Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?
Graphics Software Intel Hardware

Shortcomings Revealed in nForce4 SLI Redux 93

EconolineCrush writes "Slashdot recently covered the launch of NVIDIA's nForce 4 SLI chipset for Intel processors, and although early reviews fawned over the chipset's performance, closer examination reveals several shortcomings that the initial wave of coverage failed to document. Problems with stability, drivers, and the chipset's oft-praised hardware-accelerated firewall and Gigabit Ethernet controller escaped the scrutiny of many reviews."
This discussion has been archived. No new comments can be posted.

Shortcomings Revealed in nForce4 SLI Redux

Comments Filter:
  • I respect his kung-foo, but his last line is off the wall.

    These kinds of problems, which we readily acknowledge aren't total showstoppers, may pass muster for Athlon 64-based enthusiast systems. NVIDIA will probably find, however, that competing against Intel's chipsets requires a higher standard of competence.

    As a serious gamer and one who has built several Athlon-based machines, I can't imagine for the life of me what he's talking about here. AMD's chips are the undisputed king of the hill when it co
    • by Anonymous Coward
      I respect your kung-foo, but this line:-

      These days, AMD is the singlemost important chip maker on the planet, second only to Intel

      is off the wall.

      How can something be the singlemost important chip maker, yet still be second only to Intel?

      Your kung-foo intrigues me.
    • These days, AMD is the singlemost important chip maker on the planet, second only to Intel

      Perhaps that would make them the second-most important chip maker on the planet then?
    • by pegr ( 46683 ) * on Friday April 08, 2005 @08:17AM (#12174767) Homepage Journal
      I believe he was referring to Intel's market, not AMD's (and supporting chipset maker's) competence. To put another way, relatively minor issues in Intel's market will be perceived as bigger, and potentially more critical, issues than similar issues in AMD's space. I don't agree quite, but that's what I get out of it.
    • by Gr8Apes ( 679165 ) on Friday April 08, 2005 @08:25AM (#12174802)
      AMD certainly has the most interesting x86 technology out there. From a PC gamer's perspective, AMD probably is the most important chip maker out there. The jury's still out on businesses' opinions, however, the Opteron certainly smokes everything intel has except itanium2, which it also might smoke but I'm withholding judgement until I read better comparisons than I've seen to date.

      As for cutting edge tech, AMD wins hands down in the x86 world. They did a nice edge run around Intel's GHz GHz GHz mantra, and they're beginning to reap the rewards. The dual and multi-core chips coming soon should finish the job once they're out and in tester's hands. Intel's dual core will either burn eggs or perform sluggishly, and they're still regrouping from their P4 mistakes and trying to come up with a new tech. Their size and brand is the only thing keeping them alive at the moment IMNSHO.

      If you want to see cutting edge technology, look towards things like the Power5 (that's not a G5 btw;) and the Cell processors. One's a multi-core powerhouse, the other, well, it's an interesting amalgam of a core with multiple DSP chips to speed things up. I'm looking forward to the PS3 and its capabilities. (There are others too, but these may be the most likely to be seen by average consumers in some form or another)
      • >the Opteron certainly smokes everything intel
        >has except itanium2, which it also might smoke
        >but I'm withholding judgement until I read
        >better comparisons than I've seen to date.

        sorry, you wont be seeing any comparisons. the Itanium blows so hard that no one will buy any... so no one has them to review. Even HP, who did most of the dev work on itanium, is shipping opteron systems now.
        • I probably should have gone ahead and mentioned that Itanium is pretty much dead due to HP dropping support. However, just because it's effectively dead doesn't mean it was worse (recall Alpha?).

          I do recall seeing some benchmark results where Itanium2's performed exceedingly well, but I didn't have a chance to read through it, hence my holding back on commenting that it completely blows performance wise. It might actually do well under certain tasking, but for any normal person, AMD is the king of the hill
    • So, again, WTF is he talking about?

      Well he's not talking about AMD at all, really. He's talking about the quality of nVidia's nF4 SLI chipset/drivers.
    • by beelsebob ( 529313 ) on Friday April 08, 2005 @08:31AM (#12174825)
      Actually - I have to disagree. I think IBM are by far a more important chip maker. They do completely different things to Intel, they come up with completely new ideas - AMD just try to make chips that compete well with Intel (which they do very well at the moment)... But IBM actually try to do something different. I mean, come on, why do you think the Power5 is devastating all the opposition in the server market?
      • Please..... Power5 is dominating the market because most server apps are written already for the Power5 from ages ago. The optimized server apps development for Opteron and other x86-64 processors hasn't even started.

      • Err... What opposition? I can't think of anyone that is still around that truly competes with the Power 5 class chips these days... Well maybe Itanium does, but who really buys Itanium based systems?

        Though I think the difference here is consumer vs. bussiness (I couldn't think of better names for them). Intel mostly makes consuemr chips (Again Itanium is the exception), so does AMD (though Opteron doesn't really compete against Itanium or Power 5), IBM makes both (though consumer only through Apple systems
      • I think IBM are...They do...AMD just try to make chips...But IBM actually try to do...

        Ok. I'm sorry. I know this is nitpicky. I've been thinking about this for about a year, and I just have to say it. Someone PLEASE correct me if I am totally off here.

        IBM "are" not anything. IBM *is*. AMD just *tries*, and IBM actually *tries* to do...

        IBM, AMD, Intel, Microsoft, etc. are companies. These are SINGULAR ENTITIES. A group of people under a single umbrella is a singular entity. Just like "mankind is
    • A chipset that was designed and put into production in a month. Who would have guessed it had problems?
    • Clarification (Score:3, Informative)

      by GundamFan ( 848341 )
      The reviewer seems to be contrasting the stability of the Nforce 4 chipset with the stability of the... lets say 925XE chipset, this isn't about processors. Intel produces realy rock solid boards lately, more stable than any 3rd party board even with the same chipset. (don't buy a Supermicro desktop board... trust me)

      So... what he was saying is that an Intel user may be disapointed by the Nforce 4's shortcomings, but an AMD user is acustomed to this level of quality, still good just not the best.
    • "These days, AMD is the singlemost important chip maker on the planet, second only to Intel."

      Does not compute, systems overloading... /me dies
      • Apple free since 1990!

        how you can have a sig like this with comments that start at +1 is beyond me. you, sir, are amazing.
        if only because i'm glad to see someone talking about apple who isn't drooling over their latest product at the time of writing.
    • by ctr2sprt ( 574731 ) on Friday April 08, 2005 @08:54AM (#12174947)
      I think you are misunderstanding. The reviewer is talking about Intel chipsets, not Intel processors. Because Intel doesn't make chipsets for AMD-CPU systems, there's no direct comparison. If there were, he's arguing, people would be less excited about NVIDIA's chipsets.


      These days, AMD is the singlemost important chip maker on the planet, second only to Intel.
      It's hard to be both most important and second-most important at the same time. Yet this is apparently a feat both AMD and Intel have managed. I guess this is a byproduct of their research into quantum computers. ("Alright! We're on top! Oh shit, I just changed our importance by measuring it.")
    • These days, AMD is the singlemost important chip maker on the planet, second only to Intel.

      So, um... doesn't that make them the second most important?

    • by jtshaw ( 398319 ) on Friday April 08, 2005 @08:57AM (#12174973) Homepage
      He isn't talking about the processors....... he is talking about the CHIPSETS.

      I've found the one short coming of AMD based solutions has always been the shoddy 3rd party chipsets and motherboards out there. I have a dual Opteron system on my desk and it has been wonderful. However, I have ran into many people running Via or nVidia chipsets on brand-x motherboards that have had awful hardware difficulties.

      I'm inclided to blame the mainboard manufactures more so then the chipset manufactures because companies like Asus, MSI, and Gigabyte never seam to have trouble putting out solid mainboards based on nVidia and Via chipsets.... but the fact of the matter is there seam to be a lot of other manufactures that build absolutely terrible mainboards for AMD processors.

    • Big corporations are pretty conservative when they buy their computer systems, and they're still overwhelmingly Intel based. If you're building systems for the corporate market, your first considerations are going to be stabability and reliability. Performance is way down the list.

      An AMD fanboi, on the other hand, won't blink if you tell him that he'll need to reflash the BIOS and download the latest chipset drivers to make a certain board work. I know I didn't. It was all sleek and shiny and was sitti
    • These days, AMD is the singlemost important chip maker on the planet, second only to Intel.

      If you want your comments to be taken seriously, and not come across as an AMD fanboy, don't use sentences like these that don't make any sense. If AMD is the singlemost important chip on the planet, it's not second to anybody.

      Disclaimer: I'm an AMD fanboy, I have never used an Intel product in my life.
    • His insinuation that as an AMD fanboy, I am unable to discern the lack of power of XYZ technology is a little irritating.

      The chipset and the technology doesn't lack power. That's not the point. The point is that they weren't as reliable as they should have been.
    • What the Tech Report is saying, is that since they had some driver issues with a reference MB, on most likely Beta drivers, their not going to be competitive in the Intel Market.

      First off. everybody knows that if you have an Intel procesor, you should use an Intel chipset for the highest level of stability. I mean when the processor manufacturer makes a chipset for their own processor, it better be as stable as possible. Intel tends to focus on stability rather than performance because they are more intere
      • Not making any personal judgements here, but with the way that Intel gets roasted by most reviewers, it is good that someone is willing to see the positive things about there products.

        Same goes for ATI, a lot of the bias agenst them is based on old information perpetuated by fanboys.

        If the sites content does not work for you... don't read it, I am sure there are people out there not looking to be "converted" who want to read reviews. (not mine to say if they are right or wrong.)

        P.S. On the other side it
      • "First off. everybody knows that if you have an Intel processor, you should use an Intel chipset for the highest level of stability. I mean when the processor manufacturer makes a chipset for their own processor, it better be as stable as possible."

        I have to agree with you and frankly I think AMD is making a mistake by not making AMD chipsets and even motherboards.
        I think they would do better in the server/corporate market if they produced motherboards like Intel.
    • What he's saying is that AMD enthusiests will often put up with poor stability in exchange for better performance. I think he's referring to the general overclockers that put up with minor glitches to get those extra few frames per second.
    • Ironically, the Tech Report is usually characterized as pro-AMD.
  • by bman08 ( 239376 ) on Friday April 08, 2005 @08:16AM (#12174761)
    Early reviews are always like this. Except in rare cases when something is undeniably awful, being in on a 'sneak peak' makes a person feel like an insider... part of the team. I think that, previuosly mentioned perks of being a hardware reviewer aside; this feeling alone accounts for a lot of over positive reviews.

    The same is true of anything. I saw a pre screening of Samuel Jackson in shaft and LOVED it. Why? I don't know now.

  • by DJ-Dodger ( 169589 ) on Friday April 08, 2005 @08:19AM (#12174776) Homepage
    Their brand of in-depth, hard-hitting coverage is probably why Intel conveniently passed them over [] for the first round of Dual Core reviews; can't have any bad press at release time.
    • by sgant ( 178166 ) on Friday April 08, 2005 @08:33AM (#12174841) Homepage Journal
      You're being sarcastic, right?

      This is hardly front page Slashdot news here. Let's see, "motherboard chipmaker needs to fix some non-showstopping bugs with bios updates." yawn.

      Everything mentioned in that article is minor. His summation is even minor:

      We're pleased to see that NVIDIA has finally fixed the long-standing bug that caused its disk controller to hit a performance wall at 128 transactions per second, but the fix was a long time in coming. More notably, the ActiveArmor GigE implementation seems to have some strange problems still. We're encouraged by the fact that NVIDIA could demonstrate a fix for some of these problems with a new driver, but we're concerned about the current state of the Ethernet driver available to the public for AMD-based nForce4 boards that have been on the market for months now. These kinds of problems, which we readily acknowledge aren't total showstoppers, may pass muster for Athlon 64-based enthusiast systems. NVIDIA will probably find, however, that competing against Intel's chipsets requires a higher standard of competence.

      OK...they fixed a bug and one of the Ethernet ports is "strange".

      But that last statement seems to throw the entire review out of whack. So what he's basically saying is: "The minor Athlon fanbois may not mind this junk...but us Intel Professionals know it's crap and will stick with Intel".
  • A question... (Score:5, Insightful)

    by Silverlancer ( 786390 ) on Friday April 08, 2005 @08:21AM (#12174782)
    What the hell does the incompetence of an nVidia chipset have to do with the performance and reliability of an AMD processor? You can simply not us the nForce 4 boards--how does this give the reviewer the right to bash AMD? Or is he just an Intel fanboy?
  • Shocked... (Score:5, Funny)

    by gowen ( 141411 ) <> on Friday April 08, 2005 @08:24AM (#12174797) Homepage Journal
    I'm shocked... shocked!... that hardware review sites would make a half-assed job of their review just in order to be the first to publish.

    Whatever next!?!?
  • I wonder why (Score:4, Insightful)

    by ProfitElijah ( 144514 ) <> on Friday April 08, 2005 @08:27AM (#12174809) Homepage
    I wonder why they escaped the attentions of the early reviewers. Perhaps because the shortcomings weren't included on the press release the early reviewers regurgitated.
  • mirror (Score:4, Informative)

    by winkydink ( 650484 ) * <> on Friday April 08, 2005 @08:29AM (#12174818) Homepage Journal
    Here []
  • by Jarnis ( 266190 ) on Friday April 08, 2005 @09:02AM (#12175014)
    'Paid reviewers skip the unflattering parts' - SHOCKING!!!!

    The first review bunch of every hardware item is PAID ADVERTISING. Well, at least close to it. To get the product for such review requires signing of NDA and cooperation of the manufacturer. Trashing a product in such launch review ensures that you won't get the next shiny thing to review. Yes, some hardware reviewers are corrupt. Shocking.

    The 'active armor' firewall has never worked right on the AMD64 NF4 either. Also on AMD64 NForce4 the gigabit ethernet has it's own problems - for example, many MMOs simply disconnect you (you go linkdead) if you have the Hardware Checksum Offload feature of the LAN chip in use.

    And unsurprisingly when you compare ANY other chipset to the rock solid Intel chipsets, they look unstable. NF4 isn't the worst of the bunch, but it can't be helped. Last STABLE (rock stable) chipset on AMD platforms was AMD760. Yes, it was lacking features, but it WORKED. After that it's VIA this, nVidia that, SIS this - all suck more or less. Thankfully the suckiness has gone down over the years, and today I can say that KT800 VIA on AMD64 is usable. Still not perfect, but works. NForce 4 has bunch of quirks and unfinished drivers, but it's probably the best PCIe-based chipset so far.
    • I had one of the first nForce 4 / AMD64 boards. For weeks I couldn't figure out why I kept getting disconnected from Dark Age of Camelot. DAoC is a rock stable program. Their servers are 100%. I have 7 other computers running DAoC through the same network connection that perform flawlessly. It had to be something with the computer. After dozens of tweaks, rebuilds, driver reinstalls... It came down to the onboard network hardware and/or drivers. I ended up buying a cheap network card. The disconnect
      • Just turn off any hardware-assisted things on the NF4 LAN options. The culprit is 'Checksum Offload' - offloading some packet checksum calculation from CPU/drivers to the hardware. Turn that off and the LDs stop.
        • I agree with you. I have NForce-4 / AMD combo. I couldnt figure out why Pegasus mail kept getting hung up when send mail via SMTP. It turns out that the Checksum Offload advanced feature on the ethernet controller was the culprit!
    • I've got a few beefs with Nvidia chipsets for AMD, too, mostly on the RAID side. This is from my testing with a tier 1 motherboard, an MSI K8N.

      Problems that I've had with the Nforce3 chipset:

      * Their storage enhancements just don't work with Win2000 at all, practically speaking. Had to go with XP, not something I was thrilled about.

      * Their RAID thingy drops disks at random, for no reason. This has happened with both PATA and SATA, and with different brands of drives.

      * Their management utilities are reall
    • The reason the Hardware Checksum Offload doesnt work is that the hardware calculates the Checksum incorrectly. I warned about this many many months ago, but couldn't find a way to contact Nvidia.

      To see this in action just download Ethereal and you will see that the Checksums are incorrect.
    • I really dont understand what people mean by 'stable'.

      One of my several AMD based computers, an AthlonXP 2500+ runs Linux... and has uptimes of 20+ days.

      Another XP 2500 I had earlier ran WinXP for weeks at a stretch. And my new A64 3000 does the same.

      The only reason I turn my PC's off is to sleep without the sounds of the fans for a change. And btw all my computers are OC'd from stock speed, but not by much.

      I suspect people complain about stability because they can't use their PC's.
    • The NForce2 chipset in my Shuttle PC is "rock stable".

      I never have network, performance or stability problems. And rarely have problems with the sound. But I use Linux, so my drivers might not be all there. Haven't updated them in a while, too. There's no need. It just works.
    • ANY other chipset to the rock solid Intel chipsets, they look unstable. Last STABLE (rock stable) chipset on AMD platforms was AMD760. Yes, it was lacking features, but it WORKED.

      Amen, brother. Almost as if the people who design and build the CPUs have the best chance at making other complex chips working well with them.

      I've gone into small companies to try and solve their stability problem, to find Via (especially), SiS or ALi chipsets sitting there. Replace one with an Intel/AMD-based chipse

  • Being the 'proud' owner of a NVIDIA nForce4 board for AMD, I can only say that the sloppy drivers and support of NVIDIA irritate AMD owners A LOT (just google around or visit some forums like the one of NVIDIA themselves). Even early adopters hate it when hardware gets thrown on the market without proper and stable drivers... Peter
  • by Visaris ( 553352 ) on Friday April 08, 2005 @09:23AM (#12175179) Journal
    Just because this article mentions that the nForce4 for intel CPUs is unstable and has issues doesn't imply anything about the AMD nForce4 chipset. There are many major server vendors (Tyan comes to mind first) that are using the nVidia chipset. These vendors don't just slap anything into their motherboards you know. A lot of validation and testing going into every part they use. I am very happy with the stability and speed of Tyan's boards. If Tyan says it's good enough for them, then it is probably good enough for me. I don't see why people would even say somthing like: "The nVidia shipset sucks on Intel, I guess it's ok for AMD because people in that market are used to crap!" It just doesn't work that way. The Intel and AMD nVidia chipsets are very different.
  • nForce AMD (Score:4, Interesting)

    by Barny ( 103770 ) <> on Friday April 08, 2005 @09:53AM (#12175479) Journal
    Anyone who jumped on the 939 bandwagon at the start of its release will see this for what it is, teething problems.

    This is not a fault of nVidia chipsets (they are very good imho) and its not a fault of intel chips (again, very good, just showing their age a bit now).

    The stabillity problem smells heavily of the same sort of goings on gigabyte had with their initial flagship nf3-250 board, the k8nsnxp. Between bad temp sniffing (minor read error causeing the cpu fan to shut down because it thought the cpu was 20C below what it really was) and a huge problem getting the dual channel memory working, these boards were shunned. After much patching of bios' code they are rock stable and burning up memory benchmarks.

    Lets give these things 3 months on the market to get the bugs out then see what they can do :)
  • What a non-story. SLI is very cool tech, and if you get an early build of just about anything you are going to be updating the BIOS, heck I bought an Intel 865 motherboard and it was shipped with a Rev. 12 BIOS had to update it to Rev. 19 or 20 just to get it to work with the RAM I had. Sounds like ATYT FUD to me (not that NVDA doesn't FUD also)...
  • by tinrobot ( 314936 ) on Friday April 08, 2005 @10:43AM (#12175934)
    I reviewed the new HP xw9300 for a print magazine. Didn't find any stability problems, though I tested it mostly against 3D apps like Maya. Not too many network tests. I ran it in production for a while and it was great. In fact it's still here sitting next to my desk.

    We did request SLI, but HP sent a single card system because they told us SLI wasn't quite ready.

    I have another system on my review schedule from another vendor, and when we suggested they ship us an SLI system, they backed off.

    Looks like SLI isn't quite ready for prime time.
  • A chipset barely on the market for three months, sporting two brand-new technologies, is hounded for non-showstopper bugs that have already been fixed, or are scheduled to be fixed shortly.

    And yet, Intel doesn't get hounded because THEIR latest chipset saw a recall [] in the first two months of release?

    Sure, Intel is more stable, if you don't live on planet Earth. NOBODY is perfect.

    Oh, and I'd like to dispell another long-standing rumor, that 3rd-party chipset makers cause the A64 platform to be "unstable
  • As highlighted by Charlie over at the Inquirer, many hardware site reviewers receive certain 'incentives' from hardware manufacturers to NOT publicise these issues.
  • I've been an AMD user ever since the release of Athlon XP.

    I have never built a computer that didn't have an NVDIA chipset in it. I've owned NForce, NForce2, and NForce3 250GB systems, and I've built NForce4 systems.

    NForce, NForce2, and NForce3 250GB rock. No compatibility or stability issues, great drivers, and good performance. Excellent all-around.

    NForce4 blows. I've had compatibility problems with all four systems (two with ASUS boards, one with an MSI board, and one with a DFI board), and the drivers

"Yeah, but you're taking the universe out of context."