Slashdot Log In
AMD Says Barcelona Will Outperform Clovertown
Posted by
kdawson
on Sat Jan 27, 2007 06:59 PM
from the horse-race dept.
from the horse-race dept.
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."
Related Stories
[+]
AMD's Showcases Quad-Core Barcelona CPU 190 comments
Gr8Apes writes "AMD has showcased their new 65nm Barcelona quad-core CPU. It is labeled a quad-core Opteron, but according to Infoworld's Tom Yeager, is really a redefinition of x86. Each core has a new vector math processing unit (SSE128), separate integer and floating point schedulers, and new nested paging tables (to vastly improve hardware virtualization). According to AMD, the new vector math units alone should improve floating point operation by 80%. Some analysts are skeptical, waiting for benchmarks. Will AMD dethrone Intel again? Only time will tell."
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Full
Abbreviated
Hidden
Loading... please wait.
If only I/O speeds could also grow as fast (Score:5, Insightful)
Re: (Score:2)
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: (Score:3, Informative)
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 (Score:5, Insightful)
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 [fujitsu.com]. 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.
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.
Parent
Re:If only I/O speeds could also grow as fast (Score:5, Informative)
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.
Parent
Re: (Score:3, Informative)
By what criterion= (Score:2)
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 "a
Re: (Score:2, Redundant)
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: (Score:2)
More cores per socket, but the socket's DDR2 bandwidth stays the same. Let's hope your kernels fit in the L2/L3.
Re: (Score:2)
Re: (Score:2)
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 w
40 % faster, after a nap (Score:5, Funny)
Re:40 % faster, after a nap (Score:5, Funny)
Yeah, but if it's 40% faster, an hour-long siesta should only 35 minutes.
Parent
upcoming chipset? (Score:4, Insightful)
FTFS:
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: (Score:3, Informative)
Re: (Score:2)
The bounty of true competition (Score:5, Insightful)
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.
Imagine how great it would be... (Score:2, Interesting)
...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: (Score:2)
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: (Score:2)
Re: (Score:3, Insightful)
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: (Score:2)
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: (Score:3, Insightful)
AMD is growing (Score:3, Interesting)
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: (Score:2)
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 mo
Re: (Score:3, Informative)
Don't believe this (Score:2)
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 48
Re: (Score:2)
Re:Don't believe this (Score:5, Insightful)
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?
Parent
Re: (Score:2)
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 (Score:5, Insightful)
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.
Parent
Re: (Score:3, Insightful)
Re: (Score:2, Insightful)
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?
So.. the writer expects Intel to sit still? (Score:3, Interesting)
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: (Score:3, Insightful)
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
Re: (Score:2)
Re: (Score:2)
Intel does in fact do very stupid things from time to time. The question is, is this one of them?
what's wrong with intel (Score:3, Interesting)
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: (Score:2)
Mainly in FP (Score:5, Informative)
- 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 (Score:5, Insightful)
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).
Parent
Obvious (Score:3, Funny)
Population of Barcelona: 1,673,075 [wikipedia.org]
Population of Clovertown [google.com]: 5601 [census.gov] (or less)
Barcelona is vastly superior.
Marketing speak (Score:5, Funny)
The way they spun it, you can also claim they changed their strategy from slow to expensive.
Re: (Score:3, Informative)
Re: (Score:3, Informative)
Re: (Score:3, Informative)
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
Re:Linux on quad-core (Score:5, Funny)
Parent
Re:Linux on quad-core (Score:5, Informative)
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.
Parent
Re: (Score:3, Informative)
Re:AMD is not and never has been a serious CPU (Score:4, Insightful)
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. 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. 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.
Parent