believe that BD sucks.

It is a combination of both. It's easy to see that the process is largely inconsistent. However, in best case scenario AMD on GF SOI is 80% behind Intel's bulk 32nm process on a performance per watt standpoint. I'm not going to sit here and blame the process alone for an 80% deficit.

PD will end up being what BD should have been, 9-10 months ago.

[sergiojr]

If they couldn't get share with BD ws. the 1 and a half year old Westmere, what chances does the 10% faster PD vs. the 30-50% faster SB-E Xeons ?

But they have gained server marketshare after shipments of BD started. And predicting 10% speed increase with PD is ridiculous.