Intel's Pentium Extreme Edition 955: 65nm, 4 threads and 376M transistors
by Anand Lal Shimpi on December 30, 2005 11:36 AM EST- Posted in
- CPUs
Larger L2, but no increase in latency?
When Prescott first got a 2MB L2 cache, we noticed that along with a larger L2 came a 17% increase in access latency. The end result was a mixed bag of performance, with some applications benefitting from the larger cache while others were hampered by the increase in L2 latency. Overall, the end result was that the two performance elements balanced each other out and Prescott 2M generally offered no real performance improvement over the 1MB version.
With Presler, each core also gets an upgraded 2MB cache, as compared to the 1MB L2 cache found in Smithfield. The upgrade is similar to what we saw with Prescott, so we assumed that along with a larger L2 cache per core, Presler's L2 cache also received an increase in L2 cache latency over Smithfield.
In order to confirm, we ran ScienceMark 2.0 and Cachemem:
What we found was extremely interesting; however, Presler does have the same 27 cycle L2 cache as Prescott 2M, but so does Smithfield. We simply took for granted that Smithfield was nothing more than two Prescott 1M cores put together, but this data shows us that Smithfield actually had the same higher latency L2 cache as Prescott 2M.
Although we were expecting Presler to give us a higher latency L2 over Smithfield, it looks like Smithfield actually had a higher latency L2 to begin with. This means that, at the same clock speed, Presler will be at least as fast as Smithfield, if not faster. Normally, we take for granted that a new core means better performance, but Intel has let us down in the past; luckily, this time we're not put in such a situation.
When Prescott first got a 2MB L2 cache, we noticed that along with a larger L2 came a 17% increase in access latency. The end result was a mixed bag of performance, with some applications benefitting from the larger cache while others were hampered by the increase in L2 latency. Overall, the end result was that the two performance elements balanced each other out and Prescott 2M generally offered no real performance improvement over the 1MB version.
With Presler, each core also gets an upgraded 2MB cache, as compared to the 1MB L2 cache found in Smithfield. The upgrade is similar to what we saw with Prescott, so we assumed that along with a larger L2 cache per core, Presler's L2 cache also received an increase in L2 cache latency over Smithfield.
In order to confirm, we ran ScienceMark 2.0 and Cachemem:
| Cachemem L2 Latency (128KB block, 64-byte stride) | ScienceMark L2 Latency (64-byte stride) | |
| AMD Athlon 64 X2 4800+ | 17 cycles | 17 cycles |
| Intel Smithfield 2.8GHz | 27 cycles | 27 cycles |
| Intel Presler 2.8GHz | 27 cycles | 27 cycles |
| Intel Prescott 2M | 27 cycles | 27 cycles |
| Intel Prescott 1M | 23 cycles | 23 cycles |
What we found was extremely interesting; however, Presler does have the same 27 cycle L2 cache as Prescott 2M, but so does Smithfield. We simply took for granted that Smithfield was nothing more than two Prescott 1M cores put together, but this data shows us that Smithfield actually had the same higher latency L2 cache as Prescott 2M.
Although we were expecting Presler to give us a higher latency L2 over Smithfield, it looks like Smithfield actually had a higher latency L2 to begin with. This means that, at the same clock speed, Presler will be at least as fast as Smithfield, if not faster. Normally, we take for granted that a new core means better performance, but Intel has let us down in the past; luckily, this time we're not put in such a situation.
Facebook
Twitter
RSS
84 Comments
View All Comments
Aenslead - Saturday, December 31, 2005 - link
As J.J., from Spider-Man would say:"Ceap, crap, mega-crap!" and then toss it away.
ElJefe - Saturday, December 31, 2005 - link
well it does move very fast in games. that is nice to see finally.it would be great if the overall power draw numbers were shown as on tomshardware. even there they showed a 90 watt difference between 4800 and the new 65nm. and that wasnt on the oc'd one. The oc'd one showed 150 more watts draw.
Viditor - Saturday, December 31, 2005 - link
Agreed...if it weren't for the X2, this would be an excellent chip by comparison!
Betwon - Friday, December 30, 2005 - link
Now, anandtech begin to learn the truth. There are still many knowledge about CPU that anandtech need to learn.The resluts of tests are simple and clear, but the reasons are complex.
In past years, anandtech took many mistakes about the correct reasons.
bldckstark - Monday, January 2, 2006 - link
You do realize that none of this stuff is very important, right? Both chips work well. Nobody should be criticized for buying either one of them.I love my FIVE computers but making sure my wife and kids are healthy and happy is way more important than any electronic device, especially just one piece of it.
Your damaging and hostile statements are making it appear as if you have forgotten this and the most important thing in the world is that you make all of us geeks think Anandtech is not perfect. News update - WE ALL KNOW THAT! We still like it.
bob4432 - Friday, December 30, 2005 - link
why don't you do the gaming benchmark with bf2 fps unlocked? it appears that it is just hitting its built in lock with both the fx-57 and also P955 EE 3.46 cpus.Spacecomber - Friday, December 30, 2005 - link
I believe that they are using the timedemo feature of the game and that the frame rate max doesn't affect this. It would be nice to see more than just average frame rates reported for games, though. At least a range should be mentioned and maybe a standard deviation.Space
Betwon - Friday, December 30, 2005 - link
We see a test, where the average fps of PD is less than (about 1% - 2%) the fps of AMD's. But PD's fps is more stable than AMD's.In the case that the average fps of netburst is better than the average fps of K8, the test shows that netburst is more stable than K8.
Betwon - Friday, December 30, 2005 - link
The test isn't bf2.bob4432 - Friday, December 30, 2005 - link
any link you could give me on how to do the time demo from within bf2? is this new with the 1.12 patch?thanks