Slashdot Mirror
AMD Opteron "Hammer" Preview
Melvin Tong writes "Hardware Extreme has posted a preview of AMD's 8th-generation processor that AMD is currently developing with a few exclusive pics of the mechanical sample. AMD Athlon processors based on Hammer technology are expected to ship in the forth quater of 2002. The preview is located over at HW Extreme."
It's a pretty safe bet such systems are already in use in some TLAs, we just don't know about it. Opteron/Hammer will be a nice step forward but obviously it's not meant as competition to big iron from Sun and others for quite a while, if ever. AMD has years and years to go before they can enter such markets successfully, they are relatively small and have to concentrate on the desktop and small to mid sized servers for the
forseeable future, plus have yet to prove themselves in the higher end, attract the appropriate support and build an image they don't yet have when compared to Intel, Sun or IBM.
It also looks like they have their work cut out for them already, with not as fast clock/rampup on 0.13 micron as expected and a tight line to get the Hammer line done properly as they are pretty strung out on cash compared to Intel, while the latter seems to have no trouble in increasing clock all the time (by throwing huge gobs of money at it of course).
Strom Thurmond; the dean of the US Senate...
the deadest fart on slashdot.
I particularly liked this point:
"The AMD Opteron is designed to be scalable, reliable and compatible, which can result in lower total cost of ownership."
Gee, the whole article sounds like a lame press release. I want the real low down, not a marketing piece!
A picture of a pure copper CPU mock-up, and then a picture of an evaluation opteron. And about 4 pages of months old regurgitated AMD press releases. I wouldn't really consider this news, since AMD's been showing off the evaluation chip for a few months now.
https://www.gnu.org/philosophy/free-sw.html
As a side bonus, you can find SPEC benchmarks for Itanium and Itanium IIs on that chart (search for the word Itanium - Dell and HP have both submitted results).
Either way, It would be funny if Intel ended up having to license AMD's x86-64 technology. Even though I don't think that would happen, I suspect Intel would rather fork the 64bit platform with their Itanic (part 2) than license from AMD... but you never know!
Are you local? There's nothing for you here!
I was very happy to see the nickel cap on their new CPU. After crushing a couple of AMD chips, I became very weary of removing the heat sink after a successful mounting. More so than I probably should be, but after chipping the edge off of some $100+ CPU's, I was very nervous about picking up any of the cutting edge processors.
I look forward to lapping the cap to a shinny mirror finish!
+++ UGUCAUCGUAUUUCU
indicate the last 5 months of the year
If you're Enron or WorldCom... For most of us, the fourth quarter is the last 3 months of the year.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
Hardware Extreme has posted a preview of AMD's 8th-generation processor that AMD is currently developing...
As opposed to the 8th-generation AMD processor that Intel is developing....
(/sarcasm)
How is this a preview? This is just a preview of the marketing docs! A poorly spelt one at that.
-Spackler
PS: spelt was a joke
Clearly a blatant rip-off.
General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
I won't bother to elaborate on what several others have already mentioned, that this is a poorly edited stored pasted together from AMD press releases. The total kicker on this is the very last 'next' link takes you to a pages to buy some AMD Athlon chips!
The boundry between news and advertisement gets more porous each year...
Peace, or Not?
this picture here... on the left hand side... eh wait a minute. Sorry, just a bent pin.
Welley Corporation - SLM Scammers
This in not meant to be a troll. However, I sure as hell don't trust "sneak preview" tech specs full of typos in a article written by rumor-mongering hardware freaks half a page down from a picture of someone hitting a CPU with a giant green inflatable hammer.
Why not? The specs are cribbed from the PR kit, so they're about as trustworthy as what the company itself says.
Remember, "sneak previews" on hardware sites are like trailers in movie theaters. They're there to get you interested in the product, not to critique it.
Get off my launchpad!
Usually these previews are riddled with canned PR hype that may or may not be true. The fun part is seeing which promises come true when the product eventually hits the market, and which were totally off base.
Mmmm... nostalgia.
[PowerPoint] is a tool for capitalist presentation
I have to wonder about the lifespan of a CPU that has an integrated memory controller of any type - not just DDR, but RDRAM, or FOORAM, or NARFRAM. What happens to the family when a new RAM interface comes along?
Now, for high-integration CPUs designed for embedded style apps I can see it, but for a main-line CPU it seems to me that tying the memory controller to the CPU limits the lifespan of the design.
I realize that should POITRAM become the new speed king that the RAM controller block of the CPU can be redesigned, and I understand that putting the RAM controller in the chip can increase the memory bandwidth to the CPU.
But it does cause me to think....
www.eFax.com are spammers
These guys are designing Opteron servers, including dual Opteron 1U servers (web and render farm goodness) and quad Opteron 3U servers. Very impressive specs. The management is dominated by senior IBMers, plus a senior marketing weasel from Dell. Hmm, Dell skipped the Itanic2...
Somehow, I suspect their designs are going to get licensed by some very big vendors. Call it a hunch.
That was the longest advertisement i've ever read.
AMDZone wrote a FAQ which was a good read.
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
The NDA isn't quite up until 2400 USA (eastern? pacific?, don't ask me i don't know) time, but look at, Here
Expect reviews from the usual suspects.
AMD have modified there ratings a little so as
to keep the model numbers fair compared with
the newer faster Northwood pentium 4s. So while
the old rating system would have had 2400+ as a 1933MHz Athlon, and 2600+ as a 2066Mhz Athlon, in
fact the 2400+ is the first 2GHz Athlon while the
2600+ clocks in a 2133MHz.
We can expected newer Athlons to be released later
with 333MHz Front Side buses, and later 512MB of cache. Even when Hammer comes out, AMD will still to selling Athlons for around a year afterwoods, the Athlon will move done the low end to replace the Duron, and thats going give the celeron a real kicking. In fact Intel seems to have blown
there wad completely, with nothing to compete with
the Hammer until there Prescott strink of the
P4 in Q4 2003.
I want to see what features of Palladium have been implemented, since AMD have declared their support for it. ... the one that comes with a licensed copy of the M$'s latest and greatest.
Will the first series of Opteron prevent me from downloading mp3s, or will that be an optional extra / firmware upgrade?
Of course I expect users will be able to 'opt-out' of these new features for the next year or so, until the US government, in their infinite wisdom, decide that opt-out is no longer an option, and that there will only be one licensed implementation
That may be, but if you want to take a look at some of the serious articles on ememory & clock latency (from the CPU's perspective) you'd realize why they are adding the memory controller where they are. A 'normal' SDRAM memory controller on a VIA or AMD motherboard for instance can easily take 70+ cpu cycles before returnign the required data... So unless the cpu has other data to process (which fits into the cache) then it just sits there til it has the data requested... With a cpu built-in memory controlelr of this sort (especially if they allow tolerences for faster rated memory within the existing class) could lower the latency down to say 6 cycles...
This is great for memory intensive & system intensive tasks (from gaming to high demand servers)...
we are all invisible unless we choose otherwise
On February 28, 2002, AMD announced the support of SuSE Linux for the Opteron processors. Good news to home computer users, on April 24, 2002, Microsoft has also collaborated to further 64-bit computing. (Emphasis added.)
Yet another article implying Linux is not for the home. People read enough of these articles and they will conclude a priori Linux is not to be used in the home and never try it for themselves.
Note I'm not saying it's completely ready for home use, especially by people with extremely limited computer knowledge, but people should decide for themselves. If everything they read says or implies Linux isn't for the home, they won't even consider it an option.
I hope they dont ship them like this! (Note the bent pins on the left corner :))
CPU designs are pretty modular. It shouldn't be hard at all to swap in a new controller when the time comes. If the internal hardware interfaces weren't very clean, design would take a lot longer.
Copyright Violation:"theft, piracy"::Anti-Trust Violation:"thermonuclear price terrorism"<-Overly dramatic language.
Just don't sand off any DRM bits, or it's your ass in the slammer! The DMCA is watching you, punk.
This is only the hood, there's nothing to look at under it!
"Karma can only be portioned out by the cosmos." -Homer Simpson
How often do you upgrade the motherboard and the RAM but keep the CPU...?
As new (faster) memory becomes available, they'll simply update the memory controller on the (new) CPUs (just as they updated the FSB from 100 to 133 to 166 to 200 to 266 and soon to 333 or 400).
RMN
~~~
What happens to the family when a new RAM interface comes along?
You'd just drop in a new memory controller. Keep in mind that new memory interfaces don't come around all that often. You might get speed bumps like PC100/PC133 and the various flavors of DDR. But a single model of controller can often handle multiple speeds. Think about how many flavors of PIII/Celeron came out that used the PC66/PC100/PC133 SDR memory interface.
If this gives AMD a big performance boost, which it should, it's a good move.
That has to be the strangest looking Hammer I've ever seen. Doesn't even have a handle. On the plus side, it does seem to come with a lot of built-in nails.
RMN
~~~
I was expecting something of substance; for crists sake, we had perliminary benchmarks of the same processor months ago! All I got was a press release and a really badly done benchmark comparison("well, here's how the 800 did. For comparison, lets see what a 400 celeron did!")
Reading this truly was a waste of my time. The ad when I clicked on the final "next" link added to my frustration.
It's been a long time.
What about the criticizing of someones criticizing of someones criticizing?
Can't we all just get along?!
It's been a long time.
Generally, if you want to upgrade to a different type of memory, you want to upgrade to a different type of motherboard, and probably a different type of CPU. I don't think it's such a bad thing. "This motherboard is designed for DDR memory and a DDR compatible CPU!"
"but for a main-line CPU it seems to me that tying the memory controller to the CPU limits the lifespan of the design."
and BINGGO was his name-o
gee, then you would have to buy something even more often. Boy I bet they will cry all the way to the bank.
The Kruger Dunning explains most post on
The integrated memory controller of the Hammer series chips can be disabled and replaced with a motherboard-based memory controller. Additionally, the core of the processors were designed to make it fairly easy to swap integrated memory controllers, but "easy" is a much looser term when describing modifying a multimillion transistor multilayer CPU core.
Either way, it will not be a problem.
Computer Science is no more about computers than astronomy is about telescopes. --E. W. Dijkstra
And here's a comparison, openssl 0.9.6b (as shipped with Redhat 7.3) running on a 400 MHz
What was that about lies, damned lies and...
Australian? Join EFA
There's already a review on Ace's Hardware which concludes that the Athlon 2600+ has again leapfrogged the fastest Intel CPU. Of course, when Intel releases the 2.8GHz P4 next Monday, it will yet again leap over AMD.
Ad infinitum.
Computer Science is no more about computers than astronomy is about telescopes. --E. W. Dijkstra
A factor of four compared over the Celeron is really disappointing. It would be interesting if we know if they run the hand-coded x86 routines against GCC-compiled x86-64 code. It wouldn't be too bad, then.
or read an actual overview of the spec
"I would say that 99 per cent of what my father has written about his own life is false." - L. Ron Hubbard Jr.
I was mostly interested in the pictures, 'cause the article was terrible -- I think it was bashed together from press releases. I won't even get into the benchmarking, except to say that benchmarks should *not* compare two completely different architectures running at significantly different clock speeds with different software and OS versions. What were they trying to demonstrate?
The big image on the third page was a shocker. Ack. Sure, it's just a mechanical sample, but adding a big page showing that you bent the pins on the processor doesn't particularly add to your breathless and misleading review. Wow. That was a terrible article. If AMD wants positive press in the technical crowd, they should be giving the samples to folks who know what they're doing.
Like this you mean?
Is it just me, or does anyone else REALLY enjoy frying eggs on their computer? My current CPU isn't quite hot enough, but as soon as I get AMD's new processor, I think I'll give bacon a try... I always hear people calling computers applicances. Consider that mine runs hotter than my toaster, and eats more power than every other appliance in my house, I can see how they might get confused.
;-) ).
Seriously though... Is AMD or Intel showing any signs of reducing the power consumption and heat output of their chips? Or are they just going to gradually reduce the maximum operating temperature until you need to get a dedicated freezer just to cool your computer?
Until they get on the ball, any alternative processor suggestions? I'm willing to pay more for decent equipment, and because everything I use is in source-form, any processor will be fine. The problem is that I've never seen anything but Intel and PPC notebooks... But, even if I've got to use a different processor on a laptop than on my desktop machines, I'd be willing to. It's really time for me to change.
One hot day, I went into my BIOS and checked out the hardware section, only to find that my CPU and case were 256 degrees F, and my CPU fan was spinning at several hundred-thousand RPMs. You might instantly disregard that, but here, where room temp is often 130F, and I'd had several fans croak already, it was a coffee-spitting moment (on a related note, I need a new keyboard too
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
As I said in my previous post, I realize that the memory controller section of the chip can be redesigned. I realize that AMD would rather have a good way to sunset the current chips when a new memory design comes out. I largely wanted to make sure that I wasn't the ONLY one to realize this.
However, look at what happened during the transition from SDR to DDR, or from DDR to RDRAM - all that had to be redesigned was the external memory controller chip, which allowed the release of mobos that supported the new RAM standards fairly quickly. How quickly would they have been supported if the Celeron/Duron chips had not had external memory controllers?
Also, something that occurred to me as I slept - how do they handle memory coherency in a multiprocessor system? Does each CPU have its own memory, and they coordinate cachelines? (sort of a ccNUMA type arrangement) Or do they have a single external memory controller that all the CPUs talk to (and take the speed hit)?
If the former, that would have a pretty large impact on Linux. If the latter, then a SMP machine would take a large speed hit relative to a UMP machine due to the slower memory access.
www.eFax.com are spammers
Given recent trends, the new POITRAM would require a new chipset which would require a new processor anyway. Might as well make it a single chip.
Whoops my bad
With a cpu built-in memory controlelr of this sort (especially if they allow tolerences for faster rated memory within the existing class) could lower the latency down to say 6 cycles.
Not to be argumentative, but this IS slashdot.
I'd like to see these "serious articles" about memeory and clock latency that say that moving a memory controller from off chip to on chip will reduce latency from 70 cycles to 6.
The latency for retrieving data from main memory is an effect of current memory technology, data can only be fetched so fast from DRAM based memory. DRAM uses a capacative effect to store data and it is relatively slow especially compared to the ever faster modern processors. This is the reason for using physically more complex SRAM which stores data in much faster transistor based latches. SRAM is used for cache in modern computers.
The memory controller, which is primarily comprised of some addressing logic as well as analog stages to interface with the memory bus, must be physically positioned in between the processor and the DRAM based memory banks. Whether it is on the same piece of silicon as the CPU or on a seperate chip has only a very small effect on the latency of the CPU making requests to main memory. The reasons for positioning the controller on the die are mostly economic, and it may by a very tiny speed advantage. The IBM POWER4 processor integerates the Main memory and L3 Cache controller on the processor.
I realize I called you on your lack of references, so I should probably provide some. Unfortunately I don't know of any good web links, but I recommend reading some books on Computer Architecture and/or Computer Organization:
The Modern Computer Architecture: A common textbook in Computer Architecture/Organization classes
I'm wasn't sure which site the article was on... It was either Ace's Hardware or ArsTechnica... Unfortunately I still don't know because I haven't had the time to go check around in back articles on each site...
I do remember they were talking about the way current memory controllers (then on some VIA & AMD boards, the article itself was talking about asyncronous vs synchronous memory) and the path used to reach the memory that multiplied the base latency (those CAS/RAS latencies really) of the whole system...
having the controller on the chip & optimizing the path to the memory you drop the extra cycles it would take as the data moves thorugh the system...
we are all invisible unless we choose otherwise
Much easier (and less dangerous) to install and remove than the old clip-on type. Nearly all motherboards accept them.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10