-
Have a passion for hardware and the written word? Tom's Hardware is seeking contributors!
slvr_phoenix
Splendid
- Dec 31, 2007
- 6,223
- 1
- 25,780
From my personal experience I see the Alpha as having gone nowhere purely because they tried to press into the x86 market and failed miserably, and that in turn made a lot of people wary about Alphas.
I've still got an Alpha box sitting under my desk with an Alpha-specific version of Red Hat and Alpha patches for NT4, Dev Stuidio 6, etc. It had been intended to pave the way into the Alpha future for my company by compiling our software for Alphas. It was a complete failure.
Using the goofy x86 emulation, well over 50% of the packages that I tried to install on the Alpha box (including InstallShield of all things) wouldn't even finish their install without crashing. And whenever I would contact any company about an Alpha version of their software so that I could actually, you know, do <i>work</i>, I always got back a response of either "we don't support Alpha" or "we will support Alpha when Microsoft does". (The latter being strange since, as I said, I had an Alpha patch for NT4 and Dev Studio 6, so clearly M$ already supported Alpha.)
And I saw the effect that this had on my company's confidence in the DEC Alpha as a platform. I mean I couldn't even get Alpha versions of all of the software libraries that I needed to compile our software, and even if I could compile the software I still couldn't even make a distribution to deliver it to customers. And it was all because of a lack of direct software support and a really crappy x86 emulation.
So my personal experience of the fall of the DEC Alpha was not that it was financially unviable, but that DEC pushed the Alpha into the x86 market, but the lack of software support pushed it back out and at the same time gave the Alpha a bad taste in many people's mouths.
Had the Alpha never been pushed as an x86 replacement it might have done quite well in its true market. People really liked it until they tried to replace their Pentiums with one.
<pre><b><font color=red>"Build a man a fire and he's warm for the rest of the evening.
Set a man on fire and he's warm for the rest of his life." - Steve Taylor</font color=red></b></pre><p>
I think you are getting two things wrong and you are looking at this with a PC mindset. Alpha didn't lack industry strength software support, it ran OpenVMS and Tru64. Not exactly OS's that are known to be crappy. Those where among the best OS's for Alpha's original market, and actually they still are.
Also, any company judging Alpha on its stability or performance running x86 binaries in emulation under wintoys should fire whomever responsible. That is like buying a Mac G5, installing VirtuaPC and complaining it doesnt run the x86 version of Need for Speed III, so surely running SAP under AIX on a Power5+ server would be a terrible idea, lets get SUN boxes instead ..
No, the total absense of desktop&windows software is definately not what killed Alpha in the high end, but it is a nice illustration of three other things:
1) Alpha needed a higher volume market than just the highend Unix server market to warrant to expenses. DEC kept bleeding money on it. Itanium take note.
2)Binary software emulation is a very poor substitute for native software. Itanium, take note.
3)how hard it is to replace x86 on the desktop or even workstation market, and performance alone doesn't get you there. I think Intel already noticed that
I've never had the pleasure of toying with an Alpha Workstation myself, but collegues of mine did, and actually liked it a lot, and told me it ran most x86 binaries pretty well under FX!32. Still, there is not much point in buying a system >2x as expensive just to cripple it running emulated software which may or may not work. IPF would face the exact same challenges if ever intel wanted to crack the workstation of even desktop market with it, but with the minor difference that Alpha was roughly 2x as fast as x86 boxes back then, and often where considerably faster even emulating x86 binaries under FX!32, where IPF is roughly on par running native software.
Your experience does explain why Alpha never took over x86 on the desktop, but not why it failed in the high end where windows or x86 compatibility where total non issues.
>Had the Alpha never been pushed as an x86 replacement it
>might have done quite well in its true market
No, one has nothing to do with the other. Power(PC) failed just as miserably trying to replace x86 on the windows desktop market, but it had zero effect on its high end market.
= The views stated herein are my personal views, and not necessarily the views of my wife. =
Also, any company judging Alpha on its stability or performance running x86 binaries in emulation under wintoys should fire whomever responsible. That is like buying a Mac G5, installing VirtuaPC and complaining it doesnt run the x86 version of Need for Speed III, so surely running SAP under AIX on a Power5+ server would be a terrible idea, lets get SUN boxes instead ..
No, the total absense of desktop&windows software is definately not what killed Alpha in the high end, but it is a nice illustration of three other things:
1) Alpha needed a higher volume market than just the highend Unix server market to warrant to expenses. DEC kept bleeding money on it. Itanium take note.
2)Binary software emulation is a very poor substitute for native software. Itanium, take note.
3)how hard it is to replace x86 on the desktop or even workstation market, and performance alone doesn't get you there. I think Intel already noticed that
I've never had the pleasure of toying with an Alpha Workstation myself, but collegues of mine did, and actually liked it a lot, and told me it ran most x86 binaries pretty well under FX!32. Still, there is not much point in buying a system >2x as expensive just to cripple it running emulated software which may or may not work. IPF would face the exact same challenges if ever intel wanted to crack the workstation of even desktop market with it, but with the minor difference that Alpha was roughly 2x as fast as x86 boxes back then, and often where considerably faster even emulating x86 binaries under FX!32, where IPF is roughly on par running native software.
Your experience does explain why Alpha never took over x86 on the desktop, but not why it failed in the high end where windows or x86 compatibility where total non issues.
>Had the Alpha never been pushed as an x86 replacement it
>might have done quite well in its true market
No, one has nothing to do with the other. Power(PC) failed just as miserably trying to replace x86 on the windows desktop market, but it had zero effect on its high end market.
= The views stated herein are my personal views, and not necessarily the views of my wife. =
P4man
it funny you have mentioned SAP.
SAP on windows has been crying out for 64bit for years.
Xeon processors themselvs have been well capable of running SAP in terms of CPU utilization, however in order to get the number of required users more and more application servers had to be added purly for memory reasons.
This is one area where Itanium is starting to gain momentum. I have a number of customers who are evaluating Itanium/windows 64bit and a number who have already migrated.
A large number are also waiting for AMD64/Intel64 bit.
We have both in our labs for evaluation (AMD64 on Linux is looking very tasty)
it funny you have mentioned SAP.
SAP on windows has been crying out for 64bit for years.
Xeon processors themselvs have been well capable of running SAP in terms of CPU utilization, however in order to get the number of required users more and more application servers had to be added purly for memory reasons.
This is one area where Itanium is starting to gain momentum. I have a number of customers who are evaluating Itanium/windows 64bit and a number who have already migrated.
A large number are also waiting for AMD64/Intel64 bit.
We have both in our labs for evaluation (AMD64 on Linux is looking very tasty)
>SAP on windows has been crying out for 64bit for years.
.. but I've not yet read an announcement of SAP supporting AMD64/EM64T.. did I miss it ?
>We have both in our labs for evaluation
Both meaning operon/itanium or "Xeon 64" ?
= The views stated herein are my personal views, and not necessarily the views of my wife. =
.. but I've not yet read an announcement of SAP supporting AMD64/EM64T.. did I miss it ?
>We have both in our labs for evaluation
Both meaning operon/itanium or "Xeon 64" ?
= The views stated herein are my personal views, and not necessarily the views of my wife. =
So why am I not seeing AMD bringing in a cool 1 billion? But I will state it again this is part of the game plan. You can’t turn people to new technologies before they are ready and comfortable.
I was there for the 16-32bit switch and tell me Kelly how many years did it take them to work out the 32bit mess?
Stepping stone.
You got to spend money to make money.
Care to wager?
Is that so?
I never said it had to do with its "64-bitness", what I did say was as far as 64bit IA's go IA-64 looks to offer more.
Yaw that makes sense HP is so screwed for cash they bought Compaq, man they are certainly strapped. They dumped it because Intel's solution was better. Maybe x86 emulation module on the Itanium could get some Alpha inspiration.
Really care to show me a reputable site that shows the Itanium getting mopped up by anything from AMD. But its you Kelly so I know the x86 emulation will come up, so Ill say it now the Itanium sucks are running native x86 code.
Not a damned clue but I sure do like to debate about it.
LoL coming from a European, man you guys are sure funny. Remember our government doesn’t baby sit our free trade, so it’s a safe bet you don’t have the slightest clue how North America business works.
Hmm last I read it was a x86 replacement, when and where is unknown im not a phycic, but I am sure you are.
Is there some law I don’t know about that says you can’t spend money? Old saying still holds takes money to make money.
Got some business models and fiscal projections to support that, Otherwise we are simply both grasping at straws and man do I hate straw.
Bad timing as far as I am concerned also lack of real software/compiler/and hardware support stunted that growth.
You pointed out the obvious there, your words "MIPS and Alpha where very well entrenched in their high end markets". High end markets such as HPC do not step into the bounds of desktop users. “tried to grow into the x86 market" pretty tough to grow into a 32bit environment with a 64bit chip that can only emulate it at barely reasonable speeds. Also windows gravy market is x86 as well as 87% of all the major software developers. You don’t want to get off the gravy train before its done its course.
Got some links to back up the Intel lets other people design their products for them.
They are considering it though that’s the point.
Can’t read when, when and WHEN!!! I don’t know when that’s the point of the debate trying to pin point when Intel will bring it over to the desktop market.
Yaw ok if you say so, over that last 35 years has Intel ever designed something that did not compatibility in mind, and don’t say the Itanium because even the first one could emulate x86.
Xeon
<font color=orange>Scratch Here To Reveal Prize</font color=orange>
<font color=white>In the case of over scratching contact THG for property damage claims.</font color=white>
Ok just a few things i wanted to point out that i dont really agree with here.
I think x86-64 isnt a stepping stone for IA-64, i dont think the two are related at all really. In the future, I think x86-64 wil come to dominate, slowly leaving out x86-32 completely, and IA-64 will try to coexist, but I highly doubt that design will ever take over the desktop market, i just dont see any trends that point in that direction. I know intel has a war chest it can unleash to push tech, such as ddr2, pci-e, btx, but i dont think IA-64 is in that same league. Your talking about an entire architecture switch, I dont think even intel could spend enough to push the whole market in that direction. It would take alot more suppliers, and other manufacturers to switch the market there, and dare i say it, it would take amd switching over as well lol.
I do think x86-64 will take off quicker then IA-64 has, simply becuase of its ease of use and backward compatibility, one of the main reasons people havent run to itanium for anything other then its high end server type market where specialized apps are not uncommon.
Its very possible you could see intel slowly back off from that x86 replacement status for IA-64 and slowly swallow thier pride and accept x86-64 as the leading architecture for desktop machines in mainstream markets. Thats not to say they will give up on IA-64. They will undoubtedly push aggresively at making it work in the unix, hpc, type markets.
The step from x86-64 to IA-64 would be just as radical and difficult as a transition from x86-32 to IA-64, this middle step doesnt help that.
Also, dont forget that we may nto have seen anything from intel about x86-64 if it wasnt for amd pushing that to market. I relaly think intel wanted customers to have to go to itanium for 64bit support and had no plans to roll out x86-64 until now that they have to respond to AMD. Intel is ina balancing act right now, so itll be interesting to see how much intel plays up, or downplays, x86-64. I have a feeling they will downplay it as much as possible, while supporting it at the same time.
I think x86-64 isnt a stepping stone for IA-64, i dont think the two are related at all really. In the future, I think x86-64 wil come to dominate, slowly leaving out x86-32 completely, and IA-64 will try to coexist, but I highly doubt that design will ever take over the desktop market, i just dont see any trends that point in that direction. I know intel has a war chest it can unleash to push tech, such as ddr2, pci-e, btx, but i dont think IA-64 is in that same league. Your talking about an entire architecture switch, I dont think even intel could spend enough to push the whole market in that direction. It would take alot more suppliers, and other manufacturers to switch the market there, and dare i say it, it would take amd switching over as well lol.
I do think x86-64 will take off quicker then IA-64 has, simply becuase of its ease of use and backward compatibility, one of the main reasons people havent run to itanium for anything other then its high end server type market where specialized apps are not uncommon.
Its very possible you could see intel slowly back off from that x86 replacement status for IA-64 and slowly swallow thier pride and accept x86-64 as the leading architecture for desktop machines in mainstream markets. Thats not to say they will give up on IA-64. They will undoubtedly push aggresively at making it work in the unix, hpc, type markets.
The step from x86-64 to IA-64 would be just as radical and difficult as a transition from x86-32 to IA-64, this middle step doesnt help that.
Also, dont forget that we may nto have seen anything from intel about x86-64 if it wasnt for amd pushing that to market. I relaly think intel wanted customers to have to go to itanium for 64bit support and had no plans to roll out x86-64 until now that they have to respond to AMD. Intel is ina balancing act right now, so itll be interesting to see how much intel plays up, or downplays, x86-64. I have a feeling they will downplay it as much as possible, while supporting it at the same time.
>So why am I not seeing AMD bringing in a cool 1 billion?
As compared to Itanium <b>chips</b> ? Opteron outsold IPF by a factor 3 or 4 by chip units. Last year, with virtually no oem support yet. taking bets for this year ? my money is on somewhere between 10-1 and 25-1.
>You got to spend money to make money.
You also have to spend money to waste money.
>LoL coming from a European, man you guys are sure funny
> Remember our government doesn’t baby sit our free trade,
>so it’s a safe bet you don’t have the slightest clue how
>North America business works.
I'm sure this is very relevant for the discussion at hand. "I can't win on arguments since I don't know the first thing about big tin, but I'm american and you're not, so I know better". grow up dude.
>Really care to show me a reputable site that shows the
>Itanium getting mopped up by anything from AMD.
http://www.sap.com/benchmark
http://www.spec.org/ (specjbb and java appserver)
Not that I expect you to be able to interprete those correctly. And if you find some Itanium benchmarks on Exchange, Pro Engineer, Siebel, Navision, let me know so we can compare.
>Except that AMD64 is gaining acceptance quite a bit faster.
>Is that so?
You'd have to be blind not to see this. AMD64 is currently enjoying at least as much ISV support and momentum as IPF, while its only a year old verus 5 for IPF.
>Yaw that makes sense HP is so screwed for cash they bought
>Compaq, man they are certainly strapped. They dumped it
>because Intel's solution was better
My god you are ignorant. here are a few clues for ya:
<A HREF="http://www.hpl.hp.com/news/2001/apr-jun/itanium.html" target="_new">http://www.hpl.hp.com/news/2001/apr-jun/itanium.html</A>
<A HREF="http://www.informit.com/articles/article.asp?p=32053&seqNum=3" target="_new">http://www.informit.com/articles/article.asp?p=32053&seqNum=3</A>
HP, and no one else pioneered EPIC. And they handed it to intel for cutting costs. Intel screwed up merced, and HP took over to design McKinley. This is IPF 101.
> Maybe x86 emulation module on the Itanium could get some
>Alpha inspiration.
x86 emulation in hardware is being canned.
>Hmm last I read it was a x86 replacement
Oh you got a link for that ? I *never* read that anywhere from intel, and I assure you (unlike you) I have read quite a bit on the subject.
>Is there some law I don’t know about that says you can’t
>spend money? Old saying still holds takes money to make
>money.
So spending is good no matter what hu.. Well i'm sure using that logic the money spent on iAXP-432 micromainframe was also money well spent. Same for the gazillions invested in 80186, 960, tejas, ...
You seem to think everything intel touches turns into gold, but the harsh reality is they never made a buck on ANY of their products except x86 cpu's. Not on attempts to create risc cpu's (which all flopped miserably), not on networking cards, motherboards, sometimes a bit on flash, but more often than not even that generated (significant) losses. Read intels financial statement, it might open your eyes. x86 pays the bills, and the rest burns cash.
>Got some business models and fiscal projections to support
>that,
Yes, I did some analysis based on the numbers I had, and I already told you where I ended up. Intel needs ~50% of the unix market in cpu shipments (500k units) at their current ASPs, and only then they might break even by 2015. If you PM me your email, I might even dig up the excel sheet for ya.
>Bad timing as far as I am concerned also lack of real
>software/compiler/and hardware support stunted that growth.
Timing was perfect at the time the world moved from 16 to 32 bit OS's and x86 was trailing Risc performance by a significant margin. There could not have been a better time. "lack of real software/compiler/and hardware support" is also such a useless empty claim I will only respond by saying it was far better than what IPF currently enjoy for the workstation let alone, desktop market. And it was not only better for the high end unix market, it still is very significantly better. But what would you know , you've never even heard of Tru64.
>Got some links to back up the Intel lets other people
>design their products for them.
See above. not knowing basic stuff like this shows your colours. Why the hell do you even want to argue this, you know as much about it as about cold fusion.
>Can’t read when, when and WHEN!!! I don’t know when that’s
>the point of the debate trying to pin point when Intel will
>bring it over to the desktop market.
if, if and IF. Because the answer is more likely: never. Chances are way better 10-20 years from here we will run on Power derived hardware than IPF.
>Yaw ok if you say so, over that last 35 years has Intel
>ever designed something that did not compatibility in mind,
i432 -> flopped monumentally
i960 -> flopped miserably (well ended up in laserprinters LOL)
IPF -> may become a nice printer processor in the next decade or maybe DSP style chip for in routers and switches .
= The views stated herein are my personal views, and not necessarily the views of my wife. =
As compared to Itanium <b>chips</b> ? Opteron outsold IPF by a factor 3 or 4 by chip units. Last year, with virtually no oem support yet. taking bets for this year ? my money is on somewhere between 10-1 and 25-1.
>You got to spend money to make money.
You also have to spend money to waste money.
>LoL coming from a European, man you guys are sure funny
> Remember our government doesn’t baby sit our free trade,
>so it’s a safe bet you don’t have the slightest clue how
>North America business works.
I'm sure this is very relevant for the discussion at hand. "I can't win on arguments since I don't know the first thing about big tin, but I'm american and you're not, so I know better". grow up dude.
>Really care to show me a reputable site that shows the
>Itanium getting mopped up by anything from AMD.
http://www.sap.com/benchmark
http://www.spec.org/ (specjbb and java appserver)
Not that I expect you to be able to interprete those correctly. And if you find some Itanium benchmarks on Exchange, Pro Engineer, Siebel, Navision, let me know so we can compare.
>Except that AMD64 is gaining acceptance quite a bit faster.
>Is that so?
You'd have to be blind not to see this. AMD64 is currently enjoying at least as much ISV support and momentum as IPF, while its only a year old verus 5 for IPF.
>Yaw that makes sense HP is so screwed for cash they bought
>Compaq, man they are certainly strapped. They dumped it
>because Intel's solution was better
My god you are ignorant. here are a few clues for ya:
<A HREF="http://www.hpl.hp.com/news/2001/apr-jun/itanium.html" target="_new">http://www.hpl.hp.com/news/2001/apr-jun/itanium.html</A>
<A HREF="http://www.informit.com/articles/article.asp?p=32053&seqNum=3" target="_new">http://www.informit.com/articles/article.asp?p=32053&seqNum=3</A>
HP, and no one else pioneered EPIC. And they handed it to intel for cutting costs. Intel screwed up merced, and HP took over to design McKinley. This is IPF 101.
> Maybe x86 emulation module on the Itanium could get some
>Alpha inspiration.
x86 emulation in hardware is being canned.
>Hmm last I read it was a x86 replacement
Oh you got a link for that ? I *never* read that anywhere from intel, and I assure you (unlike you) I have read quite a bit on the subject.
>Is there some law I don’t know about that says you can’t
>spend money? Old saying still holds takes money to make
>money.
So spending is good no matter what hu.. Well i'm sure using that logic the money spent on iAXP-432 micromainframe was also money well spent. Same for the gazillions invested in 80186, 960, tejas, ...
You seem to think everything intel touches turns into gold, but the harsh reality is they never made a buck on ANY of their products except x86 cpu's. Not on attempts to create risc cpu's (which all flopped miserably), not on networking cards, motherboards, sometimes a bit on flash, but more often than not even that generated (significant) losses. Read intels financial statement, it might open your eyes. x86 pays the bills, and the rest burns cash.
>Got some business models and fiscal projections to support
>that,
Yes, I did some analysis based on the numbers I had, and I already told you where I ended up. Intel needs ~50% of the unix market in cpu shipments (500k units) at their current ASPs, and only then they might break even by 2015. If you PM me your email, I might even dig up the excel sheet for ya.
>Bad timing as far as I am concerned also lack of real
>software/compiler/and hardware support stunted that growth.
Timing was perfect at the time the world moved from 16 to 32 bit OS's and x86 was trailing Risc performance by a significant margin. There could not have been a better time. "lack of real software/compiler/and hardware support" is also such a useless empty claim I will only respond by saying it was far better than what IPF currently enjoy for the workstation let alone, desktop market. And it was not only better for the high end unix market, it still is very significantly better. But what would you know , you've never even heard of Tru64.
>Got some links to back up the Intel lets other people
>design their products for them.
See above. not knowing basic stuff like this shows your colours. Why the hell do you even want to argue this, you know as much about it as about cold fusion.
>Can’t read when, when and WHEN!!! I don’t know when that’s
>the point of the debate trying to pin point when Intel will
>bring it over to the desktop market.
if, if and IF. Because the answer is more likely: never. Chances are way better 10-20 years from here we will run on Power derived hardware than IPF.
>Yaw ok if you say so, over that last 35 years has Intel
>ever designed something that did not compatibility in mind,
i432 -> flopped monumentally
i960 -> flopped miserably (well ended up in laserprinters LOL)
IPF -> may become a nice printer processor in the next decade or maybe DSP style chip for in routers and switches .
= The views stated herein are my personal views, and not necessarily the views of my wife. =
Kelledin
Distinguished
- Mar 1, 2001
- 2,183
- 0
- 19,780
AMD doesn't have to bring in "a cool 1 billion" to make life hard for IA64.
True enough. AMD capitalized on that.
Depends on what you mean by "work out the 32bit mess."
Developers were diving head-first into "the 32bit mess" before there was even a 32-bit operating system. It didn't take much for software to wrest control of the processor from DOS and do its own thing with 32-bit registers and "flat mode."
More like comfortable resting place. Especially as in many cases, there's no real benefit to stepping "up" to IA64.
And the money is being spent on extending x86 more than on refining IA64. Itanium just doesn't make enough fast enough for a profit-driven beast to justify focusing on it.
Gladly.
Looks tend to be deceiving.
I didn't say "mopped up." I said "matched."
When AMD64 costs less, runs cooler, and supports much more software, it doesn't have to do more than match IA64 in benchmarks.
At this point, it's a lot more "if" than "when."
<i>Look! Up in the sky! It's a bird! It's a plane! It's...
...an asthmatic werehamster?
<LHGPooBaa> Well, @#!& on me.</i>
YOU:
Point is intel can't continue to waste money on IPF forever (even if it can afford it) if it can not demonstrate a likely future ROI to its shareholders. Shareholders rarely have an horizon that exceeds 3-5 years
Reply:
So have you sold your AMD stock.You get my point many corp dont make money still they are there.AMD ATI most ram\mobo manufacturer.
YOU:
Yes, at the current rate it will be where Alpha was 5 years ago by the end of the decade. Very bright future.. You still don't seem to get it. IPF is sold in a merchant model, intel doesnt sell systems.
reply:
Intel does sell itanium 4\2 way to OEM and they add they own service sold application compiler and programming tool.
You say IBM make money cuz they have high volume on Power in fact IBM microelectronic devision have loss 200 million again.Also you say EMT64 is incompatible with IA64.True but for how long
i need to change useur name.
Point is intel can't continue to waste money on IPF forever (even if it can afford it) if it can not demonstrate a likely future ROI to its shareholders. Shareholders rarely have an horizon that exceeds 3-5 years
Reply:
So have you sold your AMD stock.You get my point many corp dont make money still they are there.AMD ATI most ram\mobo manufacturer.
YOU:
Yes, at the current rate it will be where Alpha was 5 years ago by the end of the decade. Very bright future.. You still don't seem to get it. IPF is sold in a merchant model, intel doesnt sell systems.
reply:
Intel does sell itanium 4\2 way to OEM and they add they own service sold application compiler and programming tool.
You say IBM make money cuz they have high volume on Power in fact IBM microelectronic devision have loss 200 million again.Also you say EMT64 is incompatible with IA64.True but for how long
i need to change useur name.
wait, are you trying to say that EMT64 is a transition point to IA64? i just dont see how, since its in line with AMD64. EMT64 wasnt soemthing intel wanted to see used i have a feeling. This is thier answer to AMD64, to keep the marketing machine well oiled. They had to answer back with thier own solution.
I doubt they really want to make any plans around EMT64 and just want to force consumers to change over to IA64. Either way, the change over, if it ever comes, wont be as easy as that, itll be a tough time, pulling teeth to drag everyone in line.
Boy did you miss the boat. Let's leave American tax laws out of this.
"You say IBM make money cuz they have high volume on Power in fact IBM microelectronic devision have loss 200 million again."
So the Itanium like part of big blue lost money due to onshore production costs. See that division is responsible for making electronics.
On the other hand the "offshore" service division made huge profits, which the mother company could in part write down and in part discount due to the losses of the electronics devision.
You see, the profit is in service, not production. Intel is a production company, not a service company. They dont have a massively profitable service devision that requires a loss from the production devision, so they can write it off.
"You say IBM make money cuz they have high volume on Power in fact IBM microelectronic devision have loss 200 million again."
So the Itanium like part of big blue lost money due to onshore production costs. See that division is responsible for making electronics.
On the other hand the "offshore" service division made huge profits, which the mother company could in part write down and in part discount due to the losses of the electronics devision.
You see, the profit is in service, not production. Intel is a production company, not a service company. They dont have a massively profitable service devision that requires a loss from the production devision, so they can write it off.
Intel seems to be a mess right now. They need to get Itanium onto the broader workstation/high end market to start regaining invested money at a reasonable rate.
Yet they need to push out Pentium/Xeon with iAMD-64 soon to not loose marketshare to AMD, BUT such a move would hurt Itanium sales.
But then they probably cant be too picky right now AMD has taken the performance and technology crown away and most of Intel seems to be in dissaray, nobody knows what happens next as their roadmaps change by the day and thanks to the sucess of the isreali development division the americans look like a bunch of retards because of that leaky transistor mess they call prescott.
What happens next is damage control (my guess is they will start by killing further Itanium development if not already done so).
Yet they need to push out Pentium/Xeon with iAMD-64 soon to not loose marketshare to AMD, BUT such a move would hurt Itanium sales.
But then they probably cant be too picky right now AMD has taken the performance and technology crown away and most of Intel seems to be in dissaray, nobody knows what happens next as their roadmaps change by the day and thanks to the sucess of the isreali development division the americans look like a bunch of retards because of that leaky transistor mess they call prescott.
What happens next is damage control (my guess is they will start by killing further Itanium development if not already done so).
>So have you sold your AMD stock
Yes.
>You get my point many corp dont make money still they are
>there
So what ? They strive to make money. If a product no longer has the potential to make the company money or at least offer some benefit to it that outweighs its cost, it gets canned.
>Intel does sell itanium 4\2 way to OEM
Yes, In the same way AMD sells chipsets. Not because they want to, or make money on it, nor because it fits the business model, but as an enabler. Those Tiger systems also don't provide anywhere near the kindo or level of revenue I was referring to, there are just barebone boxes that help sell chips; no software, no expensive disksubsytems, no memory thats resold at 2x the marketprice, no lucrative 15%service contract, OS licences, .. Tiger systems probably have the same gross margin as the cpu's in them, they add nothing to Intel's bottom line, very much unlike Integrity, altix, superdome, Pseries,..
>You say IBM make money cuz they have high volume on Power
>in fact IBM microelectronic devision have loss 200 million
>again
Exactly. But that loss enabled IBM to sell $3.8 billion in high margin server hardware, $3.5B in software, and a large chunk of the $11.1B of global services. Quarterly numbers btw. Of course, like any company, IBM strives to make its BUs profitable, but developping Power chips makes a lot of sense even if this generates losses. Its an enabler to make it sell highly profitable products in excess of $10B/Q. Very much unlike intel.
>EMT64 is incompatible with IA64.True but for how long
Yeah, maybe Itanium will migrate to iAMD64 indeed 😛
= The views stated herein are my personal views, and not necessarily the views of my wife. =
Yes.
>You get my point many corp dont make money still they are
>there
So what ? They strive to make money. If a product no longer has the potential to make the company money or at least offer some benefit to it that outweighs its cost, it gets canned.
>Intel does sell itanium 4\2 way to OEM
Yes, In the same way AMD sells chipsets. Not because they want to, or make money on it, nor because it fits the business model, but as an enabler. Those Tiger systems also don't provide anywhere near the kindo or level of revenue I was referring to, there are just barebone boxes that help sell chips; no software, no expensive disksubsytems, no memory thats resold at 2x the marketprice, no lucrative 15%service contract, OS licences, .. Tiger systems probably have the same gross margin as the cpu's in them, they add nothing to Intel's bottom line, very much unlike Integrity, altix, superdome, Pseries,..
>You say IBM make money cuz they have high volume on Power
>in fact IBM microelectronic devision have loss 200 million
>again
Exactly. But that loss enabled IBM to sell $3.8 billion in high margin server hardware, $3.5B in software, and a large chunk of the $11.1B of global services. Quarterly numbers btw. Of course, like any company, IBM strives to make its BUs profitable, but developping Power chips makes a lot of sense even if this generates losses. Its an enabler to make it sell highly profitable products in excess of $10B/Q. Very much unlike intel.
>EMT64 is incompatible with IA64.True but for how long
Yeah, maybe Itanium will migrate to iAMD64 indeed 😛
= The views stated herein are my personal views, and not necessarily the views of my wife. =
P4man,
Opteron and Xeon64 are in our labs being evaluated, so no support announced yet. Unfortunatly we can't do much with Xeon64 bit yet until MS release the 64bit patch. But there have been some impressive interal benchmarks with Opteron on Linux. Watch this space
Opteron and Xeon64 are in our labs being evaluated, so no support announced yet. Unfortunatly we can't do much with Xeon64 bit yet until MS release the 64bit patch. But there have been some impressive interal benchmarks with Opteron on Linux. Watch this space
> Unfortunatly we can't do much with Xeon64 bit yet until MS
>release the 64bit patch
You mean you can not unstall the 64 bit beta (for opteron) on the Xeon 64 ?
> But there have been some impressive interal benchmarks
>with Opteron on Linux.
I see no reason you wouldn't be able to run the same (64 bit) Linux distro's on the Xeon64 either ? Are you saying they are "incompatible" (or maybe the OS installer just doesn't properly recognizes the chip ?)
= The views stated herein are my personal views, and not necessarily the views of my wife. =
>release the 64bit patch
You mean you can not unstall the 64 bit beta (for opteron) on the Xeon 64 ?
> But there have been some impressive interal benchmarks
>with Opteron on Linux.
I see no reason you wouldn't be able to run the same (64 bit) Linux distro's on the Xeon64 either ? Are you saying they are "incompatible" (or maybe the OS installer just doesn't properly recognizes the chip ?)
= The views stated herein are my personal views, and not necessarily the views of my wife. =
Kelledin
Distinguished
- Mar 1, 2001
- 2,183
- 0
- 19,780
I wouldn't presume to know Stimpy's situation, but I can speculate on Xeon64 readiness in general...
Some Xeon64-compatible distros are trickling out, though--Fedora ought to be working on Xeon64, as RedHat's already documenting some DMA quirkiness about it.
The "rumble in the jungle" seems to be universally agreed that Xeon64 is going to be a poor performer though. MSFT hints at that without saying anything definite, RedHat documents an ugly performance hit with the DMA quirk, and there's general rumors that the Prescott ALU drops to a little above half-speed to do 64-bit. 😱 Take all that for what it's worth.
<i>Look! Up in the sky! It's a bird! It's a plane! It's...
...an asthmatic werehamster?
<LHGPooBaa> Well, @#!& on me.</i>
IIRC the 64-bit beta WinXP got to working on Xeon64 only around build 1100 or so. MS probably needed to do a little more to cope with Xeon64's quirks.
Probably the same situation with Linux distros as with Win64. Sure there are some out for AMD64, but bootstrapping out of the chicken-and-egg scenario onto Xeon64 isn't an easy thing for end-users. It probably requires booting into 32-bit mode, setting up a cross-compiler that can target AMD64, and then cross-compiling a more recent kernel that can grok Xeon64.
Some Xeon64-compatible distros are trickling out, though--Fedora ought to be working on Xeon64, as RedHat's already documenting some DMA quirkiness about it.
The "rumble in the jungle" seems to be universally agreed that Xeon64 is going to be a poor performer though. MSFT hints at that without saying anything definite, RedHat documents an ugly performance hit with the DMA quirk, and there's general rumors that the Prescott ALU drops to a little above half-speed to do 64-bit. 😱 Take all that for what it's worth.
<i>Look! Up in the sky! It's a bird! It's a plane! It's...
...an asthmatic werehamster?
<LHGPooBaa> Well, @#!& on me.</i>
>The "rumble in the jungle" seems to be universally agreed
>that Xeon64 is going to be a poor performer though
The only thing that I (seem to remember I)'ve heard by people who's opinion/knowledge I trust, is Xeon64 might take a hit in some scenario's where more than 4 GB Ram is involved, but that wouldn't be anything major.
>there's general rumors that the Prescott ALU drops to a
>little above half-speed to do 64-bit.
I heard those, but haven't seen or heard any credible info on it. It seems to have been based on some misunderstandings, and you know how it goes with rumours like that.. for now, I wouldn't put much faith in it, but I guess we will find out soon enough.
= The views stated herein are my personal views, and not necessarily the views of my wife. =
>that Xeon64 is going to be a poor performer though
The only thing that I (seem to remember I)'ve heard by people who's opinion/knowledge I trust, is Xeon64 might take a hit in some scenario's where more than 4 GB Ram is involved, but that wouldn't be anything major.
>there's general rumors that the Prescott ALU drops to a
>little above half-speed to do 64-bit.
I heard those, but haven't seen or heard any credible info on it. It seems to have been based on some misunderstandings, and you know how it goes with rumours like that.. for now, I wouldn't put much faith in it, but I guess we will find out soon enough.
= The views stated herein are my personal views, and not necessarily the views of my wife. =
TRENDING THREADS
-
-
-
-
Question When turning PC on , randomly showing sth like " Windows boot up manager "- Started by Kapral
- Replies: 7
-
Question Games randomly crashing without performance issues and with okay temperature- Started by Koppo
- Replies: 2
-
Discussion What's your favourite video game you've been playing?- Started by amdfangirl
- Replies: 4K
Latest posts
-
Question Can I get something to improve internet speed with WiFi ?- Latest: SkyNetRising
-
-
-
-
Facebook
Twitter
Instagram