AMD Says Barcelona Will Outperform Clovertown

Dysfnctnl85 points out a ZDNet Blog posting in which AMD claims that its upcoming quad-core "Barcelona" chipset should be 40% faster than "Clovertown," Intel's quad-core Xeon 5300 line. AMD says that the introduction of Barcelona marks a shift in their strategy from emphasizing price to performance. The post goes on: "Intel is eager to claw back some of the server market share from AMD, and this is where Clovertown comes in... The Xeon 5300 line will represent excellent value for money since Intel plans on pricing them the same as its dual core Xeon 5100 processors. That could make things tough for AMD."

Comments

"native quad core"

[kimvette]

"Their native quad core 65nm processors"

I take it AMD is releasing actual quad-core processors rather than simply calling two dual-core processors stuck on a motherboard quad core?

It's about time! But, why not drop the word "native" and admit they were engaging in deceptive advertising up to this point?

Re:"native quad core"

[Compholio]

It's about time! But, why not drop the word "native" and admit they were engaging in deceptive advertising up to this point?

Nice try, AMD hasn't been claiming any sort of quad-core processors - that's Intel:

While Intel went for a design that crammed two dual-core processors into a single package, AMD have built all four cores onto a single die.

(from TFA)

Re:"native quad core"

[Anonymous Coward]

Nice try, AMD hasn't been claiming any sort of quad-core processors - that's Intel

Umm... yes they have. Do you not remember 4x4? [2x2 wold have been more appropriate].

SUNNYVALE, CALIF. -- June 26, 2006 --AMD (NYSE: AMD) announced today broad software developer support for its upcoming four-core, multi-socket enthusiast platform, codenamed 4x4.

The implication was that there were "four-cores."

"Advanced Micro Devices Inc. (AMD) launched a four-core chip platform for video gamers Thursday, continuing to battle Intel Corp. for buzz in the high-end "enthusiast" segment."

Re:"native quad core"

[Anonymous Coward]

So, two cores times two processors doesn't equal quad core? Better tell Apple!

Re:"native quad core"

[be-fan]

Yes, there were four cores. Two cores on one chip, two cores on the other. They key words from the titles you quoted are "platform" and "multi-socket", neither of which imply a single die with four cores.

Re:"native quad core"

[Anonymous Coward]

Rule #1: AMD never does anything wrong
Rule #2: When AMD does do wrong, it is justifiable
Rule #3: When AMD does do wrong, do not post about it as it goes against slashdot groupthink

Re:"native quad core"

[dramenbejs]

C'mon and set "Troll" bonus in your preferences to +3, to view interesting posts like yours, troll brother.
It works perfectly, I see 40% more subjectively interesting posts than before.

PS: Have you seen the post "linux is preemptive" modded funny??? That's funny for itself.

Re:"native quad core"

[Targon]

AMD has yet to release a quad-core CPU. The whole QuadFX(aka 4x4) is a dual-processor system that is designed to be a consumer level platform that supports two processors. The reason for QuadFX to be worth looking at as a platform is that you will be able to swap out the two dual-core processors for two quad-core processors starting around the middle of 2007.

AMD is really starting to hype their true next generation core design, not just quad-core. This is something that many people seem to be closing their eyes to about this particular issue. The current AMD processors are based on the AMD K8 design(regardless of fab process, amounts of cache, or tweaks to the design). The new stuff being released this summer is based on what is called the K8L design. Yes, it's still the same basic design as the current K8, but you could say that Intel's Core designs are based on the Pentium-M, which was based on the Pentium 3, which was based on the Pentium 2, which was based on the Pentium Pro. Obviously there will be similarities here to previous processors, but the progress between the last implementation from either company should not be ignored.

The Pentium-M was a great laptop processor, but if you tried to make it compete without a LOT of changes with the desktop processors from AMD or Intel at the time, it would NOT have done well on the desktop except in the low power usage area. So, how many changes did Intel make to turn it into the Core line of processors?

The Athlon 64 is a great processor, but AMD knows that it doesn't compete well against the Core line of chips. So, they are doing to it what Intel did to the Pentium-M, do a HUGE overhaul of the design. Intel hyped their Core chips well before release, even while the Pentium 4 was being beaten left and right by the Athlon 64, so now the shoe is on the other foot. If the new K8L design is as good as the hype indicates, then not only will the AMD quad-core beat Intel's quad-core by a good amount, but dual-core processors based on the same design should also compete very well, though probably not quite as well as the quad-core.

The reason why quad-core might compete better is based on the connections between the CPU and memory. Remember that Intel is still using the CPU-chipset-memory design while AMD connects the CPU to the memory directly. As a result, when you go from dual-core to quad-core on a native design, Intel processors will still need to go through the chipset to get to the memory. Unless the bus speed is really increased to provide that extra bandwidth, the Intel processors will end up starved for memory. In fact, that's a part of why you see so much cache on Intel processors currently, to compensate for the old design that Intel is still married to.

Intel still hasn't released a native quad-core implementation btw, they need to go through the system bus to connect each pair of cores on their "quad-core" processors.

Re:"native quad core"

[tedgyz]

There was no deception. You are a fucktard, or work for Intel.

If only I/O speeds could also grow as fast

[namityadav]

The way AMD and Intel are improving the processor speed is very impressive. I/O speed is going to become an even clearer bottleneck now.

Re:If only I/O speeds could also grow as fast

[HockeyPuck]

I guess 4Gb FibreChannel, 2.5, 10 and 20Gb/s Infiniband aren't quite fast enough?

I've managed storage for 15years and even within the largest datacenters very few host systems actually push their storage infrastructure. Most server admins say "I need linerate" b/c they have no idea how much data their applications will actually push. Plus it's a way to force the hba/hca/nic vendors to lower their prices.

Btw: 80% of servers out there could run just perfectly fine with a pair of 1Gb/s fibrechannel HBAs.

Re:If only I/O speeds could also grow as fast

[Anonymous Coward]

I think he was thinking of latency more than bandwidth, because CPUs tend to spend a lot of their time waiting for data.

Re:If only I/O speeds could also grow as fast

[suv4x4]

The way AMD and Intel are improving the processor speed is very impressive. I/O speed is going to become an even clearer bottleneck now.

Depends what you're using. I gotta say, I use PHP.. and I'm very happy with more cores coming on the server market, and CPU is quite clear bottleneck for that one technology :(

Re:If only I/O speeds could also grow as fast

[jadavis]

Latency is not a bottleneck.

You could turn it around and say that, since the disks are not using their full bandwidth, the disks spend most of their time waiting for requests.

That being said, disk latency is one of the major causes of poor performance. But "bottlenecks" only have to do with throughput.

Re:If only I/O speeds could also grow as fast

[Boone^]

I think you meant to say "memory". 1 core gets an entire set of DDR2 controllers. 4 cores get... to share the same set of DDR2 controllers.

More cores per socket, but the socket's DDR2 bandwidth stays the same. Let's hope your kernels fit in the L2/L3.

Re:If only I/O speeds could also grow as fast

[vadim_t]

Unless you mean memory (where things don't look good ATM, as they seem to be expanding bandwidth but at the cost of adding latency), you might want to consider a flash drive.

Flash has the advantage of having no seek latency, so even though I think current flash drives are slower than a hard disk as far as raw read/write speed is concerned, the seek latency cripples performance so much that a flash drive should be a massive improvement for most tasks. Most people don't do things like continously reading or writing data to disk, so eliminating latency would be a huge improvement.

Re:If only I/O speeds could also grow as fast

[that this is not und]

The GP brought up the issue of pinched I/O bottlenecks at the processor bus, and you ramble off onto something related to connections to storage on servers.

Clue: the whole world doesn't exist in 19" racks.

Re:If only I/O speeds could also grow as fast

[wirelessbuzzers]

This may be true in a very technical, pedantic sense. However, often one side or the other can't do anything while the latent operation is pending: for instance, a disk can't read or write data while it's seeking. If it spends most of its time in a "latent" state (seeking, waiting for data from RAM, fiddling with DMA protocols, whatever), and this decreases throughput, then I would say that latency is a bottleneck.

Re:If only I/O speeds could also grow as fast

[slamb]

You could turn it around and say that, since the disks are not using their full bandwidth, the disks spend most of their time waiting for requests.

Only by specious reasoning. I'll disprove by counterexample. If I continously tell the disk to seek to one extreme and read a cacheful, then seek to the other extreme and read a cacheful, it will neither be waiting for requests nor using its full bandwidth. A and not B disproves (A => B).

Latency and throughput are unrelated only if there can be infinitely many requests produced and satisfied in parallel. In the case of a hard disk, there can be only one active request per head because it can only be at one place at once. Let's consider the example of my laptop hard drive. It's rated at a data transfer rate of 150 MB/s. But look at the seek speeds - 1.5ms minimum, 12ms average read, 22 ms maximum. It can read a 1 MB file in 6.7 ms, but if that 1 MB file is fragmented into ten chunks across the drive, it'll take around 130 ms.[*] So in this case it actually transfers at 5% of its rated speed. And depending on the application, the data may be in many, many tiny chunks.

That being said, disk latency is one of the major causes of poor performance. But "bottlenecks" only have to do with throughput.

Latency limits throughput. The requestee usually can only satisfy a limited number of requests at once (see above), and the requestor may not be able to produce the next request until it's received the previous response.

Simple example: I'm performing a binary search. I need to see what's at location mid before I know if I'll next be interested in location (low+mid)/2 or location (mid+high)/2. In some cases, I can do a speculative fetch for both locations, but you can only extend that out so many generations before you've used up most of your bandwidth on data you'll never use.

Processors are smart about re-ordering instructions to keep working while they're waiting for stuff to happen, but still they frequently get to a point where they can't execute anything more because of ordering constraints - the results of some instruction are dependent on a previous instruction that hasn't completed yet because it's waiting for a value from memory. That value can be the actual instruction to be executed or an operand...either way, your shiny new processor's stuck doing nothing.

[*] - It might beat the average if it's smart about ordering. At the very least, 22 ms has to get added if one request is at one extreme and one request is at the other extreme. That brings it down to 23% of the rated speed.

Re:If only I/O speeds could also grow as fast

[Anonymous Coward]

...PHP... ...technology... BWAHAHAHAHAHAH!!!!

Re:If only I/O speeds could also grow as fast

[Joe The Dragon]

That is still better the have the cpu use the same bus and chip set that is also used by data and other cpus.

Re:If only I/O speeds could also grow as fast

[Mad Merlin]

Let's consider the example of my laptop hard drive. It's rated at a data transfer rate of 150 MB/s.

A SATA 1 interface can transfer at a maximum of 150 megs/s, but your hard drive can't. On sequential reads, you're unlikely to see much higher than 40 megs/s, even 7200 RPM desktop drives don't exceed 70 megs/s yet.

Re:If only I/O speeds could also grow as fast

[jadavis]

What you say does not make sense in the context of the entire thread. Here is how I understand the discussion (posts are obviously paraphrased):

Post #1: I/O is going to be an even greater bottleneck
Post #2: At large datacenters people often don't demand the full bandwidth of their I/O system
Post #3: I think #1 meant latency, not bandwidth
Post #4 (me): bottleneck refers to a bandwidth limitation, not latency.

Now you're talking about erratic request patterns that can decrease the performance of an I/O system. However, to measure both throughput and latency, you need to hold the requests constant. Latency is the time between when the requests arrive at the disk and the time the disk returns the first useful piece of data. Throughput is the rate at which the data arrives at the destination.

You have an I/O bottleneck (aka throughput problem) when the CPU idle time waiting for the full response to the request is much longer than the CPU idle time spent waiting for the first useful piece of data to arrive. You have a latency problem when the CPU idle time spent waiting for the first useful piece of data is much longer than the CPU idle time waiting for the full response.

I stand by my comment: bottleneck != latency problem. The seeking on a disk affects the throughput as well as the latency. The fact that you're not getting the theoretical maximum throughput for your series of random requests does not mean that you're experiencing a latency problem. A bottleneck can cause latency, and latency can cause a bottleneck (such as seeking within a disk, as you pointed out), but they are not equal.

Re:If only I/O speeds could also grow as fast

[jadavis]

If I continously tell the disk to seek to one extreme and read a cacheful, then seek to the other extreme and read a cacheful, it will neither be waiting for requests nor using its full bandwidth.

The disk is at maximum throughput for the given problem set (the requests you give it). A bandwidth rating on the box is for an ideal problem set, but if you change the problem set the bandwidth decreases.

Latency and throughput are unrelated only if there can be infinitely many requests produced and satisfied in parallel.

I never said they were unrelated, I said unequal.

Latency limits throughput.

In some cases.

A bottleneck is not latency. It's just not. A bottleneck can cause latency, and latency can cause a bottleneck, but they can happen independently as well.

Latency causes idle time on *both* devices. If we play tick-tack-toe by mail, that's a latency problem, and it's latent for both of us. If I write you a letter asking you to send me all the gold you find, that's a bandwidth problem (I can put it in my safe much faster than you can find it and send it to me).

Re:If only I/O speeds could also grow as fast

[slamb]

A SATA 1 interface can transfer at a maximum of 150 megs/s, but your hard drive can't. On sequential reads, you're unlikely to see much higher than 40 megs/s, even 7200 RPM desktop drives don't exceed 70 megs/s yet.

Yeah, you're right. I should have been more skeptical of that number, though usually the specs have real transfer speeds. When I tried it just now, I got 14.8 MiB/s reading from the beginning of the disk. Thought it was faster than that...

$ sudo dd if=/dev/disk0 of=/dev/null bs=1048576 count=1024
1024+0 records in
1024+0 records out
1073741824 bytes transferred in 69.151946 secs (15527283 bytes/sec)

But that doesn't change my point. There's a huge difference between sequential scan performance and scattered read performance. For many workloads, the device's latency is the bottleneck. (Or causes the bottleneck...split hairs about definitions and wording if you like, but it's what makes it slow.)

Re:If only I/O speeds could also grow as fast

[Anonymous Coward]

Most 7200RPM drive come out last year will read about 80MB/s.

Re:If only I/O speeds could also grow as fast

[RzUpAnmsCwrds]

It really depends on what you're doing. Wikipedia, for example, serves a tremendous number of hits with PHP and a very minimal (for Wikipedia's size) server setup.

Even a modest server should be able to hit ~ 100 pageviews/sec, depending on the database load and code complexity of your application.

If you're doing something that requires code execution speed, you're probably better off looking at Java or possibly even ASP.NET.

Re:If only I/O speeds could also grow as fast

[Anonymous Coward]

I think the person to whom you replied already knew that. That's why he wrote "pedantic sense". You just failed to understand his point.

Re:If only I/O speeds could also grow as fast

[lemaymd]

I hate to be pedantic, but my Seagate 320GB 7200RPM with perpendicular recording does 73.77 MB/s. It totally destroys my last-generation Hitachi 7200RPM. This still doesn't affect your argument, of course.

Re:If only I/O speeds could also grow as fast

[clickclickdrone]

ABout 8 years ago I was at at trade show and got jumped on by an Intel guy. We started chatting about future trends I said then I thought the PC had a huge disadvantage compared to mini's in terms of i/.o and he looked shifty and started to mumble a bit then eventually said 'yes, we know and we're working very hard on a totally new i/o system for PC's. Given the time, it may well have just been what we now have but at least they were aware that fast CPU's are worth nothing if they're waiting for the data.

Re:If only I/O speeds could also grow as fast

[jadavis]

The person to whom I replied was showing how latency internal to the drive can cause a throughput problem. However he didn't realize it was a throughput problem, he thought it was still latency.

If he had stopped after he said "This may be true in a very technical, pedantic sense," I would have stopped. However, he continued on with confusion about latency versus throughput.

Re:If only I/O speeds could also grow as fast

[Anonymous Coward]

That would certainly be true if my hard disk didn't have its own cache.

Re:If only I/O speeds could also grow as fast

[fm6]

Believe it or not, very few desktops have that kind of technology.

Re:If only I/O speeds could also grow as fast

[suv4x4]

If you're doing something that requires code execution speed, you're probably better off looking at Java or possibly even ASP.NET.

Easy to say, hard to do. When you and your team have invested years in experience and reusable code in a certain platform/technology, it won't be good news to just, switch language to get extra performance.

But faster CPU.. any time.

Re:If only I/O speeds could also grow as fast

[Anthony]

Couldn't agree more with the majority of file server NAS/SAN solutions. On the other hand, high performance Data I/O such as media rendering, supercomputing applications, high-rate transactional database management needs as much bandwidth as possible (multi GB/sec).

awesome powah

[Anonymous Coward]

their new proccesors allowed me to post first...

Re:awesome powah

[Blue0ctane]

Or not. Anyway, everyone knows that speed does not a processor make. Chances are that they'll cost an arm, a leg, and a Slashdot subscription.

Re:awesome powah

[that this is not und]

Slashdot subscriptions are cheap. You just throw them a little snippet of your positive PayPal balance you got from selling old gear on eBay.

40 % faster, after a nap

[Original Replica]

Everyone know that in Barcelona they take Ciesta. So don't plan on using you computer between noon and 1.

Re:40 % faster, after a nap

[macadamia_harold]

Everyone know that in Barcelona they take Ciesta. So don't plan on using you computer between noon and 1.

Yeah, but if it's 40% faster, an hour-long siesta should only 35 minutes.

Re:40 % faster, after a nap

[kfg]

Dude, noon to 1 is a lunch hour, not a siesta. The system will be down between 2 and 5; or maybe between 1 and 6 if some Ambar Negra gets into it.

KFG

Re:40 % faster, after a nap

[Anonymous Coward]

In Spain we use to take lunch at 2 or 3 pm so downtime will be more accurately in the range 3-5 :)

Re:40 % faster, after a nap

[12357bd]

Better Siesta and between noon and 5! :)

Re:40 % faster, after a nap

[Anonymous Coward]

Yeah,and if it's 40% faster then the 35 minute siesta should occur every 14 hours and 24 minutes

Re:40 % faster, after a nap

[fm6]

Yes, but in Clovertown they only computer when they're done with the milking...

Come to think of it, is Clovertown a real place? It's my understanding that Intel and AMD uses geographic codenames to avoid litigation. So there must be a real place called Clovertown. Except Wikipedia (which is compulsive about dumping geographic databases) doesn't have anything.

upcoming chipset?

[aczisny]

FTFS:

"Barcelona" chipset should be 40% faster than "Clovertown,"

You'd think since the blog got right that Barcelona is the upcoming processor from AMD, and since Clovertown is a processor codename from Intel, that the summary could have gotten it right too. Do submitters not read the articles either anymore?

Re:upcoming chipset?

[All_One_Mind]

ha ha. What are you talking about? Everything looks in order to me.

Re:upcoming chipset?

[Jeff DeMaagd]

Check the subject line. Clovertown and Barcelona are codenames for CPUs, not chipsets. "Chipset" is used to refer to the chips used to support the CPU, not the CPU itself.

Re:upcoming chipset?

[All_One_Mind]

ahh, good eye. I missed that. Thanks.

Re:upcoming chipset?

[digirat]

can't you call a processor line a set of chips, as in, chipset?

Re:upcoming chipset?

[aczisny]

I suppose you could, but people are going to give you funny looks. The chipset does support things for the cpu. It might have a memory management unit (like in the case of chipsets for Intel cpus), it might have graphics (if you have onboard, it's a part of the chipset. AMD has hyped up their Fusion project which will include some sort of graphics core on the die. It's been suggested before but iirc no one has finished the design and brought it to market), and of course there's an I/O hub.

A processor line is typically called a family.

Linux on quad-core

[Anonymous Coward]

Non-preemptive microarchicture like kernels don't generally do well on multi-cpu systems. I feel that as CPU's gain more core, Linux will find itself side-lined by more powerful operating systems, like Windows Vista and Solaris for example.

Re:Linux on quad-core

[be-fan]

Linux is a preemptible kernel.

Re:Linux on quad-core

[MoralHazard]

Urm, that depends. Linux CAN be a pre-emptible kernel, if you compile it to be. There are various levels of pre-emptibility, depending on your needs. The in-kernel docs say that pre-emption is intended for desktop environments where perceived latency is a big deal, but servers will probably benefit from the lessened overhead of a non-pre-emptible configuration.

But the original poster's comment is still bullshit. Windows Vista is a microkernel? What has THAT guy been smoking? Multi-core designs aren't that different from multi-CPU configurations, and we already know from experience that Linux hasn't been sidelined performance-wise.

Actually, now that I think about it, the likeliest explanation is that the OP was just trolling.

Re:Linux on quad-core

[PhrostyMcByte]

Multi-core designs aren't that different from multi-CPU configurations

While a good multi-core design will also get a perf boost from multi-cpu, it is quite different if you are aiming for top performance. Multi-core design doesn't really have to worry about NUMA but multi-cpu does, especially as the number of cpus goes up.

Re:Linux on quad-core

[dbIII]

Windows Vista is a microkernel? What has THAT guy been smoking?

This is the WinNT is really VMS stolen by the Illuminati and turned into a microkernel conspiracy theory that floats about every now and again by those that think computers operate by magic.

The bounty of true competition

[j. andrew rogers]

Whether AMD or Intel is producing the fastest, cheapest, most scalable, or most efficient processor at the moment is not terribly important.

What *is* important is that when you have two companies in genuine fierce competition at the bleeding edge of technology and performance, they extract an impressive amount of productivity and effort out of their engineering and science assets. Free markets are at their best when all the major players have a healthy fear of the capabilities of their competitors.

Re:The bounty of true competition

[Snarkhunter]

Yeah, totally unimportant. Unless you were gonna, you know... but a processor or something, but who on /. is gonna be doin that?

Re:The bounty of true competition

[Kuvter]

I have to agree with you on this, but it brought something to mind. If there wasn't fierce competition in this market we wouldn't necessarily have computers as fast as they are today, but we also wouldn't feel we need to upgrade software as much as we do. We wouldn't feel we need a faster computer and thus we wouldn't have to upgrade our computers every 7 years or less.

So yes in this situation it's great we're getting fasters products, but we're also as a society pushing for upgrading more and more and never being content with the computer and software we have currently.

Re:The bounty of true competition

[Anonymous Coward]

I have to agree with you on this, but it brought something to mind. If there wasn't fierce competition in this market we wouldn't necessarily have computers as fast as they are today, but we also wouldn't feel we need to upgrade software as much as we do. We wouldn't feel we need a faster computer and thus we wouldn't have to upgrade our computers every 7 years or less.

As a self proclaimed computer geek, I never thought I'd hear myself say this, but personally I have reached a point where my computer is "fast enough". I write audio software for Windows, I dabble in video editing, and I have my custom "media box" running Linux. My two laptops are faster than I will ever need them to be (mostly used for basic Internet and occasional /. surfing).

Granted, I don't get into gaming much, but since 1994 I always wanted more, never feeling satisfied with my computer's performance. Only recently have I become content with the performance, no longer feeling the need to have the fastest/best of everything. Other than the 1.5 GB RAM, most of my system is at least 2 years out of date. The laptop I'm typing this on even has a faster CPU than my primary desktop system...

Since I do not see any desperate need to "upgrade" to Vista in the foreseeable future, and XP does everything I need on most of my systems, I don't see any need personally to go faster, or move to 64-bits, etc. XP was the first Windows upgrade that, for me, didn't slow my system down much (going from Win2k). In my personal experience, XP boots faster than 2k did, and requires less frequent reboots. Since hardware speeds and reliability has gotten to the point that today's average system is "fast enough", for most people (not counting leading-edge gamers) there's no need to upgrade the hardware as long as it still works.

I believe this is part of the reason prices have gone down so much lately. My first brand-new laptop cost me $2200 in 2000. You can get a decent Dell laptop or desktop for around $700 these days, often with a printer/LCD panel/other accessories thrown in. Having the fastest/best/latest isn't as important these days, since any bargain PC is sufficient for surfing myspace, burning CDs/DVDs, and running the latest spyware-riddled crapware the kids are running...

Re:The bounty of true competition

[bacon55]

This isn't always true.

Resources can be wasted in competitive market grabbing that would have been better used improving the product.

I don't really want to know how much Intel spends on marketing.

AMD is growing

[adambha]

AMD says that the introduction of Barcelona marks a shift in their strategy from emphasizing price to performance.

While there are arguments both positive and negative toward the (somewhat) recent AMD/Dell alliance, this is one more indication that AMD is making even more progress in the processor market. Once considered the 'most bang for your buck' AMD is truly making a name for itself as a formidable competitor.

One of the fundamental principles of capitalism is that competition spurs growth and progress. This is a case in point.

Re:AMD is growing

[Moraelin]

While there are arguments both positive and negative toward the (somewhat) recent AMD/Dell alliance, this is one more indication that AMD is making even more progress in the processor market. Once considered the 'most bang for your buck' AMD is truly making a name for itself as a formidable competitor.

Actually, the way I remember history, AMD has only been "most bank for your buck" when it wasn't "best bang, period." As soon as it took the performance crown, an AMD computer (as in, once you also add the motherboard cost too) actually cost more than a similar performance Intel. As soon as it lost it, it cut prices and it was, "yeah, but we're best bang per buck" all over again.

Nothing against that, as it's just capitalism in action. Just pointing out that it's made the shift before, and the prices did end up higher than Intel's.

Which brings us to the other point that, you know, don't expect it to be any different this time.

Re:AMD is growing

[Anonymous Coward]

In that case the way you remember history is wrong. Ever since AMD made Intel's P4 look like a fool with their first Thunderbird CPUs they've had the best bang for your buck from low to midrange, all the way up to everything but the two or three fastest models. Since most people buying the bleeding edge CPUs don't really care about "bang for buck" but just "bang", you really don't need to include those.

That said, I will cede you the point that on a few (though rare) occasions Intel may have had a slightly better bang for buck on CPU prices alone, however the difference in costs for motherboards always more than made up for it.

And no, I'm not an Intel hater or an AMD fanboy. I'm just another geek who needs to get the most he can out of his limited income, and so I tend to watch prices and reviews fairly closely.

Re:AMD is growing

[2ms]

Then you must be new. AMD has basically been outperforming Intel ever since the 1.13Ghz PIII debacle. By Thunderbird and Thoroughbred in particular, AMD was outperforming the PIVs at about half the price. It wasn't until AMD64 that AMD started charging Intel prices. That's at least four years where AMD equaled or beat Intel performance for drastically less money (on the order of 50% price and often times less).

Don't believe this

[Nom du Keyboard]

All Intel has to do is turn up the clock the day before Barcelona ships. We already know that the Core 2 Duo chips are very overclockable, and getting another 40% -- or even 50%+ out of them -- shouldn't be a problem.

In fact I'll go further and say that buying any Intel (in my opinion, you fsking lawyers) before Barcelona launches is a Bad Move. It's seldom that performance increases by 50% in a calender year any more, as Mr. Steve Jobs found out a couple years back. This is not like the days when a 486 went from 33MHz to 66MHz in essentially one leap. As such, and you know this is coming, it is definitely worth delaying any purchase until after the Barcelona launch and see what the landscape looks like then. As much as I'm rooting for AMD, I'm surprised that C2D isn't already clocked closer to it's potential.

I feel like it's the old days with Intel right now where: We'll give you the clock speeds we've decided are best for you when we get around to it, and not before.

Re:Don't believe this

[Jeff DeMaagd]

I think the overclockability is in part due to the fact that Intel likes to have generous safety margins. It might also be that going beyond a certain level goes above their power consumption rating. Increasing the clock would mean reducing those margins.

Re:Don't believe this

[be-fan]

All Intel has to do is turn up the clock the day before Barcelona ships. We already know that the Core 2 Duo chips are very overclockable, and getting another 40% -- or even 50%+ out of them -- shouldn't be a problem.

The performance a chip can get with overclocking is way higher than what the manufacturer can deliver in final products. They have to be highly reliable at their specified clockspeed with (relatively) poor cooling, and while meeting the given voltage and thermal dissipation specifications. I've seen the Core 2 over-clock to 3.5 GHz (with conventional cooling) online, but how many of those are doing it at the stock Vcore while staying within the 65 watt TDP?

Re:Don't believe this

[Anonymous Coward]

There is no question people are getting much higher overclocks out of Core 2 Duo than they got with Pentium 4. Why do you doubt Intel can up the clock? It's quite obvious they can start binning at higher speeds with their current process. By the time Barcelona ships it will be even better.

Sure, they won't get that last drop of performance. DUH. But there's a lot of head room on Core 2 Duo.

Re:Don't believe this

[Nom du Keyboard]

but how many of those are doing it at the stock Vcore while staying within the 65 watt TDP?

You're assuming that Intel absolutely has to stay within those parameters. Expect them to step outside them the moment competitive advantage requires it. And they only have to do it with a few EE chips to claim the crown again.

Re:Don't believe this

[TheThiefMaster]

We are talking a SERVER line of cpus here. EE chips are a desktop cpu brand.

For servers TDP is incredibly important, because server rooms are air-conditioned, a room full of higher TDP cpus costs much much much more to run from an electricity point of view.

That's not to say that they won't overstep their vcore or TDP limits to get the upper hand on performance, but that wouldn't win them the performance/watt ratio crown that's the all-important stat for server cpus.

Re:Don't believe this

[be-fan]

The TDP and voltage levels are part of the platform specification. Intel can't just up them without requiring motherboards, cooling units, etc, to be upgraded to handle the new spec. They might get away with it for some consumer level stuff, but not in the server market where Clovertown and Barcelona are competing. The server folks are going to want some substantial lead-time to rejigger everything to meet higher TDP and Vcore specs.

Re:Don't believe this

[that this is not und]

Expect them to step outside them the moment competitive advantage requires it.

That would be, eight or twelve years ago. Or is it next year? Or do the facts show that they know they'd be creamed, and the PR disaster would make the Pentium floating point bug look like a company picnic?

Personally, I think they know more about what they're doing than some overclockers. Have you ever read an Intel datasheet? Have you ever read ANY IC datasheet?

Re:Don't believe this

[tomstdenis]

It may be partially because of yield, but when essentially all parts are overclockable ... it starts to smell fishy.

I've had two C2D processors [E6300 and E6600] and both were massively overclockable [1.83GHz => 2.94GHz and 2.4GHz => 3.42GHz respectively] with stable results. The only actual flakey part of the equation was the hot northbridge. If they upped the multipliers in the chips from 9 to 12 I suspect most E6600s, for instance, would be fine. Something tells me in the near future we'll see a bunch of above-3GHz chips just in the "nick of time" to beat AMD.

Tom

Re:Don't believe this

[tomstdenis]

I ran my C2Ds overclocked by at least 1GHz [both a E6300 and E6600] with stock voltage. I only turned by E6600 down to 3GHz [600MHz overclock, and yes, I can measure the diff between stock and oc rates] because the northbridge was getting a bit too warm for my liking.

I suspect staying in their TDP won't be a big stretch, especially as the design/process matures.

Tom

Re:Don't believe this

[be-fan]

Being able to overclock lower-rated chips and being able to increase the clock on your top line are two completely different things. With conventional cooling, the Core2 isn't seeing much above 3.5 GHz, and its already shipping at 2.93 GHz. It'll go up another 10-20% as the process matures, but Intel isn't holding back in any sense.

This is actually the nature of Intel's P-M and Core designs. Unlike previous designs, they're aimed at a frequency target, then tuned for power consumption. Circuits that could run faster than the given frequency target are redesigned to save power, thus reducing the frequency headroom. So in general you're not going to see huge clock-speed increases on the same process. You saw the same thing with Dothan at 90nm and Yonah at 65nm. Neither gained much more than 130 MHz during their life times.

Re:Don't believe this

[WhoBeDaPlaya]

Uhm, I'm getting 3.1GHz at stock Vcore, 3.3GHz at 1.34V. And you can't reasonably expect to up the clock frequency and not increase TDP even when maintaining stock Vcore. Gotta charge up all those lovely parasitic capacitances each activity cycle ;)

Re:Don't believe this

[sfe_software]

I've had two C2D processors [E6300 and E6600] and both were massively overclockable [1.83GHz => 2.94GHz and 2.4GHz => 3.42GHz respectively] with stable results.

I haven't done any overclocking with more recent processors, partly due to limitations in the motherboards I currently use... but I do remember a few years ago clocking my PIII 800 to over 1 GHz. I also remember the old Celeron 300a being widely known as overclockable, with some batches going as high as 466 MHz (more than 50% over the rated speed). I recall reading something about there being a market for "budget" processors, and the manufacturing process turning out more capable processors, still labeled at the lower speeds again due to market conditions...

In my own experience, I have noticed that Intel processors tend to be much more overclockable than AMDs. I can get my Athlon 2400 (clock of 1.8 GHz) to run at almost 2.0 GHz (FSB of 138), but it doesn't seem very stable (tried various voltages, has sufficient cooling, etc). Most Intel CPUs over the years, even going back to my old 233 MMX, handled higher FSB speeds stably...

My point being, Intel (judging from my own experiences) seems to provide a wider margin for error, higher tolerances, to ensure guaranteed performance at the rated speed. I recall a time when AMD beat Intel to the 1.0 GHz barrier, and many AMD processors at that time were known to overheat...

Re:Don't believe this

[be-fan]

Okay, 3.1 GHz stock Vcore. What's the Core 2 shipping at? 2.93 GHz. Wow, seems right in line, doesn't it?

So.. the writer expects Intel to sit still?

[trimbo]

Unless AMD employs completely incompetent morons as engineers, of course "Barcelona" should be faster than "Clovertown". Clovertown was released half a year to a year before Barcelona.

The tables have turned. Even though Clovertown is not a "true" quad-core (aka a single die), Intel has a huge head start on AMD on quad core. Intel will be pushing forward with their 45nm technology and pushing out yet more models by the time these arrive. With their fabrication prowess, I would expect the gap to increase over AMD. Since dumping NetBurst, Intel is finally battling AMD in an sport they can potentially win.

Re:So.. the writer expects Intel to sit still?

[Joe The Dragon]

Intel still has the FSB and that will get the way of the making the chips faster and add more IO.
Havening 2 dual-cores linked by a fsb bus will get in the way even faster as the speed of the cpu gets higher.
And a 4 cpu quad-core sever will likely choke up at the chipset to ram link as well as the chipset to chipset link.

Also intels dael quad-core workstation and the V8 only haves has the pci-e lanes for 1 x16 slot and the 8 other ones are used for the chipset to chipset link amd based ones will blow it away even more so with KL8 cpus. Right now an 2 cpu amd board has 4 pci-e x16 slots running at x16 x8 x16 x8 with 2 x4 lanes left over + each cpu can have a HTX slot or other HT based chip hook up to it.

Re:So.. the writer expects Intel to sit still?

[Aadain2001]

While all the technical reasons you listed are valid when looking towards future products, it is naive to think that Intel hasn't seen the same things you have. Intel does not employ stupid people (or at least for very long), so Intel will most likely have very good solutions to those problems by the time they are needed.

Re:So.. the writer expects Intel to sit still?

[funwithBSD]

Like say... RAMBUS? Or maybe Monterey?

Intel does in fact do very stupid things from time to time. The question is, is this one of them?

Re:So.. the writer expects Intel to sit still?

[Anonymous Coward]

I disagree.

While it's true that Intel got back in the game with Core 2 - actually pulled ahead -, what's often overlooked is that the main reason it is ahead is that Intel ships its procs with 2 to 4 Megs of cache while AMD make do with 512k to 1M. As always, Intel's real strength lies in their mastery of the process. They still haven't solved their memory bandwidth issues - clovertown is a joke as soon as you set it on non trivial apps.

The question at this point is will Intel manage to get back on the design front before AMD gets back on the process front?

Re:So.. the writer expects Intel to sit still?

[trimbo]

As always, Intel's real strength lies in their mastery of the process. They still haven't solved their memory bandwidth issues - clovertown is a joke as soon as you set it on non trivial apps.

Interesting. Have you used one? For what purpose? I never expected 8x performance on a dual clovertown box when I first fired off 8 processes of a popular 3D app to bake geometry. But that's exactly what I got when I tried it.

Re:So.. the writer expects Intel to sit still?

[evilviper]

Where the hell are the spelling and grammar Nazis when you need them?

For once, they might legitimately be able to make this post more easily comprehensible.

Re:So.. the writer expects Intel to sit still?

[Aadain2001]

Eh, everyone makes mistakes. Rambus is a great example of a something that sounded good in the begining, then quickly took a turn in the wrong direction. But as long as people (and companies) learn from their mistakes, they won't repeat them in the future.

Re:So.. the writer expects Intel to sit still?

[Sj0]

With AMDs Hammer platform, every new processor adds a hypertransport link, increasing memory bandwidth. I haven't paid any attention for 5 years and even *I* know that.

what's wrong with intel

[sarathmenon]

I am a sysadmin, and I've seen super weight systems taxed to the extreme. The best servers don't boast of the fastest clock speed; they have the best i/o buses, tight integration of the hardware and software, and more importantly, reliability. These are the reasons I've seen that amd makes a better choice than Intel. Intel is all about FUD, increasing the clock speed at any cost and in general, very unreliable systems that act strangely when pushed under heavy processor load.

I'd choose AMD over intel anyday - i've liked their strategies always, and in the server arena they are the best x86 player. But the bottom line still remains, sun's sparc line,ibm's ppc one and hp's rule. They have been in the business for quite some time, and they frankly know what they are doing.

Intel, its not late to figure out the economics. Corporations choose the best machine for the job while running their servers. No one chooses cheap when they are shopping for their new database server. The big bucks are in the hell expensive servers, and not in the mom-and-pop line. You can sell 1,00,000 cheap servers instead of 1000 expensive ones. But the margins are higher ony in the latter.

Re:what's wrong with intel

[Jeff DeMaagd]

I think you have a good point about other architectures. The problem with your post is, I have never had any problem with the reliability of Intel-based systems, even when running Windows, and even running dual Xeon systems running Netburst chips, my computers have been running cool and quietly. That, and Intel isn't necessarily about the highest clock anymore. I don't know if you've noticed, but they aren't promoting Netburst anymore either.

Imagine how great it would be...

[Anonymous Coward]

...if there was a similar competition in the OS market. You wouldn't need these mammoth processors in the first place. And having one would be a huge benefit, not a marginal one.

Re:Imagine how great it would be...

[jadavis]

...if there was a similar competition in the OS market. You wouldn't need these mammoth processors in the first place

What makes you say that? What makes you think that there are the same kind of performance gains possible from an OS? What makes you think there is not competition?

Re:Imagine how great it would be...

[drsmithy]

You wouldn't need these mammoth processors in the first place. And having one would be a huge benefit, not a marginal one.

And you base this on what, exactly ?

You mean if they made OSX for all PC's?

[Cordath]

Why doesn't Apple do this? One the one hand, they risk losing a portion of a rather small chunk of the desktop market. Their laptops are already good value for the money and would probably continue to sell just as well. However, look at the install base of normal PC's out there. Then consider how many people would run MacOS if they didn't have to buy a separate machine to do so. Getting people to switch when the price is under $200 is a lot easier than it is currently, when switching requires completely new hardware. (unless you're willing to hack things a little, which the majority of people aren't.) In the end, I think Apple would win big and so would consumers. So why is Jobs hyping gadgets like the iPhone when he has a product ready to go toe-to-toe with Microsoft for a significant share of the global PC market? I think the iPhone will be a winner, but not on the scale that an officially sanctioned PC-version of OSX would be.

Re:You mean if they made OSX for all PC's?

[the eric conspiracy]

Apple would get creamed if they did this. Most PC's out there have hardware that Apple does not have drivers for, by decoupling the OS from the hardware people would be able to pirate the OS, and of course Apple makes more money on the hardware they sell than the OS.

Re:You mean if they made OSX for all PC's?

[that this is not und]

but not on the scale that an officially sanctioned PC-version of OSX would be.

Officially sanctioned on what tiny subset of the PC hardware that's out there? Apple could never support the huge x86 hardware base out there, in fact a big part of their quality success comes from them having tight control on both the hardware and software aspects of their platform.

Also they could never handle the tech support calls. "Why doesn't my ISA-bus hand-scanner from Windows 3.1 work on OSX?"

Re:You mean if they made OSX for all PC's?

[vakuona]

No they don't. The amount of money they spend on software tends to prove that point. Apple sells the same hardware (more or less) as Dell nowadays. The software is the whole difference. That is why they make money.

Re:You mean if they made OSX for all PC's?

[Joe The Dragon]

Apple should at least come with a mid-end desktop with a real video card as install base of normal PC's out there have good monitors that people reuse on new systems.
also a $1500 laptop with real video will help them get more of the market.

It's to bad the lintel chips forced them to have so few pci-e lanes in the mac pro.

Re:You mean if they made OSX for all PC's?

[timeOday]

Why doesn't Apple do this?

Because then OSX would be just as troublesome as Windows and Linux. Apple's limited hardware support is a big part of why everything "just works."

Re:You mean if they made OSX for all PC's?

[XnavxeMiyyep]

The software is the whole difference. That is why they make money.

While the software IS the difference, Apple still makes far more money on hardware. It is simply the software that sells the hardware.

Re:You mean if they made OSX for all PC's?

[NRecob]

"why everything "just works.""

What a laugh. Sure fanboy. Be a good little MacStain and go sell your hype elsewhere.

Re:You mean if they made OSX for all PC's?

[Anonymous Coward]

I beg to differ.

When was the last time you installed a new piece of hardware and it messed your system up?
I can understand it worked/didn't work, but it definitely didn't break your system.

The main problem with windows is that everyone fscks with the kernel space. Norton, McAfee and all your anti-virus/firewalls basically do their job by modifying your kernel. Crappy programs modify dlls left and right.

Compare the times that you installed/uninstalled a program that messed your system up vs. a new hardware or a driver.

The idea that Apple is not allowing third parties in their hardware business is to maintain stability is false IMHO.
Apple does _not_ need to support the huge x86 hardware like you said, the hardware manufacturers are the ones who support Apple.

If they wanted they can say "Ok look, here are the rules. Any driver you write must do XYZ and NOT do ABC. To get your hardware running on Mac you're gonna have to pay us some serious dough to give you a certificate. If you don't have the certificate your OSX will completely block your hardware/drivers".

3rd party hardware for Apple is possible, they just don't want it.

P.S.: The rage (if any) in this post wasn't directed at you personally, I just wanted to raise my voice in the whole Mac hand/glove experience thing. Thanks :)

Re:You mean if they made OSX for all PC's?

[evilviper]

Because then OSX would be just as troublesome as Windows and Linux.

No. Not unless Apple half-asses the project.

Linux is troublesome only because it hasn't been designed to just work, and because vendor support is nil.

The BSDs are far better, in this area. OpenBSD detects all the hardware in a system on boot-up. So either it works, or it doesn't. There is no messing around, no configuration, no hdparm, no editing modules.conf, no loading modules, etc.

FreeBSD, is probably the best example, though. It detects pretty much everything except for soundcards on boot-up, but that is detected and loaded by the installation program, and you can tell it to load all the soundcard modules, without adverse effects, and be able to put that FreeBSD system in any PC hardware around. Even with that, FreeBSD is rock-solid on every bit of PC hardware out there.

Linux has two minor things over the BSDs... One is eth0... The BSDs use different names for network cards (xl0, de0, etc.), so you have to set your network addresses again when you change cards, or you have to put the information in the config file repeatedly, with the different card types you think you might use. The other advantage is X11, since the BSD's don't have programs like Kudzu, X11 needs to be configured by hand. Not that either of these two issues isn't pretty close to trivial to workaround.

So with OSes like FreeBSD out there, and working perfectly right now, the excuses of instability due to variety don't hold one bit of credibility. It's purely an attempt to do what software companies have always done... Blame hardware for their bugs.

Re:You mean if they made OSX for all PC's?

[blackest_k]

possibly
but probably not.
how about the following scenario:

There would be two phases to buying osx for pc.
A try before you buy program to test if your hardware is good enough to run osx.
perhaps it would generate a serial based on the tested hardware. If the hardware isn't upto the requirements then it would list the failing hardware and suggest solutions and maybe an equivilent macintosh.

with a serial you can make osx available to download possibly limiting hardware support to that configuration.
with the amount of linux and bsd open source drivers around its possible third party developers / mac enthusiasts could produce community based drivers.

Apple would be getting paid for this so would have an income from the osx sales for a budget to enlarge osx support for other hardware.

There is no reason to suppose they would charge the same as for regular macs either a 100% premium wouldn't be out of order after all mac hardware does subsidise osx.

however just because something is possible, doesn't mean it will happen, what would it do to apple sales?
Ensuring osx sells at a premium for generic pc hardware, would ensure that it would be cheaper to buy the equivilent mac rather than buy a 3rd party pc to put osx on.

The biggest block to this happening is Steve jobs and I doubt he could be persuaded.
Then there are your mum and dad would they buy a software only solution.

In someways its kind of similar to the TomTom GPS situation you can run the software on a pda but its a lot simpler and more reliable to just buy a dedicated unit.

Re:You mean if they made OSX for all PC's?

[vakuona]

The only reason that people pay that much money is because it comes with seriously good software. Ergo, people are paying for the software. They just don't sell it separately. Why else do you think they charge people for newer versions of their software. They want to make money of it.

Apple doesn't sell hardware without software, so there is no way to say they make money on the hardware than on the software.

Re:You mean if they made OSX for all PC's?

[evilviper]

I fail to see how this is a troll at all.

It is just some simple facts. BSD kernels auto-detect most hardware on boot-up, while Linux does not. So that makes something like FreeBSD a lot closer to the OS X "just works" motto than Linux, while still being completely stable on the vast array of hardware out there.

Re:You mean if they made OSX for all PC's?

[Anonymous Coward]

Linux has two minor things over the BSDs... One is eth0... The BSDs use different names for network cards (xl0, de0, etc.), so you have to set your network addresses again when you change cards, or you have to put the information in the config file repeatedly, with the different card types you think you might use.

I would agree if you change a network card to an already configured system. However, when there are two different network cards in a system, in BSD, I know exactly which one I am configuring, but in linux, it depends on where on the bus it is, and I'm usually wrong on the scan order. :(

Re:You mean if they made OSX for all PC's?

[evilviper]

However, when there are two different network cards in a system, in BSD, I know exactly which one I am configuring, but in linux, it depends on where on the bus it is, and I'm usually wrong on the scan order. :(

I very, very rarely end up with DIFFERENT cards in a system. If it's something like a router, I'll have 5 cards in there, but they'll all be the same brand. The naming scheme simply doesn't help differentiate them.

Plus, if you add/remove on of those cards, it might bump the numbering on all the rest of them, so it doesn't help there, either.

Even when I use the integrated NICs, more often than not, it is (coincidentally) using the same driver as the PCI NICs anyhow... There's only a few common chipsets out there.

If they wanted to create a network interface naming system that didn't clash, they could easily prefix "eth" (or whatever), and then suffix it with the last (4?) digits of the MAC address, to make it extremely unique, and vastly unlikely to clash.

But, since they AREN'T doing that, they might as well use a catch-all name like eth0. No worse in 99% of cases, and better in 20% of cases (wild guess).

Re:You mean if they made OSX for all PC's?

[Anonymous Coward]

sigh. nobody ever used the word "ergo" before the matrix. here's hoping "hence" makes a come back.

Mainly in FP

[Visaris]

This 40% faster than Clovertown claim is only referring to FP code. The integer side is not nearly as clear. Expect AMD to improve integer performance over K8, but I don't expect any miracles. Here is a small list of improvements Barcelona will have over K8:

- Double L1 cache bandwidth
- Double FP units
- Single-cycle SSE (vs K8's 2-cycle)
- More fast-path decoding
- Double TLB size
- Independent DDR channels
- More cache (L3)
- Out-of-Order loads
- New instructions (LZCNT, POPCNT, EXTRQ/INSERTQ, MOVNTSD/MOVNTSS)
- Double prefetch (from 16 bytes -> 32 bytes)
- Larger Branch Target Buffer
- Larger Out of Order (OoO) buffers
- Support for new HT standard (3.0)

Re:Mainly in FP

[Erich]

New instructions (LZCNT, POPCNT, EXTRQ/INSERTQ, MOVNTSD/MOVNTSS)

Interesting!

I can't find much information on it, but I'm guessing "LZCNT" is count-leading-zeros. This is like "find-first-one" from the other direction. It's very useful for things like finding the magnitude of an unsigned numbers. It's used quite often on architectures without FPUs (like ARM) in floating point routines for renormalization. I guess it could also be useful if you are having to do floating point emulation for numbers with enourmous precision.

I guess if you have "BSR" then LZCNT = -BSR

POPCNT is probably population count, the number of 1s in a value.

Both LZCNT and POPCNT are instructions that are a pain to do in software if you lack the instruction in the hardware, and they are relatively cheap (especially if you have BSF/BSR already).

I'm still a bit suprised that there aren't a few more of these bit-banging instructions in x86, like bit interleave/deinterleave and bit reverse. Modern processors are doing enough signal processing work that one would think you'd thow the tools in the bucket, as cheap as they are. I guess lookup tables are good enough.

What's the over/under for which SSE revision will add a galois field multiplier? 7? 8?

But seriously, the dual ported caches are probably the best improvement for most people. You can't be too rich, too thin, or have too much memory bandwitdth.

It looks like AMD has done the same thing Intel did with "Core 2"... just take a good architecture and keep making improvements... more issue width, more memory bandwidth, more flexibility in scheduling. Every bit counts.

I think we're getting to a similar point in modern CPU microarchitectures to where we are in some other industries, where drastic improvements are much more rare and it all comes down to really great implementation... like making engines. There are some innovative ideas for engines, and certainly a lot of people experiment, but really the best designs are just really well balanced and tuned. (although more cylinders is usually a good thing for horsepower).

Re:Mainly in FP

[Bj�rn]

It will be interesting to see if Intel follows AMD and copies the new instructions.

Re:Mainly in FP

[Watson Ladd]

Oh, come on. MMIX beats all other instruction sets hands down! Now we just need to get it into sillicon.

Re:Mainly in FP

[ghoul]

LZCOUNT and POPCOUNT are for making the call stack more efficient

Re:Mainly in FP

[Anonymous Coward]

LZCNT is also very useful when decoding variable length codes - zip files, audio/video codecs, ...

In Other News

[benc]

Coke says "Coke Ultra X2" will taste better than "ÜberPepsi With Pomegranate." ALERT THE MEDIA

By what criterion=

[Moraelin]

The way I remember the definition of a "bottleneck", it's basically the lowest performance part that brakes everything else. It doesn't matter _why_ it's low performance. Just that less data gets through the bottle's neck than through everything else. If it brakes performance and other sub-systems have to wait for it, it's a bottleneck and that's that.

Defining it as "throughput" is at best prestidigitation. The _real_ throughput is how much data actually gets through. No more, no less. The keywords being "actually gets through". Not theoretical throughputs in some purely SF scenarios full of conditions and assumptions are never actually true. (Everything is read sequentially, everything fits in one single cylinder so there are no seeks, etc.)

Basically think of it this way: let's say that I write a super-duper program that, dumbly enough, uses bubble-sort for its main database sorting and searching. (Assume I write my own database files.) It doesn't matter in what theoretical conditions (e.g., the database is already sorted) my algorith would have great throughput, it only matters if in practice it brakes everything else or not. That's it. If it's the #1 cause of performance loss, then it's the bottleneck, and the next thing that should be optimized. If I came to you and said, "well, see, on the ideal case (again, think already sorted database file) my sort algorithm wouldn't be bad at all and it would have great throughput, so it's not the bottleneck, so I'm going to optimize something else instead", you'd probably tell me to take a hike and get a clue.

Re:By what criterion=

[jadavis]

The way I remember the definition of a "bottleneck", it's basically the lowest performance part that brakes everything else.

Bottleneck is an analogy.

If you have a bottle of beer, you can drink the beer much faster than the neck of the bottle allows (at least in the short term, with normal gravitational forces propelling the beer at normal beer-flowing speeds and accounting for the need to exchange air back into the bottle). If you widen the thinnest part of the neck, the second-thinnest part will then become the point of constriction, and will be the new "bottleneck". This is a throughput problem.

However, something like shampoo (when there's only a little left in the bottle) is a latency problem.

mbod up

[Anonymous Coward]

cont8ibuted code lagged behind,

Because competition is good for the CONSUMER

[p3d0]

Competition only hurts a company's bottom line.

You're Right

[Anonymous Coward]

It's about time Humans upgraded their crummy I/O ports.

Obvious

[Mazin07]

Jeez, of course AMD's right. Take a look:

Population of Barcelona: 1,673,075
Population of Clovertown: 5601 (or less)

Barcelona is vastly superior.

Re:Obvious

[dbIII]

Barcelona is vastly superior.

Especially in terms of the architecture.

Mod parent up!!!

[Anonymous Coward]

How the hell is this flamebait? He's just giving his honest opinion, whether you agree with it or not is irrelevant. He's clearly not trying to antagonize anyone.

Re:Mod parent up!!!

[MrNaz]

You must be new here.

Re:Mod parent up!!!

[dfghjk]

Maybe it has something to do with the "Intel is all about FUD..." comment. That's pretty antagonistic.

Re:Mod parent up!!!

[Anonymous Coward]

How the hell is this flamebait? He's just giving his honest opinion, whether you agree with it or not is irrelevant. He's clearly not trying to antagonize anyone.

It could be because of his premise:

Intel is all about FUD, increasing the clock speed at any cost

Not true since NetBurst. We're now several generations past that.

and in general, very unreliable systems that act strangely when pushed under heavy processor load.

I've never heard Intel referred to as the "unreliable" brand, so at least one example would be required for his own statement not to be FUD.

In 4X4 Motherboards

[airencracken]

It'll be interesting to see how this chip preforms in the 4X4 system that AMD has been promoting. I imagine it will have a lot of horsepower.

Marketing speak

[suv4x4]

Intel's quad-core Xeon 5300 line. AMD says that the introduction of Barcelona marks a shift in their strategy from emphasizing price to performance

The way they spun it, you can also claim they changed their strategy from slow to expensive.

faster at what?

[80 85 83 83 89 33]

funny. but what i really want to know is: what the hell does amd mean by "40 percent faster"??? really, has somebody cornered them on it? i bet they found some wierd statisical ratio or other hogwash that with a little fuzzy math gave a number of 40%. does it roll down a hill faster, because it is more round? GOTO loops?

how about a well known or commonly used benchmark, AMD?

and since i'm posting this question a whole seven hours after the story (an eternity in /.), the odds of someone who might have more details posting a reply are near zero. oh well....

Meh.

[ColaMan]

My .sig says it all, really.

I must be getting old. Once upon a time, I drooled over a P90 and how much faster it was compared to my DX2-66.
Now, it's just a feeble wave of the index finger and a sarcastic, "Processors are getting incrementally faster? woo-hoo."

AMD is not and never has been a serious CPU

[GuyFawkes]

The day you can remove the fan and heatsink from a running AMD CPU and it will simply carry on running throttled down until the fan and heatsink are replaced, they will be ready for "professional" use.

You only have to see one or two CPUs go up in smoke because the HSF failed or whatever, taking the mobo with it, and blowing the PWN thus frying things like hard disks, to never ever ever fit anything but Intel to a "serious" machine.

I recently swapped from 3.5 GHz P4 to AMD 3200, both on A-bit, by way of experimentation.

Allegedly these are similar CPUs.

In practice, the AMD is slightly faster at some things such as gaming.

Overall the Intel, thanks to hyperthreading, felt faster.

The AMD was slightly unstable, while the intel needed the windows xp driver updates to achieve best performance, the amd needed them to achieve stability, as without them it has a *slight* tendency to reboot.

One of the reasons AMD were cheaper, bang for buck, is they left out all the extra stuff Intel did not, like on chip thermal management so it didn't catch fire when the heatsink / fan failed. Penny wise, Pound foolish.

Re:AMD is not and never has been a serious CPU

[ocbwilg]

The day you can remove the fan and heatsink from a running AMD CPU and it will simply carry on running throttled down until the fan and heatsink are replaced, they will be ready for "professional" use.

That's an oddly arbitrary statement. As an IT professional, I have to say that I have never once had the occasion to remove the HSF assembly from a CPU in a running PC (or had a user that did). In fact, it seems like an incredibly stupid thing to do. So why would that somehow be the mark of a "professional" CPU?

And more to the point, you're referring undoubtedly to that old Tom's Hardware article from 3 or 4 years ago where they performed this test to demonstrate the value of thermal monitoring circuitry. At the time they were comparing a Pentium 4 and an Athlon CPU. I don't think that it was even as current as an Athlon XP CPU, let alone an Athlon 64, X2, or Opteron (which BTW *DO* have the thermal protection circuitry in them).

I'm also curious about the constant claims of system instability on AMD-based systems. I've used AMD CPUs exclusively for years and had no issues with them. I recently outfitted my entire 250 system organization with AMD-based desktops and laptops, and have had no problem with them. Every tier 1 PC/server vendor (and most tier 2 and 3 vendors as well) have validated AMD CPUs and are selling systems based on them. Yet there's always someone claiming that AMD CPUs somehow aren't 100% compatible with Windows, or that they can't run an OS in a stable fashion. Have you ever considered that perhaps your problems have more to do with the incredibly complicated software that you're running (Windows), or your lack of PC knowledge?

Re:AMD is not and never has been a serious CPU

[SScorpio]

I think the OP is not trying to say that someone should remove the HSF while the computer is running, but the CPU should be able to handle the situation in case the fan were to fail the CPU would not just burn out from overheating.

I do agree that AMD CPU's are stable and compatible with Windows as I'm running an Athlon 64 right now. I'm not sure if AMD's CPUs do the down clocking but I wouldn't be surprised if they do. Though I'd rather not risk my system by testing it.

Re:AMD is not and never has been a serious CPU

[GuyFawkes]

Ok, I'll bite

Users removing fans and heatsinks while the CPU is operating, no, fans failing and even heatsink clips failing while the CPU is operating, yes.

Opteron does NOT have equivalent thermal monitoring and control as the equivalent Intel chip, anyone who says different is either lying or does not know the subject.

As for the stability of AMD, it's just a gut feeling.

You can blow your "incredibly complicated software" and "my lack of PC knowledge" out of your ass.

When I install an OS and software and that OS is Windows I do not have an option to compile for one CPU over another, therefore one must assume that as far as Windows is concerned they are equal. if this results in issues then they are Microsoft issues, not mine.

Lack of PC knowledge, I didn't build an Altair because a/ they were too expensive and b/ I had access to a DEC, anyway, only an idiot would assume that "lack of pc knowledge" would allow simply swapping an AMD on Abit for an Intel on Abit, which, being bloody windows will require a complete OS reinstall anyway, is going to generate some instability because I forget to realign the dilithium crystals and twizzle the foo foo valve.

In your own words, you RECENTLY outfitted your entire org with 250 AMD desktops and laptops.

RECENTLY, yet you still feel qualified to take a pop at people who have such experience under their belts, not once, and not recently.

IMHO based upon decades of experience, Intel is a better buy in the long run, more expensive, yes, at purchase time, but years down the line the quality wins out. I have never come across an Intel CPU that died in normal service, not one, ever.

I have come across a mere handful, literally, of AMD CPUs that died in normal service, two of them took the mobos and hard disks (via pwm) with them, one of them triggered the smoke / fire system in the rack, and cut power to the entire rack.

I am not at liberty to name names here, but the discussions afterwards were about the 4000 pounds saved on that rack by going amd instead of intel, versus the 9000 pounds the downtime cost.

Buying cheap you save money ONCE.

buying quality you save money every single time nothing happens.

Re:AMD is not and never has been a serious CPU

[Anonymous Coward]

As for the stability of AMD, it's just a gut feeling.

This pretty much sums it up: you've pulled this from your ass along with a couple of anecdotes.

Re:AMD is not and never has been a serious CPU

[ocbwilg]

Users removing fans and heatsinks while the CPU is operating, no, fans failing and even heatsink clips failing while the CPU is operating, yes.

Well, I've had issues with CPU fans failing and never had any hardware failure as a result. In the cases that I have seen the temp will rise, and when it reaches an unsafe threshold the system shuts down, but that hardly means that your system is fried.

You can blow your "incredibly complicated software" and "my lack of PC knowledge" out of your ass.

There's an intelligent and well thought rebuttal. I think that I may have heard something similar on a playground once. If you want to ignore the possibility that your problems may have been the result of a software glitch or incorrect configuration, then I suppose that's your choice.

anyway, only an idiot would assume that "lack of pc knowledge" would allow simply swapping an AMD on Abit for an Intel on Abit, which, being bloody windows will require a complete OS reinstall anyway, is going to generate some instability because I forget to realign the dilithium crystals and twizzle the foo foo valve.

Let me make sure that I understand this clearly. You had problems with an AMD-based system. Then you swapped out the mainboard and CPU and reinstalled Windows, and the problem went away, therefore you believe that AMD was the problem? What if it was the mainboard? What if it was a software problem or corrupt file and reinstalling Windows fixed it?

In your own words, you RECENTLY outfitted your entire org with 250 AMD desktops and laptops.

RECENTLY, yet you still feel qualified to take a pop at people who have such experience under their belts, not once, and not recently.

Yes, but also in my own words, I said that "I've used AMD CPUs exclusively for years and had no issues with them." You very conveniently left that point out. When I made the comment about "recently outfitting my entire organization with AMD-based systems" I wasn't saying that "this is my only experience with AMD," though you obviously wanted to twist it that way. I was merely giving an example of the kinds of deployments that I have done with AMD, to demonstrate that I'm not some hobbyist who builds two PCs a year for his relatives and therefore considers himself an expert on the subject.

And for the record, I'm not "taking a pop" at anyone. I am merely correcting an erroneous statement that isn't supported by anything other anecdote. I haven't made any sorts of comments against you personally. But if you really want to get into a pissing match I suppose we can whip out our IT pedigrees and you can brag about how your 40 years of IT experience trumps my 30 years of IT experience, though anyone with an ounce of sense would realize that the net difference (especially when discussing current hardware) is nil.

Of course, you are also more than welcome to expand on your "decades of experience" in deploying both Intel and AMD-based solutions in enterprise environments. I'm assuming that you do have that sort of experience, right? Because if you've been deploying and working in Intel-only environments for decades, then you're hardly qualified to comment on the difference between the two vendors, are you?

IMHO based upon decades of experience, Intel is a better buy in the long run, more expensive, yes, at purchase time, but years down the line the quality wins out. I have never come across an Intel CPU that died in normal service, not one, ever.

I have. In fact, the only CPU that I ever had die on me was a Pentium 4. Go figure. You have anecdotal experience to support your case, I have anecdotal experience that undermines it. So who is right? I'm going to put my money on what the major PC and server manufacturers think. AMD was making CPUs for the IBM PC in the early 1980's, just like Intel was. Current AMD CPUs have been validated by all of the major manufacturers and are being sold in their consumer and enterprise hardware lines. If you want to put your anecdotal experience up against the thousands of hours of validation testing that these companies do, feel free, but you'll only make yourself look sillier by comparison.

Re:AMD is not and never has been a serious CPU

[androvsky]

Since you were very snotty to someone who replied to you, I'm posting the snotty version of my reply... There is a point at which one must face facts. Since the AMD 64 line, AMD has had the superior cpu over the P4. The AMD line was, with few exceptions, cheaper, cooler, faster, and as stable, if not more so, and it did that while still including your precious protection against forgetting to put on the heatsink. Is that something you do often? Yes, the Athlon line was neck-and-neck with the P3, the Athlon XP line did well against the P4 until the end, the 64 line, like I mentioned, kicked ass until the Core 2 Duo came out, something you might not be aware of since you're posting problems with AMD cpus that haven't been an issue since 2001. Yes, the Core 2 line is awesome, but for high-end server work (expensive stuff), the AMD architecture still has some very nice things going for it.

Re:AMD is not and never has been a serious CPU

[RzUpAnmsCwrds]

The day you can remove the fan and heatsink from a running AMD CPU and it will simply carry on running throttled down until the fan and heatsink are replaced, they will be ready for "professional" use.

This just isn't true, nor is it really relavent. Intel CPUs do throttle if you have - say - a fan failure, but the throttling is not enough to keep the CPU stable without a heatsink.

I have pulled the heatsink from an old Northwood, and, let me say this - the results are not pretty. The system crashed almost immediately.

The Tom's Hardware tests you are probably referring to were pretty clearly faked.

And, more to the point, when was the last time that you saw heatsink fell of of a system while it was operating? Fan failures, yes. Heatsinks falling off - not unless the box is dropkicked.

The AMD was slightly unstable

Was it? Tell that to the people who have been running Opterons successfully for years in server environments. Tell that to Dell, to HP, to Sun, to IBM, or to the millions of people who use AMD CPUs every day.

One of the reasons AMD were cheaper, bang for buck, is they left out all the extra stuff Intel did not, like on chip thermal management so it didn't catch fire when the heatsink / fan failed.

AMD CPUs have had on-die thermal management since Athlon 64, and chipset-implemented thermal management since the Athlon XP.

Intel's thermal montior (TM1) feature has been the source of hell for lots of users. It's a good idea, poorly implemented - instead of halting the system or producing an error, the system continues to run - poorly. It makes it difficult to diagnose whether or not the heatsink is working properly, unless you use tools which detect throttling, which, unfortunately, aren't bult in to Windows.

Re:AMD is not and never has been a serious CPU

[dbIII]

The day you can remove the fan and heatsink from a running AMD CPU and it will simply carry on running throttled down until the fan and heatsink are replaced, they will be ready for "professional" use.

It is also a bad idea to run a car with no water in the radiator. It is best to read instructions and assemble things correctly, you have a lot of heat being generated in a small space. As for the comparison above, there are a lot of variables (chipsets etc) making any comparison difficult but it is a good anecdote. My anecdote about older AMD chips - I've had a lot of fans die on AMD systems and a few originally in badly designed cases that nearly worked as ovens and haven't lost very many of them. I still have five out of seven dual AMD boards that have been running since 2003, only turned off for a few days each year when no work is being done - so around 25000 hours of operation. These machines are at 100% CPU usage for a week at a time for some jobs (so would have reached the maximum possible temperature and kept at it for days), the surviving ones lived through air conditioning faults and even fan failures on some.

Both major x86 manufacturers have advantages and disadvantages - having to get two matched Intel CPUs from Ebay for an 18 month old server was annoying after one CPU died and there was nowhere I could find a new one of the correct bus speed and stepping. As for not being suitable for "professional" use - who else would use an 8 way CPU? I'm looking forward to being able to get two of these things when the price comes down and be able to have 1U sized 8 CPU machines (in two sockets on a board that isn't huge) instead of the big things you need now.

Re:AMD is not and never has been a serious CPU

[dbIII]

I have never come across an Intel CPU that died in normal service, not one, ever.

This more than anything else shows the lack of experience - things break if you look at enough of them.

Also there are major difference between a 233MHz AMD CPU that runs hot and a recent Opteron - years of casual exposure is no substitute for paying attention to recent developments and then making wild claims based on running things contrary to their design - talking about running modern CPUs without cooling shows ignorance. While it would be nice to do that the trade offs are only worth it in specific situations like the low power consumption CPUs from a variety of manufacturers (including AMD) - but even those should have a heatsink.

Re:AMD is not and never has been a serious CPU

[WuphonsReach]

1) I'd bet that the AMD chip is more picky about the memory that you have installed. Which is probably why you had reboot issues. It's one of the few downsides of having the memory controller on the CPU.

2) AFAIK, all AMD chips that came out after the AthlonXP series have on-chip sensors and thermal throttling (AMD Cool-n-quiet?). At the time of the infamous Tom's Hardware test, Intel chips had thermal throttling and AMD chips didn't. And up until Core and Core 2, Intel chips were affectionately known as space heaters, taking back the space heater crown from the old AMD Athlon chips.

3) I call bullshit on a CPU failure taking out other components.

check your facts pls

[Deternal]

Athlon cpu's have had build in heat sensoring since they started using the performance numbers for their cpu's again.

Of course it needs to be supported by the motherboard as well to work, which it is on serious boards.

Re:check your facts pls

[GuyFawkes]

This is still not the same as Intel

Intels own words

What measures the temperatures:
"There are two independent thermal sensing devices in the Pentium 4 processor on 90 nm process. One is the on-die thermal diode and the other is in the temperature sensor used for the Thermal Monitor and for THERMTRIP#. The Thermal Monitor's temperature sensor and the on-die thermal diode are independent and physically isolated devices with no defined correlation to one another. Circuit constraints and performance requirements prevent the Thermal Monitor's temperature sensor and the on-die thermal diode from being located at the same place on the silicon. The temperature distribution across the die may result in significant temperature differences between the on-die thermal diode and the Thermal Monitor's temperature sensor. This temperature variability across the die is highly dependent on the application being run. As a result, it is not possible to predict the activation of the thermal control circuit by monitoring the on-die thermal diode."
From the above statement it's clear that the Thermal Monitor is used to control the processor temperature by activating the TCC (Throttling) when the processor silicon has exceeded its maximum operating temperature. The TCC reduces processor power consumption as needed by modulating (starting and stopping) the internal processor core clocks. Other than the Thermal Monitor Intel placed a thermal diode on-die to give end users a way to monitor temperatures of the core.
How the temperature is figured:

"The processor temperature is determined through an analog thermal sensor circuit comprised of a temperature sensing diode, a factory calibrated reference current source, and a current comparator (see above image). A voltage applied across the diode induces a current flow that varies with temperature. By comparing this current with the reference current, the processor temperature can be determined."
The temperatures we see using MBM5, ASUS Probe, and other monitoring programs come from the on-die thermal diode. Thanks to a thermal sensor (ie: Super I/O chip) located on the motherboard the data from the thermal diode can be used to monitor the die temperature of the processor for thermal management. This thermal diode is separate from the Thermal Monitor's thermal sensor and cannot be used to predict the behavior of the Thermal Monitor.
What influences the reading:
Every processor has different amounts of leakages and is therefore unique. Since each processor is unique the readings between two processors should never be the same. One of the parameters used in the equation to determine the CPU temperature is the Diode Ideality Factor (A.K.A. the Non-Ideality Factor) which describes the behavior of the diode relative to a theoretically perfect diode. The ideality factor depends on the characteristics of each individual processor and will vary slightly from one chip to the next. According to the data sheet the range of ideality for the Prescott ranges from 1.008 to 1.0137. The temperature calculation should take this range in to account in order to improve the accuracy of the reading.
One other variable that influences the temperature accuracy is the "'Series Resistance", which is a measure of the resistance in the traces leading up to and away from the thermal diode. The Prescott data sheet shows that this ranges between 3.242 and 3.594 ohms. Some diode sensors have the ability to adjust for the temperature error caused by the resistance, some do not. Lastly, cross-talk from high speed signals to the thermal diode traces might also introduce an error.

Re:check your facts pls

[ben there...]

I know you were referring to P4, but while we're talking about Intel...with Core 2, there are actually two thermal protections built into the chip. THERMTRIP#, which you mentioned, is used for throttling, which can be disabled in the BIOS. The other maxes out a few degrees Celsius higher, and actually shuts off the chip to prevent it from frying. That one can't be disabled. Both are mentioned in the Core 2 thermal design whitepaper somewhere on Intel.com.

Also, most motherboard monitoring programs don't accurately read the Core 2 thermal diode. They likely use the other sensor. Core Temp and Intel TAT give the most accurate temps from the diode.

Re: Fast Enough

[TaoPhoenix]

I agree.

I am currently using a retired office machine designed to buy the 8 months left until Intel perfects their new generation coming out this year. Then I am quite sure that machine will be Fast Enough for a long time. For the projects I do, the BioWetware of my mind is the limiting factor, not the speed of the silicon when I finally click Start (foo). Also, the TeraByte drives are due at that same period, so Storage will also be Large Enough.

All that's left is to study the price breaks to get solid quality just before the price joins Richard Branson in the sky.

Right, but you forgot RAID

[tacokill]

Yep, that's exactly right. But - that's why many of us put (2) WD Raptors in RAID 0 (stripe) and then you can move about 120-145 meg/s. This is because the stripe reads/writes both drives at the same time thus increasing total throughput to the bus.

So yea, its still a bottleneck. SATA2 increases things to 300 meg/sec so that will help some.

Re:Right, but you forgot RAID

[dfghjk]

RAID 0 doesn't automatically ensure that both/all drives will read/write at the same time. In fact, RAID 0 is typically configured NOT to optimize that. Large interleave sizes, the ones typically used, optimize for multiple concurrent requests, not parallel throughput. Furthermore, each drive in a SATA system will get its own SATA link, so unless you are attaching a RAID 0 box via a single SATA link, your entire point is specious. Even if you had 2 RAID 0 drives with a small interleave attached via a single SATA 150 link, you've proven that 150 MB/sec is fast enough.

Bah

[StikyPad]

Is anyone else sick of seeing chip makers try to speculate about the merits and performance capabilities of major metropolitain areas? As if anyone could outperform a people with the power of clovers on their side!

You can't RAID a single drive

[Mad Merlin]

I'm aware of RAID, but we were only speaking of a single hard drive, which rules out any useful application of RAID.

LZCNT

[CrimsonScythe]

I can't find much information on it, but I'm guessing "LZCNT" is count-leading-zeros.

No, I'm pretty sure "LZCNT" means "lazy cunt". I'll leave "POPCNT" as a mental exercise for the reader...

Lies or falsehoods?

[eviljav]

AMD also claimed they were going to release a 35W version of their AM2 3800 processor last June, but as of yet it's still not available at any retailer. Only 7 months late so far! So, take what they claim with a large grain of salt...

Faster at calculations perhaps....

[ivan256]

...but will it boil water as quickly as the (120 watt!) clovertown?