One possibility is that Asus decided to hold back for the VIA chipset which will be out RSN - the AMD750 is an interim chipset and Asus may not have wanted to get stuck with old inventory. They may have been spooked by the chipset errata too.
OTOH, Intel hasn't done everything they could to crush AMD yet - otherwise we'd see Celery chips with 100mhz FSB - a C400/100 is faster than a 466/66 in many things.:)
That's just flamebait. For one thing, we might not even be talking about this sorta thing if it weren't for Linus... and if you look back at the beginning, he wrote Linux JUST for the x86 (x>=3) and didn't design it to be portable at all. We're not talking about NetBSD here:) I don't see any problem with Linus maintaining just 1-2 platforms, as long as there are dedicated port maintainers keeping up with it, which is true for PPC among others.
Right now only Quake III (AFAIK) supports SMP mode, in which case a dual P6 (Celery/P2/P3) >=400mhz will beat the Athlon.
Also the K7 will support the same APIC system as the Alpha, not Intel's APIC system IIRC. Also VIA puportedly is implementing an APIC in their next chipset series even for Intel (they're very nice even with 1 CPU.)
Now, a bi-athlon (or tri-athlon...) box should be really, really cool.
Re:22 million transistors?
on
Athlon Reviews
·
· Score: 1
The current mobile P2 has more than 22million transistors (for the on-die 256K L2 cache. Sound familiar?)
Humm... there should be a RealTek driver for Linux, but it is a slow chip anyway. (Grr. All the GOOD cheap NICs vanished the day Intel bought DEC Semi.) As for sound cards, I've found the Ensoniq-chip Creative boards (AudioPCI, SBPCI64(D/V)/128, _not_ PCI512) cheap ($30) and well supported. There are probably some other good cheap PCI sound cards too with native (no SB emulation) Linux drivers.
Actually the 'Basic' API isn't all that portable either - if you start from Windows NT and want to use all of it's features. For instance, I once took advantage of the 32-bit GDI coordinate space under NT4 for a map function (Considering 4Bx4B is _far_ bigger than say 1024x768, this works really well and allows smooth scrolling/zooming.) It could be modified for Win9x, but it wouldn't work nearly as well.
Same goes for many of the other cool NT features... I wouldn't be suprised if most of the things in the 1st edition of "Advanced Windows NT" wouldn't work on 95 quite right.:)
(Linux would work just as well, by the way) It was a bit hard to set up, but with the new Samba features I could even get it recognized and browsed under the domain. I orignally tried WHFC but it does not work well with machines that have multiple ethernet ports. (AARGH!) I also set up the local CD-R drive on the box, for added value.:)
I know the K6-3 can cache the full 4GB, but I'm interested in hearing about how much the AMD-related chipsets can cache... anything out there than can do >=256MB ECC for instance?:) (I've heard the FIC PA-2013 2MB can...)
Nah. Intel can't and won't do this. Intel NEEDS the volume to keep the factories up to date to produce the advanced high margin chips.
Also, they are NOT losing money on even the Celeron because their yields are very strong, even with the 18 million transistors on 'em.
Finally, they've spent _tons_ of $ on branding, etc... they wouldn't do that just to pull out of the consumer space.
As for AMD... if they can build enough Athlons, at the right speeds (having too many slow chips is bad... look at Cyrix and K6-2 350 prices:) and get chipsets and boards as good as the 440BX products from Intel et. al... they have a darned good chance of evening out the market. And that would be a Good Thing.
Yup... the P3 adds nothing to the casual Internet user, who is bandwidth limited and probably needs nothing more than a 200mhz box. SSE is really for gamers and complicated math apps. And for a casual user of those a regular Celery or K6-2 is enough...
Aargh. It ate a bit of my post. What I meant to say was that Cyrix used an LX motherboard on the Celerons below 366mhz, and a SIS 620 mobo on the faster ones. The end result is that the 333 and 433 get the SAME score because of the use of a poor chipset.:)
Something I've meant to send to The Register for a while...
If you look at the PR ratings of the Celerons that they compare to (333-433) something interesting comes out. For the So, I guess you could call the Celery 333 a PR433.:)
As for the Cyrix chips... if they had a real 433mhz chip it would be interesting, and it might even compete with the K6-3 on integer.
Now, $45 or so for a K6-2 350... now there's a deal. And you can't forget the Celerons, either, especially on a nice cheap LX or BX mobo.
Most of that free non-GPL software is in an XFree86 style license... i.e. do anything you want with it as long as you keep the (c) notice.:) So, it's even more free than GPLed code.
You can already do this with the BSD's... 'make world'. Although, this won't be as effective since for say FreeBSD since the current release version still uses GCC 2.7.2.1 for stability reasons - you need to upgrade to current.
Also, Mandrake is pre-compiled for either P5 or P6 march...
A lot of Egghead's problems stem from poor inventory presentation. It's tougher to find a product at Egghead than it is at NECX for instance.
Also their surplus side dosen't have too many great deals. The system prices are higher than Intellesale, and their parts are generally not the best stuff around.
The Onsale merger might be okay - except they're going to run a risk of annoying their retailers... they'll have to be careful.
As long as companies like ID are around I don't think this will completely happen.:) In fact John Carmack has been actively working on the Matrox G200 GLX module... it's cool to have accelerated 3D in multiple windows using a totally free, sourced solution.
Yup... from the Joe SixPack perspective DVD is copy protected by Macrovision. CSS (until it gets cracked, if that ever happens) prevents even techies from copying DVD's onto DVD RAM disks (which are currently more expensive anyway.)
However, DVD's copy protection schemes are basically user-transparent, and Macrovision is accepted (or at least tolerated) by VHS users anyway. It dosen't get in the way of Fair Use like That Other Format did.
Now, if SDMI's copy protection is NOT user-transparent a lot of people will complain. If you get enough of those ppl together the format is gonna have a REALLY tough time. (Among techies MP3 is a mostly-open format with no protection, and that's what/.ers are gonna insist on.:)
But... you can't BUY a new copy of an old version.
on
All Hail Bloatware
·
· Score: 1
This article would make sense if said companies would be willing to sell you old versions... which seems to be his take on the matter. The problem is that if you don't own copies of the old stuff (esp. before it was distributed on CD-ROM) it's a PITA to get old versions, compared to getting the current version.
Nope... Linux isn't faster for 3D gaming... _yet_. Right now it uses indirect rendering and goes through the X server, etc. Tons of context switches, lotsa memory passing, etc. Not fun. (The plus side is that you can run a 3D app on another machine and display it on something running a GLX-enhanced driver, e.g. a G200 or TNT(2))
Performance issues aside, the GLX module system is quite nifty. And XFree86 4.0 will fix that... when it comes out.
(Last comment: Something to remember is that the X system is sorta like NT 3.x's GDI in that it is a seperate process and not in-kernel. It's better in that it's network transparent though.)
Re:Athlon is one thing. Motherboards are another.
on
Carmack on the K7
·
· Score: 1
I'll blame Intel for this one - the AGP specs were kept a little bit closer-in, in comparison to the VERY widely open PCI specs*. Also, as the LX/BX chipsets were the first ones out of the gate, a lot of graphics companies only tested with the Intel chipsets. Sigh.
Other than that, IMO the Super 7 chipsets are pretty nifty. (So is the 440BX, though:)
(* - this is why I called Micro Channel a marketing blunder... IBM failed to sell it to other clone vendors. Intel et. al. OTOH made it easy for anyone to get PCI onto their motherboards, etc.)
Slashdot Mirror
One possibility is that Asus decided to hold back for the VIA chipset which will be out RSN - the AMD750 is an interim chipset and Asus may not have wanted to get stuck with old inventory. They may have been spooked by the chipset errata too.
OTOH, Intel hasn't done everything they could to crush AMD yet - otherwise we'd see Celery chips with 100mhz FSB - a C400/100 is faster than a 466/66 in many things. :)
That's just flamebait. For one thing, we might not even be talking about this sorta thing if it weren't for Linus... and if you look back at the beginning, he wrote Linux JUST for the x86 (x>=3) and didn't design it to be portable at all. We're not talking about NetBSD here :) I don't see any problem with Linus maintaining just 1-2 platforms, as long as there are dedicated port maintainers keeping up with it, which is true for PPC among others.
Guess they got /.'ed. :)
Right now only Quake III (AFAIK) supports SMP mode, in which case a dual P6 (Celery/P2/P3) >=400mhz will beat the Athlon.
Also the K7 will support the same APIC system as the Alpha, not Intel's APIC system IIRC. Also VIA puportedly is implementing an APIC in their next chipset series even for Intel (they're very nice even with 1 CPU.)
Now, a bi-athlon (or tri-athlon...) box should be really, really cool.
The current mobile P2 has more than 22million transistors (for the on-die 256K L2 cache. Sound familiar?)
Still a nice box, esp. with a G3 upgrade. :)
Humm... there should be a RealTek driver for Linux, but it is a slow chip anyway. (Grr. All the GOOD cheap NICs vanished the day Intel bought DEC Semi.) As for sound cards, I've found the Ensoniq-chip Creative boards (AudioPCI, SBPCI64(D/V)/128, _not_ PCI512) cheap ($30) and well supported. There are probably some other good cheap PCI sound cards too with native (no SB emulation) Linux drivers.
Actually the 'Basic' API isn't all that portable either - if you start from Windows NT and want to use all of it's features. For instance, I once took advantage of the 32-bit GDI coordinate space under NT4 for a map function (Considering 4Bx4B is _far_ bigger than say 1024x768, this works really well and allows smooth scrolling/zooming.) It could be modified for Win9x, but it wouldn't work nearly as well.
Same goes for many of the other cool NT features... I wouldn't be suprised if most of the things in the 1st edition of "Advanced Windows NT" wouldn't work on 95 quite right. :)
(Linux would work just as well, by the way) It was a bit hard to set up, but with the new Samba features I could even get it recognized and browsed under the domain. I orignally tried WHFC but it does not work well with machines that have multiple ethernet ports. (AARGH!) I also set up the local CD-R drive on the box, for added value. :)
I know the K6-3 can cache the full 4GB, but I'm interested in hearing about how much the AMD-related chipsets can cache... anything out there than can do >=256MB ECC for instance? :) (I've heard the FIC PA-2013 2MB can...)
Nah. Intel can't and won't do this. Intel NEEDS the volume to keep the factories up to date to produce the advanced high margin chips.
Also, they are NOT losing money on even the Celeron because their yields are very strong, even with the 18 million transistors on 'em.
Finally, they've spent _tons_ of $ on branding, etc... they wouldn't do that just to pull out of the consumer space.
As for AMD... if they can build enough Athlons, at the right speeds (having too many slow chips is bad... look at Cyrix and K6-2 350 prices :) and get chipsets and boards as good as the 440BX products from Intel et. al... they have a darned good chance of evening out the market. And that would be a Good Thing.
Yup... the P3 adds nothing to the casual Internet user, who is bandwidth limited and probably needs nothing more than a 200mhz box. SSE is really for gamers and complicated math apps. And for a casual user of those a regular Celery or K6-2 is enough...
Aargh. It ate a bit of my post. What I meant to say was that Cyrix used an LX motherboard on the Celerons below 366mhz, and a SIS 620 mobo on the faster ones. The end result is that the 333 and 433 get the SAME score because of the use of a poor chipset. :)
Something I've meant to send to The Register for a while...
If you look at the PR ratings of the Celerons that they compare to (333-433) something interesting comes out. For the So, I guess you could call the Celery 333 a PR433. :)
As for the Cyrix chips... if they had a real 433mhz chip it would be interesting, and it might even compete with the K6-3 on integer.
Now, $45 or so for a K6-2 350... now there's a deal. And you can't forget the Celerons, either, especially on a nice cheap LX or BX mobo.
Most of that free non-GPL software is in an XFree86 style license... i.e. do anything you want with it as long as you keep the (c) notice. :) So, it's even more free than GPLed code.
Humm... I wouldn't be suprised if Acer wanted their own CPU company. They already make almost everything else. :)
You can already do this with the BSD's... 'make world'. Although, this won't be as effective since for say FreeBSD since the current release version still uses GCC 2.7.2.1 for stability reasons - you need to upgrade to current.
Also, Mandrake is pre-compiled for either P5 or P6 march...
A lot of Egghead's problems stem from poor inventory presentation. It's tougher to find a product at Egghead than it is at NECX for instance.
Also their surplus side dosen't have too many great deals. The system prices are higher than Intellesale, and their parts are generally not the best stuff around.
The Onsale merger might be okay - except they're going to run a risk of annoying their retailers... they'll have to be careful.
As long as companies like ID are around I don't think this will completely happen. :) In fact John Carmack has been actively working on the Matrox G200 GLX module... it's cool to have accelerated 3D in multiple windows using a totally free, sourced solution.
Otherwise we'd all be paying cash for everything, and not get discount cards. And NOBODY would be a Costco member :)
Yup... from the Joe SixPack perspective DVD is copy protected by Macrovision. CSS (until it gets cracked, if that ever happens) prevents even techies from copying DVD's onto DVD RAM disks (which are currently more expensive anyway.)
However, DVD's copy protection schemes are basically user-transparent, and Macrovision is accepted (or at least tolerated) by VHS users anyway. It dosen't get in the way of Fair Use like That Other Format did.
Now, if SDMI's copy protection is NOT user-transparent a lot of people will complain. If you get enough of those ppl together the format is gonna have a REALLY tough time. (Among techies MP3 is a mostly-open format with no protection, and that's what /.ers are gonna insist on. :)
This article would make sense if said companies would be willing to sell you old versions... which seems to be his take on the matter. The problem is that if you don't own copies of the old stuff (esp. before it was distributed on CD-ROM) it's a PITA to get old versions, compared to getting the current version.
By not offering the source :)
Seriously, in this instance Linux and MS products have to compete on equal terms... it'll be interesting to see what happens.
Nope... Linux isn't faster for 3D gaming... _yet_. Right now it uses indirect rendering and goes through the X server, etc. Tons of context switches, lotsa memory passing, etc. Not fun. (The plus side is that you can run a 3D app on another machine and display it on something running a GLX-enhanced driver, e.g. a G200 or TNT(2))
Performance issues aside, the GLX module system is quite nifty. And XFree86 4.0 will fix that... when it comes out.
(Last comment: Something to remember is that the X system is sorta like NT 3.x's GDI in that it is a seperate process and not in-kernel. It's better in that it's network transparent though.)
I'll blame Intel for this one - the AGP specs were kept a little bit closer-in, in comparison to the VERY widely open PCI specs*. Also, as the LX/BX chipsets were the first ones out of the gate, a lot of graphics companies only tested with the Intel chipsets. Sigh.
Other than that, IMO the Super 7 chipsets are pretty nifty. (So is the 440BX, though :)
(* - this is why I called Micro Channel a marketing blunder... IBM failed to sell it to other clone vendors. Intel et. al. OTOH made it easy for anyone to get PCI onto their motherboards, etc.)