Low Latency L2 Cache

We mentioned at the start of this review that the Pentium M featured a very large, yet low latency L2 cache - at 2MB for the current 90nm Dothan based Pentium M. But how much low latency are we talking? To find out, we turned to ScienceMark 2.0 and Cachemem.

First, let's have a look at the latency of the Pentium M's 64KB L1 cache:

   Cachemem L1 Latency  ScienceMark L1 Latency
AMD Athlon 64 3 cycles 3 cycles
Intel Pentium 4 (Northwood) 1 cycle 2 cycles
Intel Pentium 4E (Prescott) 4 cycles 4 cycles
Intel Pentium M 3 cycles 3 cycles

With such a large L1 cache, it is difficult to get much lower than 3 cycles, as we see that the Pentium M has a similar L1 access latency as the Athlon 64. What is also important to note, however, is that the Pentium M does have a lower L1 access latency than the Pentium 4E.

But what we came here to look at was L2 cache latency, which matters much more in real world application performance where not everything fits into L1:

   Cachemem L2 Latency  ScienceMark L2 Latency
AMD Athlon 64 17 cycles 18 cycles
Intel Pentium 4 (Northwood) 16 cycles 16 cycles
Intel Pentium 4E (Prescott) 23 cycles 23 cycles
Intel Pentium M 10 cycles 10 cycles

Here's where things get very interesting - the Pentium M has the lowest L2 cache access time of any of the modern day desktop microprocessors. With a 10 cycle L2 latency, any application that fits within the Pentium M's 2MB cache will most definitely perform very well on the CPU. It is the 10 cycle L2 that allows the Pentium M to be competitive with much higher clocked CPUs in most mobile applications as they are normally office application tasks that are generally very cache-friendly. Keep this in mind as we look at the actual performance numbers of the Pentium M.

Understanding Pentium M Architecture Memory Latency and Bandwidth
Comments Locked

77 Comments

View All Comments

  • bob661 - Tuesday, February 8, 2005 - link

    The only problem with this chip is that the marketing is oriented towards the mobile market and therefore not a direct competitor to the A64. It would be nice if it was. It might bring some cats out of the bag on the AMD side. Competition in the marketplace is good for us all.
  • jvrobert - Tuesday, February 8, 2005 - link

    Really, AMDroids, get a grip. You're all excited because the AMD chips beat a mobile processor pretty handily, and because you are making some silly assumption that the Pentium-M in its current form is Intel's "last chance".

    First, Intel doesn't need a last chance. They make enough money to make AMD look like a Mexico City taco stand. So enough of those delusions of grandeur.

    But on a technical front, if Intel ramps the clockspeed up to the 2.8 range (easy), and releases a desktop class chipset for the Pentium M it would match or exceed any current chip. And these are _basic_ steps. What if they made more improvements?
  • jvrobert - Tuesday, February 8, 2005 - link

  • bob661 - Tuesday, February 8, 2005 - link

    #45
    You are a rock. The point of the article was to compare the P-M to desktop CPU's because most of us here wanted to know it will perform. And you know what? It performed very nicely.
  • classy - Tuesday, February 8, 2005 - link

    I just can't help but to laugh at some folks. Its a nice chip but clearly not in the A64 ballpark. Its that simple. As far as the 2.8 oc, that was only accomplished in one reveiw. All the reviews show the same thing you have oc so it can it compete. What's interesting though is most of these Intel fanboys don't want to see a comparison of an oc'ed A64 vs a Dothan. Smoke city :)
  • FrostAWOL - Tuesday, February 8, 2005 - link

    IF the Pentium-M and P4 are electrically incompatible then someone please explain this:

    HP Blade system Pentium-M with Serverworks GC-SL chipset
    http://h18000.www1.hp.com/products/servers/prolian...

    FrostAWOL
  • jae63 - Tuesday, February 8, 2005 - link

    Great review & of interest to those of us with HTPCs. Too bad the price point is so steep.

    One minor correction on page 11:
    "The Pentium M does a bit better in the document creation tests, as they are mostly using applications that will fit within the CPU's cache. However, the introduction of a voice recognition program into the test stresses the Pentium M's floating point performance, which does hamper its abilities here."

    Actually NaturallySpeaking uses almost no floating point but is very memory intensive. The performance hit that you are seeing is because it uses a lot of memory bandwidth and its dataset doesn't fit in the L2 cache.

    Here's some support for my statement, by the main architect of NaturallySpeaking, Joel Gould:
    http://tinyurl.com/6s4mh
  • segagenesis - Tuesday, February 8, 2005 - link

    #43 - I think you have the right idea here. This processor is not meant to be performance busting but rather a low energy alternative to current heat factories present inside every P4 machine. I would love to have this in a HTPC machine myself but the cost is still too damn high. Hopefully higher production will bring the cost down.
  • Aileur - Tuesday, February 8, 2005 - link

    I guess the pentium M isnt ready (yet) for a full featured gaming machine, but with that kind of power, passively cooled, it would make for one hell of an htpc.
  • PrinceGaz - Tuesday, February 8, 2005 - link

    #45- It was not an unfair review, on the contrary it seemed very well done. The reason the P-M was compared with fast P4 and A64's is because they cost about the same.

    Maybe someone else buys your computers for you, but most of us here have to spend our own money on them so cost is the best way to decide what to compare it with.

Log in

Don't have an account? Sign up now