Talk:Itanium/Archive 2
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | Archive 2 |
Automate archiving?
Does anyone object to me setting up automatic archiving for this page using MizaBot? Unless otherwise agreed, I would set it to archive threads that have been inactive for 60 days.--Oneiros (talk) 16:00, 24 December 2009 (UTC)
Done The bot should start in the next 24 hours.--Oneiros (talk) 23:02, 26 December 2009 (UTC)
Market reception
Edited and added current content showing more balanced view of the Itanium market reception. As previously written, the article focused more on the negative side, and did not present a fair case for each view as is stipulated in WP:NPOV. Removed the large quote marks, but kept the quote by Kruth. The large quote marks gave more more weight and not the proper balance to the negative side. Updated references with 2009 and 2010 citations. ITtrendhunter (talk) 15:52, 30 March 2010 (UTC)
- 2 problems - 1) copyright violation in your contribution, 2) the tone does not fit with the rest of the article. Raysonho (talk) 16:04, 30 March 2010 (UTC)
- NPOV does not require "equal weight." It does require fairness. But Negative views of Itanium are massively more common than positive views in the trade press and in the statistics. I agree with you about the big quote marks. The comment that "80% of major corporations use Itanium" is egregiously missleading: compare to "90% of all heroin addicts drink water." A major corporation is so big that chances are, soemwhere in the corporation, there is at least one computer of every type that has ever been built in the last 20 years. Therefore, this is a meaningless statistic. I strongly suspect that there are zero major corporations that have more itaniums than X86 machines, and (with the possible exception of HP) zero that have more Itaniums than Power or SPARC. This is true even when "more" is measures in aggregate computing power, and even when we remove Windows X86. With the release of the latest 8-core Xeon, Itanium is not even Intel's most powerful CPU. Arch dude (talk) 17:24, 30 March 2010 (UTC)
Coining Itanic
Mike Magee still worked for The Register back when the pun was coined and he later claimed to have been its original author. However, http://www.theregister.co.uk/1999/11/02/the_quick_guide_to_register/ claims:
Itanic Our name for Intel's IA-64 Itanium, formerly called Merced, which was kindly contributed to us by a reader.
and this article claims that it was first used in unnamed Usenet sources.JeR (talk) 00:18, 2 April 2010 (UTC)
- The earliest Usenet sighting I can find in a quick and dirty search is this 2001 one on comp.lang.java.advocacy.--NapoliRoma (talk) 02:26, 2 April 2010 (UTC)
Deleted uncited sentence in one section that is repeated elsewhere in the article and has sources
I deleted the sentence about Itanium being called Itanic in the Market Reception section as it is unsourced there and because the information is included elsewhere in the article with references: (Mentioned and cited in the History -- Development 1989-2000 section, fifth paragraph:
Left the Itanic in this History section (put in Italic for explantory purposes):
Intel announced the official name of the processor Itanium, on October 4, 1999.[20] Within hours, the name Itanic [21] had been coined on a Usenet newsgroup, a reference to Titanic, the "unsinkable" ocean liner which sank in 1912. Itanic has since often been used by The Register,[22] Scott McNealy,[23] and others,[24][25] implying that the multibillion dollar investment in Itanium—and the tremendous early hype—would be followed by its relatively quick demise.
Deleted this sentence from the Market Reception section as it is unreferenced and is repetitive to the history section that has citations (references.)-- deleted this sentence (put in Italic here) that is not sourced
Its sales were so disastrously below expectations that the appellation of "Itanic" has been applied to the franchise, invoking the ill-fated ocean liner RMS Titanic.[citation needed] JLRedperson (talk) 23:42, 3 April 2010 (UTC)
Lead and market reception changes - move towards NPOV
I've corrected what I saw as some major flaws in the article:
- I removed the notion that MS dropped Windows support because Itanium is dead (not quite what the article says but clearly what it basically meant). In fact it did so because, although Itanium does have some server customers, they almost exclusive run HP-UX.
- I separated the market reception section into high-end server and other markets. It's true that Itanium is dead as far as PCs are concerned, now that x86-64 is so successful, but this dominated the market reception section when really the server story is more important. It's now much more obvious how lacking the market reception section is with regard to servers.
- I pointed out that at the end of last year Itanium was (only just) making a profit.
- I removed the quote from the nut-job who claimed that Itanium caused the collapse of the dot-com bubble. Seriously, WTF!?!
There is still plenty wrong. Mostly more references are needed, and also the end of support for Windows needs to be added to the history section, but I don't have time to fix these up too. Quietbritishjim (talk) 11:50, 7 April 2010 (UTC)
- Though you may not agree with the John C. Dvorak article you did not give a reason for why you removed the opinion of a well known PC journalist. Also though the Itanium business started making a profit in 2009 it is quite possible that the chip itself will never make a profit for Intel. For example if they spent $30 billion on development/marketing but only get $1 billion from selling the chip they would lose $29 billion overall. That Intel started selling Itanium chips for a profit in 2009 doesn't mean much especially when you consider that before 2009 the Itanium chips were not being sold for a profit. --GrandDrake (talk) 05:31, 9 April 2010 (UTC)
- Your comment is really two separate things: (a) I didn't say why I removed John Dvorak's comment and (b) although you didn't say it explicitly, your comments about profitability suggest that I think that Itanium is a success and I'm trying to move the article towards saying that. So I'm going to respond to these separately (in reverse order).
- (b): Absolutely not! I think the Itanium has been a huge flop. It was meant to be the future replacement for x86 everywhere, and that's clearly not the case (although that's technically speculation on my part, so I'm against saying it that explicitly in the article). However – and I'm not even saying this is currently true – it's still possible for it to become a successful niche high-end computing chip for a single manufacturer, just as SPARC is for Sun. This is why I separated market reception out into two subsections. In fact the main reason I edited this article in the first place was to remove the silly claim that Microsoft dropping support for Itanium meant that it was dead for even high-end servers, which is not true (at least not for that reason).
- (a): I've never heard of this guy, but looking at his article he does seem to be quite well know (what a pity, given what he said in that article). However I was quite clear about why I removed his quote: not because he was not well known, but because he claimed that Itanium triggered the burst of the dot-com bubble. To repeat what I said above, seriously, WTF!?! If that were the case then why doesn't the dot-com bubble article mention it? Why aren't there supporting quotes from economists and other industry spectators? Because it's garbage; he was trying to write an article that would be exciting and get readers, whereas a factually accurate one would have been still negative for Intel, but basically rather dull. Anyone with half a brain can see that this article is nonsense.
- When I say "NPOV", I assure you, I'm not using this as a mask to push my POV. I'm all about contrasting viewpoints, and if most people think that Itanium was a flop, I'm happy for most quotes to say that Itanium is flop. But can they please be from people that actually care about the truth, not readership numbers? The David House quote referenced below seems much more relevant, perhaps we could swap it with that? Quietbritishjim (talk) 13:37, 9 April 2010 (UTC)
- The Dvorak quote is certainly his opinion on the Itanium but I do think that it is worth mentioning. As for whether the Itanium could remain around as a high end chip I have read that the new Xeon 7500 series contains over 20 RAS (reliability, availability, and serviceability) features including some from the Itanium. I have read several articles which say that this indicates that Intel is planning to drop support for the Itanium. --GrandDrake (talk) 21:47, 10 April 2010 (UTC)
- You've completely ignored my point. My point is that the article is clearly garbage designed to get hits. Answer honestly: do you think that Itanium caused the dot com bubble collapse? Even if you do believe this, I can't imagine you could find any reliable source to back it up. NPOV does not mean quoting nonsense from any source that sounds good without regard for its reliability. I'm restraining myself from removing the quote again to avoid an edit war, but I really think that that article is not a reliable source. Quietbritishjim (talk) 22:28, 12 April 2010 (UTC)
- The quote is "This continues to be one of the great fiascos of the last 50 years". The quote itself does not say that Itanium caused the dot.bomb. Whether or not the article is valid as a whole, the quote accurately reflects the consensus of the industry analysts, and Dvorak is a highly visible and respected industry analyst. If you can find a better quote, from a "better" article, to convey this, then please replace the quote. -Arch dude (talk) 23:51, 12 April 2010 (UTC)
- I just re-read the article. It does not state that Itamium was responsible for dot.bomb. It states that Itanium was responsible for the decline of "the computer industry." We all know that dot.bomb occurred concurrently with the problems in the computer industry, but the specific problems quoted by Dvorak are much more narrowly focused. Bacically, and according to the article, Intel's pursuit of Itanium had severe negative effects on the industry, most of which happened prior to dot.bomb. So after carefully re-reading the article with your objections in mind, I cannot reach the conclusion you assert. like Dvorak, I remember hearing thge hype as it happened. I remember the Intel de-emphasized the Pentium and moved its best an brightest developers to Itanium. I remember the trade shows where Intel never even mentioned Pentium improvements and spent the whole time hyping Itanium. I remember the promises that Pentium was basically in maintanence mode and Itanium would supercede it in all sectors. If Intel had emphasized Pentium instead of Itanium, We would have had an amd64 architecture in 2000, or maybe even in 1998. In this industry, six years is forever. Dvorak's article is almost exclusively about this, and not about the much larger dot.bomb problem. Dot.bomb was mostly about ridiculous web-based software, not about the computer hardware industry. -Arch dude (talk) 00:19, 13 April 2010 (UTC)
- The quote is "This continues to be one of the great fiascos of the last 50 years". The quote itself does not say that Itanium caused the dot.bomb. Whether or not the article is valid as a whole, the quote accurately reflects the consensus of the industry analysts, and Dvorak is a highly visible and respected industry analyst. If you can find a better quote, from a "better" article, to convey this, then please replace the quote. -Arch dude (talk) 23:51, 12 April 2010 (UTC)
- You've completely ignored my point. My point is that the article is clearly garbage designed to get hits. Answer honestly: do you think that Itanium caused the dot com bubble collapse? Even if you do believe this, I can't imagine you could find any reliable source to back it up. NPOV does not mean quoting nonsense from any source that sounds good without regard for its reliability. I'm restraining myself from removing the quote again to avoid an edit war, but I really think that that article is not a reliable source. Quietbritishjim (talk) 22:28, 12 April 2010 (UTC)
- I still read it differently from you, but clearly it isn't as clear cut as I thought. Given your argument I drop my objection to the quote being in the article. Thanks for taking the time to think about what I said and post such a reasoned reply. Quietbritishjim (talk) 21:20, 14 April 2010 (UTC)
- And thank you both for your initial insights and for your courteous discussion. There is still a great deal to be done on this article. If Itanium dies precipitously, we will need to be especially watchful to maintain a NPOV, and I hope you will be able to help. -Arch dude (talk) 22:33, 14 April 2010 (UTC)
- I still read it differently from you, but clearly it isn't as clear cut as I thought. Given your argument I drop my objection to the quote being in the article. Thanks for taking the time to think about what I said and post such a reasoned reply. Quietbritishjim (talk) 21:20, 14 April 2010 (UTC)
Parking lot for House quote
I removed the quote from David House from the lede, as it didn't really belong there -- this should be a summary of the article rather than introducing backup quotes like this. It was also uncited, but here it is with a ref I found, in case someone wants to find the right place to include it:
This led former Intel official David House to predict "This will end up being one of the world's worst investments, I'm afraid."
<ref>{{cite web |url = http://performancetrap.org/2009/12/17/intel-atom-and-itanium/ |title = Atom and Itanium: A Tale of Two Processors (or how Intel learned to avoid the performance trap) |accessdate = 2010-04-07 |author = David Wesley |date = December 17, 2009 |work = Avoiding the Performance Trap }}</ref>
--NapoliRoma (talk) 14:01, 7 April 2010 (UTC)
Ridiculous statistic quoted. Please discuss
I removed the following:
- , but does claim that more than 80 percent of the Global 100 corporations have chosen Itanium-based servers.[1]
This statement is tecnically true, but it is a grave disservice to the readers of this encyclopedia, because it is horribly miss-leading The statement tries to imply that Itanium has an 80% market share but this is completely wrong. Consider the follwing statement:
- "80 percent of the Global 100 corporations employ child abusers."
Clearly, any company with 20,000 or more employees will have hired at least two child abusers.
The List of companies by revenue shows that all of the top 100 companies have more than 20,000 employees, and that alt least 80% of them have more than 100,000. A trivial analysis shows that each has many major datacenters, each with many computers, plus dozens of smaller centers and research labs. The compnies with fewer employees are in the financial and energy sectors and tend to have a lot more computers per employee. I find it astonishing that only 80 of these companies have purchased at least one Itanium in the last 9 years. I would have thought that any company this large wold have at least one computer of each type that has been manufactured in the last 15 years!
The true NPOV statistic is that Itanium is in fourth place, after Xeon, PPC, and SPARC, in terms of deployed "UNIX-like" servers. It is in fifth place if we include "mainframes", because it falls behind IBM z-series.
Note also that Business wire (the reference for this statement) is not a secondary source: it is basically a press release. Yes, the statement was specifically attributed to Intel, but this is not in my opinion sufficient to mitigate the POV. Please discuss before re-adding this statement. -Arch dude (talk) 23:38, 14 April 2010 (UTC)
- Hmmm. In my experience, most press releases and statistics about market share are similarly flimsy. What is marketshare? Units sold per quarter? Installed base? Revenue per quarter? For the Xeon systems, it is number of units sold, since they are inexpensive systems that sell a high volume. For the z-Series, it is revenue, since they sell for millions each, but don't sell very many. It's also a little misleading to try to compare Itanium to the other processors by installed base, since HP sold both Itanium and PA-RISC until the end of last year, and Sun & IBM have been selling their CPUs for 24 & 19 years respectively.
- I completely agree with removing that stat, by the way. —Joshua Scott (LiberalFascist) 05:37, 15 April 2010 (UTC)
- BTW, using Intel's logic, it should also be pointed out that "100 percent of the Global 100 corporations have chosen x86-based computers". Raysonho (talk) 05:49, 15 April 2010 (UTC)
- Agreed. Vague marketing figures are meaningless, and useless for an encyclopaedia. HumphreyW (talk) 07:51, 15 April 2010 (UTC)
Migration from PA-RISC
On the HP-UX page, it mentions that HP-UX runs on PA-RISC and Itanium processors. As PA-RISC doesn't seem to be sold anymore, people may naturally want to upgrade to Itanium from PA-RISC. Therefore, can we see some comment of how compatible an Itanium processor is with the PA-RISC instruction set, even if "not at all" is the answer? It seems to me that people will end up at this page or the HP-UX page looking for this information, and neither page seems to have it. I think this page may be the best place, as the HP-UX page seems older/less updated. Can someone knowledgable provide that info? Thanks. Jgwinner (talk) 17:54, 29 October 2010 (UTC)
- I added a sentence to the "software support" section. However, this really belongs in the HP-UX article, where it is only mentioned in passing in, in the timeline. -Arch dude (talk) 15:26, 30 October 2010 (UTC)
- PA-RISC assembly set and coding practice and Itanium assembly ones are completely / totally different. Migration can be achieved at the source level in case when code is written at least in C or any other high level programming language. This requires at least recompilation and perhaps some code tweaking. Of course assembly language parts need to be completely rewritten as well as all systems code dealing with inner details of cpu and memory control. I believe porting process can be to some extend automated, that is HP can provide developers with porting tools.
—Preceding unsigned comment added by 83.6.111.170 (talk) 21:27, 31 March 2011 (UTC)
- A gentle reminder: this talk page is supposed to be about how to improve the article. It is not a general forum about Itanium. I took the above comment as an implicit request to address this issue in the article. I added a sentence that clarifies (I hope) that HP provided an instruction set emulator that executes PA-RISC binary files. -Arch dude (talk) 23:04, 31 March 2011 (UTC)
Visual Studio support
"However, that does not mean that [Microsoft] won't continue supporting its [Itanium] compilers in the Visual Studio kits."
No, but this would seem to mean that:
Visual Studio Dropping Support for Itanium
--NapoliRoma (talk) 00:39, 18 January 2011 (UTC)
...actually, just noticed that the Microsoft posting that's the source for the above is already a reference in the article. It says specifically that not just Windows Server, but any new releases of Visual Studio and SQL Server will not be supported on Itanium. Regards, NapoliRoma (talk) 19:28, 18 January 2011 (UTC)
- Good point; I've deleted the above assertion about VS while condensing the over-long lead. Letdorf (talk) 22:02, 19 January 2011 (UTC).
IDC
The article mentions "IDC" as a source sometimes. But what is it? IDC has quite some versions to offer. — Preceding unsigned comment added by 87.155.90.82 (talk) 20:52, 16 June 2011 (UTC)
- Linked first occurance.-Arch dude (talk) 21:23, 16 June 2011 (UTC)
Though its speed would have been impressive had it been introduced on time in 1999, it ran only half as fast as the contemporary x86-based Pentium 4
The statement in the article summary references some report of Wall Street Journal which claims "What would have been a speedy processor if introduced on time in 1999 now will run only half as fast as Intel's next version of the Pentium 4, which still uses the x86 architecture". However, there is no evidence or proof or test described in the article which could confirm the claim; it is just an unconfirmed claim. I'm not trying to put Wall Street Journal honesty at question; but there is an Itanium special feature: AFAIK first Itanium versions contained slow hardware x86 CPUs as their x86 compatibility layer (so that x86 binaries could be run on Itanium, although very slow). It is possible that Wall Street Journal by this claim cited some test or review which measured x86 performance, and thus the claim has nothing to do with Itanium performance but only describes performance of hardware x86 chip (included solely for compatibility). Also, while i'm not an Itanium expert, i find it to be extremely doubtful for Itanium with clock speed of 733-800MHz and high IPC to be 2x slower than Pentium 4 Wilamette with clock speed of 1300-2000MHz. I believe that this questionable claim should either be removed from the Wikipedia article or be backed by some other more reliable (in terms of IT) reference. Penartur (talk) 01:38, 26 March 2011 (UTC)
- When running in x86 compatability mode, Itanium executed x86 instructions at the same rate as a 100Mhz Pentium, i.e. horribly slowly. The WSJ article's number is much closer to reality: Itanium was not dramatically superior to Pentium at the time, which it the point that WSJ was making. If you can find a better source (i.e., one with a benchmark, please add it. -Arch dude (talk) 05:40, 26 March 2011 (UTC)
- I thought that hardware compatibility unit was clocked like 500MHz? Penartur (talk) 08:10, 27 March 2011 (UTC)
- Maybe the WSJ article was just comparing clock speeds? 733-800 MHz is approximately half of 1300-2000. Letdorf (talk) 20:05, 26 March 2011 (UTC).
- I believe that this is the case. The press, even the technical press, treated clock speed as the most important figure of merit at that time and all the way up until the end of the Netburst era. The WSJ statement is even valid in this regard: a 800Mhz Itanium would have been very impressive compared to an 800Mhz Pentium. But the 2Ghz Pentium/Xeon had better integer performance than the 800Mhz Itanium, for much less than half the price. This is the biggest reason Itanium failed to displace Pentium (which was Intel's original stated goal.) The WSJ quote is therefore a good, concise, verifiable (WP:V) way to capture what happened. From our post-Netburst perspective, we know that SPEC-INT and SPEC-INT/$ are better measures, but that is far too complicated for the lede, and the WSJ quote is really does capture the perceptins of the time. -Arch dude (talk) 20:30, 26 March 2011 (UTC)
- Then still we need some technical article where they actually test CPUs and have some evidence beyond just "itanium is 2x slower than something (we won't even say which model is something)". Also, IIRC, Wilamette with frequency of 2x is about as slow/fast as Coppermine with the frequency of x; that is, 2GHz Wilamette was comparable to 1GHz Coppermine. And by combining this with "a 800Mhz Itanium would have been very impressive compared to an 800Mhz Pentium" we get that Wilamette even at 2GHz is certainly not 2x faster than itanium, and maybe is not faster at all. Also, from the article it seems that Itanium was aimed exclusively at enterprises, which it did well. Penartur (talk) 08:10, 27 March 2011 (UTC)
- The tech article is referenced in the "history" section (ref 33, I think.) As you can see from the early part of the "history" section, Itanium was not originally aimed only at enterprises. In fact, Intel's stated goal was to replace Pentium. Itanium was re-targeted to enterprise-only after failing miserably in other areas, notably workstations.
- Then still we need some technical article where they actually test CPUs and have some evidence beyond just "itanium is 2x slower than something (we won't even say which model is something)". Also, IIRC, Wilamette with frequency of 2x is about as slow/fast as Coppermine with the frequency of x; that is, 2GHz Wilamette was comparable to 1GHz Coppermine. And by combining this with "a 800Mhz Itanium would have been very impressive compared to an 800Mhz Pentium" we get that Wilamette even at 2GHz is certainly not 2x faster than itanium, and maybe is not faster at all. Also, from the article it seems that Itanium was aimed exclusively at enterprises, which it did well. Penartur (talk) 08:10, 27 March 2011 (UTC)
- I believe that this is the case. The press, even the technical press, treated clock speed as the most important figure of merit at that time and all the way up until the end of the Netburst era. The WSJ statement is even valid in this regard: a 800Mhz Itanium would have been very impressive compared to an 800Mhz Pentium. But the 2Ghz Pentium/Xeon had better integer performance than the 800Mhz Itanium, for much less than half the price. This is the biggest reason Itanium failed to displace Pentium (which was Intel's original stated goal.) The WSJ quote is therefore a good, concise, verifiable (WP:V) way to capture what happened. From our post-Netburst perspective, we know that SPEC-INT and SPEC-INT/$ are better measures, but that is far too complicated for the lede, and the WSJ quote is really does capture the perceptins of the time. -Arch dude (talk) 20:30, 26 March 2011 (UTC)
- I look at it as simply a wording debate. I think it is perfectly true that the Itanium ran at about half the speed of the Pentium 4. But that says absolutely nothing about the performance or capability of each processor. Just the same as most modern cars can travel much faster than the average 30 ton lorry. But the same things happens with that statement, it says nothing about the cargo transporting performance or cargo capability of each vehicle. HumphreyW (talk) 15:51, 27 March 2011 (UTC)
- However, it makes the reader think that the performance of Wilamette is 2x better than performance of Itanium. So, if this is the case, we should either rephrase the statement or remove it completely. Penartur (talk) 12:47, 28 March 2011 (UTC)
- Do you have a reference with the comparative performance figures? I would like to see that first before I agree to any changes. We shouldn't be making changes without proper references to back them up. HumphreyW (talk) 13:05, 28 March 2011 (UTC)
- I'm not saying we should replace current article content with e.g. "Itanium is 100x faster than Pentium-IV". I don't have a reference with the comparative performance figures; neither does the article. I'm saying that maybe we should remove this unbacked statement that Itanium is 2x slower than Pentium-IV. Removal of unbacked statement does not require us to prove it: Russell's teapot Penartur (talk) 16:57, 4 April 2011 (UTC)
- Do you have a reference with the comparative performance figures? I would like to see that first before I agree to any changes. We shouldn't be making changes without proper references to back them up. HumphreyW (talk) 13:05, 28 March 2011 (UTC)
- A ref currently in the article from 2001 is at least a starting point.
- The author is clearly not biased against Itanium; in fact, he would seem to be pretty solidly on its side, although he modestly allows that it won't replace Xeon right away; that may have to wait as long as 2003.
- Executive summary: Merced, although "within the range of high-end server processors", gets beaten on SPECint2000 by "everything but Transmeta", but just wait until McKinley.--NapoliRoma (talk) 20:38, 4 April 2011 (UTC)
- It turns out the published SPEC benchmark results from 2001 are still on line and searchable. I haven't pursued this to any great length (that is, I poked around for about 2 minutes), but I did find at least one Itanium and Xeon result from about the same time. I've done no vetting whatsoever of the overall hardware configs below, so memory, software, and so on may be--almost certainly are--radically different.)
- Dell Precision 730, 800 MHz Itanium, June 2001: CINT2000 314
- Compaq Evo W6000, 1.4 GHz Xeon, Aug 2001: CINT2000 509
- --NapoliRoma (talk) 20:52, 4 April 2011 (UTC)
The statement in question should be removed in my opinion. Firstly, the WSJ article is simply too ambiguous as to what speed refers to. Is it the clock rate or performance of the processor that it is referring to? The informal, imprecise language makes its intent open for debate and that is a very bad thing. Secondly, the WSJ might not be a reliable source in the context of technical claims. I don't expect general-audience publications like the WSJ to have the intent and capability to provide accurate and insightful coverage of any science or technology topic. The performance of a processor is a technical topic, and one which is well known to be full of caveats and fine nuances. Once again, I don't expect the WSJ to excel at making solid evidence-based statements on the matter. Their sweeping generalizations and imprecise language might be suitable for their audience of Wall Street bankers, but is of no use to Wikipedia. Rilak (talk) 11:16, 15 April 2011 (UTC)
Replacement for x86 ?
At present the article makes the claim: "Intel had originally hoped to make Itanium a replacement for the original x86 architecture."
The citation points to a page on Features.techworld.com, which says in passing: "Once touted by Intel as a replacement for the x86 product line, expectations for Itanium have been throttled well back." and itself has no references for that assertion.
Ought not such a major claim about Intel's intentions point to specific *Intel* claims or announcements ?
86.25.123.51 (talk) 11:24, 8 April 2011 (UTC)
- That's actually debatable in the world of WP. Secondary sources are quite often considered preferable to primary sources. That is, "I once walked on the moon" might seem valid if you're quoting primary source Neil Armstrong, but not so much if you're quoting primary source Donald Trump, so WP prefers the secondary source: "Walter Cronkite says Neil Armstrong once walked on the moon."
- So in this case, it might be preferable to have a contemporaneous analyst or journalist saying, "Intel plans to use IA-64 as a replacement for the x86 product line." However, an Intel product roadmap from that time would appear to me to be a relevant source as well.--NapoliRoma (talk) 19:31, 8 April 2011 (UTC)
- Oh, and it of course should also be a reliable source: "Glenn Beck says Donald Trump once walked on the moon" is right out.--NapoliRoma (talk) 19:35, 8 April 2011 (UTC)
Look at reference 6. The retiring Intel executive also stated that Itanium had been expected to supplant x86. -Arch dude (talk) 21:12, 8 April 2011 (UTC)
Architecture and implementation
The article confuses architecture with implementation. For example, the article talks about how many instructions are in a bundle, and then goes on to describe the instruction (or should we say bundle) fetch mechanism. The instruction and bundle formats are architectural issues, while the instruction fetch mechanism is an implementation issue. Covering distinct concepts in the same sentence and section without distinction is a bad idea. Doing so in an article about an architecture based on VLIW concepts, where some features that are implementation issues in non-LIW architectures are architectural issues is dangerous since the reader is forced to guess what the article is saying. The original Itanium article is less confused and meatier than the treatment here. Perhaps it should be restored? Rilak (talk) 11:31, 15 April 2011 (UTC)
Compiler discussion
There should be a section devoted to compilers, so that Knuth's comment, "The Itanium approach...was supposed to be so terrific—until it turned out that the wished-for compilers were basically impossible to write" isn't left in a vaccuum. From the rest of the article, the difficulty of getting compilers to compile existing code for the Itanium sounds like its biggest single problem. 208.107.152.253 (talk) 07:50, 22 November 2011 (UTC)
Architecture and implementation
The article confuses architecture with implementation. For example, the article talks about how many instructions are in a bundle, and then goes on to describe the instruction (or should we say bundle) fetch mechanism. The instruction and bundle formats are architectural issues, while the instruction fetch mechanism is an implementation issue. Covering distinct concepts in the same sentence and section without distinction is a bad idea. Doing so in an article about an architecture based on VLIW concepts, where some features that are implementation issues in non-LIW architectures are architectural issues is dangerous since the reader is forced to guess what the article is saying. The original Itanium article is less confused and meatier than the treatment here. Perhaps it should be restored? Rilak (talk) 11:31, 15 April 2011 (UTC)
Compiler discussion
There should be a section devoted to compilers, so that Knuth's comment, "The Itanium approach...was supposed to be so terrific—until it turned out that the wished-for compilers were basically impossible to write" isn't left in a vaccuum. From the rest of the article, the difficulty of getting compilers to compile existing code for the Itanium sounds like its biggest single problem. 208.107.152.253 (talk) 07:50, 22 November 2011 (UTC)
Update for the 9500 series?
I'm no Itanium expert, but this article should be updated given that Intel finally released the 9500 series, no? — Preceding unsigned comment added by 99.159.21.154 (talk) 17:46, 26 December 2012 (UTC)
Strange layout
Hi. Is it just me, or is the layout of this article really strange? It starts with "Market reception" (marketing), then "history" (historic), "market share" (marketing), "architecture" (technical), "hardware support" (technical), "software support" (technical), competition (marketing), supercomputers (marketing), processors (mixed contents) and "timeline" (historic). Why are topics so incoherently thrown about? Isn't better to have some form of coherence?
Ideally, I think it is better to tell the reader what the subject of the article is before proceeding to describe its market reception and marketing topics, so technical topics should all follow the "History" section. Then, marketing topics can follow.
Best regards,
Codename Lisa (talk) 00:45, 8 March 2013 (UTC)
- Yes, I agree and first thought of just rearranging. Then saw it is rather long (KB) so decided to split. Might not have been the best idea. IA-64, the architecture would have been a small article and not getting any longer? It's hard to separate the architecture entirely from the implementation. Maybe some of the information about processors should say there, and most of the marketing stuff and reception here. Or someone wants to revert. I just saw afterwards that IA-64 was not just a REDIRECT but seems to have histoy of it's own on the architecture that went missing. comp.arch (talk) 17:11, 11 November 2013 (UTC)
"Software support" section is mutating
The "Software support" section, which starts with the phrase
- Itanium is supported by the following operating systems
...then pretty much instantly goes off the rails by talking about operating systems that are being ported to another architecture.
I suggest that this section should remain about Itanium-supported software, and perhaps add a separate section on OSes that are being ported to non-Itanium platforms.--NapoliRoma (talk) 05:31, 14 May 2015 (UTC)
Similarly, this sentence doesn't seem to belong in the "software" section and should probably be moved to another part of the article:
- In late September 2012, NEC announced a return from IA64 to the previous NOAH line of proprietary mainframe processors, now produced in a quad-core variant on 40 nm, called NOAH-6.
--NapoliRoma (talk) 16:06, 14 May 2015 (UTC)
External links modified
Hello fellow Wikipedians,
I have just added archive links to 3 external links on Itanium. Please take a moment to review my edit. If necessary, add {{cbignore}}
after the link to keep me from modifying it. Alternatively, you can add {{nobots|deny=InternetArchiveBot}}
to keep me off the page altogether. I made the following changes:
- Added archive http://web.archive.org/web/20110720155941/http://www.eetimes.eu/18302255 to http://www.eetimes.eu/18302255
- Added archive http://web.archive.org/web/20071014061746/http://www.theinquirer.net:80/default.aspx?article=33115 to http://www.theinquirer.net/default.aspx?article=33115
- Added archive http://web.archive.org/web/20061126193800/http://www.byteandswitch.com/document.asp?doc_id=81342 to http://www.byteandswitch.com/document.asp?doc_id=81342
When you have finished reviewing my changes, please set the checked parameter below to true to let others know.
An editor has reviewed this edit and fixed any errors that were found.
- If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
- If you found an error with any archives or the URLs themselves, you can fix them with this tool.
Cheers.—cyberbot IITalk to my owner:Online 10:43, 27 February 2016 (UTC)
- Two of them just moved and I restored them; the Byte and Switch one appears to be gone, but the Wayback Machine version works. Guy Harris (talk) 19:59, 27 February 2016 (UTC)
External links modified
Hello fellow Wikipedians,
I have just modified one external link on Itanium. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
- Added archive http://web.archive.org/web/20160303215640/http://h20341.www2.hp.com/integrity/w1/en/systems/integrity-systems-overview.html?jumpid=ex_R11294_us/en/large/tsg/go_integrity to http://h20341.www2.hp.com/integrity/w1/en/systems/integrity-systems-overview.html?jumpid=ex_R11294_us/en/large/tsg/go_integrity
When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}
).
This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}}
(last update: 5 June 2024).
- If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
- If you found an error with any archives or the URLs themselves, you can fix them with this tool.
Cheers.—cyberbot IITalk to my owner:Online 15:47, 11 July 2016 (UTC)
System Environment should be mentioned!
Both terms, IA-64 and Itanium, could be used to refer to the same thing, but Itanium is a little bit different than IA-64. Because when Intel released Itanium processors, they permit the IA-32 software could also work onto this innovative platform. In Itanium architecture manual, it describes there exist two system environment, Itanium System Environment and IA-32 System Environment, both system environment are booted from the same EFI firmware, and the former is the default. When Itanium processor working under IA-32 system environment, it works like a real x86 processor running the x86 operating systems and x86 software, 16-bit and/or 32-bit. This is quite important thing needs mentioning! — Preceding unsigned comment added by 175.19.66.153 (talk) 08:46, 10 June 2017 (UTC)
- The original Itanium could executed The IA-32 opcodes in hardware. It was so poor at this that the IA-32 instructions were removed. The last Itanium with IA-32 hardware was released in 2005. Since then, an Itanium system can execute IA-32 binaries only via emulation. Note that the Itanium was never able to execute x86-64 instructions in HW. -Arch dude (talk) 19:28, 11 June 2017 (UTC)