Slashdot Mirror
AMD Officially Rolls Out 1Ghz Athlon
spudwiser writes: "AMD has a press release on their Web page concerning shipment of the 900, 950, and 1000MHz Athlon processors. Also included are times for the live satellite interview with the CEO and VP of AMD." Check out some of the benchmarking info about the new chips as well. I wonder how Andy Grove [?] is feeling today.
All that time I spent drooling over 4mhz systems, hoping to someday be able to afford 64k of memory . . .
.
And now, the "insignificant difference" in the *cache* speed is twice the speed of my dreams, and that cache holds more memory than a room full of computers or a box of disks . .
Pass the Geritol and my cane, please . . .
I agree that the 33Mhz 32bit PCI bus can be a limitation for some things, like trying to push a gigabit ethernet at a gigabit. I don't see this as the big problem in PC class machines just yet. Maybe in two years or so.
First off you can get 66Mhz 64bit (or maybe just 33Mhz 64bit) PCI slots in motherbords that cost under $300. That is 2x or 4x the bandwidth. So if 33Mhz 32bit PCI is a "little limiting", 66Mhz 64bit PCI should have a lot of head room.
Secondly, there are very very few things you hook up to a machine that push past the PCI bandwidth (the gigabit ethernet being on of the few, an extreamly fast RAID controler being another, and it is only a little past the 33Mhz 32bit PCI bandwidth). So you need a case where you push multiple devices to their limit before PCI is the limiting factor (like I want to read off my massave disk array and send it out these two GE ports...). (3D video cards being a modest execption, because they allready have their own "bus")
Thirdly, and more important, it is main memory that is the bottleneck for most things. It has been for over a decade. It isn't getting better (main memory is getting faster, but not as fast as CPUs are).
From a technical point of view what is wrong with RAMbus? It is 1992 technology that was great then, and great for five years after. The compony that invented it charged too much for patent rights, never managed to hit the mass market, didn't get to drive costs down, and in the end didn't make enough money to plow it back into research to keep up with the entire rest of the world. But that isn't a technological problem, it is an econmic one. More importnatly RAMbus is extreamly intresting.
Plus if you want bandwith, and a little latency is Ok RAMBus kicks ass. Unfortunitly bandwith isn't really as useful as low latency for most applications, and at RAMBus's current price it is cheeper to build very wide normal memorys.
Oh, and RAMBus and PCI don't play in the same space. One is a memory system, the other is a perphrial system. It is like saying "Volvo's T5 engine is out dated, they should devlop something like the VW transmition!".
They did make something like the SBus. It is called PCI. On paper it uses the same FCode (in practice is normally uses Intel x86 assembly). It has vender and device IDs. It has auto config. It is fast (the SBus is 32bits at about 25Mhz, slightly slower then the PCI bus). PCI abandoned the SBus form factor. PCI abondened using strings for dev/manuf IDs. PCI dropped the IOMMU (which was re-invented for AGP as "Intel's innovatave new DIME"). But PCI is relitavly similar, and being designed afterwords it is even somewhat better in some ways.
Sun has dropped the SBus over the last few (3?) years. They are using PCI. Not just in the low end "PC priced" machines, but even their most expensave machines use PCI for perphrials (they use their own thing for memory and CPU boards, no big supprise).
I'm afarid you are wrong. The SBus is indeed a bus. many high end SPARC systems had multiple SBus's hooked up over what may have been a point-to-point controler (like the FHB), but i think that was just a bigger bus. Existing high end SPARCs hook backplane bords up over a point to point system, some backplane bords have CPU, others PCI busses, others memory. SGI does the same thing.
You may be thinking of Intel's NGIO (Next Gen I/O), or the Cisco/20others Future-something-or-oither. Both are extreamly fast serial-ish point to point systems that could have a fast non-blocking switch at the center. NGIO has been "in the works" since like 1995. Not sure when the other started, or when we will see anything.
Yeah, companies like Dell is the most logical explanation for why AMD is doing this so soon. Dell will have to explain why they can't supply a 1GHZ machine while Gateway can. There may be a very good explanation (the AMD 1GHZ processors are crippled compared to what will be out later this year), but how many consumers will understand that?
This puts pressure on Dell and others who haven't gone for supporting AMD products.
First, make it work, then make it right, then make it fast, then, make it bloated!
Yes, the Athlon does support Multiprocessing. It's a point to point protocol of the EV6 bus, which gives each processor in a multiprocessor setup full bandwidth, as opposed to Intel's GTL+/SMP, which shares the same bus for multiple processors, giving diminishing returns on more than a couple of processors.
Hopefully the chipsets for this should be out soon. Tyan was talking first quarter at the end of last year, but I think it will more likely be second quarter, since there's only 3 weeks left in the first quarter.
A pet peeve: People refer to the new Athlon chipsets as "SMP". They're not. That's Symetrical MultiProcessing. They're Point to Point Multiprocessing, which I guess would be PTPMP or something.
--- "So THAT's what an invisible barrier looks like!" - Time Bandits
I understand the desire of AMD to beat Intel to to 1 GHz but when Intel does release more than vaporware, the benchmarks will blow the K75 core Athlon away. The 1 GHz Athlon is basically a crippled chip, with an L2 cache running at only 1/3 of the processor speed. Until the Thunderbird core is released and they have cache running at full processor speed, expect the Athlon to significantly lag the PIII at the same speed.
I'm not convinced that AMD would ever have released this chip except to beat Intel to the punch. The significant improvements will only come as the entire architecture improves (full speed L2 cache, AMD 760 and AMD 770 chipsets with DDRAM support). Current Athlon users will have virtually no incentive to upgrade until then. I'm sticking with my K7 500 (running smooth and stable at 750 w/ 1/2 cache). I'd be willing to bet the improvement would hardly be measurable trading off 250 MHz of wait states for lowering the cache from 1/2 to 1/3.
r/
Dave
OK, but I don't know that it's fair to blame AMD for this. Some of their early offerings (K5 generation, IIRC) did emphasize good architecture - pipelining, parallel execution etc - at the expense of raw MHz. And they suffered for it, because Joe Public has never heard of pipelining or parallel execution, but knows that 400 is a bigger number than 350.
Result: AMD got the message and refocused their efforts on explicitly trying to pump clock speed as high as possible. I recall an AMD exec openly saying as much some time back. Doesn't seem to have done them any harm, you must admit.
My point is: if the mass market is too dumb to care about anything but clock speed, you can't blame AMD for giving it to them. You might blame Intel, whose advertising seems to be deliberately aiming to sow confusion and ignorance about the technology they sell, but that's another story.
There's an article over at TechWeb that says intel plans to have a 1GHz p3 out by the 8th. And HP announced it will be the first to ship a 1GHz machine. I hope they didn't sacrifice quality for a good press blurb. But then again, I'd love to see them steal more market share from Intel.
I wonder how much of an impact this will have on performance, compared to Intel's Kamati offerings with a 1/2 speed cache? Also, does anybody know what the multiplier is on Intel's newest P3's? I can't remember their stupid name.
You can see on this page some benchmarks, showing that the AMD 1Ghz just barely outperfom the Kyrotech 1Ghz chips. When you look at the cache, the kyrotech is an 800 upped to 1G, therefore the cache, which was at 320 is at 400 in the kyrotech, vs 333 in the AMD chip. What else was changed in the AMD chip to make it outperform an older version of itself with faster cache?
thanx
Well, if it wasn't for the rush to the magic number (and the record books) we never would have seen this so soon. Intel is going to be hard pressed to get their yields up to where they can keep up with AMD's deliveries. I think we're seeing the changing of the guard at this point where AMD takes over the market leadership role and is going to drive Intel's product development rather than the opposite as has been the case in the past. Even with the superior performance in a "mhz even" case for AMD, they haven't stood still to have to prove equivalent benchmarks, they've attacked Intel in the forum Intel chose, clock rates. Congrats to AMD! Lead on!
Well its good to see that x86 based processors have hit the 1GHz mark. Now i am no expert on CPU's when I did computer architecture we studied the MIPS architecture and it was pretty clean (the subset we studied) and small....
I would like to know what the clocks on many of your boxes are like. I am pretty sure Sun does not sell their Eu10K's based on the MHz rating of their CPU;s. Also, how about some info on the SGI boxes and others ? ?
Also I find it interesting how marketing has made the MHz mark so freaking important that people spend 100's of dollars to get an extra 50 Mhz and then go and get IDE drives!!!
In parting I have to say that i have been a fan of AMD for sometime... cant wait till i start working so i can actually afford a K7! GO AMD!!!
and kudos to the Engineers there to be able to keep the x86 arch. going... as i recall it was called the "Golden Handcuff..big money for backward compatability with a backward technology"
Non-Deterministic Finite Automata
The latest benchmark utility from Ziff Davis, Content Creation, is described as a system-level, application-based benchmark. Using Adobe Photoshop 5.0, Adobe Premiere 5.1, Macromedia Director 7.0, DreamWeaver 2.0, Netscape Navigator 4.6, and Sonic Foundry's Sound Forge 4.5, CC Winstone 2000 applies stress on a system's CPU to determine real-world content creation performance.
Heh, what better way to see how stressed a processor can get than to throw Netscape 4.x at it?
The competition between Intel and AMD has been good on the one hand in that it has increased processor speed, encouraged new innovation and dropped the price of the processors down. But I'm starting to wonder how many corners AMD and Intel are cutting trying to one up each other. I think they've both gotton so absorbed with processor frequency that they forget the real benchmark of processors: How fast they run applications. There are other, non-x86 processors out there that would blow an Intel/AMD processor out of the water, even running at half the clock speed. So what if I have a bajillion-kagillion megahertz processor when my Palm Pilot runs faster.
I think they need to start making the processors better, not faster. If they improve the quality of the CPU, the speed will come along naturally.
kwsNI
As you can see here, the Athlon 800 delivered a severe can of whoop-ass to the Pentium III 800 (both 133 and 100 bus speeds). And the following two points can be observed:
1) The Athlon 800 has the same cache divider as the Athlon 1Ghz.
2) The performance of the Athlon does not "severely lag" behind the Pentium, and in fact, it's a whole lot faster!
expect the Athlon to significantly lag the PIII at the same speed
Dude, either you work for Intel (FUD anyone?), or you better have some concrete information to back up your outrageous claims.