Intel: More Details on Penryn and Nehalem
by Johan De Gelas on March 28, 2007 3:00 PM EST- Posted in
- CPUs
Better Virtualization
Intel's current hardware support for virtualization in the current Core architecture is lackluster to say the least. To understand this you must understand what happens in a "pure" software-based virtualization solution such as VMware ESX 2.5.3 running on older Intel CPUs.
A technique called "ring deprivileging" is used as the guest OS cannot be allowed to run in the lowest ring 0 where it normally runs; the Virtual Machine Manager or hypervisor now runs there. That means that every time the guest application asks the help of the guest OS, which needs to run instructions which are only available in ring 0, the VMM must intercept that "SYSENTER" and emulate the normal execution. This is quite costly in performance terms.
Hardware assisted virtualization does not have that problem: both the OS and the VMM have their own ring 0. Despite this, Intel's HW assisted solutions didn't give any speed boost. It has not been discussed in detail, but Penryn speeds up virtual machine transition (entry/exit) times by 25% to 75%, and this requires no virtual machine software changes. This might be similar to AMD's nested page technology, although we don't have any clear details at present.
Last but not least, the dual core Penryn processors get a 6 MB shared cache and the quad versions get 12 MB cache. Both new designs will also come with a "higher degree of associativity". Considering the current designs are 16-way set associative, most likely the newer chips will feature a 24-way set associative L2 cache.
Intel EDAT: the End of the Multi-core Clock Speed Disadvantage?
Intel also talked about its "Enhanced Dynamic Acceleration Technology" which is effectively integrated overclocking based on load. If you are running a single threaded application (or a multi-threaded application that's predominantly using a single thread), Intel's EDAT can power down the second core and increase the frequency of the working core to maintain the same thermal envelope at all times.
Intel's EDAT could spell the end of the clock speed differential between single and multi-core processors. With all cores running workloads, the multi-core system would be clocked lower, but when some cores are idle the chip could potentially run at the same speed as a single core solution would. Single core designs have pretty much disappeared from roadmaps already, but considering there are still applications that are single threaded in nature and benefit more from clock speed improvements, future processors will offer both options in a single package.
Performance
Intel hasn't revealed too much about the performance of Penryn but Pat did leave us with a few comments. We don't know anything more about the test conditions than what we are presenting, and we didn't do the measurements ourselves, so take it for what it's worth.
Comparing a 3.2GHz Penryn (1.6GHz FSB) to a 3.0GHz Conroe (1.33GHz FSB), Intel has measured more than 20% increase in gaming performance (with no code changes). For video encoding applications, if SSE4 is utilized, the same Penryn vs. Conroe comparison can offer more than a 40% increase in performance.
Finally, Intel mentioned that in the server space, the fastest quad core Penryn available (>3GHz) vs. a 2.67GHz quad core Xeon resulted in a greater than 45% increase in performance in "bandwidth and FP intensive applications". It's incredibly vague (and oddly similar to AMD's claims of Barcelona vs. Xeon performance), but Pat mentioned that STREAM and certain benchmarks in SpecFP could be considered to be "bandwidth and FP intensive".
Again, we are just reporting what Intel told us. It will be a while before we can actually verify any of these claims or put them in the right context. Given the various enhancements that we've reported on, however, it's only reasonable to expect Penryn to be faster than Conroe, clock-for-clock. Whether that's 10% faster, 20% faster, or something else will be made clear in the future.
Intel's current hardware support for virtualization in the current Core architecture is lackluster to say the least. To understand this you must understand what happens in a "pure" software-based virtualization solution such as VMware ESX 2.5.3 running on older Intel CPUs.
A technique called "ring deprivileging" is used as the guest OS cannot be allowed to run in the lowest ring 0 where it normally runs; the Virtual Machine Manager or hypervisor now runs there. That means that every time the guest application asks the help of the guest OS, which needs to run instructions which are only available in ring 0, the VMM must intercept that "SYSENTER" and emulate the normal execution. This is quite costly in performance terms.
Hardware assisted virtualization does not have that problem: both the OS and the VMM have their own ring 0. Despite this, Intel's HW assisted solutions didn't give any speed boost. It has not been discussed in detail, but Penryn speeds up virtual machine transition (entry/exit) times by 25% to 75%, and this requires no virtual machine software changes. This might be similar to AMD's nested page technology, although we don't have any clear details at present.
Last but not least, the dual core Penryn processors get a 6 MB shared cache and the quad versions get 12 MB cache. Both new designs will also come with a "higher degree of associativity". Considering the current designs are 16-way set associative, most likely the newer chips will feature a 24-way set associative L2 cache.
Intel EDAT: the End of the Multi-core Clock Speed Disadvantage?
Intel also talked about its "Enhanced Dynamic Acceleration Technology" which is effectively integrated overclocking based on load. If you are running a single threaded application (or a multi-threaded application that's predominantly using a single thread), Intel's EDAT can power down the second core and increase the frequency of the working core to maintain the same thermal envelope at all times.
Intel's EDAT could spell the end of the clock speed differential between single and multi-core processors. With all cores running workloads, the multi-core system would be clocked lower, but when some cores are idle the chip could potentially run at the same speed as a single core solution would. Single core designs have pretty much disappeared from roadmaps already, but considering there are still applications that are single threaded in nature and benefit more from clock speed improvements, future processors will offer both options in a single package.
Performance
Intel hasn't revealed too much about the performance of Penryn but Pat did leave us with a few comments. We don't know anything more about the test conditions than what we are presenting, and we didn't do the measurements ourselves, so take it for what it's worth.
Comparing a 3.2GHz Penryn (1.6GHz FSB) to a 3.0GHz Conroe (1.33GHz FSB), Intel has measured more than 20% increase in gaming performance (with no code changes). For video encoding applications, if SSE4 is utilized, the same Penryn vs. Conroe comparison can offer more than a 40% increase in performance.
Finally, Intel mentioned that in the server space, the fastest quad core Penryn available (>3GHz) vs. a 2.67GHz quad core Xeon resulted in a greater than 45% increase in performance in "bandwidth and FP intensive applications". It's incredibly vague (and oddly similar to AMD's claims of Barcelona vs. Xeon performance), but Pat mentioned that STREAM and certain benchmarks in SpecFP could be considered to be "bandwidth and FP intensive".
Again, we are just reporting what Intel told us. It will be a while before we can actually verify any of these claims or put them in the right context. Given the various enhancements that we've reported on, however, it's only reasonable to expect Penryn to be faster than Conroe, clock-for-clock. Whether that's 10% faster, 20% faster, or something else will be made clear in the future.
Facebook
Twitter
RSS
14 Comments
View All Comments
cornfedone - Wednesday, March 28, 2007 - link
Intel knows they are about to get punished again by AMD so they are greasing the skids to get the media hype they need to delude consumers into believing Intel's products will still be competitive after Barcelona, when the facts prove otherwise.Hey if nothing else, Intel is good for a laugh. It's nice to see those criminals squirm for a change. It's only gonna get worse when they get to court and AMD proves that Intel violated law over and over and over again in their obsession to eliminate competition.
fitten - Thursday, March 29, 2007 - link
Intel couldn't, and wouldn't, eliminate AMD. In fact, Intel would probably float AMD a loan, if it came to it, in order to keep AMD afloat simply to avoid the backlash of other potential monopoly charges should AMD disappear.Plus, Penryn is the next thing up for Intel, supposedly out around the time as AMD's Barcelona (later this year - which is 2007). Nahelem is out in 2008-ish and does even more (according to the dates listed, a year before AMD's Fusion which is slated for 2009).
All I can do is laugh giddily because I have no allegiance to either CPU maker. I simply buy what is the best bang for the buck (for a long time now, it was AMD and I've had about a dozen AMD boxes) which, for the next couple/few years, looks like it's going to be Intel. If nothing else, the competition will kick AMD into overdrive and competition will get even hotter... and I benefit regardless of who is on top.
johnsonx - Wednesday, March 28, 2007 - link
ah, Cramitpal, how we missed you!Crassus - Thursday, March 29, 2007 - link
Oh year, those were the days ...amdsupport - Wednesday, March 28, 2007 - link
Congratulations! You've just constructed the 4millionth floor on the edifice of stupidity.
Seriously...there is no basis for claiming AMD's new stuff(or rather, refreshed) will exceed Intel's offerings. http://www.anandtech.com/cpuchipsets/showdoc.aspx?... ">Anandtech themselves claimed in a previous article that "AMD could come close to offering something competitive to Intel." That sounds like AMD will only bring themselves back in line with Intel's current offerings...not exceed or punish as you claim.
As far as the court proceedings go...any result from it at best would result with a simple "slap on the wrist" punishment at best.
Justin Case - Thursday, March 29, 2007 - link
If the numbers I've been hearing are right (and they probably are), AMD will get ahead of Intel in terms of floating-point performance, and will extend its lead in terms of memory performance (Intel's real Achilles's heel in HPC and servers).On the desktop, they'll be more or less matched with Barcelona, but AMD's Fusion could tip the scales in its favor for gaming systems.
Intel will probably retain its lead in mobile platforms, but I don't have much real data on Barcelona's power-saving features, so I really can't say for sure.
Penryn should put Intel back in the lead on the desktop (Fusion being the unknown factor here) and bring them pretty close in servers (Intel is still trying to push Itanium, so there are a lot of internal conflicts about server-oriented improvements, which gives AMD a bit of slack).
All in all, it's business as usual: Intel is better at manufacturing and optimization, AMD takes more risks with design (and, with all the great engineers they hired from DEC, Intel, IBM, ATI, etc., they seem quite capable of pulling it off).
Phynaz - Wednesday, March 28, 2007 - link
Is the sound of piss running down Hectors leg.HammerFan - Wednesday, March 28, 2007 - link
I really like to see this kind of information, that neither company is sitting down and saying "enough's enough." I wonder though, if Intel is going to use their own graphics systems in their all-in-one chip, or if they'll turn to nVidia/ ATi for a graphics solution. As for chipset manufacturers, how well they do will largely depend on the features they can pack into their SB's IMHO, as that's largely what I look for when I want a new MB, since performance from board to board doesn't vary that much anyway.chucky2 - Wednesday, March 28, 2007 - link
Yeah, I'd myself really like to see AMD, Intel, and nVidia adding support of native 1394b (Firewire 800) in their chipsets, so products like that can start moving forward.Having eSATA would be nice as well...
Chuck
tumbleweed - Wednesday, March 28, 2007 - link
The earlier a company discloses what their plans are, the earlier they get feedback on how the public likes or dislikes some of their plans.Think about how much money Intel could have saved if they'd realized much earlier how badly the public hated RAMBUS.