Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
Wireless Networking Intel Hardware IT

No WiFi In 'Grantsdale' Chipset 166

Posted by timothy from the better-to-sell-it-separately dept.
bizpile writes "A company spokesman confirmed Friday, Intel has decided not to enable the wireless access-point functionality in its 'Grantsdale" chipset. Intel decided not to include this feature because of the proliferation of cheap wireless access points. Spokesman Dan Snyder said, 'So many wireless APs are out there, and they're essentially free" when purchased in conjunction with DSL or cable service from an ISP. The company may still develop a custom chipset to re-enable the WiFi functionality if a large customer requests it. Also, their Centrino plans and production will be unchanged."
This discussion has been archived. No new comments can be posted.

No WiFi In 'Grantsdale' Chipset More

No WiFi In 'Grantsdale' Chipset

Comments Filter:

  • large? (Score:5, Funny)

    by LBArrettAnderson ( 655246 ) on Sunday September 26, 2004 @11:12PM (#10359716)
    The company may still develop a custom chipset to re-enable the WiFi functionality if a large customer requests it.

    7'2" 300 lbs... do I count?

  • Good (Score:5, Funny)

    by TheFlyingGoat ( 161967 ) on Sunday September 26, 2004 @11:13PM (#10359725) Homepage Journal
    This is a good thing. There's little reason to enable wi-fi in the chipset, and this eliminates the chance that something will be wrong with it. This means fewer patches and a more stable system.
    • *There's little reason to enable wi-fi in the chipset, and this eliminates the chance that something will be wrong with it. This means fewer patches and a more stable system.*

      less features enabled that are though actually built into the thing == good? really? like the 486 sx was a good idea?

    • Re:Good .... Makes me want to SCREAM... (Score:2, Interesting)

      by Anonymous Coward
      Yes, I read the article.
      And even the article is unclear about being either the Access Point component or the 802.11x in the chipset.

      My guess is that they are leaving a feature connector attached for a third party WIFI card, and disabling the Access Point features.

      Which means absolutly nothing. The third party WIFI card can act as an Access Point.

      Can a daughter card providing WIFI be considered part of the chipset (ala Centrino)?

      I really wish article authors could show an attempt to clarify ideas.

  • Rushed? (Score:5, Interesting)

    by chrispyman ( 710460 ) on Sunday September 26, 2004 @11:14PM (#10359729)
    It seems to me that it was more of a nonessential feature getting cut in order to ship ontime. They do have a point though that WAPs are cheap, but more often than not they suck. Why else would people turn certain WiFi network cards into WAPs (perhaps on a Linux server)? If there was some sort of awesome "do everything" WAP and for cheap, maybe then you'd have a valid excuse to cut this feature out.

    • Re:Rushed? (Score:5, Insightful)

      by Jeff DeMaagd ( 2015 ) on Sunday September 26, 2004 @11:47PM (#10359897) Homepage Journal
      Not a whole lot of people are turning computers into access points, it is kind of inefficient when a 12W device can route, firewall, run QoS, act as a switch AND be an access point.

      I really haven't had any problems with existing WAPs.
      • Plus it's tough to run an entire computer with PoE.
      • I have an endless problem with Linksys WAP54G. It has an Ethernet port and a RESET button, that's it. And in any combination it can't be found anywhere on the network, most of the time. Once I manage to configure it, then it kinda works. But once reset... no way. It still doesn't work. The setup s/w once detected it today, and said that its IP address is 1.192.168.1 ... guess what, it's a lie. And the default IP is not there either; the manual even talks about three different "default" IPs. It's garba
        • I have an endless problem with Linksys WAP54G.
          So solve it [openwrt.org].
        • I've seen lots of cheap router/access points go down. They're pretty much disposable. They start flaking out, usually after a power blip (am loving life in Florida right now).

        • Well, my manual [linksys.com] only mentions 192.168.1.245 as the default address. Even if it didn't mention any address, you could always hook a PC to the ethernet port and run tcpdump when you turn on the AP. I think you will see some activity that leads to the address it is using, but I don't know for sure right now.

      • Some math on an access point. vs. PC firewall (Score:5, Insightful)

        by xtal ( 49134 ) on Monday September 27, 2004 @12:43AM (#10360082)
        Let us assume electricity costs $0.15 per kWh.

        I'll assume you already have a low-end PC for use "Free", e.g. already paid for and with a market value near zero.

        I will also assume you can configure it to spin down the HD and turn off the monitor. We'll say this is around ~30W; it's actually much less if your processor is idle. My via c3 backup server consumes about 14W, the firewall a little more, an older 486 at around 20W.

        Let's stick with 30W. To be a fair comparison, it needs to run a wireless card. That's not a major addition to power, but we'll account for it.

        30W is 0.003 kW, so per day, this device costs 0.72kWh x $0.15/kWh = $0.108/day. Per year electricity cost would be roughly $40.

        An access point costs about $150 in my parts; I'll say you can get it for $100 for the sake of arguement though.

        At 12W, using the above calculations again, this access point takes 40% of the power. Or, a yearly power bill of about $16. The difference in the power bills is $24.

        So it would take about 4 years to catch up, assuming the access point doesn't die. I have enough spare parts and obsolete hardware to run a firewall indefinately for no extra expense. The PC based firewall can do a lot more stuff too - much more configurable, patchable, can run other servers, etc etc. I run OpenBSD on mine and find it more than adequate. Plus, unlike every access point I've seen in the $100 range, my ages-old USR modem I bought 10 years ago sits there doing it's job shuffling bits around. No DSL in these parts.

        At best I'd consider it a draw. You add a little polution, but can save that firewall computer from ending up as toxic waste, too.

        Myself, I run an access point and a firewall. I don't like trusting one device to do everything, and I know the firewall is very hard to beat.
        • Netgear MR814 [amazon.com] -- 802.11b access point + 4 port switch + firewall/nat for sale at Amazon for less than $20 after rebate-- in other words, barely any more than the wireless card for your gateway.

          There's really no reason to run a general purpose machine for the wireless gateway on an ordinary network. Even if you're going to have a dedicated gateway, you can use a wireless router with the routing functionality turned off.
        • There's a problem with your analogy. This is a new processor and chipset, so likely it would not be compatible with that old hardware you have lying around. So if you wanted to make use of this technology, you would have to buy a new computer, or a new motherboard too at the very least (and then possibly new RAM as well).

          Within a few years this new hardware you buy now would be back to that near-zero market value. But by then Wifi APs will likely be much cheaper than $100. Also the power consumption of the
        • An access point costs about $150 in my parts;

          Might I suggest that `in your parts' must translate to `in that place in the mall where the owner giggles as he takes my money'.

          Even over here in the Uk where electronics are very expensive I saw one the other day for 33 quid - $60.

        • The problem with that, in wanting a WAP in the chipset is that the chipset in question is for a Prescott-based system. They won't be cheap for a long time, and the leakage draw alone of most 90nm chips is more than 30W.
        • I purchased my (badass) D-Link WAP/Router/Switch for $60 CDN after main-in rebate. That comes out to probably $40 USD per month, so it would take two years to catch up.

          If you don't have a wireless NIC in your machine as well, that's more cost - another $150 CDN for the cheapest I've seen. So with wireless PCI NICs running at twice the price of a wireless router, I'm pretty convinced.

          The access point is also smaller, easier to store somewhere with good signal, is easily relocated without unplugging, produc
    • Why else would people turn certain WiFi network cards into WAPs (perhaps on a Linux server)?

      Because! It's fun editing configuration files, hunting for compatible drivers, and testing it over and over until you get the silly thing to work.

      I once spent hours trying to setup a firewire on a linux box, but in retrospect, I probably would have save a great deal of agony just purchasing a hardware router/firewall.
      • "I once spent hours trying to setup a _firewire_ on a linux box, but in retrospect, I probably would have save a great deal of agony just purchasing a hardware router/_firewall_."

        (emphasis mine).

        Perhaps you were looking up the wrong docs? :)

        • ack! It's another case of me thinking one word, and typing another. I should really stay away from slashdot at 1:00am in the morning. :P

  • sensationalism (Score:5, Insightful)

    by garbletext ( 669861 ) on Sunday September 26, 2004 @11:14PM (#10359731)
    What's the big deal? They disabled the ability to use the pre-installed chipset as an access point. This is just a cheap, consumer grade chipset for people who want wireless to come with the laptop they buy. anyone who needs to use their laptop as an access point will know about this and buy a different laptop, or just use a different Wifi card.

  • Essentially free? (Score:5, Insightful)

    by idesofmarch ( 730937 ) on Sunday September 26, 2004 @11:14PM (#10359735)
    Wireless network cards for desktops are never free. Access points are not generally free either, unless somehow bundled with the Internet service by the ISP. Maybe I am missing something.
    • I think they are trying to say that they are free from your ISP, but that isn't true either.

      It's a bad example, but my DSL ISP [cincinnatibell.com] did charge. They are now running a promotional offer though.

      Thing is, you need to pay for "professional" installation to get 'free' which costs 100 bucks, and isn't needed. Now the DSL modem is USB compatible and comes with a step by step CD. They used to charge for installation of the AP and an equipment fee every month.
    • I got a belkin 54G router, a clone of the 54G linksys, for $7.50 CDN.

      It was a $50.00 instant rebate with a $50.00 mailin rebate on a $99.00 router.

      The rebate came in 4 weeks.

      So NET I paid the tax on $50.00 or $7.50. Free and the goverment still gets it's cut.

      And the router works and there is a linux distro for it too.
    • DealMeIn.Net [dealmein.net] (ran by a fellow slashdotter) and other [dealsites.net] sites [rasputinj.com] like it frequently have low-end consumer grade wireless equipment for free or almost free after rebates.

      Josh.

  • Costs? (Score:5, Interesting)

    by zalas ( 682627 ) on Sunday September 26, 2004 @11:15PM (#10359736) Homepage
    How much would it cost Intel to enable the WiFi? I mean... if it's a small cost, then they'd get their wireless access points into computers instantly so people don't have to go out and buy wireless APs. I mean, if my computer came with a wireless access point, then I certainly wouldn't go out and buy another one (unless it was terrible). It's the same deal with firewire. Most people don't use it, but it's sure a lot more convenient if it was there when you needed it.
    Hate to be suggesting monopolistic marketing ideas, but Intel can really get a lot of their wireless AP into computers by bundling it.

    • Re:Costs? (Score:5, Insightful)

      by servognome ( 738846 ) on Sunday September 26, 2004 @11:23PM (#10359783)
      Doesn't matter if you want it or not, Intel sells to IBM, DELL, etc. The OEMs are saying they don't want it. Most likely the OEM's have their own WiFi "solution" if you want it, so they'd rather not have to worry about dealing with the WiFi daughter card.
      • This could even be useful even in corporate environments.

        A Wi-Fi deployment would just consist of deploying access-point control software to all your desktops. Then you'd have virtually no dead-zones as a laptop just needs to be within 100 yards of a desktop somewhere. No need for extra hardware. I'm sure that with software you could add real authentication as well.

        Just a thought. I'm sure it wouldn't replace the need for at least a few access points in obscure locations or in places like cafeterias t
        • Of course, you would have to be very careful to make power very low, and do get proper channel re-use. Furthermore, many laptop wifi drivers hang on to an AP for dear life once they've associated, rather than move to one with a stronger signal.

          So in theory it *should* work, and its a great idea. In practice, at present, I think it would be a nightmare.
    • By not enabling it, they probably wont test it in manufacturing which can save them a little bit of the test costs.
  • They are too busy reverse engineering the latest AMD processors to be bothered with WiFi features!

    Kidding! ;)

  • Why WAP - why not wifi adapter instead? (Score:5, Insightful)

    by stu72 ( 96650 ) on Sunday September 26, 2004 @11:20PM (#10359769)
    What I don't understand is why anyone needs a wifi access point in their desktop. I know it could be useful in some circumstances, but far more useful, IMHO would be an on board adapter so you could just log onto your wifi network w/your desktop.

  • It's the 486 all over again (Score:5, Interesting)

    by Jason1729 ( 561790 ) on Sunday September 26, 2004 @11:26PM (#10359798)
    Intel made a 486SX chip which was a 486 with the math co-processor disabled and it was slightly cheaper than the 486DX. They also sold a 487SX which they called the co-processor, but in truth it was just a 486 with a working co-pro, and when you used the 487SX, it completely disabled the 486SX and took over as the only CPU in the system.

    The whole thing seemed like a test of how gullible their customers were. It looks like they're doing the same thing again.

    Jason
    ProfQuotes [profquotes.com]
    • Back in the day a math co-processor was a big deal. This seems to be more a case of feature pruning to expedite shipment.

    • Re:It's the 486 all over again (Score:2, Informative)

      by Anonymous Coward
      Actually it was chips where the FPU failed testing and was disabled.

      Back in the DOS days, very few business apps needed a FPU, so it was a fair tradeoff for some customers.
      • very few business apps needed a FPU, so it was a fair tradeoff

        And then the 3D FPS was invented. DOOM and QUAKE really needed that FPU to work properly (QUAKE required one, IIRC). That's the real reason the FPU made it's way back into the mainstream chips, gaming made it a must-have.

        Before that, the FPU only made a difference if you were doing SERIOUS number crunching, and I'm not talking about excel.
        • QUake required one, but quake was WAY after the days of 486s, let alone 486sx. When quake was release i had a P133...
          NO game of the 486sx are needed FPU. Not doom, not Duke nukem 3d, not Rise of the Triad,... They all didnt even SUPPORT fpu ops, because at that time even the 487 was much slower than integer math and look up tables (4cyle add and 15 or so cycle mul if i remember correctly)
          • Quake came along during the last years of the 486. My brother had actually just put together a 486 DX 133 MHz when it came out. Yeah, it only ran about 15-25 fps, but only needed it to meet the campus network's system requirements, otherwise he probably could have taken the 386 along.

            I do find it funny that Rise of the Triad displays, "Buy a 486!" if you have the screen set really small. I guess it's a case in point as to games not needing FPUs.
            • The only game of the 386 or 486 area (1990-1995 or so) that really supported a FPU was falcon 3.0 with its high detail aerodynamic model on the highest realism setting.
              Too bad i always crashed with high realism, so i never used it :)
            • I used to play quake 1 on a 486DX/2-66 at 320x200 and it ran pretty well. I'm pretty sure it was more than 15 fps. It's hard to see what's going on at low resolution but the fact that it's polys-based means that when you move, you get more useful visual data from which your brain can composite an image, unlike the flat-sprites psuedo-3d id games.
        • It wasn't until Pentium that the FPU was put onto an x86 chip standard.

          I think Quake was helped because of the 486DX, before 486, the x87 chip was a separate chip and rarely sold to consumers in a consumer computer, it was often an extra $100 or something like that. I remember back in the 286 & 386 days when I insisted on the x87 (or math coprocessor) because I played with CAD and even programmed a few 3D wireframe vector programs that I wrote.
      • Actually it was chips where the FPU failed testing and was disabled.

        The 486SX sold so well that Intel made a new mask that removed the FPU entirely. This increased the yield per wafer.

    • A 486 chip that failed the FPU verification would also get the SX stamp.

      Also, having the SX chip around allowed us to standardize our hardware around 486 motherboards, which made selecting components and planning for upgrades on our corporate fleet much easier.

      Later on as our 'power users' were upgraded to pentiums, their DX2 chips and RAM would move down the line, and took about five minutes to slap into an existing box.

      It was not entirely a bad thing.
      • Wow.. we never recommended 486sx chips to our clients. At that time we were recommending 386DX40. They were cheaper and quicker the the Intels.

        BTW: wasn't that the 1st time AMD spanked Intel?

        • I bought the 386DX40 then, it was my first AMD chip, and every CPU I've ever bought since has been AMD.

          Of course, "cheaper" is relative. It was $340 for the CPU, $200 for the MB, and $800 for 16 meg of memory.

          I'd bought 2 4-meg simms for $400 only to find out the 32-bit 386 needed 4 sticks, so I could pay a 20% restocking charge on the 4-megs and buy 4 1-megs or buy 2 more 4-megs.

        • Re:It's the 486 all over again (Score:4, Informative)

          by spacefrog ( 313816 ) on Monday September 27, 2004 @02:59AM (#10360537)
          Compared especially to the very early 486SX chips, You are absolutely right, in many instances the AMD 386DX-40 would either be faster or be a much better value.

          However, the later AMD-based 386DX boards that were cheap used surface-mounted CPU's and from an upgrade-sense were foobar. The ones that were socketed could be upgraded to a Cyrix chip that was often a nightmare, between having to use utilities in the autoexec to enable the L1 cache, and having previously stable systems decide they would start locking up at random.

          A 486SX-33 on a board with 256KB cache and VL-Bus slots would cream it, though, and had a very sweet upgrade path.

          Once AMD had their 486's on the market, a lot of those boxen that didn't get hand-me-down intel DX2's got the (very affordable) AMD DX2-66.

          Going that route and buying high-quality motherboards was a major win. They could have had a third round of CPU upgrades, but the price/performance ratio on the Intel 'overdrive' CPU's was just too pathetic.
        • BTW: wasn't that the 1st time AMD spanked Intel?

          I wish it was because for a while, back when I was a kid, we were going to upgrade our 286 (at 13 MHz I think) to a 486/33. I think we decided to go for a cheaper 386 DX/40 MHz instead. Doom was barely playable on it when it came out. I even tried Doom on a 486SX/33 laptop and got to see it run about twice as fast. The 386 was the main computer for about five years, too. Shame. Luckily, we upgraded to a Pentium 133 just in time for QTest (the test release of

  • Large customers? (Score:4, Funny)

    by ShatteredDream ( 636520 ) on Sunday September 26, 2004 @11:28PM (#10359806) Homepage
    And as a complementary gift for our "large customers" we provide coupons for 20 big macs per buyer. *Lawyer whispers in Intel CEO's ear* And by large buyer we mean someone who buys a lot of our products, we would never, ever think to provide so much almost 700g of pure fat to those struggling with weight at McDonalds...

  • Dumb question (Score:5, Interesting)

    by mcc ( 14761 ) <amcclure@purdue.edu> on Sunday September 26, 2004 @11:29PM (#10359815) Homepage
    that maybe someone who's more familiar with Intel's recent chips could answer...

    When Intel says they're "disabling" this, do they mean they're going to be physically leaving it out, or permanently disabling it, or just deactivating a jumper or something? By which I mean, could overclockers re-enable the feature on chips they possessed themselves if they really, really wanted to?
    • Intel had Hyperthreading tech built into the Netburst-cored processors from day one, and merely disabled it in the earlier Pentium 4 processors. And unless you had access to a microprocessor engineering facility nearby, it stayed off. I would assume they'll handle it the same way with their core logic chipsets from now, especially after mainboard manufacturers managed to enable the "soft-off" memory enhancement feature in the i865 chipset, effectively turning it into the more expensive i875 chipset (sans

    • Re:Dumb question (Score:3, Interesting)

      by lachlan76 ( 770870 )
      If overclockers could change stuff like that on the silicon, I think they'd change the multiplier before making it into an access point, especially since they can probably afford the $150-$200 for the WAP.

  • battery friendly? (Score:4, Insightful)

    by hdd ( 772289 ) on Sunday September 26, 2004 @11:43PM (#10359880)
    haven't look into Grantsdale processor yet, assuming it will be used in a laptop, won't it be more battery friendly to let cpu handle the wireless function instead of having another power hungry minipci wifi card? It's a sure winner if this can at least add 30 minutes to the battery life.
    • I doubt it... when's the last time you saw a minipci card with a fan on it?

      CPUs are big giants, and even if they can handle the processing quickly and go back to sleep, there's a lot more things going on in these things than there needs to be for simple wifi processing (floating point unit, large cache, large DRAM intefaces etc.)

      The DSPs in wifi chip sets are well suited for the job. They can easily act as a first line of processing and determine if an incoming packet requires waking the main processor. T

    • Re:battery friendly? (Score:3, Informative)

      by pjbass ( 144318 )
      haven't look into Grantsdale processor yet, assuming it will be used in a laptop

      Grantsdale is a derivitave of the Northwood process used to make the majority of P4's (I work at Intel, trust me on the product evolution here...). Grantsdale was certainly not intended to be soley laptop-grade chipsets, in fact, it is intended to be in high-end desktops, boasting the 7.1 Dolby sound, GigE network, and Serial ATA, to name a few features. In the marketplace, it's known as i915 and i925 (as far as I'm aware ri
    • won't it be more battery friendly to let cpu handle the wireless function instead of having another power hungry minipci wifi card?

      On a long flight, there is a lot to be said for simply unplugging the AP and putting it away. Then it uses zero power guranteed.
    • But then you can't get support without drivers, which wouldn't come instantly for Linux.

      Remember what happened when someone decided to let the CPU handle processing for modems? THAT was a good idea, wasn't it?

  • Craig Barrett listen to your advice (Score:5, Insightful)

    by LinuxInDallas ( 73952 ) on Sunday September 26, 2004 @11:54PM (#10359935)
    How interesting is it that we read this after seeing the interview with Craig Barrett posted earlier where he says that Intel's future growth is dependent on merging their technology with communications. Ha!

  • It's simple economics (Score:4, Insightful)

    by DelugeDreamer ( 815695 ) on Monday September 27, 2004 @12:12AM (#10359984)
    There aren't any buyers for the feature, so it's not enabled. If a company isn't going to be able to turn a profit, then they don't waste resources. Otherwise, they cut their losses and move on.

  • Great for security (Score:5, Insightful)

    by MoceanWorker ( 232487 ) on Monday September 27, 2004 @12:25AM (#10360032) Homepage
    It's good to hear that Intel won't be including wi-fi capability in its upcoming chipset..

    With all the talk about them including DRM in their processors it would have been interesting, and scary as well, to see how it would affect their wi-fi chips (had they continued production)..

    Might as well make their job easier (and a deeper stab at privacy) by having it notify them in seconds of any sort of "violations".. *cough cough*

  • Intel vs. WEP (Score:2, Interesting)

    by SlashdotMirrorer ( 669639 )
    Perhaps this is Intel's first steps toward their own secure implementation of the wireless WEP protection. As any bearded terminal hacker knows, WEP (Wireless Encryption Protocol) is insecure in it's use of shifting between key frames, so perhaps Intel has come up with a solution for this that is more secure than the hacks that Cisco and Lucent have put into place so far.

  • Have they fixed Centrino yet? (Score:5, Informative)

    by multiplexo ( 27356 ) * on Monday September 27, 2004 @12:46AM (#10360091) Journal
    My experience with Centrino was that it didn't work. I ended up installing PCMCIA cards to get wireless support. Maybe I didn't have the right drivers, in the commercials I saw people surfing the net in the Acropolis (perhaps it's now an 802.11x hotspot), I couldn't get consistent performance 15 feet from the access point with a clear line of sight.

    Other than economics I wonder why Intel just doesn't produce a kick-ass mini-PCI card that supports the various wireless standards and then flog the Hell out of it to the PC makers. The mini-PCI approach, combined with well designed internal antennas works very well for the Macintosh.

    • When I worked at CompUSA it was great. I'd walk over to any of the Centrino laptops and look up some product spec on the internet.

      I suspect either you had a cheap wireless hub, or someone left the microwave door open.
      • When I worked at CompUSA it was great. I'd walk over to any of the Centrino laptops and look up some product spec on the internet.

        I suspect either you had a cheap wireless hub, or someone left the microwave door open. Actually I had the Linksys hub that I purchased at CompUSA. Strangely enough this hub works just fine with four different PCMCIA wireless cards and with the Linksys 802.11 bridge I have connected to my SliMP3 and to the wireless card built into my Squeezebox.

        Of course if everyone on /. ha

    • You do know that Centrino is implementation specific, right?

      I'm not sure from your post how many different models you tried. Some (many?) of them put diversity antennas in the screen, so they should have good reception and transmission.
  • Nobody mentioned yet that the article title is misleading? It's not that there's no WiFi in Grantsdale, it's just that the Grantsdale wi-fi card will be a normal client and won't be able to double as an access point.
  • If its true that its available most everywhere, and for free, why not include WIFI be default, since most people will want/use it?

    Or did i miss something obvious here?

  • According to some old Slashdot stories, wireless computing seems to be proliferating but isn't profitable [slashdot.org] and Intel had been looking into makeing chips that use cellular networks [slashdot.org] for wireless connectivity, which would have to come with a fee. Crippling the Wifi networking at the consumer level and steering them towards using cellular networks would be more agreeable to and supported by established industries.

    However, like the story says "The company may still develop a custom chipset to re-enable the WiFi

Slashdot Top Deals

The explanation requiring the fewest assumptions is the most likely to be correct. -- William of Occam

Close