Slashdot Mirror
AMD Releases Barton: Athlon 3000+
Harle writes "Today AMD has introduced a new version of the Athlon, codenamed "Barton," that features twice as much L2 cache as previous chips. Along with the increase in L2 cache comes an increase in the Athlon's performance rating -- specifically the new 2.17 GHz chip is rated at 3000+.
The clockrate is actually slighly lower than the Athlon XP 2800+'s 2.25 GHz speed, so the question becomes "Does the cache improve performance enough to counter the loss in clockspeed?" For the most part, the answer seems to be "yes," however, it doesn't unilaterally stand up to the 3.06 GHz Pentium 4.
With the recent delay of the Athlon 64 to September, this is AMD's top desktop chip for some time to come. The reviews are starting to pop up at Ace's Hardware and Extremetech." There's also reviews on The Tech Report, SimHQ, HotHardware, EarthV, in Norwegian on Hardware.no, and last but not least AMD's press release. I'm sure there's many many more links, but I'm tired of pasting them all in here, so post 'em below. *grin*
so post 'em below..
OK Then... Anandtech link
We've now got Palomino, Thoroughbread A, Thoroughbread B and Barton under the Athlon XP name. To make things worse, some of the chips are using a 133 MHz FSB (Front Side Bus), and some 166.
:)
Due to this and AMD's PR ratings you have to be real careful of what you buy, if you're aming for a specific core. Expecially since AMD doesn't plan to replace all Athlon XPs with the new core.
Just remember to do your research, and you'll be fine
.: Max Romantschuk
I concur, but I doubt this will happen, as such a spec for direct performace comparison (1) is difficult to do, as performance can be quite different depending on the intended application; and (2) would hamper each company's marketing efforts.
For an quick read on some of the issues associated with different benchmarks, you could look here.
The cache memory has to run VERY VERY fast, and VERY VERY fast memory is very expensive, so yes, there is a cost consideration. In an earlier example, the AMD's K6-3 was significantly faster than the K6-2 because it had a large cache.
Too expensive, especially in terms of yields. The cache use a _large_ area of the die. Larger core, more room for defects.
Belief is the currency of delusion.
Theres an excellent review at Bit-Tech.net. Its interesting how the P4 wins easily on the benchmarks, but in games things are much closer. The Athlon even wins several games tests.
Does the cost implications completely prohibit this or do the performance benefits tail off too quickly. SUN seem to able to achieve impressive performance with lower far lower Mhz (I know its different architecture) but I get the impression the large amounts of cache (2-4 MB) they use contributes significantly to performance.
Yes, cache is terribly expensive to place on chips in large amounts. It tends to be much harder to shrink than the rest of the transistors on the chip, and the design work necessary to scale the cache to meet the ever shrinking die size is complex and harsh. Overall, with consumer chips that need to be under a certain price threshold to be purchased, Intel and AMD have both found its far cheaper to keep increasing clock speed while decreasing die size than it is to increase cache.
My guess is that this latest move by AMD is an update to that mentality. It proceeds along their realization that they might be unable to compete solely on the grounds of clockspeed. However, with the decreasing performance returns for clockspeed increases, this is less of an issue for AMD than one would think. This new core seems to indicate its becoming cheaper for their engineers to spend more time on chip design as well as use the limited die space for cache rather than other components.
The yield drops roughly quadratically with area of the chip. So, if the L2 cache occupies 50% of the chip and we want to double the cache size, the yield would drop to roughly 44% of the original yield. To make up for this the price would have to more than double and I do not believe that most people buying desktops are willing to accept that.
You can find a thorough test of the Athlon XP 2700+, made with Linux software,
here
Hers's the [H]'s take on it, , and here's Sudhian's.
They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.
To give you an idea of the costs involved for the cache difference (from the Sun parts list):
US2-400Mhz - 2Mb cache = $3895
US2-400Mhz - 4Mb cache = $4795
US2-450Mhz - 8Mb cache = $6695
I can't seem to find any price for the "US2-400Mhz - 8 Mb" in their list, but from memory it was around $6500.
I think you would find the same kind of price difference between Intel Xeon cpus with different levels of cache.
Sun also had some very bad PR way back in 2000-2001 with the Ecache errors on the 8Mb series (the cache was not ECC ram). Those CPUs caused the systems to panic randomly and the problem was only totally fixed by shipping a new revision with mirrored caches (they first tried fixing it with an OS patch)
Here's the Link http://www.tomshardware.com/cpu/20030210/index.htm l
It's seems like they say that AMD'S 3000+ rating is extremely aggressive and they do not seem to think it should have had that rating. They also have a good point in there about price gouging essentially eliminating the comeptitive price difference. Since the chip is so hard to get the price skyrockets. It's kind of disturbing that AMD recommends testing this chip with DirectX 7 that definitely does not speak well of AMD's confidence in competing with Intel's stengths.
"Freedom of speech has always been the abstract red-headed stepchild of the Constitution"
-Suck
X-bit labs has also got some info on the new Barton procs http://www.xbitlabs.com/cpu/athlonxp-3000/
You have to tweak a memory management setting in the registry to take advantage of your L2 cache anyway. Look here... http://www.tweaktown.com/document.php?dType=guide& dId=145&dPage=6
It doesn't overlap on Athlons. This is known as an exclusive cache, as opposed to Intel's CPUs, or pretty much any other CPU out there where all data in the level one cache *always* is present in the lever two cache.
All X86 CPUs are "emulators", regardless of who manufactures them. The real guts of a P4 look nothing at all like you would expect from the opcodes you feed into it. It's emulating that 25 year old architecture.
Since AMD has been producing X86 CPUs since the days of the 8086, I would argue that they have just about as much experience in emulating X86 CPUs as intel does.
I think right now the Athlon XP 3000+ based on the Barton CPU core is the right first step, but I think the CPU that will REALLY worry Intel will likely come later this spring when we see Barton core Athlon XP's that take full advantage of DDR400 DDR-SDRAM.
Remember, under pure-CPU tests the Athlon XP 3000+ has almost the same performance as the Intel Pentium 4 3.06 GHz CPU with HyperThreading; what will happen when the Athlon XP gets the Front Side Bus speed bump necessary to support DDR400 memory?
Not much I'd imagine, the bump to a 333MHz system bus did next to nothing. The P4 is far more sensitive to memory bandwidth than the Athlon.
And in April P4s at 3.2GHz with an 800MHz system bus come out...
I have done some trial runs with the divx codec and admittingly have found it to be quite a bit faster at compressing, however in my own rather unscientific tests [ducking] I found that WM9 did a marginally better job at 1.5Mbps (with a 64Kbps audio stream which was sufficient for the audio that came from a mic on a videocamera). It seems like WM9 does a tremendous amount of analysis of the image to try to eak out every last morsel of compression and it really labours the system.
Well, the Thoroughbred's seem to run pretty cool... my Athlon 1700+ is running at 35C (95F) right now under full load with air cooling. On hotter days, it makes up to 40C or so. I never had many heat issues with my K6-3 (although it did run hotter than this). Many of the newer motherboards also have emergency temperature shutdowns if the temp gets too high.
Okay, seriously. I'll do it for you.
LINK.
HTML is easy! Kids, try it at home!
>>Self shutdown - if the heatsink falls off an Intel processor, it throttles itself down until it stops, if required, so that it doesn't burn, whereas an AMD chip just burns. >>Less heat - this is a major issue in a datacenter
--My AMD Duron 900MHz had its cooling fan STOP once. I continued using the PC for about an hour, then smelled something funny. Shutdown, reactivated the cooling fan, and I'm still using the PC right now for this post.
--However, if it had been 1GHz or better, I think the chip would have fried...
.
== WolfriderV6 == I'm willing to admit that *I just might* be wrong... Are you??
The NexGen Nx586 was the first RISC processor with a hardware based decoding unit to translate 386 instructions into its native risc instructions.
This system is still in use today, as AMD purchased NexGen in 1997.
These processors don't really "emulate" a processor, although many people thought the NexGen did since it was always recognized as a 386, even if it was far faster. One of the reasons out of order execution is so important in these processors is the fairly random X86 instructions are reordered into the most efficient RISC instructions possible.
I don't read or respond to AC posts
In most cases, it is far more cost effective to buy what you need for the moment, and upgrade later after
- prices have dropped,
- you actually need the upgrade.
Here's one example: in 1998 I bought a Pentium II 350 at about the same time my friend bought a Xeon 450. Four years later, I upgraded to a Pentium III 1.2GHz, while his Xeon slot doesn't support any cpu upgrade options. Today my system smokes his system for way less money.Of course, if you actually need the Xeon right now then you should get it, but buying in anticipation of future needs makes no sense to me when prices are falling so fast.