First Benchmarks of AMD Hammer Prototype

porciletto writes "As seen on Ace's Hardware, this article features Quake 3 benchmarks comparing an 800 MHz ClawHammer sample to Athlon MPs at 800 MHz and 1667 MHz, as well as a Willamette Pentium 4 (256 KB L2, 400 MHz FSB) at 800 MHz and 1600 MHz. The benchmark results indicate a 40% performance increase over an Athlon MP for the ClawHammer. Additionally, the 800 MHz ClawHammer manages to tie (actually outperform by 1 FPS) the 1667 MHz Willamette Pentium 4."

Comments

Intel has a Big Problem

[Anonymous Coward]

I've been saying this for many months, and I'll say it again: By far the biggest problem on Intel's horizon is the AMD Hammer series of chips. In the IA64, Intel decided to make a clean break and go to a new architecture, incurring performance hit when running IA32 code. AMD instead blew out the IA32 architecture to 64 bits.

Expect a massive FUD attack from Intel in the coming months as they try to convince the world that their chips aren't really inferior to those from AMD.

Re:Intel has a Big Problem

[Jay Carlson]

Yeah. But IA64 made a lot of sense for Intel, given their market position when the effort started.

Think back to Rambus. (Back?) Intel got a lot of options on Rambus stock, provided that Intel could ship n percent of systems using Rambus memory. If Intel had no significant chipset competition, this would be easy. But it turned out there was enough competition to give people a choice of chipsets, and hence memory technologies.

Still, the P4 seems consciously designed to play to Rambus strengths. It chews memory bandwidth like candy through prefetching, which helps cover the higher Rambus latencies. I think Intel took a performance hit relative to AMD when the market preferred DDR SDRAM.

Anyway, it's a great story for Intel if they could control the future of PC technology. Rambus gets rich, Intel gets rich, you pay more. Three cheers for AMD for breaking this.

IA64 now looks similiar. If it wasn't for the aura of inevitability associated with the Itanic, nobody would be particularly thrilled with it. The initial SPECint numbers where it barely kept up with a SPARC were the first practical warning---if you don't count the schedule slips.

If IA64 was inevitable, everybody would have to pay up to transition to it. If it was the banner Win64 platform, a lot of places would be buying them regardless of relative price/performance. But because it looks like AMD will eat IA64 from the low end, and with POWER4 staring down from the high end, there's no longer an obvious niche where IA64 dominance is inevitable.

Four cheers for AMD.

Re:Intel has a Big Problem

[ryanvm]

In the IA64, Intel decided to make a clean break and go to a new architecture [...] AMD instead blew out the IA32 architecture to 64 bits.

Right, and what's interesting is that from a pure geek perspective, Intel did the right thing - AMD did not.

People have been griping about CISC and Intel's grotesque manifestations of x86 for years now. So they finally do the right thing and sit down with HP to spend a couple years hammering out a brand new design. And what do they get from the geeks? Nothing but boohs and hisses. You guys should be ashamed of yourselves. Did you really want a Pentium V, VI, etc.?

I'm glad Intel finally quit x86 cold turkey. AMD may have bought themselves a little time with the Opteron, but the sooner we're all off x86 the better.

Oh, and don't think that IA64 won't be looking MUCH better once we start seeing properly optimized software and later iterations of it. Intel is just like Microsoft, the first implementations invariably suck, but they always get better from there.

Re:Intel has a Big Problem

[Sloppy]

Eventually you have to break backwards compatibility to move forward without making things ugly. x86 is old, it is overly complex, it is inefficient in many respects, it is time to say good bye.

Everybody has been saying that for twenty years.

Twenty years. It is far too late for x86ers to worry about "making things ugly". That sacrifice was made in the early 80s. And it paid off.

The reason Intel is still in business, is because they knew what drove the market. Superior (in performance, power use, and just plain elegance) alternatives were around all along, but x86 still got all the sales. The reason for this is that the strongest market force is the need for good compatability with The Legacy. Against this force, all other considerations are irrelevant.

That's why Intel survived (flourished) in the 90s, and why AMD is about to kick their ass. AMD's embarrassing toadying to this principle in the Hammer design, shows that they understand. Intel attempt to raise the bar, shows that they have forgotten. Intel's chip is going to be the next 68k or PPC or SPARC. It'll be a niche, where everyone says how neato it is, and yet few actually use it. And in the mean time, AMD will be selling gazillions of Hammers.

Re:Intel has a Big Problem

[ryanvm]

Or maybe you mean the sooner we're all on Intel's DRM technology the better.

Exactly what Intel-sponsored DRM are you referring to? The only technology that Intel has introduced that bothered me was the short-lived serial number fiasco, and once the press put some heat on them they dropped that quicker than a pair of wet undies.

Most of the hardware companies out there are not too keen on DRM. How does it benefit them? It's the legislation sponsored by content providers that you have to worry about.

arrg stop with the quake already

[tomstdenis]

Why is Quake the benchmark of a good processor? Maybe computers can do something other than cache intense graphics?

Gah.

Tom

Re:arrg stop with the quake already

[Fantanicity]

Why is Quake the benchmark of a good processor?

Because Quake is what will be used by people who believe 'reviews' and 'benchmarks' from sites like aceshardware.

Re:arrg stop with the quake already

[Syllepsis]

Why is Quake the benchmark of a good processor? Maybe computers can do something other than cache intense graphics?

You are right, and in fact Quake is not even a good benchmark for gaming in general. However, it is very memory intensive and was generally the P4's strong point.

Saying that the Opteron will smoke a P4 at Quake is saying that it smokes the P4 at its own game.

The test is a good indicator that if ...if... AMD can deliver at somewhere near the promised clockspeeds, Intel is going to have to ramp the P4 very high to compete.

Re:arrg stop with the quake already

[sgtsanity]

It has a standardized timedemo, is more CPU-intensive than some of the newer games, has been around forever, and outputs easy to understand, real results. Benchmarking-only programs, like 3dmark 2001, output more abstract numbers. Games, being the only reason many people upgrade their hardware, are the only programs that are used by many to tax their computers to the limits. Although, starting up Mozilla taxes my hardware to the limits rather nicely. :)

Re:Hammer's final name

[Anonymous Coward]

The Opteron name only applies to the MP variant of their new CPU, as you may have noticed if you had actually read the page at the other end of your link. Chances are that the UP CPU will carry some sort of Athlon moniker (Athlon XP+? Athlon XXL?).

Re:It would be more interesting if...

[Jugalator]

Enlighten me! :)
How many registers are there?

I've only programmed on assembly level for MIPS (32 regs) and x86 (whoa - registers gone awry) :)

In short, the MIPS was fun and an excellent "beginner's processor" to try out your noob assembly skills on. The x86 was a nightmare. :P

Re:Benchmarks

[10Ghz]

"This is a server chip. Benchmark it using a database, a web server, number crunching, etc."

In case you don't know, Clawhammer is meant for desktops/workstations. Hell, there's even a mobile version of it in the pipeline! Then we have Clawhammer DP (dual-processor) and Sledgehammer that are meant for servers.

Please, get your facts straight before opening your mouth!

The reason it's only 800MHz.

[zardie]

It's just a sample. AMD released the Clawhammer processor to manufacturers for demonistrations and testing, so they can develop the platform, so that, get this, benchmark results would not be released. Let's face it - who in their RIGHT MIND would benchmark an 800MHz CPU against the latest and greatest processors?

Obviously, these guys did. AMD will NOT be happy about this.

Also remember that the Opteron will be running at MUCH higher clock speeds upon release. I'd guess above the 2GHz range for sure, but AMD doesn't want anybody to know that. This also suggests that this lil' 800MHz sample could be very overclockable.

This is AMD's weapon that can really take a LOT of market share. Microsoft already have a Windows XP build ported to the Opteron/x86-64 platform. The Opteron runs cooler, as well.

One thing that disappoints me - I have not seen ONE PCI64 slot on any of these test boards!! I hope that this'll be worked out before release.

Re:1667?

[Jugalator]

Forgot one thing... He elaborated a bit about P4's and said that "Intel has an interesting super long pipeline in the P4's - it's gonna be interesting to see what clock speeds it requires to fill so it can be of use to 100%". :)

I guess we have an explanation of the diff in AMD/Intel clock frequencies right there...

Sign...more of the same

[Anonymous Coward]

I'll start this by saying YES, I work for Intel. Hate me...whatever.

But its SOOOOO disheartening to see my fellow nerds and /.ers so ignorant on something like the computer scene. I'm talking about all the AMD LOVE and Intel hatred posts that always follows a news article about CPU's.

I can understand the love for Linux. A group of people programming for free, fighting a giant like Microsoft. But why should AMD garner the same sort of love and respect? AMD is a giant corporation itself, willing to screw you over. They'd charge you $2000 per processor if Intel wasn't around (and yes Intel would do the same).

Last week Intel dropped the prices of its processors. AMD was forced to follow suit, dropping their prices about 2 days later. Did the Slashdot community cheer Intel?

So along comes this news...AMD Opteron 800 MHz beats a Pentium 4 1.6 GHz by one frame pre second. I guess I fail to see why everyone is so excited?

I'll wager ANYTHING that when it ships, a 800MHz Opteron will sell for at LEAST twice the price of a Pentium 4 1.6Ghz.

Why do I even bother.

Re:It would be more interesting if...

[Anonymous Coward]

The assertion that 64 bit software was used is absolutely wrong. It's obvious you didn't bother to try to read or get the article translated. The Author's of the story provided an English version at http://www.tecchannel.de/hardware/937/ . The article clearly states that 32-bit software, and only 32-bit, was used for the Benchmarking. I'm glancing through the article and they don't specifically say why, but they mention later on in the article that they only had an hour to benchmark so that may have somethign to do with it. Plus, I get the impression that they didn't have 64-bit software available suitable for Benchmarking, but I could be wrong on that. However, assuming this article is legit then only 32-bit software was used.

Based on the above and assuming this is legit, these benchmarks are impressive. An 800 Mhz ran Quake 3 faster than a Pentium 4 running over twice as fast. You can't ignore or gloss over that bit of info no matter how hard you try. There is only one direction to go at this point with this CPU and that's up (as in better). For one thing, you have to believe the clockrate will be faster. I think it's save to say that it will be out running at least 1.6 Ghz. Heck, MSI had a board at computex. They had a spec sheet listing some items. They may have let a cat out of the bag so to speak because on the sheet, they listed the board as being able to run an Opteron at 1.6 Ghz. They didn't list a range that it ran in, they mentioned that speed specifically so I think when the CPU is released it will run at least that fast. If software ran that good at 800Mhz, imagine it at 1.6. There is a real possibility the Clawhammer will be released during the last 2 or 3 months of this year. Intel is saying the P4 will be around 3 Ghz around then. It'll be interesting to see how a 1.6 Ghz CPU stacks up to a 3 Ghz model. So far, Intel didn't fare to well in this early benchmarking. But again, you have to wonder about how legit this article is... C-ya!

Ace's is pretty good, actually.

[Rui del-Negro]

You're probably thinking of [someoneelse]'s Hardware (cough*tom*cough)...

To judge real-world performance, Quake is at least as good as any synthetic benchmark. Personally, I'd like to see benchmarks for 3DS MAX, TMPGEnc or Photoshop (because those are some of the programs I use daily). But between Quake and WhateverMark2002, I prefer Quake (and I don't even play Quake).

RMN
~~~