Slashdot Mirror

http://www.amd.com/us-en/Weblets/0,,7832_11104_111 05,00.html?redir=CPVP01/

Here's what AMD says about the new security features. Lots of hype on this page but not much info.

"Protects against certain classes of viruses" ... any guesses as to which classes these might be?

Pentium 4 a weapon? Well, I need no stinkin' Pentium 4. Soon, I'll be happy with an AMD64 Opteron made in Europe . Meanwhile a PDA, katana and a crossbow are quite sufficient for my arsenal.

BTW, here I didn't seen U.S. made computer components or other electroncs for more than a decade. Everything from chips to cards, boards and displays is made in Taiwan, China, Malaysia, Korea, Ireland. Including Intel CPUs.

Yes, but a) this has caused no known problems and b) there will be a BIOS fix soon.

Interestingly, this isn't listed on AMD's errata page, which appears to be two years out of date.

I was able to find the Opteron/Athlon 64 Revision Guide from a post on one of AMD's online forums. It looks like the forums are pretty responsive, from what I saw.

Yes, but a) this has caused no known problems and b) there will be a BIOS fix soon.

Interestingly, this isn't listed on AMD's errata page, which appears to be two years out of date.

I was able to find the Opteron/Athlon 64 Revision Guide from a post on one of AMD's online forums. It looks like the forums are pretty responsive, from what I saw.

All the modern x86 CPUs I know of have bugs. It's a matter of how serious they are and whether you can fix them without a recall.

See: Prescott bugs
Itanium bugs
Opteron bugs

Just do a search of errata and the cpu you're interested in.

Most people won't encounter these bugs because the CPU makers would have tested the CPUs on a superset of what most people do. And nowadays most people don't write new code and of the code that people write, most of it is actually written by compilers, so genuinely novel machine code could be quite rare.

Thus it would be pretty poor QA to release something with a bug that prevents _boot_up_ in a significant percentage of cases. While most people may not do some rare computer operation followed by some weird stuff that doesn't do anything really useful 99.999% of the time, most people do have to _boot_ their PCs at least once in a while.

Maybe the issue did not show up on Intel's reference boards and only shows up on some 3rd party boards. If it actually showed up on Intel's boards then it is a very bad sign.

So far Intel and AMD's x86 CPUs have been pretty OK in my experience. Sun's UltraSPARC IIs were/are crap tho - 2nd lev cache probs.

I've been researching chipsets for digital TV. Here are my links to current hardware products:

STMicroelectronics System on Chip (2) Get Linux here
ATI Xilleon 220 (Products)
Sigma Designs Digital Media Processors (Products)
IBM PowerPC405 STBxx (Zarlink [2], Araneo)
Texas Instruments DM642 DSP (i3 Mood Box , X-Designs Flikit + Softier MediaLinux)
NEC EMMArchitecture2 (Galaxis + LinuxTV , PRISMIQ + Linux)
Equator Technologies BSP-15 boards
Via CN400 (Mini-ITX Board), PM800 and PM880 (w/ HDTV for Pentium 4) , ShowShifter HMN, Soyo Multimedia Ready Motherboard (with TV Tuner, $129.99)
Toshiba TX System RISC (MontaVista Linux)
Windows chipsets:
Intel 815 VisionPlus terrestrial box (Korean OEM)
AMD Geode (CoCom)
ARM (Samsung, etc.)
Digeo X-Stream (Paul Allen company)

check this out

and this

Will motherboard manufacturers come out with AMD-compatible boards that sport PCI-Express and the other (non-CPU) new features that are talked about in this article?

PCI-X and PCI-Express

From that link:

AMD expects PCI-Express to be adopted first as the next generation of graphics technology, replacing AGP 8X. In 2004, AMD expects PCI-Express graphics to debut in workstations, while predicting that PCI-X 266 will be strongly adopted in servers for higher-speed I/O card capabilities. Because the success of PCI-X 533 will likely be gated by the availability of 533MHz adapter cards and devices, AMD believes that servers requiring I/O bandwidth in excess of PCI-266's capabilities will begin to feature PCI-Express options in the second half of 2004 and into 2005.

As for DDR2, just remember that Opteron and Athlon64 have on-die memory controllers.

1) The fastest possible CPU, in *true* GHz, not in AMD's inflated "+" bogoghz.

No problem. AMD already publishes the true clock speeds of all their CPU's. The "3400+" or whatever you've seen is a model name, not a measurement of clock speed but rather of performance. AMD explains it here. Your post suggest that you are unaware of the fact that other things than clock speed have a significant impact on the performance of a CPU.

Next you'll be complaining that car makers name their cars cryptic things like "320Ci", "XC90" or "GT40" instead of naming each car according to its BHP rating.

Seasonic Super series power supplies. My UPS load meter registered a ~15% drop in PC power consumption after I switched to these from Antec. Highly, highly recommended.

Also, use AMD 64-bit CPUs and set /sys/devices/systsem/cpu/cpu0/cpufreq/scaling_sets peed to match the power/performence balance you think is best. See the Athlon 64 Processor Power and Thermal Data Sheet. For example, a current top-of-the-line Athlon 64 3800+ burns 89W at 1.5V at maximum (better than Intel, but still a lot). If you lower the clock speed by 200MHz, the chip burns 72W @ 1.4V, another 200MHz lower burns 53W @ 1.3V, and another 200MHz lower burns 39W @ 1.2V. You can cut it all the way back to 22W max, 1000MHz @ 1.1V. With the current Fedora Core 2 kernel and a power management daemon like powernowd the speed will be adjusted automagically, but if you want to run Folding @ Home without excessively spiking your electric bill it's nice to set a fixed speed manually.

The Mobile Athlon 64 3200+ (62W @ 1.4V max) is interesting if you really want to limit power consumption. I put one in my ASUS K8V Deluxe motherboard (Zalman CNPS7000A-AlCu heatsink, be VERY careful not to overtighten it and crack the unprotected core as there's no protective aluminum lid like on the desktop CPUs, not all heatsinks will fit). Drop 200MHz and get 46W, another 200MHz gets 34W, and at 800MHz a mere 13W. Given that the new Prescott-core Pentium 4's burns well north of 100W, this is pretty neat. Note that since AMD's transistors have a MUCH lower leakage level than Intel's (20% versus 50%) your idle power consumption at any clock rate is going to be pretty low. Things will get even better when the new 90nm chips come out in a few months.

Agreed. Why stop there. You could also technologies like AMD's Cool n' Quiet.

Video card manufacurers could also implement somthing similar.

I currently have 2 machines at home: A low speced linux box for daily duties (mail, web browsing, homework) and a higher spec
Pentium 4 with all the high power goodies just for gaming.

IO-0--1--2--3-IO
| |\/| |
| |/\| |
IO-4--5--6--7-IO

They are identicle in features (dual channel) except the 939 lacks one pin. That pin just happens (wink) to be a HyperTransport link that was removed.

That is not at all correct. The two sockets are actually significantly different from one another. You can find a comparison of the two layouts of the two chips bellow:

Socket 939 description.

Socket 940 description.

As you can see looking at the pin-outs (pages 26 and 22 respectively) there are a LOT more than just one pin that has changed.

This means there are not enough links to support multi-processor setups because you can't have the links to the other processors because you don't have enough.

Dual-core chips don't use any external hypertransport links anyway, so this is a non-issue. Everything is handled inside the chip. As such, a dual-core Socket 939 chips is DEFINITELY possible, though whether it will actually happen or not is another question.

This is basically a marketing move to segment the workstation market from the desktop market. There is no (techincial) reason they couldn't have used socket 940 for dual channel desktop processors.

Actually there is a good technical reason why AMD introduced Socket 939 for desktop processors, they have a LOT more power and grounding pins. This allowed motherboard makers to design 4-layer motherboards using this socket while Socket 940 required more expensive 6 or 8-layer motherboards.

So if you chop a HT link off the Opteron, you get two. That should be enough to connect to the chipset and one other processor, right? Shouldn't a 939 be technically capable of dual processor configurations (although AMD has certainly disabled it)? Just wondering.

If you have a look at the pin-out diagrams you'll see that AMD has completely removed all the pins for two of the three hypertransport links and replaced them primarily with power and grounding pins on Socket 939 vs. the old Socket 940. As such there really and truly is only one hypertransport link on this new socket and it is not capable of classic SMP configurations.

However, as mentioned above, these dual-core designs are handle all inside the chip and only require a single HT link to the outside world, so they could certainly work in a socket 939 product. Whether such a product ever makes it to market or not is another question, though there are some hints that AMD might release a dual-core Athlon64 FX chip in socket 939.

They are identicle in features (dual channel) except the 939 lacks one pin. That pin just happens (wink) to be a HyperTransport link that was removed.

That is not at all correct. The two sockets are actually significantly different from one another. You can find a comparison of the two layouts of the two chips bellow:

Socket 939 description.

Socket 940 description.

As you can see looking at the pin-outs (pages 26 and 22 respectively) there are a LOT more than just one pin that has changed.

This means there are not enough links to support multi-processor setups because you can't have the links to the other processors because you don't have enough.

Dual-core chips don't use any external hypertransport links anyway, so this is a non-issue. Everything is handled inside the chip. As such, a dual-core Socket 939 chips is DEFINITELY possible, though whether it will actually happen or not is another question.

This is basically a marketing move to segment the workstation market from the desktop market. There is no (techincial) reason they couldn't have used socket 940 for dual channel desktop processors.

Actually there is a good technical reason why AMD introduced Socket 939 for desktop processors, they have a LOT more power and grounding pins. This allowed motherboard makers to design 4-layer motherboards using this socket while Socket 940 required more expensive 6 or 8-layer motherboards.

So if you chop a HT link off the Opteron, you get two. That should be enough to connect to the chipset and one other processor, right? Shouldn't a 939 be technically capable of dual processor configurations (although AMD has certainly disabled it)? Just wondering.

If you have a look at the pin-out diagrams you'll see that AMD has completely removed all the pins for two of the three hypertransport links and replaced them primarily with power and grounding pins on Socket 939 vs. the old Socket 940. As such there really and truly is only one hypertransport link on this new socket and it is not capable of classic SMP configurations.

However, as mentioned above, these dual-core designs are handle all inside the chip and only require a single HT link to the outside world, so they could certainly work in a socket 939 product. Whether such a product ever makes it to market or not is another question, though there are some hints that AMD might release a dual-core Athlon64 FX chip in socket 939.

Here they are.

Or you may prefer AMD-64, here.

You are our new Slashdot champion. Lets' break down your +5 score:

+1 for mentioning Linux.
+1 for introducing it in a clever, offhand way (a link to kernel.org which will associate the linking phrase in google, YOU SLY DOG YOU!)
+1 for advocating a return to client/server computing in some form
+1 for using "open" and "hardware" in the same sentence.

Good job! I'd shake your hand if I could reach it.
Now for the bad news:

You don't know what the FUCK are you talking about.
The Sempron is going to be 100% like the Duron. Is the Duron in ANY way more suited then any other processor for thin client computing? AMD's new server line of Opterons with lower voltage cores might be a starting point. The Pentium-M might be a starting point. The Sempron is NOT a starting point.

Oh wait, AMD already sells an embedded x86-compatible processor Am5x86! And, WHAT'S THIS?!?! It even sells a miniature integrated platform, called Elan.

Well tie me up and violate me with a spoon. Do you enjoy being such a karma whore or are you just that ignorant?

You are our new Slashdot champion. Lets' break down your +5 score:

+1 for mentioning Linux.
+1 for introducing it in a clever, offhand way (a link to kernel.org which will associate the linking phrase in google, YOU SLY DOG YOU!)
+1 for advocating a return to client/server computing in some form
+1 for using "open" and "hardware" in the same sentence.

Good job! I'd shake your hand if I could reach it.
Now for the bad news:

You don't know what the FUCK are you talking about.
The Sempron is going to be 100% like the Duron. Is the Duron in ANY way more suited then any other processor for thin client computing? AMD's new server line of Opterons with lower voltage cores might be a starting point. The Pentium-M might be a starting point. The Sempron is NOT a starting point.

Oh wait, AMD already sells an embedded x86-compatible processor Am5x86! And, WHAT'S THIS?!?! It even sells a miniature integrated platform, called Elan.

Well tie me up and violate me with a spoon. Do you enjoy being such a karma whore or are you just that ignorant?

For thin clients, AMD has both the x86 Geode (acquired from NatSemi, basically an x86 SoC), and the Alchemy line if you'd prefer something in MIPS.

The Sempron is, as others have noticed, the 64-bit Duron (erm, in the same way that 'Duron' and 'Celeron' were budget models or binnings of modern Athlon and Pentium cores), so heat and power needs will compare to AMD's current chips. (Meanwhile, AMD's current chips are actually doing great in that department versus the P4, so yes, if you can really think of a reason to need GHz of performance in such a machine, a Sempron might just be cheaper and 'better' than any of the truly embedded stuff.)

For thin clients, AMD has both the x86 Geode (acquired from NatSemi, basically an x86 SoC), and the Alchemy line if you'd prefer something in MIPS.

The Sempron is, as others have noticed, the 64-bit Duron (erm, in the same way that 'Duron' and 'Celeron' were budget models or binnings of modern Athlon and Pentium cores), so heat and power needs will compare to AMD's current chips. (Meanwhile, AMD's current chips are actually doing great in that department versus the P4, so yes, if you can really think of a reason to need GHz of performance in such a machine, a Sempron might just be cheaper and 'better' than any of the truly embedded stuff.)

Soon this should be remedied (sorta) with FarCry for AMD64

Won't be much variety, but could be a good starting point for benchmarking.

I'm pretty sure wake-on-lan is possible

It is - see, for example, a knowledgebase article on it - but that's "wake on magic packet" (or Magic Packet(TM)) wake-on-LAN, not the more general packet matching wakeup that some network interfaces support.

I.e., the machine won't automatically wake up when you try to ssh into it; you need to send it a Magic Packet(TM) to wake it up. A packet-matching wakeup might be able to match incoming unicast packets to the machine, broadcast ARP requests asking for the MAC address corresponding to the machine's IP address, and other packets that it would need to respond to, so that attempting to ssh into it would wake it up, without making it respond to various random broadcasts and multicasts for which it wouldn't have to wake up (e.g., a broadcast ARP request for somebody else's MAC address, assuming it doesn't have to reply to that for e.g. proxy ARP purposes).

However, wake-on-Magic-Packet(TM) might be sufficient for the purposes of the person to whom you responded; I think one purpose for which it was intended was to allow administrators to wake up sleeping machines in order to do various remote administrative operations - including the remote software updates that they wanted to do.

The article also mentions a MIPS chip AMD plans to put out to be targeted at the Handheld PC market. Imagine a 1.4ghz Pocket PC?

That platform is called AU and it's been out for quite a long time now - albeit a lot slower. These are aimed at the embedded market. I've worked with those and they're quite usable System-on-Chip (SoC) devices.

The development board I had had a AU1500 I think, complete with mini-PCI and all the other interfaces you could possibly want. It was clocked at something like 300-400 MHz, but in the embedded area, that usually is not as important as you might think. In fact, often enough CPUs get their clock reduced to minimize heat development and EM interferences produced by the device.

Because of that and many other things I doubt there's many applications where speed is the biggest issue. Features (many interfaces, a good number of general-purpose IOs [GPIOs]) are an important part of the success too. Having a great many interface components implemented in the CPU itself means that designing your own carrier board is much easier and faster, and thus also a lot cheaper.

Memory bandwidth doesn't tend to scale up with CPU speed, so while you can expect a linear speed increase for executing instructions in cache, most applications are going to be hitting system memory a lot and dragging performance down.

Check AMD's white paper on XP product numbering; you'll see they actually base their numbers on a wide range of benchmarks to try to give customers a number which actually reflects performance fairly well; that's important when, say, they increase the amount of on-die cache, as with the Barton; a 2500+ Tbred has a higher clockrate than a 2500+ Barton -- can you think of a clearer way of showing that their performance is largely the same?

Benchmarks used to determine model numbering.

If you take a look at the benchmark suite used to come up with the performance ratings, you'll see a few highly-memory-performance-dependent benchmarks, whose results would not have scaled much with a higher clock speed.

No Athlon SMP, sorry.

What about the Athlon MP?

64 bit linux drivers have been out for the nforce 3 since last year. You can grab them here. Opteron 250's official AMD pricing is $851. Street pricing is near $1,000 only because availability is still low. As more vendors pick them up they will drop considerably.

The other models - 2800+, 3000+, and 3200+ are rated to operate at 65W while the 2700+ is rated to operate at 35W.

AMD Athlon 64 Processor Power and Thermal Data Sheet

Where did you find the data sheet for the new low power cores? I couldn't locate it.

Upstarts can overtake the entrenched Powers That Be!

AMD was founded in 1969. So it's an upstart that only took 35 years to lead the week in chip sales. :)

Then Sun has not much left to offer but nice support plans.

And really, really nice 64bit hardware. But, you're right, that's going to have major competition in the years going forward.

I like sun. I think there will still be a niche market for them for quite some time (they're not dying, blah blah). But, I do think they need to innovate something. If they open solaris, it's going to keep a lot of people on solaris that might have switched and just been willing to deal with mediocre hardware, but it will also probably mean that a lot of stability could find its way into linux. So, yeah, you're right. This + something new and cool could be a good business move.

~Will

Why waste all the power on a Via C3 (multiple watts) when you could use an AMD Alchemy Au1550, which consumes less than 1 watt? The development board is MUCH smaller than any uATX-like form factor.

Since when did frequency determine how fast a cpu was?

And more from Intel:
http://www.intel.com/research/silicon/leadfree.htm

And more information from AMD:
http://www.amd.com/us-en/Processors/TechnicalResou rces/0,,30_182_4040,00.html

Disclosure: I don't work for, or own stock in AMD or Intel. I haven't purchased an Intel chip since the Pentium came out.

-ted

The page you linked to (repeating here) contains only documentation regarding instructions and programming the processor, NOT the processor's architecture.

Stop misleading people. Your post does nothing other than that it shows how incompetent you are...

Although there were rumors about an Intel Yamhill 64-bit x86 part for many years, they didn't announce an 64-bit x86 architecture extension until February 18, 2004, and it was announced sheepishly as a very minor point in a press release rather than amid great fanfare as AMD had done. Intel still has not released any product incorporating this extension. Thus they've had more than 3 1/2 years to develop their own 64-bit x86 based on the AMD specifications. No need whatsoever for reverse-engineering. In fact, reverse engineering would have taken much longer, because they would have had to wait to get their hands on working AMD silicon.

Or maybe, just maybe, Intel simply downloaded the tech docs off AMDs website..

http://www.amd.com/us-en/Processors/TechnicalResou rces/0,,30_182_739,00.html

Whoa..

Barring that, Intel could have simply browsed to AMD's web page and downloaded it themselves.

In Slashdot Utopia we could mark this article as "-1, Yellow Journalism".

AMD has been working on going lead free for a while (apparently since 2001).

AMD has been working on going lead free for a while (apparently since 2001).

Moore's first law is a two-edged sword - more transistors for the same price is great for computers, but it's hell on batteries: As the processor power doubles, the power consumption also rises.
The amount of computation done per watt also rises with each generation - an AMD Opteron at 500MHz would use under 10 watts, or an amount similar to an original Pentium.

But that alone won't do it. We need to improve system layouts and cooling techniques.
Better cooling won't reduce power - it means you can burn MORE power without getting hotter. It doesn't help your battery life.

We must create better interconnects, reduce sloppy software code, eschew processors that are faster than necessary, and, of course, build better batteries.
What is wrong with things like AMD's PowerNow!, AMD's Cool'n'Quiet, or Intel's SpeedStep? They all reduce the power consumption of a processor when it isn't heavily loaded.

Moore's first law is a two-edged sword - more transistors for the same price is great for computers, but it's hell on batteries: As the processor power doubles, the power consumption also rises.
The amount of computation done per watt also rises with each generation - an AMD Opteron at 500MHz would use under 10 watts, or an amount similar to an original Pentium.

But that alone won't do it. We need to improve system layouts and cooling techniques.
Better cooling won't reduce power - it means you can burn MORE power without getting hotter. It doesn't help your battery life.

We must create better interconnects, reduce sloppy software code, eschew processors that are faster than necessary, and, of course, build better batteries.
What is wrong with things like AMD's PowerNow!, AMD's Cool'n'Quiet, or Intel's SpeedStep? They all reduce the power consumption of a processor when it isn't heavily loaded.

Short review:

Athlon 64's are THE gaming CPU. 3000+ to 3400+. Heat isn't a problem. UT2004, Far Cry, etc. have 64-bit editions coming out and WinXP 64-bit is alledged to ship when SP2 does (late June). Fedora Core 1 for AMD64 runs with the "idle=poll" kernel option workaround (HP has been alerted and will fix this in the BIOS).

The 15.4" 1680x1050 widescreen is stunning.

The 802.11g wireless range is extremely impressive thanks to the two antenna panels behind the screen.

Bluetooth works well with Microsoft's keyboard/mouse kit. Buy AA NiMH rechargable batteries for at least the mouse, you'll need them.

Quadspeed DVD burners are available, but DVD-ROM/CD-RW was the best available when I bought.

Only 4200RPM drives are offered, but swapping in a 7200RPM Hitachi 60GB drive was trivial. They include a full WinXP Home CD and separate driver and application CDs. The drivers install in 1 shot. VERY nice.

The 12-cell battery, combined with AMD PowerNow!, make for excellent battery life.

Cluestick-beating-worthy offense: they used a nVidia GeForce 440 Go 64MB video chip. On what is marketed as a gaming notebook. And it's not on a separate card so you can't upgrade. ARGH!!! This chip is fine for strategy games but it's not up to playing UT2004 at full detail.

Memory: buy from HP, buy Micron-based SODIMMs (Crucial.com is sending out Infineon-based SODIMMs lately that DO NOT WORK on Athlon 64 notebooks when paired with a second SODIMM, nor do Corsair SODIMMs), or buy from somewhere you can easily return memory to. Kingston is supposed to work. I'm working with HP to run this down. More here. You can get to the SODIMM slot that's under the keyboard with minor effort when you know how (see the previous link).

In summary: GREAT notebook for getting work done, will be great for Linux once the BIOS is fixed and it's workable now, so-so for gaming. Serious gamers will probably want eMachines' notebooks instead, or wait for Athlon 64 notebooks from ASUS, Acer, etc. If you want thin-and-light, the Athlon 64 2700+ 1.2V 35W CPU (see page 20 here) looks ideal, if there are any OEMs out there with enough brains to use it.

No, they still offer wireless 802.11b low power chips.

I don't know if you're talking about a different base station, but there was a story here on slashdot last year about a guy who pulled his apart. It had an AMD Au1500 CPU, a MIPS-compatible processor. Not x86 compatible by any stretch of the imagination :)

This is an example of a newer 802.11b card that uses the AMD AM1772 chipset.

AMD makes a lot of differnet types of semiconductors. Not as diverse as Ti, but they're not trying to compete with Intel across all markets or anything.

> C3? 8.5 watts still needs cooling.

I've got a passively cooled 533 MHz Via Eden. Heatsink only. I think the newest Via in the Nano-Itx
boards don't need fans either.

> These don't need heatsinks.

Via is popular because it democratized low power solutions with Mini-itx.A guy on the street cannot build a system using the AMD chips you pointed to.

Um, there are plenty of vendors offering tested AMD dual CPU 1U setups. And NO AMD cpu's do NOT need more power than Intel, Intel Xeon MP 3.2GHz Themal Guidline is 92W from Intel, AMD Opteron 2.2GHz Thermal Design Power is 89W from AMD. And to give you an idea where Intel is headed their 90nm 3.4GHz P4 has a Thermal Guidline of 103.0W with industry people stating that the 90nm line will end up north of 150W!

C3? 8.5 watts still needs cooling.
These don't need heatsinks.

As covered by arstechnica, there are also reviews at [H]ardOCP, Hexus, HotHardware.com, Sudhian, and The Tech Report. AMD's official announcement is here.

Why does this incorrect info keep getting posted (and modded as "informative" at that)? AMD stated several times quite publicly that their rating initially was meant to compare against the "Thunderbird" Athlon chips. More recently they've simply said that it's relative performance between the AthlonXP line and that it can "outperform it's closest competitors". Here's a direct quote from AMD's AthlonXP FAQ

Q: What does the 3200+ model mean?

A: This is a model number. AMD identifies the AMD Athlon XP processor using model numbers, as opposed to megahertz. Model numbers are designed to communicate the relative application performance among the various AMD Athlon XP processors. As additional evidence that performance is not based on megahertz alone: the AMD Athlon XP processor 3200+ operates at a frequency of 2.2GHz yet can outperform an Intel Pentium(R) 4 processor operating at 3.0GHz with an 800 FSB and HyperThreading on a broad array of real-world applications for office productivity, digital media and 3-D gaming.

AMD's model numbers not rated against Intel's P4 chips? You might want to tell AMD that!