Slashdot Mirror
Intel Ships Dual-Core Chips
Torrey Clark writes "Intel seems to be the first to ship a batch of dual core x86 64-bit processors to OEMs.
Intel's first dual-core chip is the Intel Pentium Processor Extreme Edition 840. The new processor runs at 3.2 GHz, backs Intel's Hyper-Threading and is supported by the company's 955X Express chipsets, formerly code-named Glenwood.
Dell also announced that it would be one of the first PC makers to ship Intel's new dual core Pentium Extreme." Reader wyckedone adds "AMD is set to ship their dual core Opteron processor, designed for servers, next week."
There's a bunch of interesting information about AMD's dual core offerings over at AnandTech. Very insightful read.
From the article. "If dual core Opterons do indeed have two memory controllers, the pincount of dual core Opterons will go up significantly - it will also make them incompatible with current sockets. AMD is all about maintaining socket compatibility so it is quite possible that they could only leave half of the memory controllers enabled, in order to offer Socket-940 dual core Opterons. AMD isn't being very specific in terms of implementation details, but these are just some of the options."
I think these are the "phony" dual cores. Two dice bonded together in the same package.
Sun, HP and IBM have allready got machines ready, just waiting for launch.
p teron_dualcore/
http://www.channelregister.co.uk/2005/04/08/amd_o
"If everybody is thinking alike, somebody isn't thinking" - Gen. George S. Patton
SMT is only needed if your execution units are having trouble remaining filled up, which was the problem with the NetBurst architecture due to the huge hits that it takes with a branch mis-prediction penalty. When a mis-predict happens the execution unit has to sit idling away and wait for the proper info to go be re-fetched. With SMT, the unit simply switches over to one of the other threads waiting in the wings which keeps the processor doing useful work instead of wasting cycles. This is why the software has to be re-written to take advantage of it so that the processor knows which threads to give priority to.
Intel stuck SMT into the Pentium in order to balance out the some of the negative effects the go hand-in-hand with a processor that has a LONG pipeline. AMD has a much shorter pipeline (especially when compared to the new Prescott) and therefore they don't suffer much of a penalty when a mis-predict happens. Also, if I remember correctly the Athlon was already known being extremely efficient in terms of resource allocation within the processor since AMD can't afford to just dump tons of extra cache onto the chip.
Both of these things taken together means that using up extra real estate on the die of the Athlon in order to get SMT isn't really worth it in terms of the performance it would bring. Even on the Pentium the benefits aren't all that hot and it's only in specific types of code that you see any impresive speed gains.
I tried to look through the article, but couldn't find it... does this mean that you will have 4 "virtual cores" on one single chip now?
Yes, but only for the Pentium Extreme Edition. The Pentium D, which should come out soon, won't have HyperThreading enabled.
I support the Center for Consumer Freedom
it has been speculated that they might just place two normal cores under one heatspreader.. can't know sure untill the reviews hit that open them up.
also.. it could be a paper launch for most parts(paper launch= you launch the product, ship it to reviewers.. but are unable to provide the product in significant numbers to any resellers). they're getting way too popular.
world was created 5 seconds before this post as it is.
You can also put 4 dual core processors in a normal 4 socket motherboard to make a 8-way computer. AMD has tons of information on this stuffhttp://www.amd.com/us-en/Corporate/VirtualPre ssRoom/0,,51_104_543~94936,00.html
mnewberg.com
This has been studied, but you get more like 8 cores, not 64. Look up the DEC Piranha or Sun Niagara.
Compilers would be designed to break a program up not into a paltry 2 or 3 threads, but into dozens.
Assuming that was possible, which it isn't.
Are AMD's and/or Intel's processors supposed to work in existing motherboards?
AMD: Yes
Intel: No
A64/Opterons (especially the HE et al) are decently efficient for power usage.
http://finance.yahoo.com/q?s=AMD
Before everyone gets excited about Intel's dual core - you should review some of the benchmarks at tomshardware or anandtech. The speed improvements are not that great. In fact performance in a gaming environment is not as good as their high end single core chips.
Intel is currently pandering to the gamer/workstation market. Given the Dell XPS announcement - clearly a target is the gamer. Is it really though? Would any self respecting gamer buy a Dell as their "l33t boXor"? Would they not already know that the nVidia + ATI + AMD64 platform represents the pinnacle of performance in the gamer world? Probably. I'm not sure that dell is going to see the boost to their XPS line that they'd like from the addition of this chip. Intel does get points for getting their earlier but it's not nearly the lead that AMD took a year ago with the deployment of the AMD64 3X00+ desktop processors. The bottom line however is that if you are planning to buy a new machine and actually do a little research you'll find that the Intel chips are not as fast dual core or not - as the AMD chips.
As far as workstation performance goes - the chip seems to hold it's own with some decent performance gains. The real issue with this chip and the architecture as a whole is memory bandwidth. For truly intensive processing tasks, video, audio, data processing, computationally intensive tasks, a fast memory bus makes a world of difference. Intel doesn't seem to have the architecture to support these types of task as well as it should. However, this may be the only area that AMD doesn't have covered well primarily due to the cost of their dual core offering. If you are looking for a workstation that can be programmed to handle multiple threads the Intel offering might be for you.
And then there is AMD - totally ignoring the desktop market. Instead they are going after the high end server market. Why? Itanic is dead - and there is a need, and a void for high density, but "cheap" machines. The dual core AMD provides high performance, low heat and competitive performance especially in situations where high memory bandwidth is needed. It also scales 2x further than the competing Xeon servers (The 8XX series could be called the 16XX series). Since the launch of the Opteron two years ago AMD has established the proper channels to deliver these chips to customers. Sun and HP both offer servers with these chips and will also be offering the dual core chips as well.
Pricing - which I think is most telling. AMD's products are priced at the high-end. They are the leader in 64 bit computing (Intels 64-bit approach is architecturally inferior). They will provide organizations who need this technology an upgrade path to the 64-bit world if they are not already in it. AMD chips scale better than the competing Intel technologies. Thus AMD will continue to consolidate their lead in the high end server market. Also, AMD appears confident that their customers will pay for their high end CPUs. In the workstation/gamer market Intel is trying to stoke adoption through lower prices. In some ways this is a contradiction - low priced chip in a high end segment. I'd argue it's the wrong strategy for stemming their loss of market share to AMD. Why go with an Intel dual core when you can get a dual Opteron? They are faster after all.
Though it comes down to the fact that Intel and AMD are after different things. The Opteron platform is a high-end platform. It's clear that AMD is making it a priority. A sound strategy, as Intel's blunder with the Itanium is still continuing to cause a slip in market share. Second is the AMD64 platform which offers great performance at a reasonable price. No doubt, when dual core hits the AMD64 their will be a bit of a premium to pay - it is a sound architecture which is just starting to hit it's stride. Conversely, Intel is trying to stem the bleeding on their desktop lines - clearly they see the high end gaming and workstation segment suffering. How desperate does a company need to be to tack on "Extreme Edition"
The Xeon parts to come are SIP, but it looks like the EE is a single die. Though, with no on-chip interconnect, it seems a distinction without a difference. EETimes.com - Potholes seen in latest Intel road map
Free login required. Messy to reach the article if not already logged in. The login dumps you at the the main page, not the page you wanted. Click on the link again after logging in.
I've always wondered why software companies are allowed to get away with charging more money for multi-CPU systems anyway. I mean, there's still only one copy of the software on the system, so where's the justification in charging for more than one CPU?
Afterall, you don't have to pay twice as much to run the software on a system that has a CPU with twice the clockspeed...
A good question. I've supported and tested and used per-cpu licensed software, so here's the theory:
it's how much use you get out of the software.
For example: If I have a render farm of 6 computers, I may have 6 licenses. (This example is OBSOLETE - most render-only licenses are free now).
If I upgrade my hardware so i need only new workstation to do the same amount of work, I lost 5 sales units.
So the trick is to keep sales income flat with the hardware curve.
I'm not advocating this -- just pointing out the why behind the what.
AMD *CPUS* are fine, the problem is the POS SIS, VIA and various other low-quality, buggy and incompatible chipsets that proliferate on AMD-compatible motherboards.
I don't have a problem with AMD chips, but long and bitter experience with certain chipset manufacturers (VIA in particular) has ensured I will never buy another product with their name attached to it again, nor in any way be involved in supporting such a machine. Thus, I am rather limited in my choice of AMD-compatible motherboards, particularly at the not-frighteningly-expensive end of the market.
>> using up extra real estate on the die of the Athlon in order to get SMT isn't really worth it in terms of the performance it would bring. Even on the Pentium the benefits aren't all that hot and it's only in specific types of code that you see any impresive speed gains
The real estate used is only about 5% on a P4. If you get more than 5% return in performance (as you do in many cases), then it's a win. It's really the complexity of it all that kills it for AMD - they can't afford the engineering resources to put something like that in. SMT was actually implemented on Willamette (the first P4) way back in 2000, but it was disabled until the engineers could get it to work a few years later.
It may interest you to know that many resources on an out-of-order machine are often idle. On an Athlon, you may get about 1-1.5 instructions per clock throughput on average when the peak is 3. Adding another thread can more fully utilize those idle resources. Beyond 2 threads, the resources are probably saturated (1.5 x 2 = 3).
Then you move to dual core, where your execution resources double again.
So that raises the question, why not build a wider processor with more resources but with a single thread and let the hardware find the parallelism for you INSTEAD of dual-core/SMT where the burden is on the programmer? Short answer - it takes too much power for hardware to extract paralellism.
This isn't some cop-out by "lazy" circuit designers. This is fundamental (ie - the laws of physics work that way).
Sorry coders - your lives will get harder and harder as the years go on and many-threaded processors sell a 100M units in volume each quarter. Better get to work on making parallel programming easier, or you won't see the performance gains that you've come to expect from Moore's Law.
Read the two part review of Intel's new dual core chip at www.anandtech.com. Maybe you'll put away your typical slashdot pessimism away.
... but Everyone runs a few programs at once.
... but the testers could easily flip back and forth between many different programs, each doing hard tasks, without the computer stalling or the programs going totally non-responsive, and overall more being accomplished in the same time.
... Constantly having to flip back and forth, fire up one, close another, grab a coffee while I wait.
Yes, very few programs are ready for multithreading now
They did a great multi-tasking test to compare the usability of the new dual core chip at 3.2ghz versus intel's 3.73ghz single core chip. And pointed out that Windows XP is in fact multithreaded and can benefit well on its own from a dual core chip.
Of course the single threaded programs ran about the same as usual
The 'real-world' usage tests showed a huge benefit to having dual core, with much smoother operations, far better than hyperthreading alone.
Poor non-HT AMD user's like myself dream of being able to multitask that much without waiting forever when switching between active programs.
I have no doubt that I literally could save an hour or two per day of wasted time at my job if I had a dual core processor. Two large autocad files, an ArcGIS dataset, text editors and more
George Bush + Linux = "I will not let information get in the way of the fight against Windows"
For what it's worth, the parent is a blatant ripoff from this post from January:
= 11516655
http://it.slashdot.org/comments.pl?sid=137696&cid
You have to ship them for production for it to count as a release. Shipping evaluation units to OEMs is just a stage of development. Until someone can actually buy a computer that contains this processor it hasn't been released yet.
Dual core Opterons have been shipped for production. The actual servers and workstations containing them will be available for immediate delivery on April 21. Computers containing dual core PEEs (snicker) will not be.
Afterall, you don't have to pay twice as much to run the software on a system that has a CPU with twice the clockspeed...
I guess you have never priced an Oracle database.
love is just extroverted narcissism