AMD to decimate workforce several times over? With AMD announcing lowered expectations for the third quarter and CFO Thomas Siefert shown the door or finding the exit on his own a month ago, the talk has turned to what the company will do to get to profitability - and the prospect of deep job cuts. The Wall Street Journal jumped out in front with the layoff rumors, citing …
From the Oxford dictionary:
verb
[with object]
1kill, destroy, or remove a large proportion of: the inhabitants of the country had been decimated
drastically reduce the strength or effectiveness of (something): public transport has been decimated
2 historical kill one in every ten of (a group of people, originally a mutinous Roman legion) as a punishment for the whole group: the man who is to determine whether it be necessary to decimate a large body of mutineers
I guess you like to pretend dictionaries don't include the first definition
Except it's not - quite, anyway- if you decimated several times over then you'd fire one in ten, then fire one in ten of those who survived, then fire one in ten of those who'd survived twice. (You'd end up firing 27 out of 100, over the three rounds, if my maths is correct)
That's not quite the same as just firing 30% (It would certainly be more harrowing for the staff!)
Was used on Roman Legions if they failed to perform or ran away.
Also when a Roman soldier was found asleep on watch his mates were required to beat him to death.
Though I have a question, a Roman Century was made up of 80 men (not 100 as you might expect) so was Decimation 1 in 8 or was it decimal 1 in 10?
Per the OED, decimate has had the meaning of 'devastate' since at least 1663; every single use of the word in the British National Corpus as maintained by Oxford University uses it in that sense, though they all seem to be from the 1990s or newer so that's not a fantastic argument.
It's the idea that it's being used incorrectly that's a modern invention. It's up there with the idea that the lack of a possessive apostrophe on 'its' is a special case* for false inventions that have somehow gripped the public imagination.
* it's not — check the other personal pronouns; 'one's' is the special case.
Seconded. If they'd had any intention of going x86-64, the development would have been ongoing and they wouldn't have been caught on the hop quite so thoroughly when AMD shipped the AMD64 Opterons.
The enforced need to change horses in midstream and run the Red Queen's race to get a 64-bit x86 offering out had the side effect of starving the Itaniums of development. The multicore Itanics were so late that, by the time they turned up, everyone had lost interest with the exception of HP who had bet the farm on IA64.
We don't only get to thank AMD for x86-64, we can also credit them with delivering the killing blow to IA64 as well.
Samsung? Same rationale as HP, but maybe a happier marriage.
TSMC? If they fancied competing with Intel rather than just fabbing whatever someone else tells them to.
I also wonder if AMD isn't now so weak that Intel needs to throw it a few sops? Intel really needs AMD, because if it weren't for AMD then Intel would be a monopoly and regulated as such. Far better that it has an independant competitor that's more a terrier than a wolf.
IBM also has a secure trusted DOD/government fab in East Burlington, and their Special Clients would have a huge say in the transfer or discontinuance of that facility. huge say. the other facilities in Fishkill are reasonably expected to be supporting the Vermont works, and since they still have a great need for Germanium on Silicon at IBM, I don't see them becoming fabless. if Big Blue has Intel x86 licenses, I don't see why they would want AMD in the first place for something they already have. this sounds like another "analyst" looking for multi-millions for basically doing nothing stoking the rumor mill for business.
IMPHO getting to be about time to start totally ignoring the "market analyst" on Wail Street, since mostly they are playing the old Pump & Dump.
Big Blue did a very decent x86 design themselves (the Blue Lightning) and settled with Intel as a part of ending the hostilities in the war of the clones.
~ 1996 as a demonstration of typical "Intel Innovation" Intel went out against people who were producing _BETTER_ 486 designs to ensure the survival of the rather lame and expensive early Pentiums. AMD X5, IBM Blue Lightning and UMC U5 could run circles around the first Pentium. IMHO, the U5 and BL were perceived as the biggest threat as they had a performance edge in integer already and had announced floating point enabled parts coming within the next 6-9 months. Both (to the extent of my knowledge) were using the same design which gave the Opteron an edge a few years later - emulation of x86 on top of superscalar simple "risk-like" backend.
Prior to the "war" x86 were produced by AMD, UMC, IBM, Cyrix and Geode (these are the ones I can remember, probably there were more). AMD and Cyrix had prior cross-licensing deals so they managed to fend off the attack and continue onto shipping _BETTER_ Pentium clones (at least in the case of AMD, Cyrix never quite managed to catch up). Geode descended into the SOC obscurity. IBM settled with the terms of settlement unknown. U5 got court-nuked.
That is when and where the competition in the x86 market got nuked. It was not 10 years later during the Opteron vs NetBurst P4 wars. It was long before that.
Coming back to the "Golden Chain" - IBM's "Golden Chain" may be different from what you think. It may be tied by a very specific chain which prevents it contractually from designing x86 CPUs.
If IBM doesn't try to buy AMD, Amazon may be a contender if they don't buy out TI.
Dell is a bit of a long shot, but the threat of them buying AMD would leverage all sorts of goodies from Intel.
Huawei might make a play, but the Senate would have a mild conniption fit.
Finally, Apple could be looking to control more of the value chain in tablets, and having a company that can integrate graphics and other co-processors into the CPU chip might be attractive.
Cricket anyone?!
And - quelle surprise! - after many minutes of reflection they recommended massive lay-offs. Because there you are, smart young thing at a business consultancy after doing ever-so-well in your business studies degree and pretty darn well too in maths, and the cry for help comes from some firm in trouble in an industry you know 9/10ths of bugger all about:
What's the one sure thing that would help - all else remaining equal?
-- Reduce expediture!
And what's the one sizeable expenditure every firm has in common?
-- Salaries!
No need to thank me, citizens, hi-ho Silver! away....
Mentioned Oracle, but not that they would be as likely as HP.
The love owning full top to bottom silo, and AMD would allow them to sell servers to run Oracle/Sun etc on x86 as cheaper alternative to Sun acquisitions.
I could see Samsung or Amazon as mentioned above due to the high value of the Fusion/Trinity chips for laptops and mobile devices. (HTC, Nokia, etc would be interested, but probably don't have the cash)
Microsoft or Apple could be a real option here as well for the same reason. They would basically banish AMD from the desktops as a favor to Intel and use AMD on mobile devices. Apple would be more likely candidate here, as they loved the wall garden, but M$ is trying real hard to be more like apple these days, and take more control over the windows "experience"
Google is a likely suitor as well... unlike most of the companies listed, (except Apple) they actually have the cash sitting around to do this. AMD would get rolled into the Google servers, and they would use fusion on the mobile/Motorola offerings. And the ATI side of the company is actually a good money maker still.
With a fusion/trinity APU stuffed in a microATX form factor, Google could release media theater TV's that could play HD movies, and the latest PC gaming titles, and offer something Apple can't with Apple TV.
(Course Samsung could do the same)
I don't see IBM as getting in,, they have amazingly high profit margins on all their other business, and have the clout and position to stay in those types of markets, there is just not large enough profit margin on chips these days for them to bite (not as a 2nd place x86 competitor)
If they get low enough, Nvidia could make a run at it, they no longer have to worry about monopoly as they could at least cover themselves with Intels Video integrated offerings (and Power VR and other companies are still around)
Lenovo, Hauwei etc, are not likely to bite, due to the government contracts with x86 manufactures, its not likely that they would allow the deal to go through, the fear of it (and costs of failure) would keep them away.
RedHat however has a lot cash and has not been mentioned yet. It would give them a more like Sun Like OS and Hardware integration, and they could use the merger to push RedHat branded Servers and Desktops. (Ubuntu could be interested too, but likely don't have the cash).
Really and Server maker, or for that matter Network device maker that basically sells custom purpose servers for network use (Riverbed, Juniper, etc) would make a good merger partner as well, giving them a hardware integration edge that their competition lacks.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
