Spring IDF 2006: Gelsinger Benchmarks Conroe
by Anand Lal Shimpi on March 7, 2006 1:28 PM EST- Posted in
- Trade Shows
Pat Gelsinger took the stage after Rattner's keynote was over, and he offered a high level overview of the Enterprise market. There were a few interesting morsels that Gelsinger shared during his presentation.
First off we've got Intel's array of quad-core CPUs, note that these are all two dual core die on a single package. Kentsfield is the desktop offering.
Gelsinger put up an interesting slide that showed increase in CPU performance relative to the 486, showing that the Core architecture offers the biggest increase in performance since the original move to an Out of Order execution core:
Pat teamed up with Microsoft to bring a demonstration of Office 12 running on Conroe as well as Pentium D 950. Microsoft mentioned that Office 12 would have a greater focus on multithreading, and this particular test was of a multithreaded Excel 12 task:
Conroe completed the task in less than half the time, but in general Intel is still saying a 40% improvement over Pentium 4.
Now we're off to benchmark Conroe ourselves...
Facebook
Twitter
RSS
20 Comments
View All Comments
rayo123 - Tuesday, March 7, 2006 - link
The one showing normalized bars for various Pentiums is bogus. We all know the P4 was inferior in core int/fp performance to the P3 clock by clock.DeathSniper - Tuesday, March 7, 2006 - link
Unfortunately, most people out there (Jack and Jill) don't.DeathSniper - Tuesday, March 7, 2006 - link
They just see a neat-looking and pretty graph ;)ElFenix - Tuesday, March 7, 2006 - link
i like how the performance chart is normalized to a 25 mhz 486 dx.and how is a 2.8ghz pentium d more than 2x as fast as a 1.4 ghz pentium 4?
anyone want to run specint on some things and see if intel is shoveling it?
psychobriggsy - Tuesday, March 7, 2006 - link
So Intel think a PIII (450MHz at introduction) has around 1/4 the Integer performance of a P4, which was 1.5GHz at introduction and often beaten by 1GHz PIIIs in benchmarks?Conroe will be good though. It's good to see Intel is back on track. AMD have got their work cut out the next couple of years. It's no good if AMD are still ramping 65nm and Intel start shipping 45nm. K8L had better be more than interesting.
mschira - Tuesday, March 7, 2006 - link
This is the biggest crap ever.We all know that a Pentium D (which is basicly two slow Pentium 4's on one die) is NOT faster than a Pentium 4. At least single threaded.
So why it he so much faster than a Petium 4 in that slide?
I thought Intel is dooing better with their new CPUs, why would they have to lie that bloddy bad?
I fear for them they are not doing good after all.
One thing is for sure, belive no performance word from them. No single word.
Increadable.
M.
Hulk - Tuesday, March 7, 2006 - link
If they are comparing a Pentium 4 single core with a Pentium D using a multithreaded application they can get that performance delta.Griswold - Tuesday, March 7, 2006 - link
Congrats, you found today's FUD. :)BrownTown - Tuesday, March 7, 2006 - link
the reason the Pentum D beats the P4 is becasue these are launch speeds, so its being rcompared to a 1.5A Williamette, not a 3.8 PResscottpsychobriggsy - Tuesday, March 7, 2006 - link
They're probably comparing a 1.5GHz P4 to a 3.2GHz Pentium D 840 (performance at introduction). But yes, the graph is pretty damn poor, there's not enough information on it.