AMD’s Athlon XP 2200+: Thoroughbred unbridled

THE GREAT GIVE-AND-TAKE battle between Intel and AMD for performance dominance in desktop processors took a decisive turn recently with the introduction of new chipsets for the Pentium 4 (and the corresponding P4 processors) that support a 533MHz front-side bus speed. That move, combined with a further ratchet of the Pentium 4’s top clock speed to 2.53GHz, gave Intel a decisive advantage over AMD—the biggest such lead for Intel in quite a while, in fact.

Intel’s processors have benefited greatly from the die shrink that happened back in January when the “Northwood” P4s first hit the scene. Built using Intel’s 130nm manufacturing process, Northwood runs faster and cooler than the original “Willamette” Pentium 4, even as it packs in 512K cache—double the amount in its predecessor. Since the Northwood arrived, processor watchers have been waiting impatiently for AMD to counter with a 130nm CPU of its own, a version of Athlon XP code-named “Thoroughbred.” It’s taken some time, but T-bred is finally here, running at 1.8GHz and given an AMD model number of 2200+.

But is T-bred swift enough to catch one of Intel’s 2.4 or 2.53GHz burners? Let’s take a closer look.

Thorough breeding
Let’s dispel the rumors right now. Thoroughbred is simply a die shrink of the Athlon XP. Nothing more. Well, OK, not much more. AMD says T-bred’s transistor count is 37.2 million, down a smidgen from the previous “Palomino” version of the Athlon XP, due to a more efficient layout and “lower voltage handling requirements.” That explains why the T-bred is a different shape than Palomino, too. But T-bred doesn’t include any notable performance tweaks like more cache, SSE2 instructions, or other sorts of engineering magic. The die shrink means the Athlon XP ought to be able to reach much higher clock speeds in the future, but clock for clock, a T-bred ought to perform exactly like a Palomino.

Not only that, but AMD is not yet—if ever—raising the Athlon XP’s front-side bus speed from its present speed of 266MHz. So Athlon XPs—of any flavor—will be hard pressed to take full advantage of advances in memory performance like DDR333 or dual-bank memory controllers.

Of course, none of these things mean, all by themselves, that the T-bred won’t be a screamer. The Athlon XP has scaled up fairly linearly in performance to date, and only the benchmarks will tell whether how well T-bred fares in that department.

We’ll get to benchmarks in a moment, but let’s consider T-bred’s other virtues. As you might expect with a die shrink, T-bred should be smaller, run cooler, and suck up less voltage than the Palomino. The 2200+ version requires only 1.65V, while the 1700+ needs just 1.5V. (All Athlon XP models, from 1700+ up, will transition to the T-bred core.) In fact, AMD allocated its first, limited supplies of production T-breds to computer makers for use in laptop PCs.

But the size difference is T-bred’s most striking attribute. Have a look at the difference between a Pally and a T-bred:


Throughbred’s on the left, and Palomino’s on the right


T-bred is clean underneath because the resistors have moved up top

T-bred is downright teeny. To my eye, it’s nearly half the size of the Palomino. The shrink from 180nm to 130mn is major. Officially, T-bred is 80 mm2, while Palomino is 128 mm2. By contrast the Pentium 4 is absolutely mammoth. Early Northwoods packed all 55 million of their transistors into a space 145 mm2, while ongoing process tweaks have cut the size on newer chips down to 131 mm2, according to reports.

(Also, in case you’re wondering, AMD hasn’t abandoned its plans to move its CPU packages from brown, like you see here, to green, like you can see here. Apparently the color change is just taking some time, and the Athlon XP 2200+ sample we received from AMD just happens to be brown. Eventually, minty-fresh green will engulf the entire Athlon XP lineup.)


A ‘bred in the hand… Erm. Sorry.

Of course, all of this shrinkage action has a purpose. The smaller the chips, the more chips AMD and Intel can manufacture per wafer. More chips per wafer means lower manufacturing costs, and ultimately, lower prices, too. AMD’s size advantage here is formidable, which ought to translate into a competitive advantage. However, Intel pulled a new trick out of its bag recently: it increased the size of its wafers from 200 mm2 to 300 mm2, and there’s some debate over who has the advantage in terms of manufacturing costs as a result. Whatever the case, know this: a processor price battle is coming. The latest round of price cuts has already gone mighty deep, and there’s more looming on the horizon.

Well, OK, maybe not looming on the horizon. More like hanging out over there, waiting to throw us a little party later on. With free beer and little cheese wedges with toothpicks in them. I can’t wait.

AMD should also be able to keep overall Athlon XP system costs down, because T-bred doesn’t require a new Socket or, by and large, even a new motherboard design. Usually a BIOS update will suffice; even some old KT133A boards will work with T-bred, though I’m not sure I see the point of that. AMD is quite proud of the relative stability of its Socket A platform in this respect.

Personally, I’m happy for them and everything, but I’d rather have a faster front-side bus than a killer (or probably overkill) CPU upgrade for my KT133A rig.

 

What to watch for in the test results
Now that we’ve introduced you to T-bred, it’s time to get down to business and see how this beast performs. Thing is, we already know this chip isn’t wildy different from its predecessor, so the only real difference we’d expect to see here come from the 66MHz clock speed increase between the Athlon XP 2100+ and 2200+—from 1.73GHz to 1.8GHz.

Don’t expect a light show or anything.

Meanwhile, we’ll be comparing it to Pentium 4 chips with both 400 and 533MHz front-side bus speeds, which is a little more intriguing. We know the Athlon XP seriously outperforms the Pentium 4 on a clock-for-clock basis, but the higher bus speeds improve the Pentium 4’s ability to execute instructions on a per-clock basis. The Pentium 4 is faster at 2.4GHz on a 533MHz bus than at 2.4GHz on a 400MHz bus, especially when paired with fast memory. We’ll be interested to see how the Athlon XP matches up against the Pentium 4 now that the P4 is faster at a given speed.

Beyond that, the real question most folks are probably asking about T-bred is: How does it overclock? Does the die shrink bring immedate benefits to those of us willing to run things out of spec a little? We’ll delve into that question, as well.


T-bred (left) and its bitter rival, the Pentium 4 (right)

Our testing methods
As ever, we did our best to deliver clean benchmark numbers. Tests were run at least twice, and the results were averaged.

Our test systems were configured like so:

  Athlon XP Pentium 4 845 Pentium 4 850 Pentium 4 850E
Processor AMD Athlon XP 2100+ 1.73GHz
AMD Athlon XP 2200+ 1.8GHz
Intel Pentium 4 2.4GHz Intel Pentium 4 2.4GHz Intel Pentium 4 2.4GHz
Intel Pentium 4 2.53GHz
Front-side bus 266MHz (133MHz double-pumped) 400MHz (100MHz quad-pumped) 400MHz (100MHz quad-pumped) 533MHz (133MHz quad-pumped)
Motherboard Shuttle AK35GT2/R Abit BD7-RAID Intel D850MD Intel D850EMV2
Chipset VIA KT333 Intel 845 Intel 850 Intel 850E
North bridge VT8367 82845 MCH 82850 MCH 82850E MCH
South bridge VT8233A 82801BA ICH2 82801BA ICH2 82801BA ICH2
Chipset drivers VIA 4-in-1
4.38(2)v(a)
Intel Application Accelerator 6.22 Intel Application Accelerator 6.22 Intel Application Accelerator 6.22
Memory size 512MB (2 DIMMs) 512MB (2 DIMMs) 512MB (4 RIMMs) 512MB (4 RIMMs)
Memory type Corsair XMS3000 PC2700 DDR SDRAM Corsair XMS2400 PC2100 DDR SDRAM Samsung PC800 Rambus DRAM Samsung PC800 Rambus DRAM
Graphics NVIDIA GeForce4 Ti 4600 128MB (Detonator XP 28.32 video drivers)
Sound Creative SoundBlaster Live!
Storage Maxtor DiamondMax Plus D740X 7200RPM ATA/100 hard drive
OS Microsoft Windows XP Professional
OS updates None

I want to give a big thanks to Corsair for providing us with DDR333 memory for our testing. Their XMS3000 DIMMs allowed us to run the memory on our Shuttle AK35GT2/R test motherboard at CAS2 timings at 166MHz (that’s 333MHz DDR, kids). Good RAM didn’t hurt in our overclocking attempts, either. If you’re looking to tweak out your system to the max and maybe overclock it a little, Corsair’s RAM is definitely worth considering. Using it makes life easier for us as we’re dealing with brand-new chipsets and pre-production motherboards, because we don’t have to worry so much about stability and compatibility. The stuff flat works.

The test systems’ Windows desktops were set at 1024×768 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.

We used the following versions of our test applications:

All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.

 
Memory performance
We generally kick off our benchmark suite with some memory tests, and this time out the test underscore an important point.

As you can see, the Athlon XP has hit a brick wall in memory bandwidth; raising the CPU clock speed doesn’t help. Although our test system has DDR333 memory, memory performance is limited by the Athlon XP’s 266MHz front-side bus. In fact, the T-bred at 1.8GHz comes out a little slower than the Palomino does at 1.73GHz. I’m not quite sure why that is, but I ran this test a number of different times, just to be sure, and the T-bred was consistently just a tiny bit slower.

The Pentium 4, on the other hand, fares especially well with a 533MHz bus. Although we chose not to repeat the results here, if you look at this review, you’ll see that the P4 does very well with DDR333 memory, too.

Now let’s make a graph that looks vaguely scientific, so your boss won’t mind if he catches you reading this at work.

Linpack shows, from left to right, floating-point performance when processing data stored in the L1 data cache, the L2 cache, and then main memory. (The processor has to step down the hierarchy of memory types and speeds as the size of the data matrices Linpack is feeding it grows.) The move from 1.73GHz to 1.8GHz for the Athlon XP boosts performance when accessing on-chip caches, but it does nothing to help once we pass about 320K matrix sizes, beyond the domain of the Athlon XP’s combined L1 data and L2 caches.

Notice that once we pass matrix sizes of about 100K, every variety of Pentium 4 system in our test is faster than any Athlon XP. The Pentium 4 platform has a pronounced advantage in memory bandwidth from the L2 cache out into main memory.

However, bandwidth is only one of the two key components, generally speaking, of memory performance. The other is latency, and we haven’t run any latency oriented tests here. (We will next time out, honest.) Also, memory performance is itself only one piece of the overall performance picture.

 

Business Winstone

The Athlon XP gains a couple of points from the 66MHz speed increase between the 2100+ and 2200+, but it’s not enough to catch the Pentium 4 2.4GHz.

Content Creation Winstone

Content Creation Winstone has been rewritten in its 2002 version to place more emphasis on memory performance, and it shows. The Athlon XP systems just can’t keep pace with the Pentium 4 rigs in this test, and bumping up the Athlon XP’s clock speed doesn’t help much.

 

POV-Ray 3D rendering

Here’s a test when the tide turns a little bit. Athlons have always womped Pentium 4 chips in raw x87 floating-point math performance, and that’s what POV-Ray rendering is all about. The T-bred’s higher clock speed shaves another 13 seconds off our render time versus the Athlon XP 2100+.

Lightwave 3D rendering
Lightwave is a nice demonstration of how 3D rendering performance can be enhanced using SIMD instruction set extensions. Lightwave uses Intel’s SSE2 extensions on the Pentium 4 (and on the Mac, it uses the G4’s AltiVec instructions) to speed the rendering process.

Obviously, this test isn’t really fair for the Athlon XP, but it is a real-world application we’re benchmarking, so Lightwave users won’t want to ignore these results.

I wrote NewTek, makers of Lightwave, to ask why their program doesn’t make use of SSE or 3DNow! extensions—both of which the Athlon XP supports—in addition to SSE2 and AltiVec. Unfortunately, I never got an answer out of them. I also wondered aloud about this question in our last CPU review, and surprisingly, no one wrote in with any good technical explanation why SSE or 3DNow! support wouldn’t be helpful in Lightwave. Where are all the know-it-all geeks when you need them?

 

LAME MP3 encoding
Our previous LAME test setup was simply being run over by high-speed CPUs; they were crunching through an entire 50MB audio file in about 20 seconds, with only fractions of a second separating the fastest times. So this time around, we’ve beefed things up by using a 101MB source audio file and asking LAME to encode a high-quality variable bit rate MP3. The exact command-line options we used were:

lame -v -b 128 -q 1 file.wav file.mp3

This encoding task produced the following results:

The Pentium 4 at 2.53GHz takes the top spot yet again.

DivX video encoding
Xmpeg can encode video files using the popular DivX format, which produces very high quality video in relatively small amounts of space. For this test, we took a 279MB video file, encoded in MPEG2 format at DVD quality, and converted it to a 37MB DivX file.

Xmpeg supports all the various x86 SIMD instruction sets, including MMX, 3DNow!, SSE, SSE2—even different flavors of 3DNow!, like 3DNow! Enhanced. Most importantly, perhaps, Xmpeg makes good use of the Pentium 4’s SSE2 instruction set, which offers potentially higher performance than the SSE or 3DNow! instructions supported by the Athlon XP.

One the Pentium 4 gets its SSE2 mojo going, there just no stopping it. The Athlon XP appears memory limited here; the clock speed increase is nearly useless.

 

Codecreatures Benchmark Pro
The Codecreatures benchmark is a graphical wonder; it pushes even more polygons than my 8th-grade geometry teacher, and it does so in conjuction with advanced graphics features like pixel shaders. We’ve wondered whether performance in this test is severerly limited by the graphics card. So is it?

Nope! The T-bred takes us to new heights in Codecreatures. No, the performance difference in terms of frames per second isn’t huge here, but when you’re hovering around 30 frames per second, every frame counts.

3DMark 2001 SE

The T-bred claws its way to an additional 100 points in 3DMark versus the Palomino, but the improvement isn’t quite enough to catch up with the Pentium 4 at 2.4GHz.

Serious Sam SE

The Athlon XP has owned Serious Sam benchmarks for ages. Only the jump to a 533MHz bus made the Pentium 4 competitive, and the T-bred nearly catches the P4 2.53GHz here.

Comanche 4

Comanche 4 only benefits a little bit from T-bred’s faster clock speed.

 

Speech recognition
Sphinx is a high-quality speech recognition routine that needs the latest computer hardware to run at speeds close to real-time processing. We use two different versions, built with two different compilers, in an attempt to ensure we’re getting the best possible performance.

Sphinx is very sensitive to memory performance, and it shows. Once again, T-bred isn’t much faster than Palomino, because it’s limited by the front-side bus bottleneck.

ScienceMark
We’ll close out our testing with ScienceMark, which is always interesting in a very, very geeky sort of way. ScienceMark runs a series of physics simulations in an attempt to confuse non-geeks. Err, in an attempt to measure scientific computing performance.

AMD’s top processor takes the top spot here, as usual. Now for some of the individual test scores…

As we’ve seen before, the Pentium 4 takes the memory-intensive Primordia test, while the Athlon XP takes the other two.

 

Overclocking
Well, this is the part where I tell you about all of our marvelous overclocking exploits with the new, cool-running, die-shrunk processor—how we cranked it up 50% over its initial clock speed; how the Athlon XP now has more headroom than a convertible Buick; how our 3DMark scores shot up by hundreds of points with a few simple BIOS tweaks.

However, I can’t do that.

I can’t do that because the darned thing wouldn’t overclock for us. Not by much, at least—not even by 100MHz. We were using mild bus speed overclocking, and we tried everything: core voltage tweaks, memory voltage tweaks, RAM timings more conservative than Gordon Liddy. Nothing helped enough to really matter.

At the end of the day, given the number of different ways we modified our system config and given the kinds of system crashes we were seeing, we could only come to one conclusion: the CPU just wouldn’t go any faster than about 1.89GHz. And even at that speed, it was on the ragged edge.

Now, overclocking is never a sure thing. Every chip is different, and you never know what will happen when you run a chip out of spec. So I’d better not draw any conclusions from our one-off, isolated experience with our very first T-bred sample. I shouldn’t speculate that AMD might be having trouble producing these chips with really good yields. And I really shouldn’t wonder out loud whether the Athlon XP’s 10-stage pipeline is hitting a snag at some point along the way that limits the chip’s peak clock speed. Most importantly of all, I shouldn’t mention the gossip I heard to that effect from other folks who had tested T- breds when I talked with them at Computex this past week. Especially not from engineers. I really, really shouldn’t do that.

Still, I can’t help but be a little bit worried about T-bred’s prospects given my experience. Certainly AMD could refine its fabrication process or tweak the T-bred core with a new stepping or two and make these things hum—up to 2GHz and beyond. Right now, however, our sample of this newly die-shrunk processor is running right at its clock speed limit, only 66MHz above its peak speed on AMD’s older fab process.

Conclusions
The benchmarks tell an interesting story. In some cases, despite all of its CPU and bus clock speed disadvantages, the Athlon XP is faster than any Pentium 4 processor. However, the Pentium 4 wins out in the majority of our tests. In many of the tests where the Athlon XP is slower than the P4, the 2200+ model’s 66MHz clock speed increase doesn’t deliver much more performance than the 2100+. Clearly, the Athlon XP’s 266MHz front-side bus is a big bottleneck; it can’t even keep up with the latest DDR333 memory, and DDR400 is already on the horizon.

Back when the Pentium III and Athlon were near the 1GHz mark, we saw this same problem: the Pentium III’s clock speed hit a wall at 1.13GHz, and its slower bus just couldn’t deliver extra performance from DDR memory. As a result, AMD took the performance lead and held on to it until the Pentium 4 came into its own. Now the tables have turned. The Athlon XP’s bus is a bottleneck, and we’re starting to wonder how well the chip will scale up to higher clock speeds. By contrast, the Pentium 4 is just getting started, and its newer design and platform give it a decisive edge. No wonder AMD has dedicated the bulk of its time and effort to bringing its K8 chip to market.

However, AMD has yet to relinquish the price-performance lead. AMD has led in this key category for ages, and given the Athlon XP’s solid performance—even if it’s not the fastest in every test—we’ve found it hard not to recommend an Athlon XP to just about anyone.

This time out, AMD is playing an odd game with its pricing. If you consult the AMD price list and then the Intel price list, you’ll see that the Athlon XP 2000+ lists at the exact same price as the Pentium 4 2.0AGHz: $193. Though it’s not yet listed there, the Athlon XP 2200+ will follow a similar pattern; it will be priced at $241, the same price at the Pentium 4 2.26GHz. AMD is matching its prices to Intel’s using its model-number rating system as a guide.

That’s a dangerous policy, since the Athlon XP isn’t scaling up as well as the Pentium 4. Plus, the new P4 chips with 533MHz bus speeds are faster, clock for clock, than the 400MHz bus versions. In fact, AMD’s model numbering scheme may need adjustment for future Athlon XP models. (I wish we’d have had time to benchmark the 2.2GHz and 2.26GHz variants of the Pentium 4 here today, so you could see a direct comparison between AMD’s model number and Intel’s clock speed. However, we were too busy with Computex this past week to make it happen.)


Left: Athlon XP Palomino. Right: Athlon XP Thoroughbred.

AMD really knows better than to price match Intel, however. If you check street prices on Pricewatch, for instance, you’ll find that the Pentium 4 2.0AGHz is selling for somewhere north of $197, while the Athlon XP 2000+ is available for as little as $161. AMD seems to be applying a discount in reality, even if the price list doesn’t reflect it.

Intel has its own discounts, too—especially for big system builders like Dell or HP. So if you plan on buying a pre-built PC, shop carefully. Of course, if you plan on doing that, well, ick. What are you thinking?

For those of us building our own boxes, the Athlon XP 2200+ is a pretty good value. However, if you’re looking to overclock, you’re probably better off with a lower speed grade of a Northwood Pentium 4, like the Pentium 4 1.6A. Those chips have a decent chance of hitting at least 2.13GHz on a 133MHz bus. If you’re not looking to overclock, by all means, check out the Athlon XP 2200+. But you might want to look closely at our benchmarks before you make up your mind; which CPU is better depends quite a bit on how you’re using it. 

Comments closed
    • Anonymous
    • 17 years ago

    I think this is a great article, and I just want to add my experiences to it:

    I run an Athlon XP 2200+ on a Gigabyte GA-7VAXP Ultra Mainboard with 512 MB of PC2700 DDR. My Graphics card is a Radeon 8500 DV 64 MB of ram, with an Enermax PSU. My motherboard came with an overclocking Utility called EasyTune4, and so far, I have succesfully run it at 1947 MHz. This could be an individual experience, or due to the utility I used. Check it out if you have the Gigabyte Mainboard, it comes as part of the bundled software. Also, my processor usually reads at about 45 to 48 degrees when I am running a heavy load, all the way down to 39 when I am at idle. Does anyone have similar experiences? Happy overclocking!

    • Anonymous
    • 18 years ago

    Excellent article! Well written and imformative. You’ve confirmed my suspicions that (sigh) Intel has indeed recaptured the “RAW PERFORMANCE” lead. Can’t wait to hear what you have to say when the Hammer finally arrives. TIll then, Thank you.

    • Anonymous
    • 18 years ago

    Come to think of it, has anyone thought of unlocking the XP2200+ processor and bringing the bus speed up to 166 Mhz with the PCI right dividers but setting the multiplers to bring the CPU clock speed to aroun d1.8 Ghz like the original? that might give a good indication of what the XP2200 could do with a proper FSB speed boost

    • danny e.
    • 18 years ago

    i have nothing to say…
    just tired of seeing that DUKE was the last to post.

    – danny e.
    —————
    italy vs. korea
    &
    japan vs. turkey

    lets rock.

    • shaker
    • 18 years ago

    Liam… I have a new Yum-Cha on my PIV… what a coincidence!

    🙂

    Have a good one.

    This thread is officially in post-mortem.

    • LiamC
    • 18 years ago

    Shaker, No need to apologise, logical argument is good. Irrational flames on the other hand…

    The problem I see here is we may be looking at things from two different viewpoints. I base b[

    • Anonymous
    • 18 years ago

    all XP2200+ reviews are at §[<http://www.amdboard.com/athlonxp2200special.html<]§ can't give you better address for now ;)

    • shaker
    • 18 years ago

    Yeah, I remember a K6 200 that I had, quite the little furnace…

    • Anonymous
    • 18 years ago

    Anand was talking out his Intel Contract in that quote. AMD employed heatspreadders across the entire K6/K6-x series.

    I don’t know what the K6/K6-x heatspreadder was made out of, but I do know that several people who removed them were able to get higher overclocks afterwards.

    -ludi

    • Anonymous
    • 18 years ago

    *[

    • shaker
    • 18 years ago

    Liam C:

    Sorry to extend this, but just one more post on the subject…

    (BTW, I have no credentials other than “seat of the pants” engineering experience)

    Theoretically, you are correct, every interface introduces a bottleneck, *however*, the more consistent the *first* interface (i.e., heat spreader to core) the more predictable the heat transfer. Remember, the second interface has to transfer less heat per unit area, so intimate contact isn’t as critical (such as meticulously applied ultra-thin coatings of Arctic Silver) as the first interface. As a matter of fact, a sloppy application of silicone will probably do 99% of the heat transfer as a thin coating of a more expensive compound. What does this do? It provides the chip manufacturer with predictable thermal characteristics so that they (Intel, AMD) can effectively ‘overclock’ their own processors (albeit not the ‘maximum overclock’), getting more yield from their process. Intel has taken the ‘bull by the balls’ here, and AMD is to follow.

    From AnandTech:
    g{<"The first thing we noticed about the CPUs is that they look a lot like the Socket-478 Pentium 4 processors. In fact, AMD did borrow quite a bit from Intel in designing the packaging of the CPUs as both the ClawHammer and SledgeHammer parts feature integrated heat spreaders (IHS), a technology which Intel has been using for almost two years now. This means that there will no longer be any horror stories of crushed cores from poorly installed heatsinks or badly manufactured heatsink clips. The next thing you'll notice, which is difficult to gauge using these words, is that the CPUs are very heavy. They are easily the heaviest socketed CPUs I've ever held, with the weight being due to the IHS; it's simply interesting that they are heavier than their counterparts from Intel."<}g Motorcycle engines are a good analogy. Although air cooling the engine is most direct way to acheive cooling, manufacturers are now using water cooling, despite it's extra weight (and thermal interface). Why? More consistent performance under varying conditions, allowing them to push the (albeit slightly lower) horsepower limits with predictable results. It could be that the future of overclocking will be to attempt to defeat the manufacturer's solutions by prying off heat spreaders and using specially modified heat sinks (maybe some will be aluminum with a 'special' copper plate bonded to the bottom) well, you get it... IMHO respectfully, shaker

    • LiamC
    • 18 years ago

    Thanks Forge. It’s always good to get reliable info.

    • Forge
    • 18 years ago

    LiamC – I think you’ll find that the AG is parroting a bit of wisdom from back when HSFs were aluminum and core covers were copper. Back then, a nice big flat copper plate between CPU and HSF actually did help a little, as the copper ‘wicked’ heat away from the core, more towards he outsides of the HSF. I did research on this subject at one point, and found I was getting ~50C at the center of the HSF, 35C at the edges. I put a spare copper plate from my H2O rig in between, and get 40-45C in a much more uniform dispersal across the heatsink. In my case core temperature dropped only a minute amount, though, probably because the increased cooling ability of the HSF was offset by the extra thermal junction.

    In short, you’re quite right, but assuming the AG was listening to RMBS too long isn’t necessarily true.

    • Anonymous
    • 18 years ago

    Hey #120, something smells fishy alright, and its coming from one post lower than this one.

    • Anonymous
    • 18 years ago

    You people are gay and I hope you eat a bag of fish.

    You don’t come to the city.

    The city comes to you.

    • LiamC
    • 18 years ago

    Shaker (#106)

    A heat spreader i[

    • Anonymous
    • 18 years ago

    Wesley96, the blue cores were from the Dresden die polisher(hence, were also copper). Of course they did better 🙂

    • wesley96
    • 18 years ago

    Hmm, I followed the link (from AMDZone… and I think Chris Tom’s going hyperbole again… how typical.. 🙂 seems like he never wants to concede) to AMD’s Model 8(Thoroughbred) AXP datasheet. Comparing to Model 6(Palomino)… on 1700+, which is what I have…

    Model 6 Model 8
    V: 1.75 1.50
    I_max: 36.6 32.9

    Now assuming that I_max stays constant(reducing voltage reduces heat output proportionally, so I think this is right), and that the heatsink can pull out the heat at a relatively similar rate, this means if I can run T-bred @ 1.45V like I have my Pally right now, it’d still have 10% less heat, and the temperature delta would be reduced accordingly… which would equal about 2.5C. Nice. Not very much, but still nice. BUT, if that thing can do something like 1.20V, I’m on a roll. 🙂

    C’mon, dudes, please test those T-breds to find out the minimum stable operating voltage.

    • wesley96
    • 18 years ago

    Heh, I think AMD’s moving to green colour now.. needs to clear those early brown batches first, though. Maybe the overclockability lies in the colour… [i]again[/i]. 🙂 Remember the old green-and-blue core fun? Seems like green is taking a revenge..

    • Anonymous
    • 18 years ago

    #103: That was a press photo. Every sight that actually posted an original photo or discussed the packaging, indicated “brown”. Also the “real” pictures all show stepping info, etc. on the label, the press photo doesn’t.

    #105: At least Ryan’s response is credible, but it is a little difficult to understand why he didn’t publish any data at all for that clockspeed. Given the March=>June delay and dozens of rumors about AMD’s 0.13

    • Anonymous
    • 18 years ago

    #113, if “everyone” knows that, then what of the other 5.8B organisms who think of themselves as people? It’s not nice to exclude their views, you know.

    • Anonymous
    • 18 years ago

    does it really matter? everyone knows AMD is the pewp..

    • Anonymous
    • 18 years ago

    AG110, you mean 730Mhz surely 🙂

    1800Mhz
    verses
    2530Mhz

    • shaker
    • 18 years ago

    From Page 51 of the the Intel PIV Datasheet:
    Table 29.Processor Material Properties
    Component Material
    Integrated Heat Spreader Nickel over copper
    Substrate Fiber-reinforced resin
    Substrate pins Gold over nickel

    The Heat spreader also helps distribute the mechanical load of the heat sink retention mechanism evenly.

    • Anonymous
    • 18 years ago

    you know the stupid PR i almost forgot how good the 2200+ scores are once you remember AMD is running 600mhz below the intel chips.

    • Anonymous
    • 18 years ago

    Ahh, nevermind. I didn’t know they were copper/nickel. Are you sure it’s not just aluminum?

    • Anonymous
    • 18 years ago

    Why would you want an aluminum heat spreader in the way of your copper heatsink? That obviously would not improve heat transfer efficiency.

    • droopy1592
    • 18 years ago

    Damage, T-R, your chip is a week 14 chip. Maybe that’s why it ain’t overclocking like mad… The only way we can tell is when someone get’s a retail T-bred.

    • shaker
    • 18 years ago

    Not to belabor the point, but an integrated heat spreader (copper inside, nickel plate outside to prevent oxidation) seems like a good idea for a small area die, as it gets the heat “out” from the center quickly. I’m running a PIV 1.6 @ 1.8 (BIOS and power supply limitations are keeping me there, for now) and it runs six degrees C over the case temp with an “average” aluminum heat sink and Radio Shack thermal goo. (The “1.5U” Utron case helps, too as a 60mm fan can evacuate the case volume pretty quickly, but that’s a whole ‘nother story).

    As to the 166 FSB: Is every MB/chipset out there able to run 166 stably? That may be the reason that AMD won’t spend the resources on making it “official”.

    shaker

    • Anonymous
    • 18 years ago

    Just a nother worthy point

    [q] Hey guys! Thanks for the comments — good and bad.

    I was as surprised as the rest of you on the overclocking results of a lot of the other websites out there. I talked with a couple, and some have stated that their processors came to them already unlocked, where the 2200+ I received was not. This may point to a different stepping or some other factor that is causing the better overclocking performance I saw.

    But, saying that I made the numbers up is kind of ridiculous. Because I am in San Francisco right now, I don’t have pictures or access to the hardware to give you any more information. The FSB that it was running at was 166 MHz and the multiplier was 13x. That gives me 2160 MHz, a 360 MHz overclock (that I mistakenly typed as 316 MHz in the article).

    [/q]

    Quoting Ryan’s (the guy who made the overclock) reply on their forums… I’d say that there is hope and that his CPU might have been a stepping up from the rest and Chris Tom is just pissed that he didn’t receive the same overclockable CPU 🙂

    read the thread it’s fun

    §[<http://www.amdmb.com/vb/showthread.php?s=2ce2f7c62a38d264696a0a19b5c708a4&threadid=132329&perpage=15&pagenumber=5<]§

    • Anonymous
    • 18 years ago

    [q] There has been a lot of discussion revolving around the Thoroughbred Processor Review and the overclocking achieved by Ryan during his review. I have since spoken to him about this while he is out of town in San Francisco where he was found to be on The Screen Savers discussing the best highend and lowend motherboards you can find for your AMD Processor. Since he was busy preparing for that interview on TechTV, he was not privy to all the huffing and puffing related to his overclock which seems to far out weigh other reviewers findings. Since he is out of town he doesnt have easy access to his notes, pics, and findings but assures me he stands behind his findings and looks forward to providing further insight and answering all concerns on this issue about his findings in an overclocking article within the next week or so.

    [/q]

    from AMDMB and their response to OC’ing the chip…. like all good conspiracies … we have to wait it seems 😉

    • Anonymous
    • 18 years ago

    Hey did anyone notice that AMDMB tb chip had green packaging while TR’s one and most of the others were brown, even though it says that this should have no effect on perfomance perhaps it does matter for OC, it is definitley a different batch.

    MR Chris Tom at Amdzone want’s the magic CPU shipped to him to OC it 😉

    §[<http://www.amdzone.com<]§ Via hardware had a green one too, and they couldn't get past 1900 mhz... hey but still???

    • EasyRhino
    • 18 years ago

    Unanimous Critter, thanks for laying down the smack on super-duper 3DNow being single precision

    ER

    • droopy1592
    • 18 years ago

    ag#99 that first finger looks fake on the tomshardware.com link.

    ag#100 but the other AG was telling the truth. It would be as slow as a mac. Maybe one from 1994 but it would still be as slow.

    • Anonymous Coward
    • 18 years ago

    Anonymous Gerbil #94:

    I’m saying this for what has got to be the 4th or 5th time, but there are various theories about why AMD’s process has not yielded high clock speeds at this time. See vanshardware or aceshardware. Van’s attributes it to AMD “formulating” their process for mobile processors, a claim they back up, and Aces’s observes that AMD has pretty much [i]never[/i] had a high-overclocking new core at release time. Both of these are sound explanations for why these chips don’t clock well. It does not seem especially likely that AMD is in serious trouble.

    Finally, here’s a quote from vanshardware.com (which is actually all he said on the subject for now):

    [quote]However, I have noticed a very large number of misconceptions in reviews today regarding the “low potential headroom” of AMD’s 0.13-micron Athlon XP. It is clear that AMD is currently using a process skew emphasizing higher yields and lower power dissipation, apparently to produce large numbers of mobile chips. This is logical since mobile parts command higher ASPs. Such a mobile-friendly process skew would result in a slower transistor and reduced maximum clock-speeds.

    Faster transistors have shorter gate lengths and are therefore leakier (consuming power when doing nothing at all). That AMD is using a slower transistor is demonstrated by typical current in Stop Grant, which is less than one-third of that leaked by the Intel Pentium 4 Northwood core in this sleep mode (the P4-Northwoods wastes an amazing 18A when the chip is doing nothing!). This makes the Thoroughbred a much better mobile part, but higher clock-speeds will likely have to wait until AMD remixes its recipe for faster transistors.[/quote]
    Anonymous Gerbil #99:

    Damn, Tom is one dirty, crafty guy.

    Anonymous Gerbil #92:

    [quote]Nope, it would still be slow… about as slow as a mac[/quote]
    Quick! Someone needs to run a Mac-bashing story before the AG trolls starve to death! Actually, looks like Apple is doing some new products in the next couple weeks, [b]boy to I hope the trolls make it that long[/b].

    • Anonymous
    • 18 years ago

    #98, read Tom’s quote VERY carefully. He was sensationalizing again:

    [q]Only a few select CPU coolers can be used – they must have a copper contact plate for the surface of the CPU die, or use some other kind of contact plating with adequate heat transfer capability. To put it clearly: if you want to equip your PC with a Thoroughbred Athlon XP, you can get rid of your aluminum cooler. As it is, if the new Athlon XP dies a thermal death when used together with a cheap cooler, you will get no guarantee from AMD.[/q]
    §[<http://www7.tomshardware.com/cpu/02q2/020610/thoroughbred-10.html<]§ The entire paragraph spends 60 or so words to tell you that you need a CPU cooler. He [i]suggests[/i] that a copper-base heatsink is a minimal necessity, and he manages to do it without actually saying "this is my opinion" OR attributing it to AMD. As usual with THG, you have to read between the lines, and liquor helps. -ludi

    • Anonymous
    • 18 years ago

    well tom mentioned they require copper base heat sinks now which people are failing to mention. this is bad for amd. their thermal density is high so it looks like they are having problems with their die temp. one easy test to confirm it is to try o/c ing with their recommended hs, and then try it again with a water cooled hs. barton will help them out with a bigger die but they still have problems. it is a big effort to run a board skew with a higher fsb and hammer is on top of barton. methinks they will not up their fsb to lower the risk of new hammer + new k7 platforms at the same time. that just leave some chip tweaking and better heat dissipation with the bigger barton die. going to be a rough road for them until hammer takes off…

    • Unanimous Hamster
    • 18 years ago

    EasyRhino…

    AMD’s 3DNow! (in both its original and enhanced versions) only supports single-precision floating point numbers.

    From AMD’s technical paper on 3DNow:

    [quote]
    3DNow! technology uses a packed data format. The data is
    packed in a single, 64-bit 3DNow!/MMX register or a quadword memory operand. Figure 2 shows the 3DNow! floating-point data type. D0 and D1
    each hold an [b]IEEE 32-bit single-precision, floating-point[/b] doubleword.

    . . .

    Figure 3 on page 6 shows the format of the [b]IEEE 32-bit,
    single-precision, floating-point format[/b].
    [/quote]

    This is from page 16 at:

    §[<http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/21928.pdf<]§

    • Anonymous
    • 18 years ago

    In software it could theoretically be faster, because you could just in time compile it, and then save the compiled images. But of course then you’d have to have filesystem support, which means an OS, which means the hardware would only run on platforms that bothered to write the translator, which would mean less choice in OS, which is not a good thing imo, and could still be done regardless if OS development were not busy building new whiz bang gui’s rather than robust stable OS support.

    • Blackfell
    • 18 years ago

    AG 94: Intel is doing the same thing with new steppings of Northwood. As process tech is refined, it’s common for design improvments to make their way in, like smaller feature sizes where possible. Nothing new about what either side has been doing. That bit about AMD ‘eating its future’ is just Intel FUD.

    As to why AMD’s .13 process is doing badly (I think), well, remember that this is AMD’s first release of product on .13. Intel, OTOH, got used to .13 a while ago, while making Celerons and P3’s long before Northwood was released. Those early .13 P3’s/Celerons were low-volume and pretty wretched overclockers, (~100 MHz on average, IIRC). Assuming AMD actually cares enough to put resources into their bulk .13 process, we should see improvements in a stepping or two.

    All just my 2 cents. I’ve had a long day, so if I ended up talking out of my ass, feel free to correct me.

    • Anonymous
    • 18 years ago

    Remember when Intel accused AMD of “mortgaging its future” by using 0.13 micron features in selected parts of Palomino? I guess we’re seeing the end result now. There evidently wasn’t much to be gained by shrinking the rest of the chip.

    On the other hand, this should improve AMD’s margins on the Athlon a fair bit due to the smaller chip. That’s more important than a lot of people here realize.

    • Anonymous
    • 18 years ago

    AG92, you forgot to say “…from 1994”

    which is actually about how fast Itanium is for x86 code anyways..

    • Anonymous
    • 18 years ago

    Nope, it would still be slow… about as slow as a mac

    • Coldfirex
    • 18 years ago

    Isnt it running in software?

    • Anonymous
    • 18 years ago

    *[

    • Anonymous Coward
    • 18 years ago

    wesley96:

    Yeah I [i]just[/i] saw that over at Van’s myself, very interesting. Was about to post it here in fact.

    Also interesting how low a voltage your chip will run at… I’ll have to try that myself (get that thing to generate less heat!). I’m even thinking some about one of these T-breds just to lower the heat output (well, that and I like playing with computer guts).

    • wesley96
    • 18 years ago

    Van’s Hardware says T-bred doesn’t overclock well because the process skew enables better power management (less power leak) and low-voltage operation while sacrificing operating speed, as the transistors operate relatively slowly. If that’s the case, would someone check to see how [b]low[/b] the voltage goes on the default clock, or make some chart about the lowest voltage attained for each speed grade. I sure as hell would be willing to do that if I had one of these babies…

    See, my Palomino 1700+ operates at 1.45V fine, and my friends say they can do that, too. I’m mainly rooting for low-voltage operation because I set all my fans at 5V…

    • Jerry S.
    • 18 years ago

    [quote] Most of the population doesn’t know Hammer is coming. Any of them that see benchmarks of AMD’s latest and greatest T-bred are going to think AMD is dead. [/quote]

    I wonder what percentage of the population actually fits into what you call most, being that they must:

    1) Not know anything about the Hammer (or Opteron).

    2) Read and understand CPU benchmarks (and even care).

    I can’t see it being a very large slice. 🙂

    b{< Then again, [b]I[/b] may just be naive! <}b 😛

    • LiamC
    • 18 years ago

    AG #82

    §[<http://www.tech-report.com/news_reply.x/3671<]§ post #35 AMD have known about this situation for about a year (I and several other were speculating that this might happen in September b[

    • Tantric
    • 18 years ago

    autonomous gerbil,

    [q]Can anyone (and I mean ANYONE!) come up with a reason we’re not seeing 166 as an (if not THE) official bus speed for T-bred?[/q]

    Money, time, people, equipment, marketing, and hundreds of other resources that it takes to pull off a massive change in infrastructure like that. Which would you do, raise the fsb to 166 costing a great deal of money and time for a few $ in ASP increase, or use the money and resources to get hammer out asap and sell those at 3x the price of the Tbred/bartons? I could care less about the Tbred fsb. If you want 166 buy a KT333 board and overclock the shit out of it. I’ll just wait patiently for Hammer. Well kinda patiently……

    • Anonymous
    • 18 years ago

    [q]It uses a CISC setup, which generates lots of heat….It’s still a RISC chip[/q]

    Itanium is no more a CISC chip than Hammer is a RISC chip. That is to say, neither chip uses either architecture.

    Itainium uses a VLIW architecture which, while CISC-y in appearance, is very different.

    Hammer is what’s called a “post-RISC” archeticture. It’s has a CISC (x86) front end for hybred RISC/SIMD functional units.

    More information can be found in various articles in the “CPU Theory and Praxis” section at Arstechnica.com, especially:

    §[<http://www.arstechnica.com/cpu/4q99/risc-cisc/rvc-1.htm<]§ §[<http://www.arstechnica.com/cpu/2q00/x86future/isa-future-1.html<]§ and §[<http://www.arstechnica.com/cpu/1q99/ia-64-preview-1.html<]§

    • Anonymous
    • 18 years ago

    When people talk about AMD or Intel “improving their yields” and getting more/more reliable chips, what kind of tweaks does that actually involve?

    • Anonymous
    • 18 years ago

    *[

    • Anonymous Coward
    • 18 years ago

    AG #79:

    [quote]He said he and AMD wanted to fundamentally change the 2/4 market place. “People are going to use the 64-bit for Hammers,” he said. No one, a year ago, would have considered a $1,000 64-bit system possible, he added.[/quote]
    That’s silly. Sun has been selling 64-bit systems at that price for more than a year. Course they’re slow 500mhz Ultrasparc IIe machines… but at least they have an on-die memory controller. [i]Oh you mean AMD wasn’t first there either?[/i] Doh.

    • Xylker
    • 18 years ago

    I think that there is a lot of worrying for nothing. These chips are also about two months old. Look here: §[<http://images.anandtech.com/reviews/cpu/amd/athlonxp/2200P/opn.jpg<]§ The date is from the 15th week of 2002, or 8-12 April. It is already June and the process is surely getting better every day. As to Hammer, I would really like to hear about its performance in a controlled environment like TR... ;-) Or, failing that, any other reputable web-site.

    • Anonymous
    • 18 years ago

    [quote]He said he and AMD wanted to fundamentally change the 2/4 market place. “People are going to use the 64-bit for Hammers,” he said. No one, a year ago, would have considered a $1,000 64-bit system possible, he added.

    “By late 2004/2005, we’ll have a good chance to change the workstation and cluster market,” he said. “Four way boards wouldn’t be $20,000 boards.

    Heye also said that whether Intel is developing a “Yamhill” processor or not, based on AMD designs, it would have a tough time positioning it in the marketplace.
    [/quote]

    Conv between Richard Heye and The Inquirer

    §[<http://www.theinquirer.net/03060205.htm<]§

    • Thresher
    • 18 years ago

    My point is that AMD has put it all on the line. I am not really all that concerned that the Hammer will fail, for all the reasons you list.

    What I do worry about is production problems, yields, and market acceptence. If there are production problems, that equals higher prices. It also means unavailability. This could lead to problems with market and OEM acceptence.

    I am usually considered an AMD fanboy, so don’t get me wrong on this.

    As for your performance numbers, I don’t think we have seen anything official on that, so I am a bit wary of what I am seeing on some of the less reliable websites. I am not dismissing them, but I would like to see some real testing.

    • R2P2
    • 18 years ago

    Small mistake, Intelmole — Itanic can run 32-bit x86 code, it just runs it really, really badly. 🙂

    • Anonymous Coward
    • 18 years ago

    IntelMole:

    [quote]Meanwhile, in 64-bit mode you can do impressive things like access 18 exabytes of RAM (how big is that again, 450,000,000 times 4 gigabytes?). [/quote]
    Actually, going from 32 to 64 bits would be more like 4-gigabytes of 4-gigabytes, if each 4-gigabytes was represented as a byte. Did that make sense? Anyway, it the old 32 bit maximum [i]squared[/i].

    However, I don’t think AMD will support full 64-bit addressing, cause thats insane. I’ve heard runors about 40-bit, which would still be 64 gigs of RAM.

    Anyway, about your other post, RISC chips actually tend to generate less heat than CISC, all other things being equal… modern x86 chips require quite a few extra transistors running just to decode the x86 crap into a RISC dialect… more transistors != less heat.

    • IntelMole
    • 18 years ago

    Yer I knew that danny e., just it is in effect acting like a RISC chip… anyways, that wasn’t one of my major points 🙂

    The point was, if you want a space heater, forget AMD, get an Itanic server :-),
    IntelMole

    • danny e.
    • 18 years ago

    intermole.. i think you have CISC and RISC confused.
    Hammer at best is a hybred.. takes the old CISC type instructions and converts them to its own micro-(RISC type) instructions.

    – danny e.

    • IntelMole
    • 18 years ago

    Oh, and Itanium has “failed” because of two things…

    One: It’s not backwards compatible with 32-bit, so that’s 95% of the market that can’t use it…

    Two: It uses a CISC setup, which generates lots of heat, stopping it from going above 800MHz or so in the last incarnation. Software has to be specifically written in IA64, and holding developers to ransom like that doesn’t work…

    Hammer succeeds in both of these areas…

    One: It’s backwards compatible with 32-bit, so everyone can use it

    Two: It’s still a RISC chip, so 99% of all software that has ever been written will still work with it. Software writers now have the option of putting 64-bit [b]extensions[/b] into their software, rather than whole new programs…

    Happy now?,
    IntelMole

    • IntelMole
    • 18 years ago

    Thresher, what the hell you smoking? It’s clouding your memory 🙂

    Hammer outperforms the T-bred by about 25% at the same clock speed… in Quake 3 Arena it outperforms the MP by about 40% 🙂 This is despite having a pipeline two stages longer, all thanks to the onboard memory controller…

    So it’ll clock higher, yet still be faster clock for clock? Sounds like an Apple wet dream 🙂

    Hammer is AMD’s 32-bit strategy, they don’t call it x86-64 for nothing remember? it’s backwards-compatible with 32bit and even 16bit operating modes…

    Meanwhile, in 64-bit mode you can do impressive things like access 18 exabytes of RAM (how big is that again, 450,000,000 times 4 gigabytes?).

    Oh and if memory serves correctly you can also do SSE2 instructions on it, let’s see how well it does on Lightwave 7b now!

    AMD just seem to have all the bases covered with this chip, and sorry for the mac-bashing, it had to go in there :-),
    IntelMole

    • Anonymous
    • 18 years ago

    Seriously, dahling, it is revolting at the mere mention of T-bred, beer, and cheese in the same article. A correction should be put forth in the review that reads, “i[free wine] and cheese. ” And make it a cabernet, while you’re at it.

    That is the only proper way to display one’s higher breeding, especially when discussing such fine tastes as Thoroughbreds and cheese.

    BM

    • danny e.
    • 18 years ago

    Thresher: go here..

    §[<http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/x86-64_wp.pdf<]§ page 7. so you dont have to rely on the voices in my head. - danny e. ------------- shut up brain or i'll stab you with a q-tip -homer

    • Anonymous Coward
    • 18 years ago

    Thresher:

    [q]The T-Bird/T-Bred appears to be pretty close to its limits.[/q]
    Did you read what ace’s had to say about that? Sure, without a faster FSB and/or more cashe the Athlon is getting near its peak, but I believe that poor bin splits on new AMD chips are not uncommon, such as what we saw when the Palomino was released (Athlon MP 1.0 and 1.2).

    So I’m not sure if we’re talking about different things, if you disagree with me, or if you are ignoring me… I’ve mentioned this once or twice already.

    [q]Furthermore, if the Hammer strategy does not pay off, they have no 32bit strategy after the Barton.[/q]
    I’ll just call you “Chicken Little” from now on. Seriously, what is it about Hammer that would prevent it from pretending to be a 32-bit CPU forever and ever? It’s not very big or expensive. Noone needs to use it’s 64-bit-ness. Microsoft is supporting it anyway, and so is Linux. How can AMD possibly go wrong?

    [q]Another thing to keep in mind is that Hammer, even in Athlon guise, will need a 64bit OS to have any advantages[/q]
    See, this is the root of your problem. The Hammer is apparently faster than equal clocked Athlons, even in 32-bit code. As far as wether or not it needs support from the OS, I think the only possible hangup is in the NUMA memory controller(s), but even that may not be an issue for a single-CPU system. In any case, M$ is well along with their support of that feature (and so is Linux).

    Now cmon, stop worrying.

    • danny e.
    • 18 years ago

    [q]I do have a question though that I have not seen answered. Will Hammer run Windows 32bit versions? If so, is there a performance hit? Do you HAVE to have a 64bit OS just to get the thing running? I know it will run 32bit apps, but I’m not sure about the OS. [/q]

    yes it will run 32 bit windows fine.

    – danny e.
    ——————–
    thats what the voice in my head keeps saying.

    • Thresher
    • 18 years ago

    AG65

    That is basically what I said. They will revert back to being the low cost provider if their strategy does not pay off.

    The thing is, you assume that they will be fast in comparison to intel. At present, intel is scaling the P4 core pretty darn well. The T-Bird/T-Bred appears to be pretty close to its limits. Barton will help, but by that time, the P4 should be hitting 3 GHz easily. So, for the next 6 months or so, AMD has nothing that really can compete with intel. Intel is not stupid, so you can expect another price war to cause even more pain to AMD.

    Furthermore, if the Hammer strategy does not pay off, they have no 32bit strategy after the Barton. If there is limited overhead for the remaining K7 line, this could put them in a world of hurt in the consumer market.

    Another thing to keep in mind is that Hammer, even in Athlon guise, will need a 64bit OS to have any advantages. Unless AMD wins some in the OEM market, 64 bit programs are going to be far and few between, which means that there will be quite a bit of time between introduction and payoff for the customer.

    I do have a question though that I have not seen answered. Will Hammer run Windows 32bit versions? If so, is there a performance hit? Do you HAVE to have a 64bit OS just to get the thing running? I know it will run 32bit apps, but I’m not sure about the OS.

    • Anonymous
    • 18 years ago

    It’s probably best to wait until Q1 2003 to upgrade. Neither Thoroughbred, Barton or anything else Intel has right now seems exciting.
    ————
    Amen.

    • Anonymous
    • 18 years ago

    Guys, before you start saying AMD is down the tubes, remember, they’ve been there before.

    The K6-2 was the first slightly competitive processor, mostly because of 3dnow.

    The K7 (which we all know and love) was a huge step for AMD, and created a competitor to Intel. Remember how AMD and cyrix used to be the slow, bargain chips? Who’s the slow bargain now? (Via’s C3, built on cyrix tech, that’s who.)

    Even if the hammer bombs, AMD will still be able to make cheap, slower chips. In fact, that’s what they’ve been doing all along, except they’re not too slow anymore…

    -The non-registered highlandr

    • Anonymous Coward
    • 18 years ago

    [q]Heck you might even be able to make a Mac look good![/q]
    Not hard, since I guess only SSE2 and Alvitech are supported. Just pick the right test and run against an Athlon… should be easy. If you really do it well, perhaps the whole scene will fit the the G4’s L3. Man, that would make a nice, pretty graph.

    Of course I think you’d have to not run against a P4… that would probably be bad. I [i]really[/i] don’t think a G4 can pull 2.5x the IPC of a P4 that’s in SSE2 code.

    On this topic, TR should be applauded for mentioning a Mac in the article without deriding it in any way.

    • Tantric
    • 18 years ago

    Damage,

    About the lightwave tests look at this review of the 2.53 at Xbit where they do a different scene.

    §[<http://www.xbitlabs.com/cpu/pentium4-2530/<]§ About 3/4 of the way down. It paints things in a different light. It is a typical 3d rendering benchmark. You can make the results say what you want by the scene chosen according to the strength of each processor. Heck you might even be able to make a Mac look good!

    • Thresher
    • 18 years ago

    Liquidsquid,

    It’s not the current predicament that is the problem. It is the trend, AMD performance measured against intel is trending downward and will be for the next 6 months or so. If they get too far behind and bet the farm on technology that isn’t widely accepted (Hammer), then they could go down for the count.

    • liquidsquid
    • 18 years ago

    Arg. A whole lot of speculating going on here on so little info. I love this enthusiast crowd! Somehow AMD is going down the tubes because Intel i[

    • Anonymous Coward
    • 18 years ago

    [q]Intel is scheduled to release Prescott on .09 micron process in Q2 of next year. with current P4’s on .13 scaling beyond 3Ghz .. Prescott will probably be go upwards of 4 ghz fairly rapidly.[/q]
    So who thinks that Intel will accidentally run over its own Itanium while trying to keep ahead of AMD? 🙂 [i]Me! Me![/i]

    • danny e.
    • 18 years ago

    [q]AMD is in trouble folks, make no mistake. They need Clawhammer and Opteron within a few months, they do not have a year. While it would not cause them to go out of business, it may put them back in the situation they were in pre-Athlon [/q]

    i agree. despite what Forge says.. AMD is having some problems, and i am convinced part of it is the .18 to .13 transistion.
    I am not so much worried about AMD’s current position.. even though they are beginning to lag alot. I am worried more about long term.. next year and beyond. If the Hammer isnt everything we all hope it is and more.. then AMD will probably never again be on top. Intel just has too much money.. and i think they have learned their lesson. As long as AMD is around Intel will probably not fall asleep again.
    the thing to look for is how fast AMD can get to .09 process… and i am thinking not too fast. Intel released chips on .13 process in January.. AMD in June.. 5 months behind. Intel is scheduled to release Prescott on .09 micron process in Q2 of next year. with current P4’s on .13 scaling beyond 3Ghz .. Prescott will probably be go upwards of 4 ghz fairly rapidly.
    I really hope AMD stays around. and i hope “hammer” is everything we all hope and dream, but i am really starting to have those nervouse/uneasy feelings that something is amis.

    – danny e.
    ————
    spammer #1

    • Anonymous Coward
    • 18 years ago

    Thresher:

    So I hate to keep mentioning aceshardware.com, but their handy review did speak about this whole clock speed deal.

    [i]Retorical question:[/i] how well did the first Palomino’s overclock? Remember the Athlon MP 1.0 and 1.2? [b]Exactly[/b], they sucked wanker, yet that core is now running spiffy as an Athlon XP.

    The sky [b]is not[/b] falling.

    • JohnnyQ
    • 18 years ago

    OMG! Danny e. remembered his password! welcome back man! 😉

    • Thresher
    • 18 years ago

    They should have skipped T-Bred and went straight for Barton.

    For the past few years, AMD has been able to keep the performance crown for most of the time, but not now.

    T-Bred just doesn’t offer any real benefit to owners of current AthlonXPs, other than some minor heat differences, but certainly not as much as I had hoped.

    This is very disappointing to me. AMD is taking a HUGE risk betting everything on the Hammer line at the expense of the current K7 architecture. While I appreciate that there could be a huge potential windfall for the company, it could also fail miserably (Itanium anyone?).

    The one thing that AMD has done right versus intel is that they have catered to the DIY and overclocker scene, carefully cultivating word of mouth. This has allowed AMD to expand into markets where it never really had much representation. It has been the enthusiast crowd that has convinced some of the major OEMs that this was in fact a viable product and it could help them make a dent in the server market as well.

    For the past 6 months or so, AMD has been lagging in performance. That’s not a big deal in the grand scheme of things. However, what really concerns me is that intel has a lot of headroom left in the P4, the T-Bred doesn’t seem to have much, and AMD does not have a serious upgrade in the pipeline for another 6 months or so. In the meantime, intel has been cutting their prices, RAMBUS prices have dropped considerably, and the 533MHz bus has been a really big improvement.

    AMD has a window of about 6 months. If they do not have the Clawhammer in the hands of enthusiasts within that time, they could lose all the gains they have made over the past few years in terms of market penetration and the “good will” of the enthusiast crowd. If they are not careful, they could revert back to the “value” chipmaker they were back in the days of the K6. I think by the time the Barton comes out, intel will be too far ahead for a simple re-jiggering of the K7 core to cut it.

    If intel is smart, during this time, they will reduce the cost of their chips to a parity level with AMD’s performance numbers. This would really put a crimp in AMD’s cash flow in the meantime.

    AMD is in trouble folks, make no mistake. They need Clawhammer and Opteron within a few months, they do not have a year. While it would not cause them to go out of business, it may put them back in the situation they were in pre-Athlon.

    • Anonymous
    • 18 years ago

    It’s probably best to wait until Q1 2003 to upgrade. Neither Thoroughbred, Barton or anything else Intel has right now seems exciting.

    • Anonymous Coward
    • 18 years ago

    corrosive23:

    [q]AG#12

    read here. A smaller die does translate into more heat[/q]
    First, note I’m not an AG. Second, I maintain that while the chip may run hotter, it is not generating more heat. It is therefore not proper to say that a smaller die leads to more heat. More [i]concentrated[/i] heat, perhaps. Also note that versions of the new core that run at the proper low voltage are pretty cool, apparently generating up to 28% less heat than an equivelent speed on the last core.

    danny e. #49:

    [q]”scientists” need to study logic.[/q]
    You trying to start a flame-war here? If I were to make a guess about your views and started bashing certain religious figures and beliefs, I’m sure you’d be delighted. It would of course be radically off topic, as was your stupid little commentary.

    • danny e.
    • 18 years ago

    [q]danny e./spammer, you need to either lay off the coffee, or drink more, I’m not sure which 🙁 [/q]

    well.. at first i needed more, but then i had too much and just overdid it. i realize i made 5 posts about the same thing.. but for lack of an edit option what was i to do? its so fun to post!!

    – danny e.
    ——————
    being called a spammer is the worst insult ever!
    *sulks*

    • droopy1592
    • 18 years ago

    r{

    • Anonymous
    • 18 years ago

    *[http://tech-report.com/reviews/2002q2/abit-bg7/index.x?pg=6<]§ ..for instance. DDR333 is often faster. [/q] Did you read the HardOCP review of the XP2200+? Given those results, it\'s hard to ignore that the P4 2.5ghz doesn\'t look so hot when paired with DDR.

    • Anonymous
    • 18 years ago

    danny e./spammer, you need to either lay off the coffee, or drink more, I’m not sure which 🙁

    • danny e.
    • 18 years ago

    <sarcasm>
    isnt it interesting how hard it is to design and build complex cpu’s.. considering the human mind evolved by accident over billions of years?
    </sarcasm>

    – danny e.
    ——————
    “scientists” need to study logic.

    • Anonymous
    • 18 years ago

    *[

    • corrosive23
    • 18 years ago

    AG#12

    read here. A smaller die does translate into more heat

    §[<http://www.overclockers.com/tips013/<]§

    • danny e.
    • 18 years ago

    hmmm little anand thinks its due to the difficulty of running wires from the bottom to the smaller core. i am not convinced. then again i am stubborn old a**.

    – danny e.
    ————-
    a dew an hour gives me coding power.

    • danny e.
    • 18 years ago

    this is what i am talking about.. take a look at the pic.

    §[<http://images.anandtech.com/reviews/cpu/amd/athlonxp/2200P/athlonfamily.jpg<]§ t'bred on the left.. palomino then thunderbird. doesnt it just seem a little strange they would revert to old packaging? - danny e. -------------------- yes i know i am delusional

    • Ryu Connor
    • 18 years ago

    §[<http://www.spodesabode.com/content/article/crackcel<]§ §[<http://www.spodesabode.com/content/article/friedcel<]§ Relates to the Mendocino Celeron, but does a better job of showing than my vague memory from post #42. Copper plate with a nickel top seems to be how it was designed.

    • danny e.
    • 18 years ago

    hmmm and no code on the core? looks blank to me.
    i’m telling you .. there is something odd about this puppy.
    i really am curious as to where they are produced.

    – danny e.
    ———————-
    hammer better have a steel handle.
    not a crappy wood one.

    • Ryu Connor
    • 18 years ago

    [q]why would a standard generic aluminum heat spreader[/q]

    Actually, I believe they are made of nickel. Or more so there is a nickel plate underneath the aluminum.

    • danny e.
    • 18 years ago

    see… if you will remember.. AMD changed to putting everything on the bottom not so many months ago…

    §[<http://www.tech-report.com/reviews/2001q4/athlonxp/index.x?pg=2<]§ and they have now decided that was a mistake so they are putting it all back on top? humbug. i dont buy it. something smells fishy. - danny e. ----------------------- is that fish in your pocket ?

    • danny e.
    • 18 years ago

    something fishy goin on. the new t’bred looks alot like older athlons with all the crap on top instead of bottom… I’m personally wondering if t’bred hasnt been done for a long long time, and they are now just releasing them. Thus they havent even switched over to the newer packaging with everything on bottom. hmmm what i want to know is what fab do they come from??

    – danny e.
    —————–
    its all rumors and speculation

    • Anonymous
    • 18 years ago

    *[

    • shaker
    • 18 years ago

    Maybe AMD should have incorporated (barring any patent infringement) a heat-spreader like the PIV… reduces the “heat bottleneck” and could allow better (or at least more consistent)overclocking.

    g{<"T-bred (right) and its bitter rival, the Pentium 4 (left) Hotdamn, pretty please with sugar on top, change the #%$ text under the picture."<}g Or, just reverse the photo... heh...

    • Anonymous
    • 18 years ago

    What’s the problem with faster FSB? Can’t we just get a slower Thoroughbred ( 1700+ ) and crank it up to 166MHz FSB? Cheaper than 2200+ and faster!

    • Aphasia
    • 18 years ago

    What i heard that microscope was actually quite a good one for the price…..

    • wesley96
    • 18 years ago

    This is what AMD really is… too small to spread resources out to many directions. But then again, Intel spreads stuff out too much. USB microscope for kids, good god.

    • Anonymous
    • 18 years ago

    *[

    • IntelMole
    • 18 years ago

    Nice, a post screen, so, how much do you have to turn it down to get into windows then?,
    IntelMole

    • EasyRhino
    • 18 years ago

    Welcome back Damage!

    Unanimous, thanks for chiming in with the SSE/3Dnow stuff. I believe that AMD has a Sooper-dooper 3DNow, that might support double-precision math. But no one except driver makers ever supports that stuff.

    Man, I can remember when a 266mhz FSB was /fast/.

    ER

    • TheCollective
    • 18 years ago

    *Yawn* Oh, did I miss something? Hammer needs to deliver this fall. No doubt about that in my mind at all. If Hammer doesn’t make it by the fall then AMD is screwed.

    • Anonymous
    • 18 years ago

    So 2160 is tops now, at the Inqwell there is a post screen with 2314 ghz §[<http://www.theinquirer.net/07060214.htm<]§ so letrs see how it works out 1800 ghz +2200 1866 +2300 1933 +2400 2000 +2500 2066 +2600 2133 +2700 - THAT IS WHERE AMDMB IS 2200 +2800 2266 +2900 2333 +3000 - MAYBE JUST MAYBE ;) and perhaps extra 256kb of L2 cache can help too.

    • Unanimous Hamster
    • 18 years ago

    I’m not familiar with the specifics of Lightwave’s rendering, never having used this particular piece of software myself, but high-end 3D sofrware, when doing final rendering, typically uses double-precision arithmatic to obtain the highest image quality.

    SSE2 supports double-precision math, while 3DNow! and SSE don’t.

    (However, Altivec also doesn’t do double-precision math, perhaps the Mac version doesn’t have quite the level of Altivec optimizations that the P4/SSE2 version does…?)

    • Anonymous
    • 18 years ago

    # 21 ; Devnull;

    uh… saw that too.

    T-bred (right) and its bitter rival, the Pentium 4 (left)

    Hotdamn, pretty please with sugar on top, change the #%$ text under the picture.
    That mistake is not even funny.

    Esp

    • Anonymous Coward
    • 18 years ago

    Quoteing Johan from Aceshardware board:

    [q]Initially is indeed the word. What was the point of Palomino at first? AMD delayed the desktop version because it hardly ran more than 1333 MHz. When I tried to overclock the Palomino 1200 (Athlon MP 1200) I got about 1.35 GHz out of it, and at the same time the Athlon Thunderbird 1333 in the labs was doing 1.5 GHz without any trouble.

    Anyway, AMD couln’t wait any longer for binsplits to improve and pushed a slightly overclocked Tbred on the market. This gives AMD a little more time to figure the process problems out.[/q]
    So I have hope.

    • Anonymous Coward
    • 18 years ago

    wumpus #18, AG #20:

    Take note of the nice graphic that THG provided that shows how the heat compares. While its true that at the top end, the new chip isn’t that cool, at the low end where the voltage is lower, like it should be, it is more like 27%-28% better. Also, aceshardware.com has a nice review on this chip. In fact, aces asserts that un-overclockability is pretty common for AMD on a new process or chip.

    [quote]Anyway, it’s depressing, because with 166mhz FSB and 512kb on-die L2 cache, there is no doubt in my mind that we’d have a neck-and-neck race again. It’s even likely the performance would be in AMD’s favor again. Who knows?[/quote]
    I would be surprised if AMD gained as much from another 256k of cashe as Intel did, and we’ve already seen 333mhz FSB tests, so we know that alone won’t turn the tables. At least we’ll get the 512k L2 latter this year (it seems).

    • Anonymous
    • 18 years ago

    For chrissakes, why can’t motherboards read the internal temp diode of the goddamn athlons by now?
    ———-
    The Soltek SL-75DRV5 does.

    • Anonymous
    • 18 years ago

    Damage – Nice review, as usual!

    Wumpus #2 – Extra validation time, for one thing. AMD has all of its spare resources focused on Hammer atm.

    …and W #5: I wish AMDMB had disclosed the core voltage they were using to get that result.

    -ludi

    • Anonymous
    • 18 years ago

    The A7V333 does…….

    • Anonymous
    • 18 years ago

    g{

    • Anonymous
    • 18 years ago

    *[

    • Anonymous
    • 18 years ago

    g{

    • Anonymous
    • 18 years ago

    *[http://www.tomshardware.com/cpu/02q2/020610/thoroughbred-05.html<]§ [/q] Interesting. Looks like the 2000+ palomino and 2000+ thoroughbred differ in wattage by about 16 percent, but the hothardware \'real world\' temp differences show only half that. Of course no guaranteeing those temp numbers from the mobo are accurate. For chrissakes, why can\'t motherboards read the internal temp diode of the goddamn athlons by now? Bah.

    • Anonymous
    • 18 years ago

    *[http://www.hothardware.com<]§ [q]While overclocking, we paid close attention to CPU temperatures and noticed that the \"Thoroughbred\" did not run much cooler than a similarly clocked \"Palomino\". As an experiment, we installed an Athlon XP 2100+ onto the Gigabyte motherboard we used, and raised the FSB to 138MHz to bring the clockspeed up to 1800MHz. Using the same Taisol cooler, we saw temperatures ranging from 47 - 55 degrees Celsius on the \"Palomino\" based 2100+. With the \"Thoroughbred\' based 2200+, at it\'s default clock speed, we saw temperatures ranging from 44 - 53 degrees Celsius, a difference of approximately 7%. We were hoping to see much lower temperatures than this but with the much smaller contact area, dissipating heat efficiently is going to be the challenge. It\'s going to be interesting to see what type of cooler AMD ultimately ships with their boxed processors. Perhaps a nice all Copper model will absorb heat from the core a little better? [/q]

    • Anonymous
    • 18 years ago

    I’m laughing my ass off at the price comparisons on THG

    §[<http://www.tomshardware.com/cpu/02q2/020610/thoroughbred-11.html#price_comparison_amd_athlon_xp_vs_intel_pentium_4<]§ I'm in England, and AthlonXP 1800+'s work out to $109, that's a BIG drop from the $160 they are going on about, surely they aren't dumb enough to compare list price to list price?

    • droopy1592
    • 18 years ago

    g{

    • droopy1592
    • 18 years ago

    Most of Alanis Morrisette’s song “Isn’t it Ironic” ain’t even about irony… Don’t believe me, listen to it again. It’s should be called “Isn’t it Fucked up” or “Ain’t I a dumb bitch for smoking so much and not having a dictionary.”

    “A free ride when you’ve already paid.”

    Dumb ass.

    But it is ironic that AMDs T-bred was produced at 130nm to reduce heat and increase clock speed but it has not proved fruitful so far with the overclocks. Or is it not?

    Death to Intel!

    • Anonymous
    • 18 years ago

    #11, Pull up the Reflective_Radiosity scene and disable radiosity for the renderer.

    I don’t know what will happen, but given how explicit Intel’s benchmarking instructions are, I imagine interesting things might happen.

    Oh, is your version of the Lightwave 7b benchmarking guide much different than this one?

    §[<http://www.intel.com/procs/perf/Pentium4/compute/docs/LightWave_Eval_Guide_v1.0.pdf<]§

    • Anonymous Coward
    • 18 years ago

    Ah hah, Tom is my hero: [b]HEAT OUTPUTS:[/b]

    §[<http://www.tomshardware.com/cpu/02q2/020610/thoroughbred-05.html<]§

    • Anonymous Coward
    • 18 years ago

    wumpus #8:

    [quote]Yeah, it’s cooler, but the core is also smaller– so you have about the same heat dispersion problem in the end. As Alanis Morrisette once sang, “isn’t it ironic?”[/quote]
    I don’t follow the logic there… density of heat generated does not make more heat… and after a heatsink is attached, all we have to worry about is the heat going from the heatsink to the outside of the case. And that all-important flow of heat [i]is[/i] reduced.

    If these suckers produce enough less heat, I’d consider getting one for my machine. I’m really not a fan of hot-air-blowing power supplies or any of that.

    • Damage
    • 18 years ago

    #10: I asked the question in the context of the review, explaining fully what the story was with code optimizations. If you are right, that’s a serious problem, obviously. But it’s not a possible outcome I didn’t anticipte or lead my readers to believe wasn’t possible.

    Wumpus: RDRAM isn’t always fastest. Check here:

    §[<http://tech-report.com/reviews/2002q2/abit-bg7/index.x?pg=6<]§ ..for instance. DDR333 is often faster.

    • Anonymous
    • 18 years ago

    Scott Wasson —

    The answer to your Lightwave MMX/3DNOW/SSE question is pretty simple. Intel uses Lightwave as a showcase for SSE2 technology. They partnered with Newtek and produced some SSE2 optimized code. A result of their partnership most likely precludes Newtek incorporating any non-SSE2 optimizations into their 7b rendering code.

    The ironic part is that only certain aspects of Lightwaves rendering technology have these SSE2 optimizations (like radiosity.)

    Notice how the i[

    • Anonymous
    • 18 years ago

    *[

    • Anonymous
    • 18 years ago

    *[1.33@1.53 for a Thoroughbred. [/q]
    Yeah, it\’s cooler, but the core is also smaller– so you have about the same heat dispersion problem in the end. As Alanis Morrisette once sang, \”isn\’t it ironic?\”

    • Anonymous Coward
    • 18 years ago

    I imagine that AMD’s apparent production difficulties will be corrected once they have sufficient time to troubleshoot.

    But, like cRock, what I really wanted to see was how cool this thing ran. [b]*snif*[/b]

    • Anonymous
    • 18 years ago

    *[

    • Anonymous
    • 18 years ago

    *[http://www.amdmb.com/article-display.php?ArticleID=187&PageID=4<]§ [q]Using a processor unlocking kit that was sent to me by the fine folks at HighSpeed PC, I was able to work some magic on the Athlon XP 2200+ processor and push it above its 1.8 GHz clock speed. At first, I had nothing but problems trying to modify the multiplier when using the CPU, but after playing with the L1 bridges, I was able to get it to work well enough to run some testing on the overclocking potential of the CPU with air cooling. I was using the Vantec CCK-6040 copper heatsink and a high speed fan for all the tests I ran. [b]Using the Epox 8K3A+ motherboard that I just recently reviewed, I was able to push the FSB and lower the multiplier to get a finishing speed of 2160 MHz or 2.16 GHz. This is a very nice 316 MHz overclock on the processor, that ran perfectly stable in all our burn-in tests in Windows XP.[/b] Look for more on overclocking the Thoroughbred core in a future article. [/q]

    • Anonymous
    • 18 years ago

    *[

    • cRock
    • 18 years ago

    Man o’ man, Damage is getting so sarcastic I’m begining to wonder if he’s reviewing hardware or practicing his standup routines! Unfortunately, the t-bred seems to be the brunt of the joke. The slow bus, unchanged cache, and lack of overclocking make me say “ho-hum” too.

    My only question would be how cool it runs. If it’s as cool as a PIII-S then AMD may have opened up a new niche – small rackservers. If these chips still run warm, AMD has truly failed to deliver anything compelling.

    • Anonymous
    • 18 years ago

    *[http://www.xbitlabs.com/cpu/athlonxp-166/<]§ Why AMD didn\'t enable 166mhz bus on the tbred is a mystery to me. Clearly the processor benefits from it-- dramatically.

    • Anonymous
    • 18 years ago

    Finally it’s here. The long wait is over. It’s time to ditch my massive heat producing Thunderbird 1.33@1.53 for a Thoroughbred.

Pin It on Pinterest

Share This